Key Reviews

์„ค์ •ํ•œ ํ‚ค์›Œ๋“œ๋ณ„๋กœ, ๋Œ€ํ‘œ review ๋…ผ๋ฌธ๋“ค๊ณผ ๋‚ด wiki ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋ฅผ ์ข…ํ•ฉํ•ด ์„ธ๋ถ€ ๋ถ„์•ผ๋ณ„ ์‹ฌํ™” ๋ฆฌ๋ทฐ๋ฅผ ์ž๋™ ์ž‘์„ฑํ•ฉ๋‹ˆ๋‹ค. ํ‚ค์›Œ๋“œ์™€ ์„ธ๋ถ€ ๋ถ„์•ผ๋ฅผ ๋ˆŒ๋Ÿฌ ํŽผ์ณ ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

_์—…๋ฐ์ดํŠธ: 2026-09-12T20:04:04Z_

NSCLC โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 30ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
Precision Medicine and Molecular Biomarkers in NSCLC

๋น„์†Œ์„ธํฌํ์•”(NSCLC)์˜ ์น˜๋ฃŒ ํŒจ๋Ÿฌ๋‹ค์ž„์€ ์กฐ์งํ•™์  ๋ถ„๋ฅ˜๋ฅผ ๋„˜์–ด ๋ถ„์ž์ƒ๋ฌผํ•™์  ํ‘œ์ ์— ๊ธฐ๋ฐ˜ํ•œ ์ •๋ฐ€์˜ํ•™์œผ๋กœ ๊ธ‰๊ฒฉํžˆ ์ „ํ™˜๋˜์—ˆ๋‹ค. ์ด ๋ถ„์•ผ์˜ ํ•ต์‹ฌ์€ Epidermal Growth Factor Receptor (EGFR) ๋Œ์—ฐ๋ณ€์ด์ด๋ฉฐ, ์ด๋Š” ํŠนํžˆ ์„ ์•”(adenocarcinoma) ๋ณ‘๋ฆฌํ•™์—์„œ ๋†’์€ ๋นˆ๋„๋กœ ๊ด€์ฐฐ๋œ๋‹ค. ์ „ ์„ธ๊ณ„ ๋ฉ”ํƒ€๋ถ„์„์— ๋”ฐ๋ฅด๋ฉด NSCLC ํ™˜์ž ์ค‘ EGFR ๋Œ์—ฐ๋ณ€์ด ๋ณด์œ ๋ฅ ์€ ํ‰๊ท  32.3%๋กœ, ์•„์‹œ์•„ ์ง€์—ญ์—์„œ๋Š” 47%๊นŒ์ง€ ์ƒ์Šนํ•˜๋Š” ๋ฐ˜๋ฉด ์œ ๋Ÿฝ์—์„œ๋Š” 14.1%์— ๊ทธ์น˜๋Š” ์ง€๋ฆฌ์  ๋ฐ ์ธ์ข…์  ํŽธ์ฐจ๊ฐ€ ๋šœ๋ ทํ•˜๋‹ค [21, 22]. ๋˜ํ•œ ์—ฌ์„ฑ(43.7%)๊ณผ ๋น„ํก์—ฐ์ž(49.3%)์—์„œ ๋Œ์—ฐ๋ณ€์ด ๋นˆ๋„๊ฐ€ ํ˜„์ €ํžˆ ๋†’์œผ๋ฉฐ, ์ด๋Š” ์ž„์ƒ์ ์œผ๋กœ Tyrosine Kinase Inhibitors (TKIs) ํˆฌ์—ฌ ๊ฒฐ์ •์˜ ๊ทผ๊ฑฐ๊ฐ€ ๋œ๋‹ค [22]. EGFR ๋Œ์—ฐ๋ณ€์ด ์–‘์„ฑ ํ™˜์ž์—์„œ๋Š” ์ดˆ๊ธฐ ์น˜๋ฃŒ๋กœ ํ™”ํ•™์š”๋ฒ• ๋Œ€์‹  TKI๋ฅผ ์‚ฌ์šฉํ•˜๋Š” ๊ฒƒ์ด ์ƒ์กด์œจ ํ–ฅ์ƒ์— ๋” ํšจ๊ณผ์ ์ž„์ด ์ž…์ฆ๋˜์–ด, ๋ชจ๋“  NSCLC ํ™˜์ž์— ๋Œ€ํ•œ ๋Œ์—ฐ๋ณ€์ด ๊ฒ€์‚ฌ๊ฐ€ ํ•„์ˆ˜์ ์ธ ํ‘œ์ค€ ๊ด€ํ–‰์œผ๋กœ ์ž๋ฆฌ ์žก์•˜๋‹ค [15]. ์ด๋Ÿฌํ•œ ๋ถ„์ž ํ‘œ์  ์น˜๋ฃŒ์˜ ๊ฒฝ์ œ์  ํƒ€๋‹น์„ฑ๋„ ์ค‘์š”ํ•œ ๊ณ ๋ ค ์‚ฌํ•ญ์ด๋‹ค. Next-Generation Sequencing (NGS)์„ ํ™œ์šฉํ•œ ๋‹ค์ค‘ ์œ ์ „์ž ํŒจ๋„ ๊ฒ€์‚ฌ๋Š” ๋‹จ์ผ ์œ ์ „์ž ๊ฒ€์‚ฌ๋‚˜ ์ˆœ์ฐจ์  ๊ฒ€์‚ฌ ๋Œ€๋น„ ๋น„์šฉ ํšจ์œจ์„ฑ์ด ๋†’๋‹ค๋Š” ๊ฒฝ์ œํ•™์  ํ‰๊ฐ€๋“ค์ด ๋Œ€๋ถ€๋ถ„ ์ง€์ง€ํ•˜๊ณ  ์žˆ์œผ๋‚˜, ์•„์‹œ์•„ ์ผ๋ถ€ ์„ค์ •์—์„œ๋Š” ๋น„์šฉ ํŽธ์ต ๋ถ„์„ ๊ฒฐ๊ณผ๊ฐ€ ๋ถ€์ •์ ์ผ ์ˆ˜ ์žˆ์–ด ์ง€์—ญ๋ณ„ ์˜๋ฃŒ ์‹œ์Šคํ…œ ์ฐจ์ด๋ฅผ ๊ณ ๋ คํ•ด์•ผ ํ•œ๋‹ค [27].

EGFR ์™ธ์—๋„ ๋ฏธ์„ธ์œ„์„ฑ ๋ถˆ์•ˆ์ •์„ฑ(Microsatellite Instability, MSI) ๋ฐ DNA mismatch repair (dMMR) ๊ฒฐ์†์€ ๋ฉด์—ญ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก์˜ ์ค‘์š”ํ•œ biomarker๋กœ ๋ถ€์ƒํ•˜๊ณ  ์žˆ๋‹ค. ESMO ๊ถŒ๊ณ ์•ˆ์— ๋”ฐ๋ฅด๋ฉด, MLH1, MSH2, MSH6, PMS2 ๋‹จ๋ฐฑ์งˆ์˜ ์ƒ์‹ค์„ ํ™•์ธํ•˜๋Š” ๋ฉด์—ญ์กฐ์งํ™”ํ•™(IHC) ๊ฒ€์‚ฌ๊ฐ€ dMMR/MSI ํ‰๊ฐ€์˜ 1์ฐจ ๋ฐฉ๋ฒ•์œผ๋กœ ๊ถŒ์žฅ๋˜๋ฉฐ, PCR ๊ธฐ๋ฐ˜ ๋ฏธ์„ธ์œ„์„ฑ ๋ณ€์ด ๋ถ„์„์ด๋‚˜ NGS๋ฅผ ํ†ตํ•œ Tumor Mutational Burden (TMB) ๋ฐ PD-L1 ๋ฐœํ˜„๋Ÿ‰๊ณผ์˜ ์—ฐ๊ณ„ ๋ถ„์„์ด ์ •๋ฐ€์˜๋ฃŒ ๊ตฌํ˜„์„ ์œ„ํ•œ ํ•ต์‹ฌ ๋„๊ตฌ๋กœ ์ œ์‹œ๋œ๋‹ค [10]. TMB๋Š” ํŠนํžˆ Checkpoint Inhibitors (CPIs)์— ๋Œ€ํ•œ ๋ฏผ๊ฐ์„ฑ์„ ์˜ˆ์ธกํ•˜๋Š” ๊ฐ•๋ ฅํ•œ ์ง€ํ‘œ๋กœ, clonal TMB๊ฐ€ ์ด TMB๋ณด๋‹ค ๋” ์šฐ์ˆ˜ํ•œ ์˜ˆ์ธก ์ธ์ž์ž„์ด pan-cancer ๋ถ„์„์„ ํ†ตํ•ด ํ™•์ธ๋˜์—ˆ๋‹ค [16]. ๋˜ํ•œ, RNA sequencing (RNA-seq) ๊ธฐ์ˆ ์˜ ๋ฐœ์ „์€ ์œ ์ „์ž ์œตํ•ฉ(gene fusions) ๋ฐ ๋น„์•”ํ˜ธํ™” RNA(non-coding RNAs)๋ฅผ ํฌํ•จํ•œ ์ „์‚ฌ์ฒด ์ˆ˜์ค€์˜ ์ง„๋‹จ์  ์ •๋ณด๋ฅผ ์ œ๊ณตํ•˜๋ฉฐ, ์ด๋Š” ์กฐ์ง ์ƒ๊ฒ€์˜ ํ•œ๊ณ„๋ฅผ ๋ณด์™„ํ•˜๋Š” liquid biopsy ํ”Œ๋žซํผ์œผ๋กœ ํ™•์žฅ๋˜๊ณ  ์žˆ๋‹ค [19]. ์ตœ๊ทผ liquid biopsy๋Š” circulating tumor DNA (ctDNA)๋ฟ๋งŒ ์•„๋‹ˆ๋ผ extracellular vesicles (EVs) ๋ฐ DNA fragmentomics์™€ ๊ฐ™์€ ์ฐจ์„ธ๋Œ€ ๊ธฐ์ˆ ์„ ํ†ตํ•ฉํ•˜์—ฌ ์‹ค์‹œ๊ฐ„ ๋ถ„์ž ๋ชจ๋‹ˆํ„ฐ๋ง์„ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•˜๋ฉฐ, AI ๋ฐ machine learning๊ณผ์˜ ๊ฒฐํ•ฉ์„ ํ†ตํ•ด ์ง„๋‹จ ์ •ํ™•๋„๋ฅผ ํ•œ์ธต ๋†’์ด๋Š” ๋ฐฉํ–ฅ์œผ๋กœ ์ง„ํ™”ํ•˜๊ณ  ์žˆ๋‹ค [25].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ด๋ฏธ TMB radiomic biomarker๋ฅผ ์ด์šฉํ•œ ๋ฉด์—ญ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก, PD-L1 ๋ฐœํ˜„๋Ÿ‰ ๋ฐ ์ƒ์กด ๊ฒฐ๊ณผ์— ๋Œ€ํ•œ ์˜์ƒ ๊ธฐ๋ฐ˜ biomarker ์—ฐ๊ตฌ, ๊ทธ๋ฆฌ๊ณ  Nivolumab plus Ipilimumab ์กฐํ•ฉ์—์„œ์˜ ๊ณ  TMB ํ™˜์ž๊ตฐ ์น˜๋ฃŒ ํšจ๊ณผ ๋“ฑ ๋ถ„์ž ํ‘œ์  ๋ฐ ๋ฉด์—ญ ๋ฐ”์ด์˜ค๋งˆ์ปค์™€ ๊ด€๋ จ๋œ ๋‹ค์ˆ˜์˜ ๋…ผ๋ฌธ์„ ํฌํ•จํ•˜๊ณ  ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ EGFR ๋Œ์—ฐ๋ณ€์ด์˜ ์ง€๋ฆฌ์ /์ธ์ข…์  ๋นˆ๋„ ์ฐจ์ด์— ๋”ฐ๋ฅธ ๊ธ€๋กœ๋ฒŒ ์น˜๋ฃŒ ์ ‘๊ทผ๋ฒ•์˜ ํ‘œ์ค€ํ™” ๋ฌธ์ œ, ๋˜๋Š” NGS ๊ธฐ๋ฐ˜ ๊ฒ€์‚ฌ์˜ ๊ฒฝ์ œ์  ํƒ€๋‹น์„ฑ์— ๋Œ€ํ•œ ์ง€์—ญ๋ณ„ ๋น„๊ต ๋ถ„์„(ํŠนํžˆ ์•„์‹œ์•„ vs ์„œ๊ตฌ)์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ๋…ผ์˜๋Š” ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ๋‚ด์—์„œ ๋ช…์‹œ์ ์œผ๋กœ ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š์€ gap์œผ๋กœ ๋ณด์ธ๋‹ค.

Immune Checkpoint Inhibitors: Efficacy, Resistance, and Safety Profile

Immune Checkpoint Inhibitors (ICIs), specifically anti-CTLA-4 and anti-PD-1/PD-L1 agents, have revolutionized the treatment landscape of advanced NSCLC by harnessing the host immune system to target tumor cells. The clinical impact of ICIs is profound, with significant extensions in overall survival (OS) observed across various cancer types, including NSCLC [3]. However, the efficacy of ICIs is not universal; resistance mechanisms remain a critical barrier. Systematic pan-tumor analyses indicate that clonal Tumor Mutational Burden (TMB) is the strongest predictor of CPI response, followed by total TMB and CXCL9 expression in the tumor microenvironment [16]. Furthermore, specific genomic alterations such as 9q34 (TRAF2) loss are associated with response, while CCND1 amplification correlates with resistance [16]. The interaction between ICIs and other systemic therapies or comedications also modulates efficacy. For instance, concomitant use of statins has been associated with improved ICI effectiveness (pooled HR for OS: 0.8) due to mevalonate pathway inhibition enhancing antitumor immunity, whereas Proton Pump Inhibitors (PPIs) consistently reduce ICI efficacy (pooled HR ~1.18) likely through gut microbiome dysbiosis [28]. Corticosteroid use presents a more complex picture, where baseline high-dose usage may negatively impact outcomes, necessitating careful management of immune-related adverse events (irAEs) without compromising therapeutic benefit [28].

The safety profile of ICIs is characterized by unique immune-related adverse events (irAEs) that require distinct management strategies compared to traditional chemotherapy toxicities. Renal toxicity is a significant concern, with acute interstitial nephritis (AIN) being the primary adverse effect associated with PD-1 inhibitors, occurring later (3-10 months) than CTLA-4 inhibitor-induced renal injury (2-3 months) [1]. The incidence of renal toxicities may be higher than initially thought, ranging from 9.9% to 29%, highlighting the need for vigilant monitoring and early intervention with steroids [1]. Musculoskeletal irAEs, particularly ICI-induced arthritis, affect up to 7% of patients, with lung cancer being a common underlying malignancy (27.7% of cases in one review) [30]. These rheumatic irAEs pose diagnostic challenges and require specific therapeutic implications to manage joint symptoms without discontinuing life-prolonging immunotherapy [30]. Additionally, the timing and type of ICI administration influence toxicity; for example, adjuvant ICIs across solid tumors show benefits in recurrence-free survival but carry risks of attrition due to toxicity, necessitating careful patient selection and monitoring protocols [29].

In the context of NSCLC specifically, the integration of ICIs into treatment algorithms has expanded from metastatic settings to earlier stages. Neoadjuvant combinations of Nivolumab plus chemotherapy have demonstrated significant improvements in pathological complete response (pCR) and overall survival in resectable lung cancer [31]. Similarly, adjuvant Durvalumab after chemoradiotherapy in Stage III NSCLC has become a standard of care, with ongoing research focusing on de-escalating treatment duration to minimize toxicity while maintaining efficacy [29]. However, phenomena such as hyperprogressive disease and pseudoprogression complicate the assessment of treatment response, requiring advanced imaging techniques and biomarkers to distinguish true progression from inflammatory changes [31]. The interplay between tumor-intrinsic factors (like TMB) and microenvironmental features (such as CXCL9 expression and T-cell infiltration markers like CCR5 and CXCL13) underscores the complexity of predicting ICI sensitivity, suggesting that future strategies must combine multiple biomarkers for optimal patient stratification [16].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ด๋ฏธ Durvalumab consolidation therapy, Neoadjuvant Nivolumab plus Chemotherapy, ๊ทธ๋ฆฌ๊ณ  Hyperprogressive disease ๋ฐ Pseudoprogression๊ณผ ๊ด€๋ จ๋œ ๋…ผ๋ฌธ์„ ํฌ๊ด„์ ์œผ๋กœ ์ˆ˜์ง‘ํ•˜๊ณ  ์žˆ๋‹ค. ๋˜ํ•œ, Immune-related adverse events (irAEs)์™€ Durvalumab ํšจ๋Šฅ ๊ฐ„์˜ ์—ฐ๊ด€์„ฑ ์—ฐ๊ตฌ๋„ ํฌํ•จ๋˜์–ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ICI ์น˜๋ฃŒ ์ค‘ ๋™๋ฐ˜ ์•ฝ๋ฌผ(Statins, PPIs, Corticosteroids)์ด ๋ฉด์—ญ์น˜๋ฃŒ ํšจ๋Šฅ ๋ฐ ๋…์„ฑ์— ๋ฏธ์น˜๋Š” ๊ตฌ์ฒด์ ์ธ ์ƒํ˜ธ์ž‘์šฉ ๋ฉ”์ปค๋‹ˆ์ฆ˜๊ณผ ์ž„์ƒ์  ๊ด€๋ฆฌ ๊ฐ€์ด๋“œ๋ผ์ธ์— ๋Œ€ํ•œ ์ข…ํ•ฉ์ ์ธ ๋ถ„์„์€ ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ๋ช…์‹œ์ ์œผ๋กœ ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š์€ ๋ถ€๋ถ„์œผ๋กœ, ์ด๋Š” ์‹ค์ œ ์ž„์ƒ ํ˜„์žฅ์—์„œ์˜ polypharmacy ๊ด€๋ฆฌ์— ์ค‘์š”ํ•œ gap์ด๋‹ค.

Local Therapies: Surgery and Radiochemotherapy Strategies

Local therapies remain the cornerstone of curative intent treatment for early-stage and locally advanced NSCLC. In early-stage disease, Video-Assisted Thoracic Surgery (VATS) lobectomy has emerged as a preferred surgical approach over open lobectomy due to its comparable oncological outcomes with reduced morbidity. Meta-analyses of randomized and non-randomized studies indicate that VATS lobectomy does not differ significantly from open surgery in terms of postoperative mortality, prolonged air leak, arrhythmia, or pneumonia [23]. Crucially, VATS is associated with a reduced rate of systemic recurrence (P = .03) and improved 5-year mortality rates (P = .04), suggesting potential long-term survival benefits without compromising locoregional control [23]. For patients who are not candidates for lobectomy, sublobar resection, Stereotactic Body Radiotherapy (SBRT), and thermal ablation offer alternative local treatment options, with systematic reviews evaluating their efficacy in early-stage NSCLC management [Library Item: Sublobar resection...]. The choice of surgical extent is further guided by the presence of occult lymph node involvement, which can be quantified to predict prognosis in clinically node-negative patients [Library Item: Quantifying the rate...].

For locally advanced NSCLC (Stage III), the integration of radiotherapy and chemotherapy is critical. A landmark meta-analysis of individual patient data demonstrated that concomitant radiochemotherapy provides a significant survival benefit over sequential therapy, with a hazard ratio (HR) for overall survival of 0.84 (95% CI, 0.74 to 0.95; P = .004) [5]. This translates to an absolute survival benefit of 5.7% at 3 years and 4.5% at 5 years [5]. Concomitant treatment also significantly reduces locoregional progression (HR, 0.77; P = .01), although it does not differ from sequential therapy in preventing distant progression [5]. However, this survival advantage comes at the cost of increased acute toxicity, particularly grade 3-4 esophageal toxicity, which rises from 4% to 18% (Relative Risk 4.9) with concomitant treatment [5]. Despite this, there is no significant difference in acute pulmonary toxicity between the two approaches [5]. The timing of treatment initiation is also crucial; delays in cancer treatment are associated with increased mortality, with each four-week delay increasing the hazard ratio for overall survival significantly across various indications, emphasizing the need for efficient care pathways to minimize wait times [6].

Preoperative (neoadjuvant) chemotherapy has also been shown to improve outcomes in resectable NSCLC. A systematic review and meta-analysis of individual participant data from 15 randomized controlled trials revealed a significant survival benefit with preoperative chemotherapy, yielding an HR of 0.87 (95% CI 0.78-0.96; p=0.007) [18]. This corresponds to a 13% reduction in the relative risk of death and an absolute 5-year survival improvement from 40% to 45% [18]. The benefit was consistent regardless of the chemotherapy regimen, platinum agent used, or whether postoperative radiotherapy was administered [18]. More recently, the addition of immunotherapy to neoadjuvant chemotherapy has further enhanced pathological complete response (pCR) rates and survival outcomes, as seen in trials combining Nivolumab with chemotherapy [Library Item: Neoadjuvant Nivolumab...]. The integration of these local and systemic therapies requires precise staging and assessment of lymph node status, including the prognostic impact of micrometastasis and isolated tumor cells, which can influence treatment decisions in Stage I-IIIA NSCLC [Library Item: Systematic review... micrometastasis].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” VATS lobectomy์˜ ์•ˆ์ „์„ฑ ๋ฐ ํšจ๋Šฅ, Neoadjuvant chemotherapy ๋ฐ immunotherapy ์กฐํ•ฉ, ๊ทธ๋ฆฌ๊ณ  Lymph node involvement์™€ Micrometastasis์˜ ์˜ˆํ›„์  ์ค‘์š”์„ฑ์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์„ ์ด๋ฏธ ํฌํ•จํ•˜๊ณ  ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ Concomitant versus Sequential radiochemotherapy์˜ ๋…์„ฑ ํ”„๋กœํŒŒ์ผ(ํŠนํžˆ ์‹๋„์—ผ)๊ณผ ์ƒ์กด ์ด๋“ ๊ฐ„์˜ ์ •๋Ÿ‰์  ๋น„๊ต ๋ถ„์„, ๋˜๋Š” ์น˜๋ฃŒ ์ง€์—ฐ(delay)์ด NSCLC ํŠน์ • ํ•˜์œ„ ๊ทธ๋ฃน์—์„œ์˜ ์‚ฌ๋ง๋ฅ ์— ๋ฏธ์น˜๋Š” ๊ตฌ์ฒด์ ์ธ ์˜ํ–ฅ(HR ๊ฐ’ ๋“ฑ)์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ๋ฉ”ํƒ€๋ถ„์„ ๊ฒฐ๊ณผ๋Š” ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ๋ช…์‹œ์ ์œผ๋กœ ์ •๋ฆฌ๋˜์ง€ ์•Š์€ gap์œผ๋กœ ๋ณด์ธ๋‹ค.

Emerging Technologies: Radiomics, AI, and Liquid Biopsy

The integration of advanced imaging analytics and artificial intelligence (AI) is transforming the diagnostic and prognostic landscape of NSCLC. Radiomics, which involves the extraction of quantitative features from medical images, offers a non-invasive method to characterize tumor heterogeneity and predict treatment response. However, the clinical utility of radiomic models is contingent upon their reproducibility and repeatability. Systematic reviews highlight that radiomic features are sensitive to variations in image acquisition settings, reconstruction algorithms, preprocessing steps, and software used for feature extraction [17]. First-order features tend to be more reproducible than shape or textural metrics, with entropy being a consistently stable feature [17]. To address these challenges, deep learning-based segmentation models have been developed to improve the accuracy and efficiency of image analysis. For instance, U-Net family models and transformer-based architectures demonstrate high performance in segmenting tumor structures, with Dice scores ranging from 0.82 to 0.83 across different imaging modalities like CT and MRI [32]. These AI-driven tools are increasingly being applied to predict pathological complete response (pCR) to neoadjuvant chemoimmunotherapy, utilizing delta-radiomics features combined with hematological indices to enhance predictive accuracy [Library Item: Delta-radiomics...].

Liquid biopsy represents another frontier in NSCLC management, offering real-time molecular insights without the need for invasive tissue biopsies. Emerging platforms leverage circulating tumor DNA (ctDNA), extracellular vesicles (EVs), and other biofluids to monitor disease progression and treatment response. EVs, as natural drug delivery systems, possess unique characteristics that make them promising vehicles for targeted therapy and diagnostic biomarkers [8]. Recent reviews emphasize the expansion of liquid biopsy applications in lung cancer, including its integration with machine learning and multi-omics approaches [25]. Technologies such as label-free circulating tumor cell (CTC) isolation, DNA fragmentomics, and end-motif profiling are being explored to overcome the limitations of traditional tissue-based diagnostics [25]. Furthermore, RNA sequencing (RNA-seq) is being translated into clinical diagnostics to detect disease-associated mutations, gene expression disruptions, and non-coding RNAs, providing a comprehensive view of the tumor's molecular landscape [19]. The combination of radiomics and liquid biopsy data through AI algorithms holds potential for creating robust predictive models that can guide personalized treatment strategies.

Despite these advancements, challenges remain in standardizing these technologies for routine clinical use. The reproducibility of radiomic features across different institutions and imaging protocols needs to be rigorously validated [17]. Similarly, the sensitivity and specificity of liquid biopsy assays must be improved to ensure reliable detection of low-abundance biomarkers [25]. Economic evaluations suggest that while NGS-based testing is generally cost-effective, variations in healthcare systems and prevalence of actionable alterations can impact its feasibility [27]. Therefore, future research should focus on developing standardized protocols for image processing and liquid biopsy analysis, as well as conducting large-scale prospective trials to validate the clinical utility of these emerging technologies in NSCLC care.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ด๋ฏธ Radiomics features์˜ universality, Deep learning์„ ์ด์šฉํ•œ pCR ์˜ˆ์ธก, Immune-pathology informed radiomics model, ๊ทธ๋ฆฌ๊ณ  Delta-radiomics์™€ hematological index์˜ ์กฐํ•ฉ์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์„ ํฌํ•จํ•˜๊ณ  ์žˆ๋‹ค. ๋˜ํ•œ, Liquid biopsy์˜ ์ „๋ฐ˜์ ์ธ ๋ฐœ์ „ ๋ฐฉํ–ฅ์— ๋Œ€ํ•œ ๋…ผ์˜๋„ ์ผ๋ถ€ ์กด์žฌํ•  ์ˆ˜ ์žˆ์œผ๋‚˜, Radiomic features์˜ ์žฌํ˜„์„ฑ(reproducibility)์— ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š” ๊ตฌ์ฒด์ ์ธ ๊ธฐ์ˆ ์  ๋ณ€์ˆ˜(์ด๋ฏธ์ง€ ํš๋“ ์„ค์ •, ์ „์ฒ˜๋ฆฌ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ๋“ฑ)์— ๋Œ€ํ•œ ์ฒด๊ณ„์ ์ธ ๊ฒ€ํ†  ๊ฒฐ๊ณผ๋‚˜, EVs๋ฅผ ์ด์šฉํ•œ ์•ฝ๋ฌผ ์ „๋‹ฌ ์‹œ์Šคํ…œ์˜ ๊ตฌ์ฒด์  ๋ฉ”์ปค๋‹ˆ์ฆ˜ ๋ฐ ์ž„์ƒ ์ ์šฉ ๊ฐ€๋Šฅ์„ฑ์— ๋Œ€ํ•œ ์‹ฌ์ธต ๋ถ„์„์€ ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ๋ช…์‹œ์ ์œผ๋กœ ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š์€ gap์œผ๋กœ ๋ณด์ธ๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1Adverse Renal Effects of Immune Checkpoint Inhibitors: A Narrative ReviewAmerican Journal of Nephrology ยท 201715291PubMed ยท DOI
2The Next Generation of Platinum Drugs: Targeted Pt(II) Agents, Nanoparticle Delivery, and Pt(IV) ProdrugsChemical Reviews ยท 20162482PubMed ยท DOI
3Immune Checkpoint Inhibitors for the Treatment of Cancer: Clinical Impact and Mechanisms of Response and ResistanceAnnual Review of Pathology Mechanisms of Disease ยท 20202263PubMed ยท DOI
4The KEAP1-NRF2 System: a Thiol-Based Sensor-Effector Apparatus for Maintaining Redox HomeostasisPhysiological Reviews ยท 20181955PubMed ยท DOI
5Meta-Analysis of Concomitant Versus Sequential Radiochemotherapy in Locally Advanced Nonโ€“Small-Cell Lung CancerJournal of Clinical Oncology ยท 20101897PubMed ยท DOI
6Mortality due to cancer treatment delay: systematic review and meta-analysisBMJ ยท 20201742PubMed ยท DOI
7Advances in immunotherapy for hepatocellular carcinomaNature Reviews Gastroenterology & Hepatology ยท 20211382PubMed ยท DOI
8Extracellular vesicles as drug delivery systems: Why and how?Advanced Drug Delivery Reviews ยท 20201236PubMed ยท DOI
9Network pharmacology: curing causal mechanisms instead of treating symptomsTrends in Pharmacological Sciences ยท 20211206PubMed ยท DOI
10ESMO recommendations on microsatellite instability testing for immunotherapy in cancer, and its relationship with PD-1/PD-L1 expression and tumour mutational burden: a systematic review-based approachAnnals of Oncology ยท 20191079PubMed ยท DOI
11PrognoScan: a new database for meta-analysis of the prognostic value of genesBMC Medical Genomics ยท 20091073PubMed ยท DOI
12Prognostic Significance of Tumor-Associated Macrophages in Solid Tumor: A Meta-Analysis of the LiteraturePLoS ONE ยท 20121051PubMed ยท DOI
13Notch Signaling in Development, Tissue Homeostasis, and DiseasePhysiological Reviews ยท 20171023PubMed ยท DOI
14Prognostic value of tumor-infiltrating FoxP3+ regulatory T cells in cancers: a systematic review and meta-analysisScientific Reports ยท 20151012PubMed ยท DOI
15EGFR Mutations and Lung CancerAnnual Review of Pathology Mechanisms of Disease ยท 2011990PubMed ยท DOI
16Meta-analysis of tumor- and T cell-intrinsic mechanisms of sensitization to checkpoint inhibitionCell ยท 2021915PubMed ยท DOI
17Repeatability and Reproducibility of Radiomic Features: A Systematic ReviewInternational Journal of Radiation Oncology*Biology*Physics ยท 2018913PubMed ยท DOI
18Preoperative chemotherapy for non-small-cell lung cancer: a systematic review and meta-analysis of individual participant dataThe Lancet ยท 2014888PubMed ยท DOI62159-5)
19Translating RNA sequencing into clinical diagnostics: opportunities and challengesNature Reviews Genetics ยท 2016857PubMed ยท DOI
20The PI3K/AKT Pathway as a Target for Cancer TreatmentAnnual Review of Medicine ยท 2015850PubMed ยท DOI
21EGFR mutation incidence in non-small-cell lung cancer of adenocarcinoma histology: a systematic review and global map by ethnicity (mutMapII).PubMed ยท 2015849PubMed
22The prevalence of EGFR mutation in patients with non-small cell lung cancer: a systematic review and meta-analysisOncotarget ยท 2016811PubMed ยท DOI
23Systematic Review and Meta-Analysis of Randomized and Nonrandomized Trials on Safety and Efficacy of Video-Assisted Thoracic Surgery Lobectomy for Early-Stage Nonโ€“Small-Cell Lung CancerJournal of Clinical Oncology ยท 2009740PubMed ยท DOI
24PRISMA for Abstracts: Reporting Systematic Reviews in Journal and Conference AbstractsPLoS Medicine ยท 2013738PubMed ยท DOI
25Consolidated Evidence and New Frontiers of Liquid Biopsy in Lung Cancer: A Narrative ReviewCells ยท 20260DOI
26Artificial intelligence-based computer-aided detection systems for adenomas during colonoscopy: a systematic review and Bayesian network meta-analysisFrontiers in Artificial Intelligence ยท 20260DOI
27Economic evaluations of next-generation sequencing for targeted therapy in non-small cell lung cancer: a systematic reviewFrontiers in Public Health ยท 20260DOI
28A narrative review on immune checkpoint inhibitor drug interactions with statins, proton pump inhibitors and corticosteroidsDiscover Medicine ยท 20260DOI
29Adjuvant immune checkpoint inhibitors across solid tumors: a meta-analysis of phase II-III randomized trialsFrontiers in Oncology ยท 20260DOI
30Profiling immune check point inhibitors โ€“ Induced arthritis: A systematic review and pooled analyses of cohort studiesJournal of Autoimmunity ยท 20260PubMed ยท DOI
31Immune checkpoint inhibitors plus chemotherapy for early-stage or locally โ€“advanced triple-negative breast cancer: a systematic review and Bayesian network meta-analysis of randomized trialsFrontiers in Oncology ยท 20260DOI
32Deep learning-based segmentation of abdominal aortic aneurysm: a systematic review and meta-analysis of performance and clinical applicabilityThe Egyptian Journal of Radiology and Nuclear Medicine ยท 20260DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1Management of locally advanced non-small cell lung cancer: State of the art and future directions.Cancer communications (London, England) ยท 2024-JanPubMed ยท DOI
2Befotertinib: First Approval.Drugs ยท 2023-OctPubMed ยท DOI
3Exploring mitochondrial ribosomal protein S12 as a novel target for non-small cell lung cancer.NPJ precision oncology ยท 2025-Nov-14PubMed ยท DOI
4Circ6834 suppresses non-small cell lung cancer progression by destabilizing ANHAK and regulating miR-873-5p/TXNIP axis.Molecular cancer ยท 2024-Jun-18PubMed ยท DOI
5SLC25A40 promotes NSCLC growth by enhancing NADPH-mediated lipid synthesis and suppressing ROS accumulation-induced ferroptosis.Experimental cell research ยท 2025-Sep-01PubMed ยท DOI
6Analysis of outcomes in resected early-stage NSCLC with rare targetable driver mutations.Therapeutic advances in medical oncology ยท 2024PubMed ยท DOI
7SPOCK1 acts as a negative regulator of ferroptosis in non-small cell lung cancer.NPJ precision oncology ยท 2026-May-14PubMed ยท DOI
8Global research trends on ALK-TKIs in non-small cell lung cancer: a bibliometric analysis.Frontiers in pharmacology ยท 2025PubMed ยท DOI
9Bilateral Adrenal Hemorrhage Heralds Bronchogenic Carcinoma.Cureus ยท 2024-JanPubMed ยท DOI
10NUP155 and NDC1 interaction in NSCLC: a promising target for tumor progression.Frontiers in pharmacology ยท 2024PubMed ยท DOI
Deep learning โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 11ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
์˜๋ฃŒ ์˜์ƒ ๋ถ„์„ ๋ฐ ์ง„๋‹จ ๋ณด์กฐ (Medical Image Analysis & Diagnostic Assistance)

Deep learning์€ ์˜๋ฃŒ ์˜์ƒ ๋ถ„์„ ๋ถ„์•ผ์—์„œ ์ „ํ†ต์ ์ธ hand-crafted feature extraction ๋ฐฉ์‹์˜ ํ•œ๊ณ„๋ฅผ ๊ทน๋ณตํ•˜๊ณ , ๋ฐ์ดํ„ฐ๋กœ๋ถ€ํ„ฐ ๊ณ„์ธต์  ํŠน์ง• ํ‘œํ˜„(hierarchical feature representations)์„ ํ•™์Šตํ•จ์œผ๋กœ์จ ์ง„๋‹จ ์„ฑ๋Šฅ์„ ํ˜์‹ ์ ์œผ๋กœ ํ–ฅ์ƒ์‹œ์ผฐ์Šต๋‹ˆ๋‹ค [6]. ํŠนํžˆ ํ‰๋ถ€ X์„ (Chest radiographs)๊ณผ CT ์Šค์บ”์—์„œ ํ๊ฒฐ์ ˆ์˜ ์•…์„ฑ ํ™•๋ฅ  ์ถ”์ •, 10๊ฐ€์ง€ ์ด์ƒ์˜ ์ผ๋ฐ˜์  ์ด์ƒ ์ง•ํ›„ ๊ฐ์ง€, ๊ทธ๋ฆฌ๊ณ  ๊ฐ„์„ธํฌ์•”(HCC) ์ ˆ์ œ์ˆ  ํ›„ ์ƒ์กด ์˜ˆ์ธก ๋“ฑ ๋‹ค์–‘ํ•œ ์ž„์ƒ ์‹œ๋‚˜๋ฆฌ์˜ค์—์„œ ๋†’์€ ์ •ํ™•๋„๋ฅผ ๋‹ฌ์„ฑํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [6]. ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ํฌํ•จ๋œ ์—ฐ๊ตฌ๋“ค์€ ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์ด ๋‹จ์ผ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ๋ฅผ ๋„˜์–ด ๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ(Multimodal) ํ†ตํ•ฉ์œผ๋กœ ์ง„ํ™”ํ•˜๊ณ  ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ํ‰๋ถ€ X์„  ์˜์ƒ๊ณผ ์ž„์ƒ ๋งค๊ฐœ๋ณ€์ˆ˜๋ฅผ ๊ฒฐํ•ฉํ•œ Transformer ๊ธฐ๋ฐ˜ ๋ชจ๋ธ์€ ์ง„๋‹จ์˜ ์ •๋ฐ€๋„๋ฅผ ๋†’์˜€์œผ๋ฉฐ [6], CT ๊ธฐ๋ฐ˜ deep learning ์•Œ๊ณ ๋ฆฌ์ฆ˜์€ ๋น„์†Œ์„ธํฌํ์•”(NSCLC) ํ™˜์ž์—์„œ neoadjuvant chemoimmunotherapy์— ๋Œ€ํ•œ ์ฃผ์š” ๋ณ‘๋ฆฌํ•™์  ๋ฐ˜์‘(major pathological response)์„ ์˜ˆ์ธกํ•˜๋Š” ๋ฐ ์„ฑ๊ณต์ ์œผ๋กœ ์ ์šฉ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก์ด ๋‹จ์ˆœํ•œ ์˜์ƒ ๋ถ„์„์„ ๋„˜์–ด ์ž„์ƒ ๋ฐ์ดํ„ฐ์™€์˜ ์œตํ•ฉ์„ ํ†ตํ•ด ๋” ๊ฐ•๋ ฅํ•œ prognostic tool๋กœ ์ž๋ฆฌ ์žก๊ณ  ์žˆ์Œ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค.

๊ฐ„ ์งˆํ™˜ ์ง„๋‹จ ๋ถ„์•ผ์—์„œ๋„ deep learning์˜ ์—ญํ• ์ด ๋‘๋“œ๋Ÿฌ์ง‘๋‹ˆ๋‹ค. ๋Œ€์‚ฌ ์ด์ƒ ๊ด€๋ จ ์ง€๋ฐฉ์„ฑ ๊ฐ„์งˆํ™˜(MASLD) ๋ฐ ๋น„์•Œ์ฝ”์˜ฌ์„ฑ ์ง€๋ฐฉ๊ฐ„์—ผ(MASH)์˜ ๊ฒฝ์šฐ, ์นจ์Šต์ ์ธ liver biopsy ๋Œ€์‹  AI ๊ธฐ๋ฐ˜ ์˜์ƒ ๋ชจ๋ธ(์ดˆ์ŒํŒŒ, CT, MRI, ํƒ„์„ฑ์˜์ƒ)์ด steatosis ์ง„๋‹จ์—์„œ AUROC 0.85~0.99, fibrosis staging์—์„œ 0.82~0.97์˜ ์šฐ์ˆ˜ํ•œ ์„ฑ๋Šฅ์„ ๋ณด์˜€์Šต๋‹ˆ๋‹ค [29]. ๋˜ํ•œ ๋””์ง€ํ„ธ ๋ณ‘๋ฆฌํ•™(Digital pathology)์„ ํ†ตํ•ด liver biopsy ์ด๋ฏธ์ง€๋ฅผ deep learning์œผ๋กœ ๋ถ„์„ํ•œ ๊ฒฐ๊ณผ, ๊ฐ„๋ณ‘๋ฆฌ ์ „๋ฌธ์˜(hepatopathologist)์™€ ๋™๋“ฑํ•œ ์ง„๋‹จ ์ •ํ™•๋„๋ฅผ ๋‹ฌ์„ฑํ•˜๊ธฐ๋„ ํ–ˆ์Šต๋‹ˆ๋‹ค [29]. ์ด๋Š” ์˜์ƒ ๊ธฐ๋ฐ˜ AI๊ฐ€ ์กฐ์งํ•™์  gold standard๋ฅผ ๋Œ€์ฒดํ•˜๊ฑฐ๋‚˜ ๋ณด์™„ํ•  ์ˆ˜ ์žˆ๋Š” ์ž ์žฌ๋ ฅ์„ ๊ฐ€์ง€๊ณ  ์žˆ์Œ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค. ํŠนํžˆ HCC ์œ„ํ—˜ ๋ถ„๋ฅ˜ ๋ฐ ๊ฐ์‹œ(surveillance)์—์„œ๋„ AI ๊ธฐ์ˆ ์ด ์ค‘์š”ํ•œ ๋„๊ตฌ๋กœ ๋ถ€์ƒํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [29].

๊ทธ๋Ÿฌ๋‚˜ ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์˜ ์ž„์ƒ์  ๋ฒˆ์—ญ(clinical translation)์„ ์œ„ํ•ด์„œ๋Š” ํ‘œ์ค€ํ™”์™€ ๊ฒ€์ฆ์ด ํ•„์ˆ˜์ ์ž…๋‹ˆ๋‹ค. Radiomics Quality Score 2.0๊ณผ ๊ฐ™์€ ํ”„๋ ˆ์ž„์›Œํฌ๋Š” radiomics ์—ฐ๊ตฌ์˜ ์ค€๋น„ ์ˆ˜์ค€(readiness levels)์„ ํ‰๊ฐ€ํ•˜์—ฌ ๊ฐœ์ธ ๋งž์ถคํ˜• ์˜ํ•™(personalized medicine)์œผ๋กœ์˜ ์ดํ–‰์„ ์ด‰์ง„ํ•ฉ๋‹ˆ๋‹ค [6]. ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” ํ๊ฒฐ์ ˆ ์•…์„ฑ๋„ ์ถ”์ •, HCC ์ƒ์กด ์˜ˆ์ธก, ๊ฐ„์งˆ์„ฑ ํ์งˆํ™˜(fibrosing interstitial lung disease)์˜ ์ง„ํ–‰ ์œ„ํ—˜ ์ธ์ž ๋ถ„์„ ๋“ฑ ์ด๋ฏธ ๋‹ค์ˆ˜์˜ deep learning ๊ธฐ๋ฐ˜ ์˜์ƒ ์—ฐ๊ตฌ๊ฐ€ ํฌํ•จ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์‚ฌ์šฉ์ž๊ฐ€ ์˜๋ฃŒ ์˜์ƒ ์ง„๋‹จ ๋ฐ ์˜ˆํ›„ ์˜ˆ์ธก ๋ถ„์•ผ์— ๋Œ€ํ•œ ํฌ๊ด„์ ์ธ ์ดํ•ด๋ฅผ ๊ฐ€์ง€๊ณ  ์žˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค. ๋‹ค๋งŒ, ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” MASLD/MASH์™€ ๊ฐ™์€ ๋Œ€์‚ฌ์„ฑ ๊ฐ„์งˆํ™˜์— ํŠนํ™”๋œ ์˜์ƒ AI ์—ฐ๊ตฌ๋‚˜, ๋””์ง€ํ„ธ ๋ณ‘๋ฆฌํ•™๊ณผ deep learning์˜ ๊ฒฐํ•ฉ์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๋…ผ๋ฌธ์ด ๋ช…์‹œ์ ์œผ๋กœ ๋ณด์ด์ง€ ์•Š์•„, ์ด ๋ถ€๋ถ„์—์„œ์˜ gap์ด ์กด์žฌํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ํ†ตํ•ฉ ๋ฐ ์˜ˆ์ธก ๋ชจ๋ธ๋ง (Multimodal Integration & Predictive Modeling)

Deep learning์˜ ๋˜ ๋‹ค๋ฅธ ํ•ต์‹ฌ ์˜์—ญ์€ ๋‹ค์–‘ํ•œ ๋ฐ์ดํ„ฐ ์†Œ์Šค(์˜์ƒ, ์ž„์ƒ ๊ธฐ๋ก, ์œ ์ „์ฒด, ์ƒํ™œ์Šต๊ด€ ๋“ฑ)๋ฅผ ํ†ตํ•ฉํ•˜์—ฌ ์งˆ๋ณ‘์˜ ์ง„๋‹จ, ์˜ˆํ›„, ์น˜๋ฃŒ ๋ฐ˜์‘์„ ์˜ˆ์ธกํ•˜๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๋‹จ์ผ ๋ฐ์ดํ„ฐ ์œ ํ˜•์— ์˜์กดํ•˜๋Š” ์ „ํ†ต์ ์ธ ์ ‘๊ทผ๋ฒ•๊ณผ ๋‹ฌ๋ฆฌ, deep learning์€ ๊ตฌ์กฐํ™”๋˜์ง€ ์•Š์€ ๋ฐ์ดํ„ฐ(unstructured data)์™€ ๊ตฌ์กฐํ™”๋œ ๋ฐ์ดํ„ฐ(structured data) ๊ฐ„์˜ ๋ณต์žกํ•œ ๋น„์„ ํ˜• ๊ด€๊ณ„๋ฅผ ํฌ์ฐฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค [6]. ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” "Multimodal deep learning for integrating chest radiographs and clinical parameters: a case for transformers"๋ผ๋Š” ์ œ๋ชฉ์˜ ๋…ผ๋ฌธ์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด, ์˜์ƒ๊ณผ ์ž„์ƒ ๋ฐ์ดํ„ฐ๋ฅผ ๊ฒฐํ•ฉํ•œ Transformer ์•„ํ‚คํ…์ฒ˜์˜ ์ค‘์š”์„ฑ์„ ๊ฐ•์กฐํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ๋‹จ์ˆœํ•œ ์˜์ƒ ๋ถ„์„์„ ๋„˜์–ด ํ™˜์ž์˜ ์ „์ฒด์ ์ธ ์ž„์ƒ ๋งฅ๋ฝ์„ ๊ณ ๋ คํ•œ ์ •๋ฐ€ ์˜ˆ์ธก ๋ชจ๋ธ์„ ๊ตฌ์ถ•ํ•˜๋ ค๋Š” ์ตœ์‹  ํŠธ๋ Œ๋“œ๋ฅผ ๋ฐ˜์˜ํ•ฉ๋‹ˆ๋‹ค.

์˜ˆ์ธก ๋ชจ๋ธ์˜ ๊ฐœ๋ฐœ๊ณผ ๊ฒ€์ฆ ๊ณผ์ •์—์„œ๋Š” ๋ฐฉ๋ฒ•๋ก ์  ์—„๊ฒฉ์„ฑ์ด ์š”๊ตฌ๋ฉ๋‹ˆ๋‹ค. CHARMS(Critical Appraisal and Data Extraction for Systematic Reviews of prediction Modelling Studies) ์ฒดํฌ๋ฆฌ์ŠคํŠธ์™€ PROBAST ๋„๊ตฌ๋Š” ์˜ˆ์ธก ๋ชจ๋ธ ์—ฐ๊ตฌ์˜ ํŽธํ–ฅ ์œ„ํ—˜์„ ํ‰๊ฐ€ํ•˜๊ณ  ์ฒด๊ณ„์  ๋ฌธํ—Œ๊ณ ์ฐฐ์„ ์ˆ˜ํ–‰ํ•˜๋Š” ๋ฐ ํ•„์ˆ˜์ ์ธ ๊ธฐ์ค€์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค [23]. COVID-19 ์ง„๋‹จ ๋ฐ ์˜ˆํ›„ ์˜ˆ์ธก ๋ชจ๋ธ์— ๋Œ€ํ•œ ์ฒด๊ณ„์  ๊ฒ€ํ† ์—์„œ๋Š” 232๊ฐœ์˜ ์˜ˆ์ธก ๋ชจ๋ธ์ด ๋ถ„์„๋˜์—ˆ์œผ๋ฉฐ, ๊ฐ€์žฅ ๋นˆ๋ฒˆํ•˜๊ฒŒ ์‚ฌ์šฉ๋œ ์˜ˆ์ธก ์ธ์ž(predictors)๋Š” vital signs, age, comorbidities, ๊ทธ๋ฆฌ๊ณ  image features์˜€์Šต๋‹ˆ๋‹ค [11]. ์ด๋Š” deep learning ๊ธฐ๋ฐ˜ ์˜์ƒ ํŠน์ง•(image features)์ด ์ž„์ƒ ๋ณ€์ˆ˜์™€ ํ•จ๊ป˜ ํ†ตํ•ฉ๋  ๋•Œ ์ง„๋‹จ ๋ฐ ์˜ˆํ›„ ์˜ˆ์ธก์˜ ์ •ํ™•๋„๊ฐ€ ํฌ๊ฒŒ ํ–ฅ์ƒ๋  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ, NSCLC์—์„œ tumor mutational burden(TMB) radiomic biomarker๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๋ฉด์—ญ์น˜๋ฃŒ(immunotherapy) ๋ฐ˜์‘ ์˜ˆ์ธก์„ ์‹œ๋„ํ•œ ์—ฐ๊ตฌ๋Š” ์œ ์ „์ฒด ์ •๋ณด์™€ ์˜์ƒ ํŠน์ง•(radiomics)์„ ๊ฒฐํ•ฉํ•œ ๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ์ ‘๊ทผ๋ฒ•์˜ ์ž„์ƒ์  ์œ ์šฉ์„ฑ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค [6].

์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” "Predicting response to immunotherapy in advanced non-small-cell lung cancer using tumor mutational burden radiomic biomarker" ๋ฐ "Non-invasive prediction for pathologic complete response to neoadjuvant chemoimmunotherapy in lung cancer using CT-based deep learning"๊ณผ ๊ฐ™์€ ๋…ผ๋ฌธ๋“ค์ด ํฌํ•จ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์‚ฌ์šฉ์ž๊ฐ€ ์ด๋ฏธ ์˜์ƒ ๊ธฐ๋ฐ˜ biomarker์™€ ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ์—ฐ๊ตฌ๋ฅผ ์ˆ˜์ง‘ํ•˜๊ณ  ์žˆ์Œ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” MASLD/MASH ์ง„๋‹จ์„ ์œ„ํ•œ machine learning ์•Œ๊ณ ๋ฆฌ์ฆ˜(ํ˜ˆ์•ก ๊ฒ€์‚ฌ, ๋Œ€์‚ฌ์ฒดํ•™, ์žฅ๋‚ด ๋ฏธ์ƒ๋ฌผ๊ตฐ์ง‘ ๋ฐ์ดํ„ฐ ํ†ตํ•ฉ)์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๋…ผ๋ฌธ์ด ๋ช…์‹œ์ ์œผ๋กœ ๋ณด์ด์ง€ ์•Š์Šต๋‹ˆ๋‹ค [29]. ์ด๋Š” ๋Œ€์‚ฌ์„ฑ ๊ฐ„์งˆํ™˜ ๋ถ„์•ผ์—์„œ ๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ๋ฐ์ดํ„ฐ(์˜์ƒ+์ƒํ™”ํ•™์  ์ง€ํ‘œ+omics)๋ฅผ ํ†ตํ•ฉํ•œ ์˜ˆ์ธก ๋ชจ๋ธ ์—ฐ๊ตฌ๊ฐ€ ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ์ƒ๋Œ€์ ์œผ๋กœ ๋ถ€์กฑํ•  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค.

์ž์—ฐ์–ด ์ฒ˜๋ฆฌ ๋ฐ ์ƒ์„ฑํ˜• AI์˜ ์˜๋ฃŒ ์ ์šฉ (NLP & Generative AI in Healthcare)

์ตœ๊ทผ Deep learning ๋ถ„์•ผ์—์„œ๋Š” Transformer ์•„ํ‚คํ…์ฒ˜๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ํ•œ ๋Œ€๊ทœ๋ชจ ์–ธ์–ด ๋ชจ๋ธ(LLM, Large Language Models)๊ณผ ์ƒ์„ฑํ˜• AI(Generative AI)์˜ ๋ฐœ์ „์ด ์ฃผ๋ชฉ๋ฐ›๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [31]. ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์€ ํ…์ŠคํŠธ ์ƒ์„ฑ, ๋ฌธํ—Œ ๊ฒ€ํ† , ์ฝ”๋“œ ์ƒ์„ฑ, ๋ฐ์ดํ„ฐ ๋ถ„์„ ๋“ฑ ๋‹ค์–‘ํ•œ ์˜๋ฃŒ ์—ฐ๊ตฌ ๋ฐ ์‹ค๋ฌด ์˜์—ญ์—์„œ ํ™œ์šฉ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ChatGPT์™€ ๊ฐ™์€ LLM์€ ์˜๋ฃŒ ๊ต์œก, ์—ฐ๊ตฌ, ์‹ค๋ฌด์—์„œ ๊ณผํ•™์  ๊ธ€์“ฐ๊ธฐ ๊ฐœ์„ , ์—ฐ๊ตฌ ํ˜•ํ‰์„ฑ ์ฆ์ง„, ์›Œํฌํ”Œ๋กœ์šฐ ๊ฐ„์†Œํ™”, ๋น„์šฉ ์ ˆ๊ฐ, ๋ฌธ์„œํ™” ์ง€์›, ๊ฐœ์ธํ™”๋œ ์˜ํ•™(personalized medicine) ๊ตฌํ˜„, ๊ทธ๋ฆฌ๊ณ  health literacy ํ–ฅ์ƒ ๋“ฑ์— ๊ธฐ์—ฌํ•  ์ˆ˜ ์žˆ๋Š” ์ž ์žฌ๋ ฅ์„ ๊ฐ€์ง€๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [12]. ํŠนํžˆ, LLM์€ ๋ณต์žกํ•œ ์˜๋ฃŒ ๋ฐ์ดํ„ฐ๋ฅผ ํšจ์œจ์ ์œผ๋กœ ๋ถ„์„ํ•˜๊ณ , ๋ฌธํ—Œ ๊ฒ€ํ†  ์‹œ๊ฐ„์„ ๋‹จ์ถ•ํ•˜๋ฉฐ, ์‹คํ—˜ ์„ค๊ณ„์— ์ง‘์ค‘ํ•  ์ˆ˜ ์žˆ๋„๋ก ๋•๋Š” ๋„๊ตฌ๋กœ ๋ถ€์ƒํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [12].

๊ทธ๋Ÿฌ๋‚˜ ์ƒ์„ฑํ˜• AI์˜ ์˜๋ฃŒ ์ ์šฉ์—๋Š” ์‹ฌ๊ฐํ•œ ์šฐ๋ ค์‚ฌํ•ญ๋„ ์กด์žฌํ•ฉ๋‹ˆ๋‹ค. ChatGPT ์‚ฌ์šฉ๊ณผ ๊ด€๋ จ๋œ 60๊ฑด์˜ ์—ฐ๊ตฌ ์ค‘ 96.7%๊ฐ€ ์œค๋ฆฌ์ , ์ €์ž‘๊ถŒ, ํˆฌ๋ช…์„ฑ, ๋ฒ•์  ๋ฌธ์ œ, ํŽธํ–ฅ ์œ„ํ—˜, ํ‘œ์ ˆ, ๋…์ฐฝ์„ฑ ๋ถ€์กฑ, ํ™˜๊ฐ(hallucination)์œผ๋กœ ์ธํ•œ ๋ถ€์ •ํ™•ํ•œ ๋‚ด์šฉ, ์ง€์‹์˜ ํ•œ๊ณ„, ์ž˜๋ชป๋œ ์ธ์šฉ, ์‚ฌ์ด๋ฒ„ ๋ณด์•ˆ ๋ฌธ์ œ, ๊ทธ๋ฆฌ๊ณ  ์ •๋ณด ์—ญ๋ฅ˜(infodemics) ์œ„ํ—˜ ๋“ฑ์„ ์ง€์ ํ–ˆ์Šต๋‹ˆ๋‹ค [12]. ์ด๋Ÿฌํ•œ ์šฐ๋ ค๋Š” ์˜๋ฃŒ ๋ถ„์•ผ์—์„œ AI ๋ชจ๋ธ์˜ ์‹ ๋ขฐ์„ฑ๊ณผ ์„ค๋ช… ๊ฐ€๋Šฅ์„ฑ(explainability)์ด ์–ผ๋งˆ๋‚˜ ์ค‘์š”ํ•œ์ง€๋ฅผ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค. ํŠนํžˆ, IoT ํ™˜๊ฒฝ์—์„œ์˜ ์นจ์ž… ํƒ์ง€ ์‹œ์Šคํ…œ(IDS) ๊ฐœ๋ฐœ์—์„œ๋„ Explainable AI(XAI) ์ ‘๊ทผ๋ฒ•์ด ๋„์ž…๋˜์–ด, ๋ชจ๋ธ์˜ ์˜์‚ฌ๊ฒฐ์ • ๊ณผ์ •์„ ํˆฌ๋ช…ํ•˜๊ฒŒ ๋งŒ๋“ค๊ณ  ์‹ ๋ขฐ์„ฑ์„ ๋†’์ด๋Š” ๋…ธ๋ ฅ์ด ์ด๋ฃจ์–ด์ง€๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [26]. ์ด๋Š” ์˜๋ฃŒ AI ์—ญ์‹œ '๋ธ”๋ž™๋ฐ•์Šค' ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ณ  ์ž„์ƒ ์˜์‚ฌ์—๊ฒŒ ์„ค๋ช… ๊ฐ€๋Šฅํ•œ ๊ฒฐ๊ณผ๋ฅผ ์ œ๊ณตํ•ด์•ผ ํ•จ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค.

์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” ๋ช…์‹œ์ ์œผ๋กœ NLP๋‚˜ ์ƒ์„ฑํ˜• AI ๊ด€๋ จ ๋…ผ๋ฌธ์ด ๋ณด์ด์ง€ ์•Š์Šต๋‹ˆ๋‹ค. ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ฃผ๋กœ ์˜์ƒ ๋ถ„์„(Diagnostic imaging)๊ณผ ์˜ˆ์ธก ๋ชจ๋ธ(Predictive modeling)์— ์ง‘์ค‘๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ๋”ฐ๋ผ์„œ, ์˜๋ฃŒ ๊ธฐ๋ก์˜ ์ž์—ฐ์–ด ์ฒ˜๋ฆฌ, ์ž„์ƒ ๋…ธํŠธ ์ž๋™ ์š”์•ฝ, ํ™˜์ž-์˜์‚ฌ ์†Œํ†ต ๋ณด์กฐ, ๋˜๋Š” ์—ฐ๊ตฌ ๋ฌธํ—Œ์˜ ์ž๋™ ๋ถ„์„์„ ์œ„ํ•œ LLM/Generative AI ์ ์šฉ ์‚ฌ๋ก€๋Š” ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ์•„์ง ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š์€ gap์œผ๋กœ ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ํ–ฅํ›„ ์˜๋ฃŒ AI ์—ฐ๊ตฌ๊ฐ€ ์˜์ƒ ์ค‘์‹ฌ์—์„œ ํ…์ŠคํŠธ ๋ฐ ๋Œ€ํ™”ํ˜• AI๋กœ ํ™•์žฅ๋˜๋Š” ์ถ”์„ธ๋ฅผ ๋ฐ˜์˜ํ•˜์ง€ ๋ชปํ•˜๊ณ  ์žˆ์„ ์ˆ˜ ์žˆ์Œ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.

๊ธฐ์ˆ ์  ๊ธฐ๋ฐ˜ ๋ฐ ๋ฐฉ๋ฒ•๋ก ์  ์ง„ํ™” (Technical Foundations & Methodological Evolution)

Deep learning์˜ ์„ฑ๊ณต์€ ๊ทผ๋ณธ์ ์œผ๋กœ NumPy์™€ ๊ฐ™์€ ๋ฐฐ์—ด ํ”„๋กœ๊ทธ๋ž˜๋ฐ(Array programming) ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์™€ ๊ฐ™์€ ๊ฐ•๋ ฅํ•œ ๊ณ„์‚ฐ ์ธํ”„๋ผ์— ์˜์กดํ•ฉ๋‹ˆ๋‹ค [1]. NumPy๋Š” Python ์ƒํƒœ๊ณ„์˜ ํ•ต์‹ฌ์œผ๋กœ, ๋ฌผ๋ฆฌํ•™, ํ™”ํ•™, ์ฒœ๋ฌธํ•™, ์ƒ๋ฌผํ•™ ๋“ฑ ๋‹ค์–‘ํ•œ ๊ณผํ•™ ๋ถ„์•ผ์—์„œ ๋ฐ์ดํ„ฐ ์กฐ์ž‘ ๋ฐ ๋ถ„์„์˜ ํ‘œ์ค€์ด ๋˜์—ˆ์œผ๋ฉฐ, deep learning ํ”„๋ ˆ์ž„์›Œํฌ(PyTorch, TensorFlow ๋“ฑ)์˜ ๊ธฐ๋ฐ˜์ด ๋ฉ๋‹ˆ๋‹ค [1]. ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์  ํ† ๋Œ€ ์œ„์—์„œ, deep learning ์•Œ๊ณ ๋ฆฌ์ฆ˜์€ ์›๊ฒฉ ๊ฐ์ง€(Remote sensing), ์œ ์ฒด ์—ญํ•™(Fluid mechanics), ์‹ฌ์ง€์–ด ํ† ์งˆ ๊ณตํ•™(Soil stabilization)๊นŒ์ง€ ๋‹ค์–‘ํ•œ ๋ถ„์•ผ์— ์ ์šฉ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [13, 19, 28]. ์ด๋Š” deep learning์ด ํŠน์ • ๋„๋ฉ”์ธ์— ๊ตญํ•œ๋˜์ง€ ์•Š๊ณ , ๋ฐ์ดํ„ฐ ํŒจํ„ด ์ธ์‹์˜ ๋ฒ”์šฉ ๋„๊ตฌ๋กœ ์ง„ํ™”ํ•˜๊ณ  ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค.

์˜๋ฃŒ ๋ถ„์•ผ์—์„œ๋„ ์ด๋Ÿฌํ•œ ๋ฐฉ๋ฒ•๋ก ์  ์ง„ํ™”๊ฐ€ ๋‘๋“œ๋Ÿฌ์ง‘๋‹ˆ๋‹ค. ์ดˆ๊ธฐ์˜ convolutional neural networks(CNNs)์—์„œ๋ถ€ํ„ฐ ์ตœ๊ทผ์˜ Transformer ์•„ํ‚คํ…์ฒ˜๊นŒ์ง€, ๋ชจ๋ธ์˜ ๋ณต์žก์„ฑ๊ณผ ์„ฑ๋Šฅ์€ ์ง€์†์ ์œผ๋กœ ํ–ฅ์ƒ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [6, 31]. ํŠนํžˆ, "Multimodal deep learning for integrating chest radiographs and clinical parameters: a case for transformers"๋ผ๋Š” ๋…ผ๋ฌธ ์ œ๋ชฉ์—์„œ ์•Œ ์ˆ˜ ์žˆ๋“ฏ์ด, Transformer๋Š” ์˜์ƒ๊ณผ ํ…์ŠคํŠธ/๊ตฌ์กฐํ™” ๋ฐ์ดํ„ฐ๋ฅผ ํ†ตํ•ฉํ•˜๋Š” ๋ฐ ์žˆ์–ด ์ƒˆ๋กœ์šด ํ‘œ์ค€์œผ๋กœ ์ž๋ฆฌ ์žก๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ, Cryo-electron microscopy์™€ ๊ฐ™์€ ๊ณ ํ•ด์ƒ๋„ ์ด๋ฏธ์ง• ๊ธฐ์ˆ ์˜ ๋ฐœ์ „์€ ๋ถ„์ž ์ˆ˜์ค€์˜ ๊ตฌ์กฐ ๋ถ„์„์„ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•˜์—ฌ, drug discovery ๋ฐ ๋ณ‘๋ฆฌ ๊ธฐ์ „ ์ดํ•ด์— ๊ธฐ์—ฌํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค [20].

์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” "Radiomics Quality Score 2.0: towards radiomics readiness levels and clinical translation for personalized medicine"๊ณผ ๊ฐ™์€ ๋ฐฉ๋ฒ•๋ก ์  ํ‘œ์ค€ํ™” ์—ฐ๊ตฌ๊ฐ€ ํฌํ•จ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์‚ฌ์šฉ์ž๊ฐ€ deep learning ๋ชจ๋ธ์˜ ์ž„์ƒ ์ ์šฉ์„ ์œ„ํ•œ ํ’ˆ์งˆ ๊ด€๋ฆฌ ๋ฐ ๊ฒ€์ฆ ํ”„๋กœ์„ธ์Šค์— ๊ด€์‹ฌ์ด ์žˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” NumPy๋‚˜ Python ์ƒํƒœ๊ณ„ ์ž์ฒด์— ๋Œ€ํ•œ ๊ธฐ์ˆ ์  ๊ธฐ๋ฐ˜ ๋…ผ์˜, ๋˜๋Š” Transformer ์•„ํ‚คํ…์ฒ˜์˜ ๊ตฌ์ฒด์ ์ธ ๋ฐฉ๋ฒ•๋ก ์  ๋น„๊ต ์—ฐ๊ตฌ๊ฐ€ ๋ช…์‹œ์ ์œผ๋กœ ๋ณด์ด์ง€ ์•Š์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ, IoT ๋ณด์•ˆ์ด๋‚˜ Explainable AI(XAI)์™€ ๊ฐ™์€ ๋ถ€์ˆ˜์ ์ธ ๊ธฐ์ˆ ์  ์ด์Šˆ(์˜ˆ: [26]์—์„œ ์–ธ๊ธ‰๋œ XAI-based IDS)๋„ ์˜๋ฃŒ AI ๋งฅ๋ฝ์—์„œ ์ถฉ๋ถ„ํžˆ ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š์€ ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. ์ด๋Š” ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๊ฐ€ ์ฃผ๋กœ '์‘์šฉ ๊ฒฐ๊ณผ'์— ์ง‘์ค‘๋˜์–ด ์žˆ์œผ๋ฉฐ, '๊ธฐ์ˆ ์  ๊ธฐ๋ฐ˜ ๋ฐ ๋ฐฉ๋ฒ•๋ก ์  ์„ธ๋ถ€ ์‚ฌํ•ญ'์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ๋…ผ์˜๊ฐ€ ์ƒ๋Œ€์ ์œผ๋กœ ๋ถ€์กฑํ•  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1Array programming with NumPyTUScholarShare (Temple University) ยท 202018813-
2Systematic review of research on artificial intelligence applications in higher education โ€“ where are the educators?International Journal of Educational Technology in Higher Education ยท 20196565DOI
3Nitric Oxide and Peroxynitrite in Health and DiseasePhysiological Reviews ยท 20076355PubMed ยท DOI
4Health literacy and public health: A systematic review and integration of definitions and modelsBMC Public Health ยท 20126293PubMed ยท DOI
5Natural products in drug discovery: advances and opportunitiesNature Reviews Drug Discovery ยท 20215340PubMed ยท DOI
6Deep Learning in Medical Image AnalysisAnnual Review of Biomedical Engineering ยท 20174929PubMed ยท DOI
7Where Is the Semantic System? A Critical Review and Meta-Analysis of 120 Functional Neuroimaging StudiesCerebral Cortex ยท 20094239PubMed ยท DOI
8Head and neck squamous cell carcinomaNature Reviews Disease Primers ยท 20204201PubMed ยท DOI
9Emerging threats and persistent conservation challenges for freshwater biodiversityBiological reviews/Biological reviews of the Cambridge Philosophical Society ยท 20183624PubMed ยท DOI
10A systematic review of immersive virtual reality applications for higher education: Design elements, lessons learned, and research agendaComputers & Education ยท 20193461DOI
11Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisalBMJ ยท 20203302PubMed ยท DOI
12ChatGPT Utility in Healthcare Education, Research, and Practice: Systematic Review on the Promising Perspectives and Valid ConcernsHealthcare ยท 20232906PubMed ยท DOI
13Machine Learning for Fluid MechanicsAnnual Review of Fluid Mechanics ยท 20192813DOI
14ExosomesAnnual Review of Biochemistry ยท 20192797PubMed ยท DOI
15Cholangiocarcinoma 2020: the next horizon in mechanisms and managementNature Reviews Gastroenterology & Hepatology ยท 20202543PubMed ยท DOI
16A systematic literature review of blockchain-based applications: Current status, classification and open issuesTelematics and Informatics ยท 20182536DOI
17The Social Determinants of Health: Coming of AgeAnnual Review of Public Health ยท 20112475PubMed ยท DOI
18A Systematic Review of the Literature on Digital Transformation: Insights and Implications for Strategy and Organizational ChangeJournal of Management Studies ยท 20202427DOI
19Deep learning in remote sensing applications: A meta-analysis and reviewISPRS Journal of Photogrammetry and Remote Sensing ยท 20192417DOI
20Cryo-electron microscopy of vitrified specimensQuarterly Reviews of Biophysics ยท 19882401PubMed ยท DOI
21A Systematic Review of the Prevalence of SchizophreniaPLoS Medicine ยท 20052202PubMed ยท DOI
22Work Group DiversityAnnual Review of Psychology ยท 20062195PubMed ยท DOI
23Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies: The CHARMS ChecklistPLoS Medicine ยท 20142062PubMed ยท DOI
24The KEAP1-NRF2 System: a Thiol-Based Sensor-Effector Apparatus for Maintaining Redox HomeostasisPhysiological Reviews ยท 20181955PubMed ยท DOI
25Executive function in adolescents with obesity: A systematic review and meta-analysisMPG.PuRe (Max Planck Society) ยท 20260-
26Explainable AI-Based Intrusion Detection Systems for IoT Environments: A Systematic Literature ReviewSensors ยท 20260DOI
27Geometry learning trajectories: A systematic review and the development of the GEO-INTECH frameworkSocial Sciences & Humanities Open ยท 20260DOI
28A systematic review on sustainable soil stabilization using machine learning and deep learningDiscover Geoscience ยท 20260DOI
29AI In Diagnosing Metabolic Dysfunction- Associated Steatotic Liver Disease (Masld): A Narrative ReviewZenodo (CERN European Organization for Nuclear Research) ยท 20260DOI
30The triple burden of malnutrition and the role of machine learning: A systematic reviewTechnology and Health Care ยท 20260PubMed ยท DOI
31Advanced deep learning methods for text generation in 2022-2024 literature: a systematic review20260DOI
32AI In Diagnosing Metabolic Dysfunction- Associated Steatotic Liver Disease (Masld): A Narrative ReviewZenodo (CERN European Organization for Nuclear Research) ยท 20260DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1Deep learning in preclinical antibody drug discovery and development.Methods (San Diego, Calif.) ยท 2023-OctPubMed ยท DOI
2Deep learning methods in metagenomics: a review.Microbial genomics ยท 2024-AprPubMed ยท DOI
3Molecular geometric deep learning.Cell reports methods ยท 2023-Nov-20PubMed ยท DOI
4Antibody design using deep learning: from sequence and structure design to affinity maturation.Briefings in bioinformatics ยท 2024-May-23PubMed ยท DOI
5Geometric deep learning methods and applications in 3D structure-based drug design.Drug discovery today ยท 2024-JulPubMed ยท DOI
6Prediction of drug-target binding affinity based on deep learning models.Computers in biology and medicine ยท 2024-MayPubMed ยท DOI
7Simplifying protein engineering with deep learning.Cell ยท 2025-Aug-21PubMed ยท DOI
8Deep learning for platelet transfusion.Blood ยท 2023-Dec-28PubMed ยท DOI
9Deep Learning-Based HLA Allele Imputation Applicable to GWAS.Methods in molecular biology (Clifton, N.J.) ยท 2024PubMed ยท DOI
10Artificial neural networks and deep learning.American journal of orthodontics and dentofacial orthopedics : official publication of the American Association of Orthodontists, its constituent societies, and the American Board of Orthodontics ยท 2024-FebPubMed ยท DOI
tumor microenvironment โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 6ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
๋ฉด์—ญ ์กฐ์ ˆ ๋ฐ ๋ฉด์—ญ๊ด€๋ฌธ ์–ต์ œ์ œ ์น˜๋ฃŒ์˜ ์—ญํ•™

์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(Tumor Microenvironment, TME)์—์„œ ๊ฐ€์žฅ ํ™œ๋ฐœํžˆ ์—ฐ๊ตฌ๋˜๋Š” ๋ถ„์•ผ๋Š” ์ข…์–‘ ์„ธํฌ๊ฐ€ ๊ตฌ์ถ•ํ•œ ๋ฉด์–ต์ œ ๋„คํŠธ์›Œํฌ์™€ ์ด๋ฅผ ํƒ€ํŒŒํ•˜๊ธฐ ์œ„ํ•œ ๋ฉด์—ญ๊ด€๋ฌธ ์–ต์ œ์ œ(Immune Checkpoint Inhibitors, ICIs)์˜ ์ž„์ƒ์  ์ ์šฉ ๋ฐ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ์ด๋‹ค. ICIs๋Š” CTLA-4, PD-1, PD-L1 ๋“ฑ์„ ํ‘œ์ ์œผ๋กœ ํ•˜์—ฌ ํ™˜์ž์˜ ๊ณ ์œ  ๋ฉด์—ญ ์‹œ์Šคํ…œ์„ ์ข…์–‘์— ๋Œ€ํ•ญํ•˜๋„๋ก ํ™œ์„ฑํ™”์‹œํ‚ด์œผ๋กœ์จ melanoma๋ฅผ ๋น„๋กฏํ•œ ๋‹ค์–‘ํ•œ ์•…์„ฑ ์ข…์–‘์˜ ์น˜๋ฃŒ ํŒจ๋Ÿฌ๋‹ค์ž„์„ ๊ทผ๋ณธ์ ์œผ๋กœ ๋ณ€ํ™”์‹œ์ผฐ๋‹ค [1][14]. ํŠนํžˆ anti-PD-1/PD-L1 ํ•ญ์ฒด๋Š” ์žฌ๋ฐœ์„ฑ ๋˜๋Š” ์ „์ด์„ฑ head and neck squamous cell carcinoma (HNSCC) ๋ฐ ์œ„์•”/์œ„์‹๋„์ ‘ํ•ฉ๋ถ€ ์•”(GEJ adenocarcinoma)์—์„œ ํ™”ํ•™์š”๋ฒ•๊ณผ์˜ ๋ณ‘์šฉ ์š”๋ฒ•์œผ๋กœ ์Šน์ธ๋˜์—ˆ์œผ๋ฉฐ, ์ด๋Š” ๊ฐ๊ด€์  ๋ฐ˜์‘๋ฅ (Objective Response Rate, ORR)์„ ์œ ์˜๋ฏธํ•˜๊ฒŒ ํ–ฅ์ƒ์‹œํ‚ค๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ๋‹ค [4][29]. ์ตœ๊ทผ meta-analysis์— ๋”ฐ๋ฅด๋ฉด, advanced gastric/GEJ cancer ํ™˜์ž 6,517๋ช…์„ ๋Œ€์ƒ์œผ๋กœ ํ•œ ๋ถ„์„์—์„œ PD-1/PD-L1 ์–ต์ œ์ œ์™€ ํ™”ํ•™์š”๋ฒ•์˜ ๋ณ‘์šฉ์€ ์ „์ฒด ์ƒ์กด์œจ(Overall Survival, OS)์˜ Hazard Ratio(HR)์„ 0.79 (95% CI: 0.75โ€“0.85)๋กœ ๋‚ฎ์ถ”์—ˆ์œผ๋ฉฐ, ์ง„ํ–‰ ๋ฌด์ƒ์กด ๊ธฐ๊ฐ„(Progression-Free Survival, PFS)์—์„œ๋„ HR 0.78์˜ ์ด์ ์„ ๋ณด์˜€๋‹ค [29]. ๋˜ํ•œ ์•„์‹œ์•„ไบบ็พค ์ค‘์‹ฌ ์—ฐ๊ตฌ์—์„œ๋Š” PFS ๊ฐœ์„  ํšจ๊ณผ๊ฐ€ ๋” ๋‘๋“œ๋Ÿฌ์ง„ ๊ฒƒ์œผ๋กœ ๊ด€์ฐฐ๋˜์—ˆ๋‹ค(HR 0.66 vs 0.78; P=0.02) [29].

๊ทธ๋Ÿฌ๋‚˜ ICI ์น˜๋ฃŒ์˜ ์„ฑ๊ณต์€ TME ๋‚ด ๋ฉด์–ต์ œ ๊ธฐ์ „์˜ ๋ณต์žก์„ฑ์— ์˜ํ•ด ์ œํ•œ๋ฐ›๋Š”๋‹ค. ์ข…์–‘ ์„ธํฌ๋Š” ํ•ญ์› ์ œ์‹œ ๊ธฐ๋Šฅ ์ €ํ•˜, ์Œ์„ฑ ๊ณต๋™์ž๊ทน ์‹ ํ˜ธ ํ™œ์„ฑํ™”, ๊ทธ๋ฆฌ๊ณ  regulatory T cells (Tregs), immature dendritic cells ๋“ฑ์˜ ๋ฉด์—ญ์กฐ์ ˆ ์„ธํฌ๊ตฐ์„ ๋™์›ํ•˜์—ฌ ๊ด€์šฉ์„ฑ ๋ฏธ์„ธํ™˜๊ฒฝ์„ ์กฐ์„ฑํ•œ๋‹ค [18]. ์ด๋Ÿฌํ•œ ๋ฉด์–ต์ œ ์ƒํƒœ๋Š” chemokine ์‹œ์Šคํ…œ์„ ํ†ตํ•œ ๋ฉด์—ญ์„ธํฌ์˜ ์œ„์น˜ ๊ฒฐ์ • ๋ฐ ์ด๋™์„ ๋ฐฉํ•ดํ•˜๋ฉฐ, ๊ฒฐ๊ณผ์ ์œผ๋กœ ํšจ๊ณผ์ ์ธ ํ•ญ์ข…์–‘ ๋ฉด์—ญ ๋ฐ˜์‘์ด ์–ต์ œ๋œ๋‹ค [15]. ํฅ๋ฏธ๋กญ๊ฒŒ๋„ ํŠน์ • ์œ ์ „์  ๋ณ€์ด๋Š” ICI ๋ฐ˜์‘์„ฑ๊ณผ ๋ฐ€์ ‘ํ•œ ๊ด€๋ จ์ด ์žˆ๋‹ค. metastatic urothelial carcinoma (mUC)์—์„œ FGFR3 alteration์€ ICI ์น˜๋ฃŒ์— ๋Œ€ํ•œ ๋‚ด์„ฑ ์ง€ํ‘œ๋กœ ์ž‘์šฉํ•  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•œ๋‹ค. meta-analysis ๊ฒฐ๊ณผ, FGFR3 ๋ณ€์ด๊ฐ€ ์žˆ๋Š” ํ™˜์ž๊ตฐ์€ ์งˆ๋ณ‘ ์กฐ์ ˆ๋ฅ (Disease Control Rate, DCR)์ด ๋‚ฎ์•˜์œผ๋ฉฐ(RR 0.74), OS(HR 1.25)์™€ PFS(HR 1.63)๊ฐ€ ๋ชจ๋‘ ๋‹จ์ถ•๋˜๋Š” ๊ฒƒ์œผ๋กœ ํ™•์ธ๋˜์—ˆ๋‹ค [32]. ์ด๋Š” FGFR3 ๊ฒฝ๋กœ๊ฐ€ TME์˜ ๋ฉด์—ญ ์ƒํƒœ๋ฅผ ์กฐ์ ˆํ•˜์—ฌ ICI ํšจ๋Šฅ์„ ์ €ํ•ดํ•  ๊ฐ€๋Šฅ์„ฑ์„ ์‹œ์‚ฌํ•œ๋‹ค.

ICI ์น˜๋ฃŒ์˜ ๋˜ ๋‹ค๋ฅธ ์ค‘์š”ํ•œ ์ธก๋ฉด์€ ๋ฉด์—ญ๊ด€๋ จ ๋ถ€์ž‘์šฉ(irAEs), ํŠนํžˆ ์‹ ์žฅ ๋…์„ฑ์ด๋‹ค. ์ดˆ๊ธฐ์—๋Š” ๋“œ๋ฌธ ๊ฒƒ์œผ๋กœ ์—ฌ๊ฒจ์กŒ์œผ๋‚˜, ์ตœ๊ทผ ์—ฐ๊ตฌ๋“ค์€ ipilimumab (anti-CTLA-4) ๋ฐ pembrolizumab/nivolumab (anti-PD-1)์— ์˜ํ•œ ์‹ ์žฅ ์†์ƒ์˜ ๋ฐœ์ƒ๋ฅ ์ด 9.9%์—์„œ 29%๊นŒ์ง€ ๋†’์„ ์ˆ˜ ์žˆ์Œ์„ ์ง€์ ํ•œ๋‹ค [1]. CTLA-4 ์–ต์ œ์ œ ๊ด€๋ จ ์‹ ์žฅ ์†์ƒ์€ ์น˜๋ฃŒ ์‹œ์ž‘ ํ›„ 2-3๊ฐœ์›” ๋‚ด์— ์กฐ๊ธฐ์— ๋ฐœ์ƒํ•˜๋Š” ๋ฐ˜๋ฉด, PD-1 ์–ต์ œ์ œ ๊ด€๋ จ ์†์ƒ์€ 3-10๊ฐœ์›” ํ›„์— ๋Šฆ๊ฒŒ ๋‚˜ํƒ€๋‚˜๋Š” ๊ฒฝํ–ฅ์ด ์žˆ๋‹ค [1]. ์ฃผ์š” ๋ณ‘๋ฆฌํ•™์  ์†Œ๊ฒฌ์€ ๊ธ‰์„ฑ ๊ฐ„์งˆ์„ฑ ์‹ ์—ผ(Acute Interstitial Nephritis, AIN)์ด๋ฉฐ, podocytopathy๋‚˜ hyponatremia๋„ ๊ด€์ฐฐ๋œ๋‹ค [1]. ์ด๋Ÿฌํ•œ ๋ถ€์ž‘์šฉ์˜ ๊ด€๋ฆฌ์—๋Š” ์Šคํ…Œ๋กœ์ด๋“œ ์น˜๋ฃŒ๊ฐ€ ํšจ๊ณผ์ ์ž„์ด ๋ณด๊ณ ๋˜์—ˆ๋‹ค [1]. ๋˜ํ•œ, cytokine storm๊ณผ ๊ฐ™์€ ๊ณผ๋„ํ•œ ์—ผ์ฆ ๋ฐ˜์‘์€ ๊ฐ์—ผ์„ฑ ์งˆํ™˜๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ๋ฉด์—ญ์น˜๋ฃŒ ์ค‘์—๋„ ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ๋Š” ์œ„ํ—˜ ์š”์†Œ๋กœ, ์‚ฌ์ดํ† ์นด์ธ ์‹ ํ˜ธ ์ „๋‹ฌ์˜ ์ค‘๋ณต์„ฑ๊ณผ ๋ณต์žก์„ฑ์ด ์ž„์ƒ์  ๊ฒฐ๊ณผ๋ฅผ ์˜ˆ์ธกํ•˜๊ธฐ ์–ด๋ ต๊ฒŒ ๋งŒ๋“ ๋‹ค [16].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ด๋ฏธ NSCLC์—์„œ์˜ radiomics์™€ immune profile ์ƒ๊ด€๊ด€๊ณ„, PD-L1/PD-L2 ๋ฐœํ˜„๊ณผ TILs/TAMs์˜ ์—ฐ๊ด€์„ฑ, ๊ทธ๋ฆฌ๊ณ  neoadjuvant immunotherapy์˜ ํšจ๋Šฅ ์˜ˆ์ธก์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์„ ํฌํ•จํ•˜๊ณ  ์žˆ์–ด, ICI ์น˜๋ฃŒ์˜ ์ง„๋‹จ์  biomarker ๋ฐ ๋ชจ๋‹ˆํ„ฐ๋ง ์ธก๋ฉด์—์„œ ๊ฐ•๋ ฅํ•œ ๊ธฐ๋ฐ˜์„ ๊ฐ–์ถ”๊ณ  ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ FGFR3์™€ ๊ฐ™์€ ํŠน์ • ์œ ์ „์ž ๋ณ€์ด๊ฐ€ ICI ๋ฐ˜์‘์„ฑ์— ๋ฏธ์น˜๋Š” ์˜ํ–ฅ, ๋˜๋Š” CTLA-4์™€ PD-1 ์–ต์ œ์ œ ๊ฐ„์˜ ์‹ ์žฅ ๋…์„ฑ ๋ฐœ์ƒ ์‹œ๊ธฐ์™€ ๋ณ‘๋ฆฌํ•™์  ์ฐจ์ด์™€ ๊ฐ™์€ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ์˜ ์„ธ๋ถ„ํ™”๋œ ๋ฐ์ดํ„ฐ๋Š” ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ช…์‹œ์ ์œผ๋กœ ํฌํ•จ๋˜์–ด ์žˆ์ง€ ์•Š์•„, ํ–ฅํ›„ ์น˜๋ฃŒ ์•ˆ์ „์„ฑ ํ”„๋กœํŒŒ์ผ๋ง์„ ์œ„ํ•œ ์ถ”๊ฐ€์ ์ธ ๋ฌธํ—Œ ์ˆ˜์ง‘์ด ํ•„์š”ํ•˜๋‹ค.

์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ์˜ ๊ตฌ์กฐ์  ๊ตฌ์„ฑ ์š”์†Œ: ์„ฌ์œ ์•„์„ธํฌ, ํ˜ˆ๊ด€ ๋ฐ ์„ธํฌ ๊ฐ„ ํ†ต์‹ 

TME๋Š” ๋‹จ์ˆœํžˆ ์ข…์–‘ ์„ธํฌ์˜ ์ง‘ํ•ฉ์ฒด๊ฐ€ ์•„๋‹ˆ๋ผ, cancer-associated fibroblasts (CAFs), ๋น„์ •์ƒ์ ์ธ ํ˜ˆ๊ด€ ๋„คํŠธ์›Œํฌ, ๊ทธ๋ฆฌ๊ณ  extracellular matrix (ECM) ์žฌ๊ตฌ์„ฑ ๋“ฑ ๋‹ค์–‘ํ•œ ๊ตฌ์กฐ์  ์š”์†Œ๋“ค์ด ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š” ์—ญ๋™์ ์ธ ์ƒํƒœ๊ณ„์ด๋‹ค. CAFs๋Š” TME์˜ ํ•ต์‹ฌ ๊ตฌ์„ฑ ์š”์†Œ๋กœ์„œ ECM ์นจ์ฐฉ ๋ฐ ์žฌํ˜•์„ฑ, ์ข…์–‘ ์„ธํฌ์™€์˜ ์‹ ํ˜ธ ๊ตํ™˜, ๊ทธ๋ฆฌ๊ณ  ์นจ์œค ๋ฐฑํ˜ˆ๊ตฌ์™€์˜ ๊ต์‹ (crosstalk)์„ ํ†ตํ•ด ์ข…์–‘ ์ง„ํ–‰์„ ์ด‰์ง„ํ•œ๋‹ค [5]. ๊ทธ๋Ÿฌ๋‚˜ CAFs์˜ ๊ธฐ์›๊ณผ ๊ธฐ๋Šฅ์  ์ด์งˆ์„ฑ(heterogeneity)์€ ์น˜๋ฃŒ ํ‘œ์ ์œผ๋กœ ์‚ผ๊ธฐ ์œ„ํ•œ ์ฃผ์š” ์žฅ์• ๋ฌผ์ด๋‹ค. ์ผ๋ถ€ CAFs๋Š” ํ•ญ์ข…์–‘ ํŠน์„ฑ์„ ์œ ์ง€ํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ, ๋ชจ๋“  CAFs๋ฅผ ์ œ๊ฑฐํ•˜๋Š” ๊ฒƒ์ด ํ•ญ์ƒ ์œ ๋ฆฌํ•œ ๊ฒƒ์€ ์•„๋‹ˆ๋ฉฐ, ์„ ํƒ์ ์ธ ์กฐ์ ˆ ์ „๋žต์ด ํ•„์š”ํ•˜๋‹ค [5].

ํ˜ˆ๊ด€ ์ •์ƒํ™”(Vascular Normalization) ๊ฐœ๋…์€ TME์˜ ๊ตฌ์กฐ์  ์ด์ƒ์„ ๊ต์ •ํ•˜์—ฌ ์น˜๋ฃŒ ํšจ๋Šฅ์„ ๋†’์ด๋Š” ์ ‘๊ทผ๋ฒ•์ด๋‹ค. ์ข…์–‘ ๋‚ด ํ˜ˆ๊ด€์€ VEGF ๋“ฑ์˜ ํ”„๋กœ-ํ˜ˆ๊ด€์ƒ์„ฑ ์‹ ํ˜ธ์™€ ํ•ญ-ํ˜ˆ๊ด€์ƒ์„ฑ ์‹ ํ˜ธ์˜ ๋ถˆ๊ท ํ˜•์œผ๋กœ ์ธํ•ด ํ™•์žฅ๋˜๊ณ , ๊ตฌ๋ถˆ๊ตฌ๋ถˆํ•˜๋ฉฐ, ๊ณผ๋‹ค ํˆฌ๊ณผ์„ฑ์„ ๋ณด์ธ๋‹ค [22]. ์ด๋Ÿฌํ•œ ๋น„์ •์ƒ์ ์ธ ํ˜ˆ๊ด€์€ ์ข…์–‘ ๋‚ด๋ถ€์˜ ํ˜ˆ๋ฅ˜์™€ ์‚ฐ์†Œ ๊ณต๊ธ‰์˜ ์‹œ๊ฐ„์ ยท๊ณต๊ฐ„์  ์ด์งˆ์„ฑ์„ ์ดˆ๋ž˜ํ•˜๊ณ , ์ข…์–‘ ๊ฐ„์งˆ์•ก ์••๋ ฅ์„ ์ฆ๊ฐ€์‹œ์ผœ ํ™”ํ•™์š”๋ฒ•, ๋ฐฉ์‚ฌ์„  ์š”๋ฒ• ๋ฐ ๋ฉด์—ญ์š”๋ฒ•์˜ ํšจ๋Šฅ์„ ์ €ํ•˜์‹œํ‚จ๋‹ค [22]. anti-VEGF ์š”๋ฒ•์€ ๋‹จ์ˆœํžˆ ์ข…์–‘์„ '๊ตถ๊ฒจ์„œ' ํ‡ดํ–‰์‹œํ‚ค๋ ค๋Š” ์ดˆ๊ธฐ ๋ชฉํ‘œ์™€ ๋‹ฌ๋ฆฌ, ํ˜ˆ๊ด€์„ ๋” ์„ฑ์ˆ™ํ•˜๊ณ  ์ •์ƒ์ ์ธ ํ‘œํ˜„ํ˜•์œผ๋กœ ๋ณ€ํ™”์‹œํ‚ด์œผ๋กœ์จ ๊ณผ๋‹ค ํˆฌ๊ณผ์„ฑ์„ ๊ฐ์†Œ์‹œํ‚ค๊ณ  pericyte coverage๋ฅผ ์ฆ๊ฐ€์‹œํ‚จ๋‹ค [22]. ์ด 'ํ˜ˆ๊ด€ ์ •์ƒํ™”'๋Š” ์ข…์–‘ ์ €์‚ฐ์†Œ์ฆ(hypoxia)์„ ์™„ํ™”ํ•˜๊ณ  ๊ฐ„์งˆ์•ก ์••๋ ฅ์„ ๋‚ฎ์ถค์œผ๋กœ์จ ๋‹ค๋ฅธ ์น˜๋ฃŒ์ œ๋“ค์˜ ์ข…์–‘ ๋‚ด ์นจํˆฌ๋ฅผ ์šฉ์ดํ•˜๊ฒŒ ํ•˜๋ฉฐ, ์ด๋Š” anti-VEGF์™€ ํ™”ํ•™์š”๋ฒ•์˜ ๋ณ‘์šฉ ์š”๋ฒ•์ด ๋‹จ์ผ ์š”๋ฒ•๋ณด๋‹ค ์ƒ์กด์œจ์„ ๊ฐœ์„ ํ•˜๋Š” ์—ญ์„ค์ ์ธ ๊ฒฐ๊ณผ๋ฅผ ์„ค๋ช…ํ•œ๋‹ค [22].

์„ธํฌ ๊ฐ„ ํ†ต์‹ ์€ exosomes๊ณผ ๊ฐ™์€ ์„ธํฌ์™ธ ์†Œํฌ์ฒด๋ฅผ ํ†ตํ•ด ์ด๋ฃจ์–ด์ง€๋ฉฐ, ์ด๋Š” TME์˜ ์žฌํ˜•์„ฑ๊ณผ ์‹ ํ˜ธ ์ „๋‹ฌ์— ์ค‘์š”ํ•œ ์—ญํ• ์„ ํ•œ๋‹ค. Exosomes์€ ์ง€๋ฆ„ 30-200nm์˜ ์ž‘์€ ๋ง‰์„ฑ ์†Œ๊ธฐ๊ด€์œผ๋กœ, ๋‹จ๋ฐฑ์งˆ, ์ง€์งˆ, ํ•ต์‚ฐ ๋“ฑ์„ ํฌํ•จํ•˜์—ฌ ๋‹ค๋ฅธ ์„ธํฌ๋กœ ์‹ ํ˜ธ์™€ ๋ถ„์ž๋ฅผ ์ „๋‹ฌํ•œ๋‹ค [10]. ์ด๋“ค์€ ECM ์žฌํ˜•์„ฑ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ๋ฉด์—ญ ๋ฐ˜์‘ ์กฐ์ ˆ, ์กฐ์ง ํ•ญ์ƒ์„ฑ ์œ ์ง€, ๊ทธ๋ฆฌ๊ณ  ์ข…์–‘ ์ง„ํ–‰์— ๊ด€์—ฌํ•œ๋‹ค [10]. ๋˜ํ•œ, microglia์™€ ๊ฐ™์€ ์ค‘์ถ”์‹ ๊ฒฝ๊ณ„(CNS)์˜ ๊ณ ์œ  ๊ฑฐ๋Œ€ ์‹์„ธํฌ๋Š” ๋‡Œ ๋ณ‘๋ฆฌํ•™์˜ ๋ฏผ๊ฐํ•œ ๊ฐ์ง€์ž๋กœ์„œ, ํ™œ์„ฑํ™” ์‹œ ๋‹ค์–‘ํ•œ ๋ฌผ์งˆ์„ ๋ถ„๋น„ํ•˜์—ฌ ์ฃผ๋ณ€ ์„ธํฌ์— ์œ ์ตํ•˜๊ฑฐ๋‚˜ ํ•ด๋กœ์šด ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ๋‹ค [6][9]. ๋น„๋ก microglia ์—ฐ๊ตฌ๊ฐ€ ์ฃผ๋กœ ์‹ ๊ฒฝํ‡ดํ–‰์„ฑ ์งˆํ™˜์— ์ดˆ์ ์„ ๋งž์ถ”๊ณ  ์žˆ์ง€๋งŒ, ๊ทธ ํ™œ์„ฑํ™” ๋ฉ”์ปค๋‹ˆ์ฆ˜๊ณผ ์‚ฌ์ดํ† ์นด์ธ ๋ถ„๋น„ ํŒจํ„ด์€ TME ๋‚ด ๋ฉด์—ญ ์„ธํฌ์˜ ํ–‰๋™ ์–‘์‹์„ ์ดํ•ดํ•˜๋Š” ๋ฐ ์œ ์‚ฌํ•œ ํ†ต์ฐฐ๋ ฅ์„ ์ œ๊ณตํ•œ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” NSCLC์˜ TME์—์„œ PD-L1 ๋ฐœํ˜„๊ณผ TILs/TAMs์˜ ๊ด€๊ณ„๋ฅผ ๋‹ค๋ฃจ๊ณ  ์žˆ์œผ๋‚˜, CAFs์˜ ์ด์งˆ์„ฑ๊ณผ ์น˜๋ฃŒ์  ํ‘œ์ ํ™” ์ „๋žต, ๋˜๋Š” ํ˜ˆ๊ด€ ์ •์ƒํ™”๊ฐ€ ์•ฝ๋ฌผ ์ „๋‹ฌ ํšจ์œจ์— ๋ฏธ์น˜๋Š” ๊ตฌ์ฒด์ ์ธ ๊ธฐ์ „, ๊ทธ๋ฆฌ๊ณ  exosomes์„ ํ†ตํ•œ ์„ธํฌ ๊ฐ„ ํ†ต์‹  ๋„คํŠธ์›Œํฌ์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ๋ถ„์„์€ ์•„์ง ๋ถ€์กฑํ•˜๋‹ค. ํŠนํžˆ CAFs์˜ ๊ธฐ๋Šฅ์  ํ•˜์œ„ ์ง‘๋‹จ์„ ๊ตฌ๋ถ„ํ•˜์—ฌ ์น˜๋ฃŒ ๋ฐ˜์‘์„ ์˜ˆ์ธกํ•˜๋Š” biomarker ๊ฐœ๋ฐœ์ด๋‚˜, ํ˜ˆ๊ด€ ์ •์ƒํ™” ๊ธฐ๊ฐ„(window of normalization)์„ ์ตœ์ ํ™”ํ•˜๋Š” ์ž„์ƒ ์ „๋žต์— ๊ด€ํ•œ ๋ฌธํ—Œ์ด ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์ถ”๊ฐ€๋œ๋‹ค๋ฉด TME ๊ตฌ์กฐ์  ์กฐ์ ˆ์— ๋Œ€ํ•œ ์ดํ•ด๊ฐ€ ํ•œ์ธต ๊นŠ์–ด์งˆ ๊ฒƒ์ด๋‹ค.

๋Œ€์‚ฌ ์žฌํ”„๋กœ๊ทธ๋ž˜๋ฐ ๋ฐ ์ €์‚ฐ์†Œ์ฆ: ์ข…์–‘ ์ง„ํ–‰์˜ ์ƒํ™”ํ•™์  ๊ธฐ๋ฐ˜

์ข…์–‘ ์„ธํฌ๋Š” ๋น ๋ฅธ ์ฆ์‹์„ ์œ„ํ•ด ๋Œ€์‚ฌ๋ฅผ ์žฌํ”„๋กœ๊ทธ๋ž˜๋ฐํ•˜๋ฉฐ, ์ด๋Š” TME์˜ ์‚ฐ์†Œ ๋†๋„์™€ ์˜์–‘๋ถ„ ๊ฐ€์šฉ์„ฑ์— ํฌ๊ฒŒ ์˜์กดํ•œ๋‹ค. Aerobic glycolysis (Warburg effect)๋Š” ์‚ฐ์†Œ๊ฐ€ ์กด์žฌํ•จ์—๋„ ๋ถˆ๊ตฌํ•˜๊ณ  ๊ธ€๋ฆฌ์ฝœ๋ฆฌ์‹œ์Šค ์†๋„๊ฐ€ ๋†’์€ ํ˜„์ƒ์œผ๋กœ, ์ด๋Š” ๋ฏธํ† ์ฝ˜๋“œ๋ฆฌ์•„ ํ˜ธํก ๊ฒฐํ•จ์ด ์•„๋‹Œ, ์„ธํฌ ๋ถ„์—ด์— ํ•„์š”ํ•œ ์ƒํ•ฉ์„ฑ ์ค‘๊ฐ„์ฒด(biosynthetic intermediates)๋ฅผ ๊ณต๊ธ‰ํ•˜๊ธฐ ์œ„ํ•œ ์ ์‘ ๊ธฐ์ „์œผ๋กœ ์ดํ•ด๋œ๋‹ค [7]. ๊ธ€๋ฆฌ์ฝœ๋ฆฌ์‹œ์Šค๋Š” ์ƒˆ๋กœ์šด ์„ธํฌ ๊ตฌ์„ฑ ์š”์†Œ์ธ biomass ์ƒ์„ฑ์„ ์œ„ํ•œ ํƒ„์†Œ ์›์ฒœ์„ ์ œ๊ณตํ•˜๋ฉฐ, ์ด๋Š” ๋น ๋ฅด๊ฒŒ ์ฆ์‹ํ•˜๋Š” ์„ธํฌ์—์„œ ์„ ํƒ์ ์œผ๋กœ ์œ ์ง€๋˜๋Š” ํŠน์ง•์ด๋‹ค [7]. ๋˜ํ•œ, pentose phosphate pathway (PPP)๋Š” NADPH๋ฅผ ์ƒ์„ฑํ•˜์—ฌ ์‚ฐํ™” ์ŠคํŠธ๋ ˆ์Šค์— ๋Œ€ํ•ญํ•˜๊ณ , ํ•ต์‚ฐ ๋ฐ ์•„๋ฏธ๋…ธ์‚ฐ ํ•ฉ์„ฑ์„ ์œ„ํ•œ ์ „๊ตฌ์ฒด๋ฅผ ๊ณต๊ธ‰ํ•จ์œผ๋กœ์จ ์ข…์–‘ ์„ธํฌ์˜ ๋Œ€์‚ฌ ํ•ญ์ƒ์„ฑ๊ณผ ์„ฑ์žฅ ์š”๊ตฌ๋ฅผ ์ถฉ์กฑ์‹œํ‚จ๋‹ค [21].

์ €์‚ฐ์†Œ์ฆ(Hypoxia)์€ TME์˜ ๋˜ ๋‹ค๋ฅธ ํŠน์ง•์  ์š”์†Œ๋กœ, hypoxia-inducible factor-1 alpha (HIF-1ฮฑ)์˜ ํ™œ์„ฑํ™”๋ฅผ ํ†ตํ•ด ์ข…์–‘์˜ ๊ณต๊ฒฉ์ ์ธ ํ‘œํ˜„ํ˜•์„ ์œ ๋„ํ•œ๋‹ค. cervical cancer ๋ฐ cervical intraepithelial neoplasia์—์„œ HIF-1ฮฑ๋Š” ํ˜ˆ๊ด€์ƒ์„ฑ(angiogenesis), ์นจ์œค(invasion), epithelial-mesenchymal transition (EMT)์„ ์ด‰์ง„ํ•˜๊ณ , ๋ฉด์—ญ ๋ฏธ์„ธํ™˜๊ฒฝ์„ ์กฐ์ ˆํ•˜๋ฉฐ, ๋ฐฉ์‚ฌ์„ /ํ™”ํ•™์š”๋ฒ•์— ๋Œ€ํ•œ ๋‚ด์„ฑ์„ ์ฆ๊ฐ€์‹œํ‚จ๋‹ค [25]. HIF-1ฮฑ๋Š” VEGF, PDGF-ฮฒ, MMP-2/9 ๋“ฑ์˜ ํ”„๋กœ-ํ˜ˆ๊ด€์ƒ์„ฑ ๋ฐ ์นจ์Šต ๊ด€๋ จ ์œ ์ „์ž ๋ฐœํ˜„์„ไธŠ่ฐƒํ•˜๋ฉฐ, HPV oncogenes์™€ ์‹œ๋„ˆ์ง€ ํšจ๊ณผ๋ฅผ ์ผ์œผ์ผœ ์ข…์–‘ ์ง„ํ–‰์„ ๊ฐ€์†ํ™”ํ•œ๋‹ค [25]. ์ด๋Ÿฌํ•œ ์ €์‚ฐ์†Œ ๋ฏธ์„ธํ™˜๊ฒฝ์€ ์ „์ด์„ฑ ํ‘œํ˜„ํ˜•์˜ ํš๋“์„ ์„ ํ˜ธํ•˜๋ฉฐ, miR-143๊ณผ ๊ฐ™์€ ํ•ญ์ข…์–‘ ์ธ์ž์™€ ํ”„๋กœ์ข…์–‘ ์ธ์ž ๊ฐ„์˜ ๊ท ํ˜•์ด ์ข…์–‘์˜ ์šด๋ช…์„ ๊ฒฐ์ •ํ•œ๋‹ค [25].

๋Œ€์‚ฌ ๋ฐ ์ €์‚ฐ์†Œ์ฆ ์‹ ํ˜ธ๋Š” ๋˜ํ•œ ์„ธํฌ ์‚ฌ๋ฉธ ํ˜•ํƒœ์™€๋„ ์—ฐ๊ฒฐ๋œ๋‹ค. triple-negative breast cancer (TNBC)์—์„œ Chinese medicine-based interventions (CMBIs)๋Š” ferroptosis (์ฒ  ์˜์กด์„ฑ ์ง€์งˆ ๊ณผ์‚ฐํ™”์— ์˜ํ•œ ์„ธํฌ ์‚ฌ๋ฉธ)๋ฅผ ์œ ๋„ํ•˜๋Š” ๊ฒƒ์œผ๋กœ preclinical ์—ฐ๊ตฌ์—์„œ ๋ณด๊ณ ๋˜์—ˆ๋‹ค [30]. Ferrostatin-1์— ๋ฏผ๊ฐํ•œ ์ด ๊ณผ์ •์€ TNBC์˜ ๊ณต๊ฒฉ์„ฑ์„ ์–ต์ œํ•  ์ž ์žฌ๋ ฅ์„ ๋ณด์ด์ง€๋งŒ, ์•„์ง ์ž„์ƒ์  ๊ฒ€์ฆ์ด ๋ถ€์กฑํ•˜๋‹ค [30]. ์ด๋Š” TME ๋‚ด ๋Œ€์‚ฌ ์ŠคํŠธ๋ ˆ์Šค์™€ ์‚ฐํ™” ์ŠคํŠธ๋ ˆ์Šค๊ฐ€ ์ข…์–‘ ์„ธํฌ ์‚ฌ๋ฉธ์„ ์œ ๋„ํ•  ์ˆ˜ ์žˆ๋Š” ์ƒˆ๋กœ์šด ์น˜๋ฃŒ ํ‘œ์ ์ด ๋  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•œ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” NSCLC์˜ immune profiling์— ์ดˆ์ ์„ ๋งž์ถ”๊ณ  ์žˆ์œผ๋‚˜, TME ๋‚ด ๋Œ€์‚ฌ ์žฌํ”„๋กœ๊ทธ๋ž˜๋ฐ(glycolysis, PPP)๊ณผ ์ €์‚ฐ์†Œ์ฆ(HIF-1ฮฑ)์ด ๋ฉด์—ญ ๋ฐ˜์‘ ๋ฐ ์น˜๋ฃŒ ๋‚ด์„ฑ์— ๋ฏธ์น˜๋Š” ๊ตฌ์ฒด์ ์ธ ์ƒํ™”ํ•™์  ๊ธฐ์ „์— ๋Œ€ํ•œ ๋ฌธํ—Œ์€ ํฌํ•จ๋˜์ง€ ์•Š์•˜๋‹ค. HIF-1ฮฑ ์–ต์ œ์ œ๋‚˜ ๋Œ€์‚ฌ ํ‘œ์  ์š”๋ฒ•์ด ICI์™€ ๋ณ‘์šฉ๋  ๋•Œ์˜ ์‹œ๋„ˆ์ง€ ํšจ๊ณผ์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๋Š” ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์˜ blind spot์ด๋ฉฐ, ์ด๋ฅผ ๋ณด์™„ํ•œ๋‹ค๋ฉด TME์˜ ๋Œ€์‚ฌ-๋ฉด์—ญ ์ถ•(metabolic-immune axis)์— ๋Œ€ํ•œ ํฌ๊ด„์ ์ธ ์ดํ•ด๊ฐ€ ๊ฐ€๋Šฅํ•ด์งˆ ๊ฒƒ์ด๋‹ค.

์ „์ด, EMT ๋ฐ ๋ฏธ์„ธํ™˜๊ฒฝ ์žฌํ˜•์„ฑ์˜ ์—ญ๋™์  ๊ณผ์ •

์ „์ด(Metastasis)๋Š” ์•” ๊ด€๋ จ ์‚ฌ๋ง์˜ ์ฃผ์š” ์›์ธ์ด๋ฉฐ, epithelial-mesenchymal transition (EMT)์€ ์ด ๊ณผ์ •์—์„œ ํ•ต์‹ฌ์ ์ธ ์—ญํ• ์„ ํ•œ๋‹ค. EMT๋Š” ์ƒํ”ผ ์„ธํฌ๊ฐ€ ๊ฐ„์—ฝ ์„ธํฌ ํ‘œํ˜„ํ˜•์œผ๋กœ ์ „ํ™˜๋˜๋Š” ๋™์  ๊ณผ์ •์œผ๋กœ, ์„ธํฌ ์ด๋™์„ฑ, ์นจ์Šต์„ฑ, ๊ทธ๋ฆฌ๊ณ  ์•„ํฌํ† ์‹œ์Šค ์ž๊ทน์— ๋Œ€ํ•œ ์ €ํ•ญ์„ฑ์„ ์ฆ๊ฐ€์‹œํ‚จ๋‹ค [13][20]. ๋˜ํ•œ, EMT๋ฅผ ๊ฒช์€ ์ข…์–‘ ์„ธํฌ๋Š” stem cell ํŠน์„ฑ์„ ํš๋“ํ•˜๊ณ  ์น˜๋ฃŒ ๋‚ด์„ฑ์„ ๋‚˜ํƒ€๋‚ธ๋‹ค [20]. ๊ทธ๋Ÿฌ๋‚˜ EMT๊ฐ€ ์ „์ด์— ํ•„์ˆ˜์ ์ธ์ง€ ์—ฌ๋ถ€๋Š” in vivo์—์„œ ๊ธฐ์ˆ ์ ์œผ๋กœ ์ž…์ฆํ•˜๊ธฐ ์–ด๋ ค์› ์œผ๋ฉฐ, ์ตœ๊ทผ ์—ฐ๊ตฌ๋“ค์€ lineage tracing ์‹œ์Šคํ…œ๊ณผ ๋™์  in vivo imaging์„ ํ†ตํ•ด EMT์™€ ์ „์ด ๊ฐ„์˜ ์ง์ ‘์ ์ธ ์ธ๊ณผ๊ด€๊ณ„๋ฅผ ๊ทœ๋ช…ํ•˜๋ ค๋Š” ๋…ธ๋ ฅ์„ ๊ธฐ์šธ์ด๊ณ  ์žˆ๋‹ค [20]. TEMTIA (The EMT International Association)๋Š” EMT ์—ฐ๊ตฌ์˜ ์šฉ์–ด์™€ ์ •์˜๋ฅผ ํ‘œ์ค€ํ™”ํ•˜์—ฌ ๋ฐ์ดํ„ฐ ํ•ด์„์˜ ์˜คํ•ด๋ฅผ ์ค„์ด๊ณ  ํ•™์ œ ๊ฐ„ ํ˜‘๋ ฅ์„ ์ด‰์ง„ํ•˜๊ธฐ ์œ„ํ•œ ๊ฐ€์ด๋“œ๋ผ์ธ์„ ์ œ์‹œํ–ˆ๋‹ค [13].

EMT์™€ ์œ ์‚ฌํ•˜๊ฒŒ, wound healing ๊ณผ์ •์—์„œ์˜ ์„ธํฌ ์ƒํ˜ธ์ž‘์šฉ๋„ TME ์žฌํ˜•์„ฑ๊ณผ ๊ด€๋ จ์ด ๊นŠ๋‹ค. ์ƒ์ฒ˜ ์น˜์œ ๋Š” hemostasis, inflammation, growth, re-epithelialization, remodeling ๋‹จ๊ณ„์—์„œ ๋‹ค์–‘ํ•œ ์„ธํฌ ์œ ํ˜•์˜ ๊ณต๊ฐ„์ ยท์‹œ๊ฐ„์  ๋™๊ธฐํ™”๋ฅผ ํ•„์š”๋กœ ํ•œ๋‹ค [8]. ๋งŒ์„ฑ ์ƒ์ฒ˜๋‚˜ ๋น„ํ›„์„ฑ ํ‰ํ„ฐ์™€ ๊ฐ™์€ ๋ณ‘๋ฆฌ์  ์ƒํƒœ์—์„œ๋Š” ๋ฏธ์„ธํ™˜๊ฒฝ์˜ ๋ณ€ํ™”(๊ธฐ๊ณ„์  ํž˜, ์‚ฐ์†Œ ์ˆ˜์ค€, chemokine, ECM ๋ฐ ์„ฑ์žฅ ์ธ์ž ํ•ฉ์„ฑ์˜ ๋ณ€ํ™”)๊ฐ€ ์„ธํฌ ๋ชจ์ง‘๊ณผ ํ™œ์„ฑํ™”๋ฅผ ๋ฐฉํ•ดํ•œ๋‹ค [8]. ๋‹จ์ผ ์„ธํฌ ๊ธฐ์ˆ (single cell technologies)์€ ์ด๋Ÿฌํ•œ ์„ธํฌ ์œ ํ˜•์˜ ํ‘œํ˜„ํ˜• ๋ฐ ๊ธฐ๋Šฅ์  ์ด์งˆ์„ฑ์„ ํ•ด๋…ํ•˜๋Š” ๋ฐ ํ•„์ˆ˜์ ์ด๋‹ค [8].

๋˜ํ•œ, microbiome๊ณผ TME์˜ ์ƒํ˜ธ์ž‘์šฉ๋„ ์ฃผ๋ชฉ๋ฐ›๊ณ  ์žˆ๋‹ค. thyroid carcinoma (TC)์—์„œ ์žฅ๋‚ด ๋ฏธ์ƒ๋ฌผ๊ตฐ(gut microbiome)์€ ๋ฉด์—ญ ๋ฏธ์„ธํ™˜๊ฒฝ์„ ์กฐ์ ˆํ•˜์—ฌ ์ข…์–‘ ์ง„ํ–‰์— ์˜ํ–ฅ์„ ๋ฏธ์นœ๋‹ค [26]. Dysbiosis, lipopolysaccharides (LPS), short-chain fatty acids (SCFAs) ๋“ฑ์˜ ๋Œ€์‚ฌ๋ฌผ์งˆ์€ ๊ฐ‘์ƒ์„  ๊ธฐ๋Šฅ๊ณผ ๋ฉด์—ญ ๋ฐ˜์‘์„ ์–‘๋ฐฉํ–ฅ์œผ๋กœ ์กฐ์ ˆํ•˜๋ฉฐ, ์ด๋Š” TC์˜ ์ „์ด์„ฑ ํ‘œํ˜„ํ˜• ํš๋“ ๋ฐ ์น˜๋ฃŒ ๋ฐ˜์‘(ํ™”ํ•™์š”๋ฒ•, ๋ฐฉ์‚ฌ์„  ์š”๋ฒ•, ํ‘œ์  ์š”๋ฒ•)์— ์˜ํ–ฅ์„ ์ค„ ์ˆ˜ ์žˆ๋‹ค [26]. ๋น„๋ก intratumor microbiome์˜ ์—ญํ• ์€ ์•„์ง ๋ช…ํ™•ํ•˜์ง€ ์•Š์ง€๋งŒ, ๋ฏธ์ƒ๋ฌผ๊ตฐ ๊ธฐ๋ฐ˜์˜ ๋ฉด์—ญ ์กฐ์ ˆ์€ ๊ฐœ์ธ ๋งž์ถคํ˜• ์น˜๋ฃŒ ์ „๋žต์œผ๋กœ ๋ถ€์ƒํ•˜๊ณ  ์žˆ๋‹ค [26].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” NSCLC์—์„œ์˜ immune profiling๊ณผ neoadjuvant immunotherapy ๋ชจ๋‹ˆํ„ฐ๋ง์„ ๋‹ค๋ฃจ๊ณ  ์žˆ์œผ๋‚˜, EMT์˜ ์—ญ๋™์  ๊ณผ์ •, ์ „์ด ๋ฉ”์ปค๋‹ˆ์ฆ˜, ๊ทธ๋ฆฌ๊ณ  microbiome-TME ์ƒํ˜ธ์ž‘์šฉ์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๋ฌธํ—Œ์€ ํฌํ•จ๋˜์ง€ ์•Š์•˜๋‹ค. ํŠนํžˆ EMT์™€ ์ „์ด์˜ ์ธ๊ณผ๊ด€๊ณ„๋ฅผ ๊ทœ๋ช…ํ•˜๋Š” ์ตœ์‹  in vivo ๋ชจ๋ธ ์—ฐ๊ตฌ๋‚˜, ์žฅ๋‚ด ๋ฏธ์ƒ๋ฌผ์ด ๋ฉด์—ญ ์น˜๋ฃŒ ๋ฐ˜์‘์— ๋ฏธ์น˜๋Š” ์˜ํ–ฅ์— ๋Œ€ํ•œ ๋ถ„์„์€ ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์˜ gap์œผ๋กœ ๋‚จ์•„ ์žˆ์œผ๋ฉฐ, ์ด๋ฅผ ์ฑ„์›Œ๋„ฃ์œผ๋ฉด TME์˜ ๊ณต๊ฐ„์ ยท์‹œ๊ฐ„์  ๋ณ€ํ™” ๋ฐ ์™ธ๋ถ€ ํ™˜๊ฒฝ ์š”์ธ์˜ ์˜ํ–ฅ์„ ๋” ์ž˜ ์ดํ•ดํ•  ์ˆ˜ ์žˆ์„ ๊ฒƒ์ด๋‹ค.

์ง„๋‹จ ๊ธฐ์ˆ ์˜ ์ง„ํ™”: ๋”ฅ๋Ÿฌ๋‹, ๋ฐฉ์‚ฌ์„ ์ฒดํ•™ ๋ฐ ์˜ค๊ฐ„ ์˜จ ์นฉ

TME์˜ ๋ณต์žกํ•œ ํŠน์„ฑ์„ ๋น„์นจ์Šต์ ์œผ๋กœ ํ‰๊ฐ€ํ•˜๊ณ  ์น˜๋ฃŒ ๋ฐ˜์‘์„ ์˜ˆ์ธกํ•˜๊ธฐ ์œ„ํ•œ ์ง„๋‹จ ๊ธฐ์ˆ ์€ ๋น ๋ฅด๊ฒŒ ๋ฐœ์ „ํ•˜๊ณ  ์žˆ๋‹ค. Deep learning (DL)์€ ์˜๋ฃŒ ์˜์ƒ ๋ถ„์„์—์„œ state-of-the-art๊ฐ€ ๋˜์—ˆ์œผ๋ฉฐ, ๊ณ„์ธต์  ํŠน์ง• ํ‘œํ˜„(hierarchical feature representations)์„ ํ†ตํ•ด ํ•ด๋ถ€ํ•™์  ๊ตฌ์กฐ, ์„ธํฌ ๊ตฌ์กฐ, ์กฐ์ง ๋ถ„ํ•  ๋ฐ ์งˆ๋ณ‘ ์ง„๋‹จ/์˜ˆํ›„๋ฅผ ์ž๋™์œผ๋กœ ์‹๋ณ„ํ•œ๋‹ค [3]. ํŠนํžˆ ์œ„์•”/GEJ adenocarcinoma์—์„œ DL์€ routine H&E ์Šฌ๋ผ์ด๋“œ๋กœ๋ถ€ํ„ฐ MSI/dMMR, EBV, HER2, PD-L1, CLDN18.2์™€ ๊ฐ™์€ ์ž„์ƒ์  biomarker ์ƒํƒœ๋ฅผ ์ง์ ‘ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๋Š” ๊ฐ€๋Šฅ์„ฑ์„ ๋ณด์—ฌ์ค€๋‹ค [28]. ์ด๋Š” ๋น„์šฉ์ด ๋งŽ์ด ๋“ค๊ณ  ๋ณดํŽธ์ ์œผ๋กœ ์ด์šฉ๋˜์ง€ ์•Š๋Š” ๋ถ„์ž/๋ฉด์—ญ์กฐ์งํ™”ํ•™(IHC) ๊ฒ€์‚ฌ๋ฅผ ๋Œ€์ฒดํ•˜๊ฑฐ๋‚˜ ์„ ๋ณ„(screening) ๋„๊ตฌ๋กœ ํ™œ์šฉ๋  ์ˆ˜ ์žˆ์–ด, TME์˜ ๋ฉด์—ญ ๋ฐ ๋ถ„์ž ํ”„๋กœํŒŒ์ผ๋ง์„ ๋ฏผ์ฃผํ™”ํ•˜๋Š” ๋ฐ ๊ธฐ์—ฌํ•œ๋‹ค [28].

Radiomics๋Š” NSCLC์—์„œ TME๋ฅผ ํ•ด๋…ํ•˜๋Š” ๋˜ ๋‹ค๋ฅธ ๊ฐ•๋ ฅํ•œ ๋„๊ตฌ์ด๋‹ค. ์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ํฌํ•จ๋œ "Deciphering the tumor microenvironment through radiomics in non-small cell lung cancer" ๋…ผ๋ฌธ์€ ๋ฐฉ์‚ฌ์„ ์ฒดํ•™ ํŠน์ง•์ด immune profiles๊ณผ ์ƒ๊ด€๊ด€๊ณ„๊ฐ€ ์žˆ์Œ์„ ๋ณด์—ฌ์ฃผ๋ฉฐ, ์ด๋Š” ๋น„์นจ์Šต์ ์œผ๋กœ TME์˜ ๋ฉด์—ญ ์ƒํƒœ๋ฅผ ์ถ”์ •ํ•  ์ˆ˜ ์žˆ์Œ์„ ์˜๋ฏธํ•œ๋‹ค. ๋˜ํ•œ, machine learning (ML)์€ mRNA ๋ฐฑ์‹  ๋ฐ ๊ธฐํƒ€ ์น˜๋ฃŒ์ œ์˜ pharmacovigilance์—์„œ๋„ ํ™œ์šฉ๋˜๋ฉฐ, adverse event ์˜ˆ์ธก ๋ชจ๋ธ์—์„œ AUC 0.85-0.87์„ ๋‹ฌ์„ฑํ•˜๋Š” ๋“ฑ ๋Šฅ๋™์ ์ด๊ณ  ์ง€๋Šฅ์ ์ธ ๋ชจ๋‹ˆํ„ฐ๋ง ์‹œ์Šคํ…œ์œผ๋กœ ์ง„ํ™”ํ•˜๊ณ  ์žˆ๋‹ค [27].

๋™๋ฌผ ๋ชจ๋ธ์˜ ํ•œ๊ณ„๋ฅผ ๊ทน๋ณตํ•˜๊ธฐ ์œ„ํ•ด human organs-on-chips ๊ธฐ์ˆ ์ด ๋“ฑ์žฅํ–ˆ๋‹ค. Organ chip์€ ๋ฏธ์„ธ์œ ์ฒด ์žฅ์น˜ ๋‚ด์— ์ƒ์ฒด ์„ธํฌ๋ฅผ ๋ฐฐ์–‘ํ•˜์—ฌ ์žฅ๊ธฐ ์ˆ˜์ค€์˜ ์ƒ๋ฆฌ ๋ฐ ๋ณ‘๋ฆฌ๋ฅผ ๋†’์€ ์ถฉ์‹ค๋„๋กœ ์žฌํ˜„ํ•œ๋‹ค [24]. ์ด๋Š” ๋ณต์žกํ•œ ์งˆ๋ณ‘ ๋ชจ๋ธ๋ง, ์•ฝ๋ฌผ ๊ฐœ๋ฐœ, ๊ทธ๋ฆฌ๊ณ  ๊ฐœ์ธ ๋งž์ถคํ˜• ์˜ํ•™์—์„œ host-microbiome ์ƒํ˜ธ์ž‘์šฉ์ด๋‚˜ ์ „์‹  ์žฅ๊ธฐ๊ด€ ๊ฐ„ ์ƒ๋ฆฌ๋ฅผ ๋ชจ์‚ฌํ•˜๋Š” ๋ฐ ์‚ฌ์šฉ๋œ๋‹ค [24]. ์ด๋Ÿฌํ•œ ํ”Œ๋žซํผ์€ TME์˜ ๋ฏธ์„ธํ™˜๊ฒฝ ์กฐ๊ฑด(์˜ˆ: ํ˜ˆ๋ฅ˜, ์‚ฐ์†Œ ๋†๋„, ์„ธํฌ ๊ฐ„ ์ ‘์ด‰)์„ ์ •๋ฐ€ํ•˜๊ฒŒ ์ œ์–ดํ•˜์—ฌ in vivo์™€ ์œ ์‚ฌํ•œ ์น˜๋ฃŒ ๋ฐ˜์‘์„ ์˜ˆ์ธกํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•œ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ด๋ฏธ NSCLC์—์„œ์˜ radiomics์™€ immune profile ์ƒ๊ด€๊ด€๊ณ„์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์„ ํฌํ•จํ•˜๊ณ  ์žˆ์–ด, ์˜์ƒ ๊ธฐ๋ฐ˜ TME ๋ถ„์„์˜ ์„ ๋‘๋ฅผ ๊ฑท๊ณ  ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ DL์„ ํ†ตํ•œ H&E ์Šฌ๋ผ์ด๋“œ biomarker ์˜ˆ์ธก [28], organ-on-chip์„ ์ด์šฉํ•œ TME ๋ชจ๋ธ๋ง [24], ๊ทธ๋ฆฌ๊ณ  ML ๊ธฐ๋ฐ˜ pharmacovigilance [27]์™€ ๊ฐ™์€ ์ตœ์‹  ๊ณ„์‚ฐ ๋ฐ ๊ณตํ•™์  ์ ‘๊ทผ๋ฒ•์€ ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ช…์‹œ์ ์œผ๋กœ ํฌํ•จ๋˜์–ด ์žˆ์ง€ ์•Š๋‹ค. ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ๋“ค์€ TME์˜ ์ •๋Ÿ‰์  ํ‰๊ฐ€์™€ ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก์˜ ์ •ํ™•๋„๋ฅผ ๋†’์ด๋Š” ๋ฐ ํ•„์ˆ˜์ ์ด๋ฏ€๋กœ, ํ–ฅํ›„ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ™•์žฅ์˜ ์ฃผ์š” ๋ฐฉํ–ฅ์ด ๋  ๊ฒƒ์ด๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1Adverse Renal Effects of Immune Checkpoint Inhibitors: A Narrative ReviewAmerican Journal of Nephrology ยท 201715291PubMed ยท DOI
2The Senescence-Associated Secretory Phenotype: The Dark Side of Tumor SuppressionAnnual Review of Pathology Mechanisms of Disease ยท 20105271PubMed ยท DOI
3Deep Learning in Medical Image AnalysisAnnual Review of Biomedical Engineering ยท 20174929PubMed ยท DOI
4Head and neck squamous cell carcinomaNature Reviews Disease Primers ยท 20204201PubMed ยท DOI
5A framework for advancing our understanding of cancer-associated fibroblastsNature reviews. Cancer ยท 20203939PubMed ยท DOI
6Physiology of MicrogliaPhysiological Reviews ยท 20113550PubMed ยท DOI
7Aerobic Glycolysis: Meeting the Metabolic Requirements of Cell ProliferationAnnual Review of Cell and Developmental Biology ยท 20113108PubMed ยท DOI
8Wound Healing: A Cellular PerspectivePhysiological Reviews ยท 20183029PubMed ยท DOI
9Microglia Function in the Central Nervous System During Health and NeurodegenerationAnnual Review of Immunology ยท 20172924PubMed ยท DOI
10ExosomesAnnual Review of Biochemistry ยท 20192797PubMed ยท DOI
11Cholangiocarcinoma 2020: the next horizon in mechanisms and managementNature Reviews Gastroenterology & Hepatology ยท 20202543PubMed ยท DOI
12The Next Generation of Platinum Drugs: Targeted Pt(II) Agents, Nanoparticle Delivery, and Pt(IV) ProdrugsChemical Reviews ยท 20162482PubMed ยท DOI
13Guidelines and definitions for research on epithelialโ€“mesenchymal transitionNature Reviews Molecular Cell Biology ยท 20202367PubMed ยท DOI
14Immune Checkpoint Inhibitors for the Treatment of Cancer: Clinical Impact and Mechanisms of Response and ResistanceAnnual Review of Pathology Mechanisms of Disease ยท 20202263PubMed ยท DOI
15Chemokines and Chemokine Receptors: Positioning Cells for Host Defense and ImmunityAnnual Review of Immunology ยท 20142133PubMed ยท DOI
16Into the Eye of the Cytokine StormMicrobiology and Molecular Biology Reviews ยท 20122017PubMed ยท DOI
17The KEAP1-NRF2 System: a Thiol-Based Sensor-Effector Apparatus for Maintaining Redox HomeostasisPhysiological Reviews ยท 20181955PubMed ยท DOI
18Immunosuppressive Strategies that are Mediated by Tumor CellsAnnual Review of Immunology ยท 20061716PubMed ยท DOI
19An overview of nanoparticles commonly used in fluorescent bioimagingChemical Society Reviews ยท 20151611PubMed ยท DOI
20Epithelial Mesenchymal Transition in Tumor MetastasisAnnual Review of Pathology Mechanisms of Disease ยท 20181605PubMed ยท DOI
21The return of metabolism: biochemistry and physiology of the pentose phosphate pathwayBiological reviews/Biological reviews of the Cambridge Philosophical Society ยท 20141603PubMed ยท DOI
22Normalization of the Vasculature for Treatment of Cancer and Other DiseasesPhysiological Reviews ยท 20111602PubMed ยท DOI
23Pathways of Antigen ProcessingAnnual Review of Immunology ยท 20131547PubMed ยท DOI
24Human organs-on-chips for disease modelling, drug development and personalized medicineNature Reviews Genetics ยท 20221526PubMed ยท DOI
25Role of hypoxia-inducible factor โ€“ 1 alpha on the progression of cervical intraepithelial neoplasia and cervical cancer: a narrative reviewFrontiers in Oncology ยท 20260DOI
26Crosstalk between the microbiome and immune microenvironment in the pathogenesis and treatment of thyroid carcinoma: a narrative reviewFrontiers in Immunology ยท 20260DOI
27Machine learning-assisted mRNA vaccine pharmacovigilance: a systematic review of multi-source real-world dataBMC Medical Informatics and Decision Making ยท 20260PubMed ยท DOI
28Deep learning prediction of MSI/dMMR, EBV, HER2, PD-L1, and CLDN18.2 status directly from hematoxylin and eosin-stained slides in gastric and gastroesophageal junction adenocarcinoma: a systematic reviewFrontiers in Oncology ยท 20260DOI
29Survival benefit and toxicity trade-offs of first-line chemoimmunotherapy in advanced gastric and gastroesophageal junction cancer: a meta-analysis of randomized trialsFrontiers in Oncology ยท 20260DOI
30Chinese medicine-based interventions in triple-negative breast cancer: mechanistic evidence, clinical findings, and translational challenges-a narrative reviewFrontiers in Pharmacology ยท 20260DOI
31Coumarins in musculoskeletal health: a systematic reviewFrontiers in Nutrition ยท 20260DOI
32Fibroblast growth factor receptor 3 (FGFR3) alterations and response to immune checkpoint inhibition in metastatic urothelial carcinoma: a systematic review and meta-analysisClinical & Translational Oncology ยท 20260PubMed ยท DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1Advances in targeting tumor microenvironment for immunotherapy.Frontiers in immunology ยท 2024PubMed ยท DOI
2Programming immune escape.Nature reviews. Cancer ยท 2024-MayPubMed ยท DOI
3The plasticity of cancer-associated fibroblasts.Trends in cancer ยท 2025-AugPubMed ยท DOI
4The role of RNA methylation in tumor immunity and its potential in immunotherapy.Molecular cancer ยท 2024-Jun-20PubMed ยท DOI
5Turning cold into hot: emerging strategies to fire up the tumor microenvironment.Trends in cancer ยท 2025-FebPubMed ยท DOI
6Nanobodies targeting the tumor microenvironment and their formulation as nanomedicines.Molecular cancer ยท 2025-Mar-04PubMed ยท DOI
7Q&A with Ilaria Elia.Cell reports ยท 2024-Nov-26PubMed ยท DOI
8Q&A with Ping Gao.Cell reports ยท 2024-Nov-26PubMed ยท DOI
9Cleanup on IL-2.Science signaling ยท 2024-May-07PubMed ยท DOI
10Targeting pyroptosis for cancer immunotherapy: mechanistic insights and clinical perspectives.Molecular cancer ยท 2025-May-03PubMed ยท DOI
ILD โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 8ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
์ง„๋‹จ ๋ฐ ์˜์ƒ ํ‰๊ฐ€ (Diagnosis and Imaging Assessment)

ILD์˜ ์ •ํ™•ํ•œ ์ง„๋‹จ์€ ์ž„์ƒ์ , ๋ฐฉ์‚ฌ์„ ํ•™์ , ๋ณ‘๋ฆฌํ•™์  ์†Œ๊ฒฌ์„ ์ข…ํ•ฉํ•˜๋Š” ๋‹คํ•™์ œ์  ์ ‘๊ทผ(Multidisciplinary Discussion, MDD)์— ์˜์กดํ•˜๋ฉฐ, ์ตœ๊ทผ์—๋Š” ์นจ์Šต์  ์กฐ์ง ๊ฒ€์‚ฌ์˜ ์•ˆ์ „์„ฑ๊ณผ ์ง„๋‹จ ์ˆ˜์œจ(diagnostic yield)์— ๋Œ€ํ•œ ์žฌํ‰๊ฐ€๊ฐ€ ํ™œ๋ฐœํžˆ ์ด๋ฃจ์–ด์ง€๊ณ  ์žˆ๋‹ค. ์ „ํ†ต์ ์œผ๋กœ ์™ธ๊ณผ์  ํ ์ƒ๊ฒ€(Surgical Lung Biopsy, SLB)์ด ๊ธˆ์ค€์œผ๋กœ ์—ฌ๊ฒจ์กŒ์œผ๋‚˜, ๋†’์€ ํ•ฉ๋ณ‘์ฆ๋ฅ (ํ๊ธฐ์ข… 20% ์ด์ƒ, ์‚ฌ๋ง๋ฅ  2.7%)๋กœ ์ธํ•ด ๊ฒฝ๊ธฐ๊ด€์ ˆ ํ ๋ƒ‰์ƒ๊ฒ€(Transbronchial Lung Cryobiopsy, TBLC)์˜ ์—ญํ• ์ด ๋ถ€๊ฐ๋˜๊ณ  ์žˆ๋‹ค [12]. TBLC๋Š” SLB์— ๋น„ํ•ด ์ž…์› ๊ธฐ๊ฐ„์ด ํ˜„์ €ํžˆ ์งง์œผ๋ฉฐ(2.6์ผ vs 6.1์ผ), ์‚ฌ๋ง๋ฅ ์€ 0.3%๋กœ ๋งค์šฐ ๋‚ฎ๊ฒŒ ๋ณด๊ณ ๋˜์—ˆ๋‹ค [12]. ์ง„๋‹จ ์ˆ˜์œจ ์ธก๋ฉด์—์„œ TBLC๋Š” ์•ฝ 82.8~88.3%๋ฅผ ๋ณด์ด๋ฉฐ, SLB์˜ 98.7%์—๋Š” ๋ฏธ์น˜์ง€ ๋ชปํ•˜์ง€๋งŒ, ์ดˆ๊ธฐ ์ง„๋‹จ ์ ‘๊ทผ๋ฒ•์œผ๋กœ์„œ์˜ ํƒ€๋‹น์„ฑ์ด ์ž…์ฆ๋˜์—ˆ๋‹ค [12, 25]. ํŠนํžˆ ์ตœ๊ทผ ์—ฐ๊ตฌ์—์„œ๋Š” ์†Œ์ง๊ฒฝ cryoprobe(1.1โ€“1.7 mm)์™€ ํ‘œ์ค€ ์ง๊ฒฝ probe(1.9 mm) ๊ฐ„์˜ ์ง„๋‹จ ์ˆ˜์œจ์— ์œ ์˜ํ•œ ์ฐจ์ด๊ฐ€ ์—†์Œ์„(RR 0.95; 95% CI 0.88โ€“1.03) ๋ณด์—ฌ์ฃผ์—ˆ์œผ๋ฉฐ, ์ด๋Š” ์žฅ๋น„ ์„ ํƒ์˜ ์œ ์—ฐ์„ฑ์„ ๋†’์ด๋Š” ๊ทผ๊ฑฐ๊ฐ€ ๋œ๋‹ค [25].

์˜์ƒ ์ง„๋‹จ ์ธก๋ฉด์—์„œ ๊ณ ํ•ด์ƒ๋„ CT(HRCT) ํŒจํ„ด์€ ILD ํ•˜์œ„ ๋ถ„๋ฅ˜์˜ ํ•ต์‹ฌ ์ง€ํ‘œ์ด๋‹ค. ๊ฒฐํ•ฉ์กฐ์ง๋ณ‘(CTD) ๊ด€๋ จ ILD(CTD-ILD)์˜ ๊ฒฝ์šฐ, ๋ฅ˜๋งˆํ‹ฐ์Šค ๊ด€์ ˆ์—ผ(RA)์—์„œ๋Š” Usual Interstitial Pneumonia (UIP) ํŒจํ„ด์ด ๊ฐ€์žฅ ํ”ํ•œ ๋ฐ˜๋ฉด(46%), ์ „์‹ ์„ฑ ๊ฒฝํ™”์ฆ(SSc), ํŠน๋ฐœ์„ฑ ์—ผ์ฆ์„ฑ ๊ทผ๋ณ‘์ฆ(IIM), Sjรถgren ์ฆํ›„๊ตฐ ๋“ฑ ๋‹ค๋ฅธ CTD ํ•˜์œ„ ์œ ํ˜•์—์„œ๋Š” Nonspecific Interstitial Pneumonia (NSIP) ํŒจํ„ด์ด ์šฐ์„ธํ•˜๋‹ค(27โ€“76%) [20]. ์ด๋Ÿฌํ•œ ์˜์ƒํ•™์  ์ด์งˆ์„ฑ์€ CTD-ILD๋ฅผ ๋‹จ์ผ ์งˆํ™˜์œผ๋กœ ๋ณด๊ธฐ ์–ด๋ ต๊ฒŒ ํ•˜๋ฉฐ, ๊ฐ CTD subtype์— ๋”ฐ๋ฅธ ๋งž์ถคํ˜• ๋ชจ๋‹ˆํ„ฐ๋ง ์ „๋žต์˜ ํ•„์š”์„ฑ์„ ์‹œ์‚ฌํ•œ๋‹ค [20]. ๋˜ํ•œ, SARS-CoV-2 ๊ฐ์—ผ ํ›„ ๋ฐœ์ƒํ•˜๋Š” ํ์„ฌ์œ ์ฆ(Post-COVID pulmonary fibrosis) ์—ญ์‹œ HRCT ์ƒ์—์„œ ์ง€์†์ ์ธ ๊ฐ„์งˆ์„ฑ ๋ณ€ํ™”๋ฅผ ๋ณด์ด๋ฉฐ, ์ด๋Š” ์žฅ๊ธฐ์ ์ธ ๊ธฐ๋Šฅ์  ์ €ํ•˜์™€ ์—ฐ๊ด€๋˜์–ด ์žˆ๋‹ค [19, 27].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” 'Update of the international multidisciplinary classification of the interstitial pneumonias: an ERS/ATS statement'๊ฐ€ ํฌํ•จ๋˜์–ด ์žˆ์–ด, ILD์˜ ๋ถ„๋ฅ˜ ์ฒด๊ณ„์— ๋Œ€ํ•œ ์ตœ์‹  ๊ฐ€์ด๋“œ๋ผ์ธ์€ ์ด๋ฏธ ํ™•๋ณด๋˜์–ด ์žˆ๋‹ค. ๋˜ํ•œ 'Risk factors and prognostic indicators for progressive fibrosing interstitial lung disease: a deep learning-based CT quantification approach'์™€ 'park-2026-quantitative-ct-imaging-in-progressive-pulmonary-fibrosis'๋Š” ์ •๋Ÿ‰์  CT ๋ถ„์„ ๋ฐ ๋”ฅ๋Ÿฌ๋‹ ๊ธฐ๋ฐ˜์˜ ์˜์ƒ ํ‰๊ฐ€์— ์ดˆ์ ์„ ๋งž์ถ”๊ณ  ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ œ๊ณต๋œ review ์ž๋ฃŒ๋“ค์€ TBLC์˜ ์•ˆ์ „์„ฑ ํ”„๋กœํŒŒ์ผ๊ณผ CTD-ILD์˜ ๊ตฌ์ฒด์ ์ธ HRCT ํŒจํ„ด ๋ถ„ํฌ์— ๋Œ€ํ•œ ๋ฉ”ํƒ€๋ถ„์„ ๋ฐ์ดํ„ฐ๋ฅผ ๊ฐ•์กฐํ•˜๋ฏ€๋กœ, ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์˜ ์ •๋Ÿ‰์  AI ๋ชจ๋ธ ์—ฐ๊ตฌ์™€ ๊ฒฐํ•ฉํ•˜์—ฌ '์˜์ƒ ํŒจํ„ด์˜ ์ •๋Ÿ‰์  ํ•ด์„'๊ณผ '์นจ์Šต์  ์ง„๋‹จ์˜ ์œ„ํ—˜-ํŽธ์ต ๋ถ„์„'์„ ํ†ตํ•ฉํ•œ ์ง„๋‹จ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ๊ฐœ๋ฐœ์ด ํ–ฅํ›„ ์ค‘์š”ํ•œ gap์œผ๋กœ ๋‚จ๋Š”๋‹ค.

๋ณ‘ํƒœ์ƒ๋ฆฌ ๋ฐ ์ด์งˆ์„ฑ (Pathophysiology and Heterogeneity)

ILD๋Š” ๋‹จ์ˆœํ•œ ํ ์กฐ์ง์˜ ์„ฌ์œ ํ™”๋ฅผ ๋„˜์–ด, ๋‹ค์–‘ํ•œ ์œ ๋ฐœ ์ธ์ž์™€ ๋ณต์žกํ•œ ๋ถ„์ž ๊ธฐ์ „์„ ํฌํ•จํ•˜๋Š” ์ด์งˆ์ ์ธ ์งˆํ™˜๊ตฐ์ด๋‹ค. ์ „์‹ ์„ฑ ๊ฒฝํ™”์ฆ(SSc)๊ณผ ๊ฐ™์€ ์ž๊ฐ€๋ฉด์—ญ ์งˆํ™˜์—์„œ ILD ๋ฐœ์ƒ์€ ๋ฉด์—ญ ํ™œ์„ฑํ™”, ํ˜ˆ๊ด€๋ณ‘์ฆ ๋ฐ ์ง„ํ–‰์„ฑ ์„ฌ์œ ํ™”์˜ ์ƒํ˜ธ์ž‘์šฉ ๊ฒฐ๊ณผ์ด๋ฉฐ, ์ด๋Š” '๊ฐ€์†ํ™”๋œ ๋…ธํ™”(Accelerated Aging)'์˜ ํŠน์ง•์„ ๊ณต์œ ํ•œ๋‹ค [17, 26]. COPD์™€ ์œ ์‚ฌํ•˜๊ฒŒ, SSc-ILD ํ™˜์ž์—์„œ๋„ ํ…”๋กœ๋ฏธ์–ด ๋‹จ์ถ•, ์„ธํฌ ๋…ธํ™”(cellular senescence), PI3 kinase-mTOR ์‹ ํ˜ธ ์ „๋‹ฌ ํ™œ์„ฑํ™”, ๋ฏธํ† ์ฝ˜๋“œ๋ฆฌ์•„ ๊ธฐ๋Šฅ ์žฅ์•  ๋ฐ ๋งŒ์„ฑ ์ €๋“ฑ๊ธ‰ ์—ผ์ฆ(inflammaging)์ด ๊ด€์ฐฐ๋œ๋‹ค [17]. ์ด๋Ÿฌํ•œ ๊ธฐ์ „์€ SSc์˜ ์ƒ์กด์œจ์— ์ง์ ‘์ ์ธ ์˜ํ–ฅ์„ ๋ฏธ์น˜๋ฉฐ, ์ง„๋‹จ ์‹œ์ ์˜ ์—ฐ๋ น(>60์„ธ), ํ™•์‚ฐ์„ฑ ํ”ผ๋ถ€ ๋ณ‘๋ณ€(diffuse cutaneous SSc), ํ๊ณ ํ˜ˆ์••(PH), ๊ทธ๋ฆฌ๊ณ  DLCO < 70% ๋“ฑ์ด ๋ถˆ๋Ÿ‰ํ•œ ์˜ˆํ›„์™€ ์œ ์˜ํ•˜๊ฒŒ ์—ฐ๊ด€๋œ๋‹ค [22].

๊ฐ์—ผ ์š”์ธ๊ณผ์˜ ์—ฐ๊ด€์„ฑ๋„ ์ฃผ๋ชฉํ•  ๋งŒํ•˜๋‹ค. ์‚ฌ๋ฅด์ฝ”์ด๋“œ์ฆ(Sarcoidosis)์˜ ๋ฐœ๋ณ‘ ๊ธฐ์ „์—์„œ ๋งˆ์ด์ฝ”๋ฐ•ํ…Œ๋ฆฌ์•„(Mycobacteria)์˜ ์—ญํ• ์ด ์˜ค๋žซ๋™์•ˆ ๋…ผ์Ÿ๋˜์–ด ์™”์œผ๋‚˜, ๋ฉ”ํƒ€๋ถ„์„ ๊ฒฐ๊ณผ ์‚ฌ๋ฅด์ฝ”์ด๋“œ์ฆ ํ™˜์ž์˜ ์กฐ์ง ์ƒ˜ํ”Œ์—์„œ ๋งˆ์ด์ฝ”๋ฐ•ํ…Œ๋ฆฌ์•„ ์กด์žฌ ๋น„์œจ์€ 26.4%๋กœ, ๋Œ€์กฐ๊ตฐ์— ๋น„ํ•ด ์œ ์˜ํ•˜๊ฒŒ ๋†’์•˜๋‹ค(Odds Ratio 9.67โ€“19.49) [21]. ์ด๋Š” ๊ฐ์—ผ์›์ด ์ž๊ฐ€๋ฉด์—ญ ๋ฐ˜์‘์„ ์œ ๋ฐœํ•˜๋Š” ํŠธ๋ฆฌ๊ฑฐ๊ฐ€ ๋  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•œ๋‹ค. ๋˜ํ•œ, COVID-19 ํŒฌ๋ฐ๋ฏน ์ดํ›„ SARS-CoV-2 ๊ฐ์—ผ์ด ํ ์ƒํ”ผ์„ธํฌ์™€ ์„ฌ์œ ์•„์„ธํฌ๋ฅผ ํ†ตํ•ด ์ง€์†์ ์ธ ์„ฌ์œ ํ™” ๋ณ€ํ™”๋ฅผ ์œ ๋„ํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ์ฆ๊ฑฐ๊ฐ€ ์ถ•์ ๋˜๊ณ  ์žˆ์œผ๋ฉฐ, ์ด๋Š” Post-Acute Sequelae of SARS-CoV-2 Infection (PASC)์˜ ์ค‘์š”ํ•œ ๊ตฌ์„ฑ ์š”์†Œ๋กœ ์ž๋ฆฌ ์žก๊ณ  ์žˆ๋‹ค [19, 27].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” 'Cancer-induced systemic pre-conditioning of distant organs: building a niche for metastatic cells'๊ฐ€ ํฌํ•จ๋˜์–ด ์žˆ์–ด, ์ „์‹ ์  ๋ฏธ์„ธํ™˜๊ฒฝ ๋ณ€ํ™”์— ๋Œ€ํ•œ ์ดํ•ด๋Š” ์–ด๋А ์ •๋„ ์กด์žฌํ•œ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ILD ํŠน์ด์ ์ธ '๋…ธํ™” ๊ด€๋ จ ๊ธฐ์ „(Aging mechanisms)'๊ณผ '๊ฐ์—ผ-๋ฉด์—ญ ๊ต์ฐจ(Cross-talk between infection and immunity)'์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๋ถ„์ž ๊ฒฝ๋กœ ์—ฐ๊ตฌ๋Š” ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ๋ช…์‹œ์ ์œผ๋กœ ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š๊ณ  ์žˆ๋‹ค. ํŠนํžˆ SSc-ILD์—์„œ์˜ ์„ธํฌ ๋…ธํ™” ๊ธฐ์ „๊ณผ ์‚ฌ๋ฅด์ฝ”์ด๋“œ์ฆ์—์„œ์˜ ๊ฐ์—ผ์› ์—ญํ• ์€ ๊ธฐ์กด ์•” ์ค‘์‹ฌ์˜ ์ „์‹  ์กฐ๊ฑดํ™”(pre-conditioning) ๊ฐœ๋…์„ ํ™•์žฅํ•˜์—ฌ, ILD์˜ ์ด์งˆ์„ฑ์„ ์„ค๋ช…ํ•˜๋Š” ์ƒˆ๋กœ์šด ๋ณ‘ํƒœ์ƒ๋ฆฌํ•™์  ํ”„๋ ˆ์ž„์›Œํฌ๋ฅผ ๊ตฌ์ถ•ํ•  ์ˆ˜ ์žˆ๋Š” ์ž ์žฌ๋ ฅ์„ ๊ฐ€์ง„๋‹ค.

์น˜๋ฃŒ ์ „๋žต ๋ฐ ์•ฝ๋ฌผ ๋ฐ˜์‘ (Therapeutic Strategies and Drug Response)

ILD์˜ ์น˜๋ฃŒ๋Š” ์งˆํ™˜์˜ ๊ธฐ์›๊ณผ ์ค‘์ฆ๋„์— ๋”ฐ๋ผ ๋ฉด์—ญ์–ต์ œ์ œ, ํ•ญ์„ฌ์œ ํ™”์ œ, ๊ทธ๋ฆฌ๊ณ  ์ฆ์ƒ ์กฐ์ ˆ์ œ๋ฅผ ํฌํ•จํ•œ ๋‹ค๊ฐ์ ์ธ ์ ‘๊ทผ์ด ํ•„์š”ํ•˜๋‹ค. CTD-ILD, ํŠนํžˆ SSc-ILD์˜ ์ผ์„  ์น˜๋ฃŒ์ œ๋กœ Mycophenolate Mofetil (MMF)๊ณผ Methotrexate (MTX)๊ฐ€ ๋„๋ฆฌ ์‚ฌ์šฉ๋œ๋‹ค. Scleroderma Lung Study II (SLS II)๋ฅผ ๋น„๋กฏํ•œ ์—ฌ๋Ÿฌ ์—ฐ๊ตฌ๋Š” MMF๊ฐ€ ๊ฒฝ๊ตฌ Cyclophosphamide์™€ ๋น„๊ตํ•ด ์œ ์‚ฌํ•œ ํšจ๋Šฅ์„ ๋ณด์ด์ง€๋งŒ ๋‚ด์•ฝ์„ฑ์ด ๋” ์šฐ์ˆ˜ํ•จ์„ ์ž…์ฆํ–ˆ๋‹ค [26]. ๋ฐ˜๋ฉด, MTX๋Š” ์ดˆ๊ธฐ ํ™•์‚ฐ์„ฑ ํ”ผ๋ถ€ ๋ณ‘๋ณ€์ด๋‚˜ ๊ทผ๊ณจ๊ฒฉ๊ณ„ ์ฆ์ƒ์— ์œ ์šฉํ•  ์ˆ˜ ์žˆ์œผ๋‚˜, ํ ๊ธฐ๋Šฅ ๊ฐœ์„ ์— ๋Œ€ํ•œ ์ฆ๊ฑฐ๋Š” ์ œํ•œ์ ์ด๋‹ค [26]. ๋”ฐ๋ผ์„œ ํ˜„์žฌ ๊ฐ€์ด๋“œ๋ผ์ธ์€ ์ž„์ƒ ํ‘œํ˜„ํ˜•(phenotype-driven approach)์— ๋”ฐ๋ผ MMF๋ฅผ SSc-ILD์˜ ์šฐ์„  ์น˜๋ฃŒ์ œ๋กœ ๊ถŒ์žฅํ•œ๋‹ค [26].

์ค‘์žฌ์  ์น˜๋ฃŒ ์˜ต์…˜์œผ๋กœ Tripterygium glycosides (Tg)์˜ ์‚ฌ์šฉ์ด ์ฃผ๋ชฉ๋ฐ›๊ณ  ์žˆ๋‹ค. ๋ฉ”ํƒ€๋ถ„์„ ๊ฒฐ๊ณผ, Tg๋ฅผ ๋ฉด์—ญ์–ต์ œ์ œ ๋˜๋Š” ๊ธ€๋ฃจ์ฝ”์ฝ”๋ฅดํ‹ฐ์ฝ”์ด๋“œ์™€ ๋ณ‘์šฉํ–ˆ์„ ๋•Œ ์ด ์œ ํšจ๋ฅ ์ด ๋Œ€์กฐ๊ตฐ(๋‹จ๋… ๋ฉด์—ญ์–ต์ œ์ œ/์Šคํ…Œ๋กœ์ด๋“œ)๋ณด๋‹ค ์œ ์˜ํ•˜๊ฒŒ ๋†’์•˜๋‹ค(RR 1.16; 95% CI 1.09โ€“1.24) [29]. ํŠนํžˆ Tg + ๊ธ€๋ฃจ์ฝ”์ฝ”๋ฅดํ‹ฐ์ฝ”์ด๋“œ ์กฐํ•ฉ์ด Tg + ๋ฉด์—ญ์–ต์ œ์ œ ์กฐํ•ฉ๋ณด๋‹ค ํ†ต๊ณ„์ ์œผ๋กœ ๋” ๋†’์€ ์œ ํšจ๋ฅ ์„ ๋ณด์˜€์œผ๋‚˜(P=0.03), ์ด๋Š” ์ฃผ๋กœ ์•„์‹œ์•„๊ถŒ ์—ฐ๊ตฌ์—์„œ ๋ณด๊ณ ๋œ ๊ฒฐ๊ณผ๋กœ, ์„œ์–‘ ํ™˜์ž๊ตฐ์—์„œ์˜ ์ผ๋ฐ˜ํ™” ๊ฐ€๋Šฅ์„ฑ์€ ์ถ”๊ฐ€ ๊ฒ€์ฆ์ด ํ•„์š”ํ•˜๋‹ค [29]. ๋˜ํ•œ, ์•ฝ๋ฌผ ์œ ๋ฐœ์„ฑ ๊ฐ„์งˆ์„ฑ ํ์งˆํ™˜(DIILD)์˜ ๊ฒฝ์šฐ, ์•” ์น˜๋ฃŒ์ œ, ๋ฅ˜๋งˆํ‹ฐ์Šค์•ฝ, Amiodarone, ํ•ญ์ƒ์ œ๊ฐ€ ์ฃผ์š” ์›์ธ์ด๋ฉฐ, ์‚ฌ๋ง๋ฅ ์ด 50% ์ด์ƒ์œผ๋กœ ๋งค์šฐ ๋†’๊ฒŒ ๋ณด๊ณ ๋œ๋‹ค [9]. DIILD์˜ ์น˜๋ฃŒ๋Š” ์ฃผ๋กœ ๊ธ€๋ฃจ์ฝ”์ฝ”๋ฅดํ‹ฐ์ฝ”์ด๋“œ(GCs)์— ์˜์กดํ•˜์ง€๋งŒ, ์ „ํ–ฅ์  ์—ฐ๊ตฌ๊ฐ€ ๋ถ€์กฑํ•˜์—ฌ ์ตœ์ ์˜ ์น˜๋ฃŒ ์ „๋žต์ด ํ™•๋ฆฝ๋˜์ง€ ์•Š์€ ์ƒํƒœ์ด๋‹ค [9].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” NSCLC์—์„œ์˜ ๋ฉด์—ญํ•ญ์•”์ œ(Durvalumab, Atezolizumab) ๊ด€๋ จ ๋…ผ๋ฌธ๋“ค์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด, ๋ฉด์—ญ์น˜๋ฃŒ์˜ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ์— ๋Œ€ํ•œ ๋ฐฐ๊ฒฝ ์ง€์‹์ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ILD ํŠน์ด์ ์ธ ์•ฝ๋ฌผ, ์ฆ‰ MMF, MTX, Tripterygium glycosides, ๊ทธ๋ฆฌ๊ณ  DIILD ๊ด€๋ฆฌ์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ์น˜๋ฃŒ ์•Œ๊ณ ๋ฆฌ์ฆ˜์€ ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ถ€์žฌํ•˜๋‹ค. ํŠนํžˆ CTD-ILD์—์„œ์˜ Phenotype-driven drug selection๊ณผ DIILD์˜ ๊ณ ์œ„ํ—˜๊ตฐ ์‹๋ณ„ ๋ฐ ์Šคํ…Œ๋กœ์ด๋“œ ๋ฐ˜์‘ ์˜ˆ์ธก ๋ชจ๋ธ์€ ์ž„์ƒ์ ์œผ๋กœ ์‹œ๊ธ‰ํ•œ gap์œผ๋กœ ๋‚จ์•„ ์žˆ๋‹ค.

๊ธฐ๋Šฅ์  ํ‰๊ฐ€ ๋ฐ ์‚ถ์˜ ์งˆ (Functional Assessment and Quality of Life)

ILD ํ™˜์ž์˜ ๊ด€๋ฆฌ์—์„œ ํ๊ธฐ๋Šฅ ๊ฒ€์‚ฌ(PFT)๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์ผ์ƒ์ƒํ™œ์—์„œ์˜ ๊ธฐ๋Šฅ์  ๋Šฅ๋ ฅ๊ณผ ์‹ฌ๋ฆฌ์  ์›ฐ๋น™์„ ํ‰๊ฐ€ํ•˜๋Š” ๊ฒƒ์ด ์˜ˆํ›„ ํŒ์ •์— ์ค‘์š”ํ•˜๋‹ค. 6๋ถ„ ๋ณดํ–‰ ๊ฒ€์‚ฌ(6MWT), ์ฆ๋Ÿ‰ ์…”ํ‹€ ๋ณดํ–‰ ๊ฒ€์‚ฌ(ISWT), ์ง€๊ตฌ๋ ฅ ์…”ํ‹€ ๋ณดํ–‰ ๊ฒ€์‚ฌ(ESWT)๋Š” ๋งŒ์„ฑ ํ˜ธํก๊ธฐ ์งˆํ™˜ ํ™˜์ž์˜ ๊ธฐ๋Šฅ์  ์šด๋™ ๋Šฅ๋ ฅ์„ ํ‰๊ฐ€ํ•˜๋Š” ์‹ ๋ขฐํ•  ์ˆ˜ ์žˆ๋Š” ๋„๊ตฌ์ด๋‹ค [3]. 6MWD๋Š” ํ”ผํฌ ์ž‘์—… ์šฉ๋Ÿ‰(r=0.59โ€“0.93) ๋ฐ ์‹ ์ฒด ํ™œ๋™๋Ÿ‰(r=0.40โ€“0.85)๊ณผ ๊ฐ•ํ•˜๊ฒŒ ์ƒ๊ด€๊ด€๊ณ„๊ฐ€ ์žˆ์œผ๋‚˜, ํ๊ธฐ๋Šฅ(FVC, DLCO)๊ณผ๋Š” ์ƒ๋Œ€์ ์œผ๋กœ ์•ฝํ•œ ์ƒ๊ด€๊ด€๊ณ„๋ฅผ ๋ณด์ธ๋‹ค(r=0.10โ€“0.59) [3]. ์ด๋Š” 6MWT๊ฐ€ ๋‹จ์ˆœํ•œ ํ๊ธฐ๋Šฅ ์ง€ํ‘œ ์ด์ƒ์˜ ์ „์‹ ์  ๊ธฐ๋Šฅ ์ƒํƒœ๋ฅผ ๋ฐ˜์˜ํ•จ์„ ์˜๋ฏธํ•˜๋ฉฐ, ์šด๋™ ํ›ˆ๋ จ ๊ฐœ์ž…์— ๋Œ€ํ•œ ๋ฐ˜์‘์„ฑ(responsiveness)๋„ ๋†’๊ฒŒ ๋ณด๊ณ ๋˜์—ˆ๋‹ค [3].

์‹ฌ๋ฆฌ์  ํ•ฉ๋ณ‘์ฆ์€ ILD ํ™˜์ž์—์„œ ํ”ํžˆ ๊ฐ„๊ณผ๋˜์ง€๋งŒ ์˜ˆํ›„์— ์ค‘๋Œ€ํ•œ ์˜ํ–ฅ์„ ๋ฏธ์นœ๋‹ค. ๋ถˆ์•ˆ๊ณผ ์šฐ์šธ์ฆ์€ ILD ํ•˜์œ„ ์œ ํ˜• ์ „๋ฐ˜์—์„œ ๋†’์€ ์œ ๋ณ‘๋ฅ ์„ ๋ณด์ด๋ฉฐ, ์ด๋Š” ๋งŒ์„ฑ ์ €์‚ฐ์†Œ์ฆ, ๊ธฐ๋Šฅ์  ์žฅ์• , ์น˜๋ฃŒ ๊ด€๋ จ ์ŠคํŠธ๋ ˆ์Šค ๋ฐ ์ƒ๋ฌผํ•™์  ์š”์ธ(์œ ์ „์  ๊ฐ์ˆ˜์„ฑ ๋“ฑ)์˜ ๋ณตํ•ฉ์  ์ƒํ˜ธ์ž‘์šฉ์œผ๋กœ ๋ฐœ์ƒํ•œ๋‹ค [31]. ์ •์‹ ๊ณผ์  ํ•ฉ๋ณ‘์ฆ์ด ์žˆ๋Š” ํ™˜์ž๋Š” ์ฆ์ƒ ์•…ํ™”, ์‹ ์ฒด ํ™œ๋™ ๊ฐ์†Œ, ๊ฑด๊ฐ• ๊ด€๋ จ ์‚ถ์˜ ์งˆ(HRQoL) ์ €ํ•˜์™€ ์œ ์˜ํ•˜๊ฒŒ ์—ฐ๊ด€๋˜์–ด ์žˆ๋‹ค [31]. ๋˜ํ•œ, ILD ํ™˜์ž์—์„œ ์‚ฌ๋ฅด์ฝ”ํŽ˜๋‹ˆ์•„(Sarcopenia)์˜ ์œ ๋ณ‘๋ฅ ์€ 26.8%๋กœ, ๋น„์งˆํ™˜ ๋Œ€์กฐ๊ตฐ(13.3%)์— ๋น„ํ•ด ํ˜„์ €ํžˆ ๋†’์œผ๋ฉฐ, ์ด๋Š” ํ˜ธํก๊ธฐ ์งˆํ™˜๊ณผ ๋…ธํ™” ๊ด€๋ จ ๊ทผ์œก ๊ฐ์†Œ์ฆ์˜ ๋ฐ€์ ‘ํ•œ ์—ฐ๊ด€์„ฑ์„ ๋ณด์—ฌ์ค€๋‹ค [13].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” 'The Eighth Edition AJCC Cancer Staging Manual' ๋“ฑ ์•” ๋ถ„๋ฅ˜ ๋ฐ ์Šคํ…Œ์ด์ง•์— ๊ด€ํ•œ ๋ฌธ์„œ๊ฐ€ ์žˆ์œผ๋‚˜, ILD ํ™˜์ž์˜ ๊ธฐ๋Šฅ์  ํ‰๊ฐ€(6MWT, ISWT)๋‚˜ ์‹ฌ๋ฆฌ์  ์›ฐ๋น™, ์‚ฌ๋ฅด์ฝ”ํŽ˜๋‹ˆ์•„์™€ ๊ฐ™์€ ์ „์‹ ์  ํ•ฉ๋ณ‘์ฆ์— ๋Œ€ํ•œ ์ž๋ฃŒ๋Š” ํฌํ•จ๋˜์–ด ์žˆ์ง€ ์•Š๋‹ค. ํŠนํžˆ 'Psychiatric comorbidities in patients with interstitial lung diseases' [31]์™€ 'Prevalence of sarcopenia as a comorbid disease' [13]์—์„œ ์ œ์‹œ๋œ ์ฆ๊ฑฐ๋“ค์€ ILD ๊ด€๋ฆฌ๊ฐ€ ํ ๊ธฐ๋Šฅ ๊ฐœ์„ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ๋‹คํ•™์ œ์  ์ ‘๊ทผ(์žฌํ™œ, ์ •์‹ ๊ฑด๊ฐ•, ์˜์–‘)์„ ํ•„์š”๋กœ ํ•จ์„ ๊ฐ•์กฐํ•œ๋‹ค. ๋”ฐ๋ผ์„œ ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ง„๋‹จ ๋ฐ ์น˜๋ฃŒ ์ค‘์‹ฌ์ด๋ฉฐ, ํ™˜์ž์˜ ์ „์ธ์  ๊ฑด๊ฐ•(Holistic health)๊ณผ ๊ธฐ๋Šฅ์  ์˜ˆํ›„ ์ธ์ž์— ๋Œ€ํ•œ ํ‰๊ฐ€ ๋„๊ตฌ๊ฐ€ ๊ฒฐ์—ฌ๋˜์–ด ์žˆ๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1An updated roadmap for the integration of metalโ€“organic frameworks with electronic devices and chemical sensorsChemical Society Reviews ยท 20171226PubMed ยท DOI
2A systematic review including meta-analysis of work environment and depressive symptomsBMC Public Health ยท 2015990PubMed ยท DOI
3An official systematic review of the European Respiratory Society/American Thoracic Society: measurement properties of field walking tests in chronic respiratory diseaseEuropean Respiratory Journal ยท 2014956PubMed ยท DOI
4Global incidence and mortality of idiopathic pulmonary fibrosis: a systematic reviewEuropean Respiratory Journal ยท 2015952PubMed ยท DOI
5Systematic Review of Psychosocial Factors at Work and Private Life as Risk Factors for Back PainSpine ยท 2000916PubMed ยท DOI
6Lipid Droplet BiogenesisAnnual Review of Cell and Developmental Biology ยท 2017799PubMed ยท DOI
7Replicability, Robustness, and Reproducibility in Psychological ScienceAnnual Review of Psychology ยท 2021767PubMed ยท DOI
8A systematic review of the use of opioids in the management of dyspnoeaThorax ยท 2002644PubMed ยท DOI
9Drug-Induced Interstitial Lung Disease: A Systematic ReviewJournal of Clinical Medicine ยท 2018426PubMed ยท DOI
10Update on Prevalence of Pain in Patients with Cancer 2022: A Systematic Literature Review and Meta-AnalysisCancers ยท 2023425PubMed ยท DOI
11Dynamics of Well-BeingAnnual Review of Organizational Psychology and Organizational Behavior ยท 2015402DOI
12Safety and Diagnostic Yield of Transbronchial Lung Cryobiopsy in Diffuse Parenchymal Lung Diseases: A Comparative Study versus Video-Assisted Thoracoscopic Lung Biopsy and a Systematic Review of the LiteratureRespiration ยท 2016381PubMed ยท DOI
13Prevalence of sarcopenia as a comorbid disease: A systematic review and meta-analysisExperimental Gerontology ยท 2019336PubMed ยท DOI
14A Gold Standard Publication Checklist to Improve the Quality of Animal Studies, to Fully Integrate the Three Rs, and to Make Systematic Reviews More FeasibleAlternatives to Laboratory Animals ยท 2010314PubMed ยท DOI
15Delivering safer immunotherapies for cancerAdvanced Drug Delivery Reviews ยท 2017306PubMed ยท DOI
16Impact of clinical registries on quality of patient care and clinical outcomes: A systematic reviewPLoS ONE ยท 2017304PubMed ยท DOI
17Senescence in COPD and Its ComorbiditiesAnnual Review of Physiology ยท 2016269PubMed ยท DOI
18Pulmonary fibrosis: from pathogenesis to clinical decision-makingTrends in Molecular Medicine ยท 2023249PubMed ยท DOI
19COVIDโ€19 and pulmonary fibrosis: A potential role for lung epithelial cells and fibroblastsImmunological Reviews ยท 2021243PubMed ยท DOI
20Prevalence, imaging patterns and risk factors of interstitial lung disease in connective tissue disease: a systematic review and meta-analysisEuropean Respiratory Review ยท 2023236PubMed ยท DOI
21Molecular evidence for the role of mycobacteria in sarcoidosis: a meta-analysisEuropean Respiratory Journal ยท 2007236PubMed ยท DOI
22Survival and prognosis factors in systemic sclerosis: data of a French multicenter cohort, systematic review, and meta-analysis of the literatureArthritis Research & Therapy ยท 2019222PubMed ยท DOI
23Innovative pedagogical principles and technological tools capabilities for immersive blended learning: a systematic literature reviewEducation and Information Technologies ยท 2022221PubMed ยท DOI
24Systematic review and economic modelling of effectiveness and cost utility of surgical treatments for men with benign prostatic enlargementHealth Technology Assessment ยท 2008221PubMed ยท DOI
25Diagnostic yield and safety of small versus standard cryoprobes in transbronchial lung cryobiopsy for interstitial lung disease: a systematic review and meta-analysisEgyptian Journal of Bronchology ยท 20260DOI
26Methotrexate Versus Mycophenolate Mofetil as First-Line Therapy in Systemic Sclerosis: Evidence from Clinical Trials and Real-World Studiesโ€”A Narrative ReviewSclerosis ยท 20260DOI
27Pulmonary sequelae of SARS-CoV-2 infection: a narrative review of long-term functional, radiological, and clinical outcomes in COVID-19 survivors, with a regional focus on Saudi ArabiaFrontiers in Physiology ยท 20260DOI
28School staff experiences and perceptions of roles, responsibilities, skills and training of teaching assistants supporting students with disabilities in Australian schools: A systematic reviewReview of Education ยท 20260DOI
29Meta-analysis of tripterygium glycosides combined with immunosuppressants or glucocorticoids in the treatment of various types of connective tissue disease-interstitial lung diseaseLetters in Drug Design & Discovery ยท 20260DOI
30The Multiorgan Complications of Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis: A Systematic ReviewClinical Cosmetic and Investigational Dermatology ยท 20260PubMed ยท DOI
31Psychiatric comorbidities in patients with interstitial lung diseases: A narrative review of epidemiology, clinical impact, and integrated care strategiesTherapeutic Advances in Respiratory Disease ยท 20260DOI
32Factors Associated with Systemic Lupus Erythematosus-Associated Interstitial Lung Disease and Clinical Outcomes: A Systematic Review and Meta-AnalysisMedicina ยท 20260DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1Interstitial Lung Disease: A Review.JAMA ยท 2024-May-21PubMed ยท DOI
2Current State of Fibrotic Interstitial Lung Disease Imaging.Radiology ยท 2025-JulPubMed ยท DOI
3Etiology and Pathogenesis of Rheumatoid Arthritis-Interstitial Lung Disease.International journal of molecular sciences ยท 2023-Sep-25PubMed ยท DOI
4Identification, Monitoring, and Management of Rheumatoid Arthritis-Associated Interstitial Lung Disease.Arthritis & rheumatology (Hoboken, N.J.) ยท 2023-DecPubMed ยท DOI
5Systemic autoimmune rheumatic diseases-associated interstitial lung disease: a pulmonologist's perspective.Breathe (Sheffield, England) ยท 2025-AprPubMed ยท DOI
6Macrophages as determinants and regulators of systemic sclerosis-related interstitial lung disease.Journal of translational medicine ยท 2024-Jun-27PubMed ยท DOI
7Treatment of rheumatoid arthritis-associated interstitial lung disease: An appraisal of the 2023 ACR/CHEST guideline.Current treatment options in rheumatology ยท 2024-DecPubMed ยท DOI
8The Incidence of Abemaciclib-induced Interstitial Lung Disease: A Single-center Retrospective Study in Japan.The Kobe journal of medical sciences ยท 2025-Aug-07PubMed ยท DOI
9A screening tool to detect interstitial lung disease in systemic sclerosis: the ILD-RISC score.Rheumatology (Oxford, England) ยท 2025-Dec-01PubMed ยท DOI
10Naringin nanoparticles alleviate RA-ILD pulmonary fibrosis by targeting 14-3-3ฮถ to inhibit LYVE1+ macrophages TGF-ฮฒ1 secretion.Materials today. Bio ยท 2026-AprPubMed ยท DOI
Interstitial lung disease โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 8ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
์ง„๋‹จ ๋ฐ ์˜์ƒ ๋ถ„์„ ๊ธฐ์ˆ ์˜ ์ง„ํ™”

Interstitial lung disease (ILD)์˜ ์ •ํ™•ํ•œ ์ง„๋‹จ์€ ์ž„์ƒ์  ๊ฒฝ๊ณผ์™€ ์น˜๋ฃŒ ์„ ํƒ์— ๊ฒฐ์ •์ ์ธ ์—ญํ• ์„ ํ•˜๋ฉฐ, ์ตœ๊ทผ์—๋Š” ์นจ์Šต์  ์ƒ๊ฒ€์˜ ๋Œ€์•ˆ์œผ๋กœ ๋น„์นจ์Šต์  ์˜์ƒ ๊ธฐ์ˆ ๊ณผ ๋ฏธ์„ธ์นจ์Šต์  ์‹œ์ˆ ์ด ๋น ๋ฅด๊ฒŒ ๋ฐœ์ „ํ•˜๊ณ  ์žˆ๋‹ค. ์ „ํ†ต์ ์œผ๋กœ ํ‰๋ถ€ X์„ (CXR)์ด ํ๋ ด ๋“ฑ์˜ ํ˜ธํก๊ธฐ ์งˆํ™˜ ์ง„๋‹จ ํ‘œ์ค€์ด์—ˆ์œผ๋‚˜, ๋ฐฉ์‚ฌ์„  ๋…ธ์ถœ๊ณผ ์ง„๋‹จ ๋ณ€๋™์„ฑ ๋ฌธ์ œ๋กœ ์ธํ•ด Point-of-care ultrasound (POCUS)์— ๋Œ€ํ•œ ๊ด€์‹ฌ์ด ๋†’์•„์ง€๊ณ  ์žˆ๋‹ค. ์†Œ์•„ ํ™˜์ž์—์„œ์˜ community-acquired pneumonia (CAP) ์ง„๋‹จ ์ •ํ™•๋„๋ฅผ ๋น„๊ตํ•œ ์ฒด๊ณ„์  ๋ฌธํ—Œ๊ณ ์ฐฐ ๋ฐ ๋ฉ”ํƒ€๋ถ„์„ ๊ฒฐ๊ณผ, POCUS๋Š” ๋ฏผ๊ฐ๋„ 91% (95% CI: 89โ€“93%)์™€ ํŠน์ด๋„ 86% (95% CI: 83โ€“88%)๋ฅผ ๋ณด์˜€์œผ๋ฉฐ, ์ด๋Š” CXR์˜ ๋ฏผ๊ฐ๋„ 88%์™€ ํŠน์ด๋„ 76%๋ณด๋‹ค ์šฐ์ˆ˜ํ•œ ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ๋‹ค [25]. ํŠนํžˆ SROC ๋ถ„์„์—์„œ POCUS์˜ AUC๋Š” 0.9569๋กœ CXR์˜ 0.9291์„ ์ƒํšŒํ•˜์—ฌ, ํ ์‹ค์งˆ ์‘๊ณ (pulmonary consolidation)์™€ ํ‰๋ง‰ ์‚ผ์ถœ์•ก(pleural effusion) ๊ฒ€์ถœ์—์„œ ๋” ์ผ๊ด€๋œ ์„ฑ๋Šฅ์„ ์ž…์ฆํ–ˆ๋‹ค [25]. ์ด๋Ÿฌํ•œ ๋น„์นจ์Šต์  ์˜์ƒ ๊ธฐ์ˆ ์˜ ์ •๋ฐ€๋„ ํ–ฅ์ƒ์€ ILD ์ดˆ๊ธฐ ์Šคํฌ๋ฆฌ๋‹ ๋ฐ ๋ชจ๋‹ˆํ„ฐ๋ง์— ์ค‘์š”ํ•œ ์‹œ์‚ฌ์ ์„ ์ œ๊ณตํ•œ๋‹ค.

ํ•œํŽธ, ์กฐ์งํ•™์  ์ง„๋‹จ์ด ํ•„์ˆ˜์ ์ธ ILD ํ™˜์ž๋“ค์—๊ฒŒ Transbronchial lung cryobiopsy (TBLC)๋Š” ์ˆ˜์ˆ ์  ํ ์ƒ๊ฒ€(surgical lung biopsy)๋ณด๋‹ค ๋œ ์นจ์Šต์ ์ธ ๋Œ€์•ˆ์œผ๋กœ ์ž๋ฆฌ ์žก๊ณ  ์žˆ๋‹ค. ์ตœ๊ทผ ์—ฐ๊ตฌ์—์„œ๋Š” TBLC ์ˆ˜ํ–‰ ์‹œ ์‚ฌ์šฉ๋˜๋Š” cryoprobe์˜ ์ง๊ฒฝ(์†Œํ˜• vs ํ‘œ์ค€ํ˜•)์— ๋”ฐ๋ฅธ ์ง„๋‹จ ์ˆ˜์œจ(diagnostic yield)๊ณผ ์•ˆ์ „์„ฑ ์ฐจ์ด๋ฅผ ํ‰๊ฐ€ํ–ˆ๋‹ค. 5๊ฐœ ์—ฐ๊ตฌ(232๋ช…)๋ฅผ ํฌํ•จํ•œ ๋ฉ”ํƒ€๋ถ„์„ ๊ฒฐ๊ณผ, ์†Œํ˜• probe (1.1โ€“1.7 mm)์™€ ํ‘œ์ค€ probe (1.9 mm) ๊ฐ„์˜ ์ง„๋‹จ ์ˆ˜์œจ์—๋Š” ํ†ต๊ณ„์ ์œผ๋กœ ์œ ์˜ํ•œ ์ฐจ์ด๊ฐ€ ์—†์—ˆ๋‹ค (RR 0.95; 95% CI 0.88 to 1.03; P = 0.212) [31]. ๋˜ํ•œ ์ถœํ˜ˆ(RR 0.67)๊ณผ ๊ธฐํ‰(pneumothorax, RR 1.44) ๋ฐœ์ƒ๋ฅ ์—์„œ๋„ ์œ ์˜ํ•œ ์ฐจ์ด๊ฐ€ ๊ด€์ฐฐ๋˜์ง€ ์•Š์•„, ์†Œํ˜• probe๊ฐ€ ์ง„๋‹จ ์ •ํ™•๋„๋ฅผ ํฌ์ƒํ•˜์ง€ ์•Š์œผ๋ฉด์„œ๋„ ์•ˆ์ „์„ฑ์„ ์œ ์ง€ํ•  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•œ๋‹ค [31]. pooled single-arm yield๋Š” 88.3% (95% CI 78.2 to 94.1)๋กœ ๋ณด๊ณ ๋˜์—ˆ๋‹ค [31].

์˜์ƒ ๋ถ„์„ ๋ถ„์•ผ์—์„œ๋Š” Deep learning ๊ธฐ์ˆ ์ด ์˜๋ฃŒ ์˜์ƒ ํ•ด์„์˜ ์ •๋ฐ€๋„๋ฅผ ํ˜์‹ ์ ์œผ๋กœ ๋†’์ด๊ณ  ์žˆ๋‹ค. ๊ณ„์ธต์  ํŠน์ง• ํ‘œํ˜„(hierarchical feature representations)์„ ํ†ตํ•ด ๋ฐ์ดํ„ฐ ์ž์ฒด์—์„œ ํ•™์Šต๋œ ํŒจํ„ด์„ ์‹๋ณ„ํ•˜๋Š” deep learning์€ ์กฐ์ง ๋ถ„ํ• (tissue segmentation), ์งˆ๋ณ‘ ์ง„๋‹จ ๋ฐ ์˜ˆํ›„ ์˜ˆ์ธก ๋“ฑ์—์„œ state-of-the-art ์„ฑ๋Šฅ์„ ๋ณด์ด๊ณ  ์žˆ๋‹ค [3]. ํŠนํžˆ ILD ๋ถ„์•ผ์—์„œ๋Š” Deep learning ๊ธฐ๋ฐ˜์˜ CT ์ •๋Ÿ‰ํ™” ์ ‘๊ทผ๋ฒ•์ด progressive fibrosing interstitial lung disease์˜ ์œ„ํ—˜ ์š”์ธ๊ณผ ์˜ˆํ›„ ์ง€ํ‘œ ๋ถ„์„์— ํ™œ์šฉ๋˜๊ณ  ์žˆ์œผ๋ฉฐ, ์ด๋Š” ์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์ด๋ฏธ ์กด์žฌํ•˜๋Š” "Risk factors and prognostic indicators for progressive fibrosing interstitial lung disease: a deep learning-based CT quantification approach" ๋ฐ "park-2026-quantitative-ct-imaging-in-progressive-pulmonary-fibrosis" ๋…ผ๋ฌธ๋“ค๊ณผ ๋งฅ์„ ๊ฐ™์ด ํ•œ๋‹ค. ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์  ์ง„๋ณด๋Š” ILD์˜ ์ด์งˆ์ ์ธ ๋ณ‘๋ฆฌ ์ƒ๋ฆฌ๋ฅผ ์ •๋Ÿ‰์ ์œผ๋กœ ํ‰๊ฐ€ํ•˜๊ณ , ๊ฐœ์ธํ™”๋œ ์น˜๋ฃŒ ์ „๋žต ์ˆ˜๋ฆฝ์— ๊ธฐ์—ฌํ•  ๊ฒƒ์œผ๋กœ ๊ธฐ๋Œ€๋œ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” Deep learning ๊ธฐ๋ฐ˜ CT ์ •๋Ÿ‰ํ™”๋ฅผ ํ†ตํ•œ ์ง„ํ–‰์„ฑ ํ์„ฌ์œ ์ฆ ์œ„ํ—˜ ์š”์ธ ๋ถ„์„ ๋…ผ๋ฌธ์ด ์ด๋ฏธ ํฌํ•จ๋˜์–ด ์žˆ์–ด, ์˜์ƒ ์ง„๋‹จ์˜ ์ •๋Ÿ‰ํ™” ์ธก๋ฉด์—์„œ๋Š” ๋น„๊ต์  ์ž˜ ๊ตฌ์ถ•๋˜์–ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ POCUS์™€ ๊ฐ™์€ ์ดˆ์ŒํŒŒ ๊ธฐ๋ฐ˜ ๋น„์นจ์Šต์  ์ง„๋‹จ์˜ ILD ์ ์šฉ ๊ฐ€๋Šฅ์„ฑ, ๋˜๋Š” TBLC probe ํฌ๊ธฐ ์„ ํƒ์— ๋Œ€ํ•œ ์ตœ์‹  ์ž„์ƒ ๊ฐ€์ด๋“œ๋ผ์ธ ๋ฐ˜์˜์€ ์•„์ง ๋ถ€์กฑํ•˜๋ฏ€๋กœ, ์ด ๋ถ€๋ถ„์˜ ๋ฌธํ—Œ์„ ์ถ”๊ฐ€ํ•˜๋ฉด ์ง„๋‹จ ์•Œ๊ณ ๋ฆฌ์ฆ˜์˜ ์™„์„ฑ๋„๊ฐ€ ๋†’์•„์งˆ ๊ฒƒ์ด๋‹ค.

๋ฉด์—ญ ์กฐ์ ˆ ๋ฐ ์ „์‹  ์น˜๋ฃŒ์˜ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ

ILD์™€ ๊ด€๋ จ๋œ ์ „์‹  ์น˜๋ฃŒ, ํŠนํžˆ ๋ฉด์—ญ ๊ด€๋ฌธ ์–ต์ œ์ œ(Immune Checkpoint Inhibitors, ICIs)์˜ ์‚ฌ์šฉ ์ฆ๊ฐ€๋กœ ์ธํ•ด ๋ฉด์—ญ ๊ด€๋ จ ๋ถ€์ž‘์šฉ(irAEs)์— ๋Œ€ํ•œ ์ดํ•ด๊ฐ€ ํ•„์ˆ˜์ ์ด๋‹ค. ICIs๋Š” melanoma ๋“ฑ ๋‹ค์–‘ํ•œ ์•…์„ฑ ์ข…์–‘ ์น˜๋ฃŒ์—์„œ ํ™˜์ž์˜ ๋ฉด์—ญ ์‹œ์Šคํ…œ์„ ํ™œ์„ฑํ™”ํ•˜์—ฌ ์ข…์–‘์„ ๊ณต๊ฒฉํ•˜๋Š” ํ˜์‹ ์ ์ธ ์น˜๋ฃŒ๋ฒ•์œผ๋กœ ์ž๋ฆฌ ์žก์•˜์œผ๋‚˜, ์‹ ์žฅ ๋…์„ฑ(renal toxicities)๊ณผ ๊ฐ™์€ ์‹ฌ๊ฐํ•œ ๋ถ€์ž‘์šฉ์„ ์œ ๋ฐœํ•  ์ˆ˜ ์žˆ๋‹ค [1]. Ipilimumab (anti-CTLA-4), pembrolizumab, nivolumab (anti-PD-1) ๋“ฑ ์ฃผ์š” ICIs์— ๋Œ€ํ•œ ๋ฌธํ—Œ ๊ฒ€ํ†  ๋ฐ FDA ๋ณด๊ณ  ๋ฐ์ดํ„ฐ ๋ถ„์„ ๊ฒฐ๊ณผ, ๊ธ‰์„ฑ ๊ฐ„์งˆ์„ฑ ์‹ ์—ผ(Acute interstitial nephritis, AIN), podocytopathy, ์ €๋‚˜ํŠธ๋ฅจํ˜ˆ์ฆ(hyponatremia) ๋“ฑ์ด ์ฃผ์š” ์‹ ์žฅ ๋ถ€์ž‘์šฉ์œผ๋กœ ํ™•์ธ๋˜์—ˆ๋‹ค [1]. ํŠนํžˆ PD-1 ์–ต์ œ์ œ ๊ด€๋ จ ์‹ ์žฅ ์†์ƒ์€ CTLA-4 ๊ธธํ•ญ์ œ์— ๋น„ํ•ด ๋ฐœ์ƒ ์‹œ๊ธฐ๊ฐ€ ๋Šฆ์€ ํŽธ(3-10๊ฐœ์›” vs 2-3๊ฐœ์›”)์ด๋ฉฐ, ์ด์‹ ํ™˜์ž์—์„œ ์‹ ์žฅ ๊ฑฐ๋ถ€ ๋ฐ˜์‘(kidney rejection)๊ณผ๋„ ์—ฐ๊ด€๋˜์–ด ์žˆ๋‹ค [1]. ์ดˆ๊ธฐ์—๋Š” ํฌ๊ท€ํ•˜๋‹ค๊ณ  ์—ฌ๊ฒจ์กŒ์œผ๋‚˜, ์ตœ๊ทผ ์—ฐ๊ตฌ์—์„œ๋Š” ์‹ ์žฅ ๋…์„ฑ์˜ ๋ฐœ์ƒ๋ฅ ์ด 9.9%์—์„œ 29%๊นŒ์ง€ ๋†’์„ ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•˜๊ณ  ์žˆ์–ด, ICIs ์น˜๋ฃŒ ์ค‘ ์‹ ๊ธฐ๋Šฅ ๋ชจ๋‹ˆํ„ฐ๋ง์˜ ์ค‘์š”์„ฑ์ด ๊ฐ•์กฐ๋œ๋‹ค [1].

๋ฉด์—ญ ๋ฐ˜์‘๊ณผ ์—ผ์ฆ ์กฐ์ ˆ์€ ILD์˜ ๋ณ‘๋ฆฌ ์ƒ๋ฆฌ์™€๋„ ๋ฐ€์ ‘ํ•œ ์—ฐ๊ด€์ด ์žˆ๋‹ค. Chemokines ๋ฐ ๊ทธ ์ˆ˜์šฉ์ฒด๋Š” ๋ฉด์—ญ ์„ธํฌ์˜ ์ด๋™๊ณผ ์œ„์น˜ ๊ฒฐ์ •์„ ํ†ต์ œํ•˜๋ฉฐ, ๊ธ‰์„ฑ ์—ผ์ฆ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์ ์‘์„ฑ ๋ฉด์—ญ ๋ฐ˜์‘(priming of naive T cells, effector/memory cell differentiation)์—๋„ ํ•ต์‹ฌ์ ์ธ ์—ญํ• ์„ ํ•œ๋‹ค [8]. ๋˜ํ•œ, ์ˆ˜๋ฉด๊ณผ ๋ฉด์—ญ ์‹œ์Šคํ…œ ๊ฐ„์˜ ์–‘๋ฐฉํ–ฅ ์ƒํ˜ธ์ž‘์šฉ(sleep-immune crosstalk)์€ ๊ฑด๊ฐ•๊ณผ ์งˆ๋ณ‘ ๋ชจ๋‘์—์„œ ์ค‘์š”ํ•œ ์กฐ์ ˆ ๊ธฐ์ „์ด๋‹ค. ๊ฐ์—ผ ์‹œ ๋ฉด์—ญ ํ™œ์„ฑํ™”๋Š” ์ˆ˜๋ฉด ์ง€์† ์‹œ๊ฐ„๊ณผ ๊ฐ•๋„๋ฅผ ์ฆ๊ฐ€์‹œํ‚ค๋ฉฐ, ์ด๋Š” ์ˆ™์ฃผ ๋ฐฉ์–ด(host defense)๋ฅผ ๊ฐ•ํ™”ํ•˜๋Š” ํ”ผ๋“œ๋ฐฑ ๋ฉ”์ปค๋‹ˆ์ฆ˜์œผ๋กœ ์ž‘์šฉํ•œ๋‹ค [11]. ๋ฐ˜๋ฉด, ๋งŒ์„ฑ์ ์ธ ์ˆ˜๋ฉด ๋ถ€์กฑ์€ ์ „์‹ ์„ฑ ์ €๋“ฑ๊ธ‰ ์—ผ์ฆ(chronic systemic low-grade inflammation)์„ ์œ ๋ฐœํ•˜์—ฌ ๋‹น๋‡จ๋ณ‘, ๋™๋งฅ๊ฒฝํ™”์ฆ, ์‹ ๊ฒฝํ‡ดํ–‰์„ฑ ์งˆํ™˜ ๋“ฑ ์—ผ์ฆ ์„ฑ๋ถ„์ด ๊ด€๋ จ๋œ ๋‹ค์–‘ํ•œ ์งˆํ™˜์˜ ์œ„ํ—˜์„ ๋†’์ธ๋‹ค [11]. ILD ํ™˜์ž์—์„œ ์ด๋Ÿฌํ•œ ๋ฉด์—ญ-์ˆ˜๋ฉด ์ถ•์˜ ๋ถˆ๊ท ํ˜•์ด ํ ์„ฌ์œ ํ™” ์ง„ํ–‰์— ๋ฏธ์น˜๋Š” ์˜ํ–ฅ์€ ์ถ”๊ฐ€ ์—ฐ๊ตฌ๊ฐ€ ํ•„์š”ํ•œ ์˜์—ญ์ด๋‹ค.

๋˜ํ•œ, Sepsis์™€ Acute respiratory distress syndrome (ARDS)๊ณผ ๊ฐ™์€ ์ค‘์ฆ ํ˜ธํก๊ธฐ ๋ฐ ์ „์‹  ์—ผ์ฆ ์ƒํƒœ์—์„œ๋Š” Extracorporeal blood purification techniques (ECCOโ‚‚R, CytoSorbยฎ hemoadsorption)์˜ ์—ญํ• ์ด ์ฃผ๋ชฉ๋ฐ›๊ณ  ์žˆ๋‹ค. ECCOโ‚‚R์€ ์ด์‚ฐํ™”ํƒ„์†Œ ์ œ๊ฑฐ๋ฅผ ์šฉ์ดํ•˜๊ฒŒ ํ•˜๊ณ , CytoSorb๋Š” ์ˆœํ™˜ํ•˜๋Š” ์‚ฌ์ดํ† ์นด์ธ๊ณผ ์—ผ์ฆ ๋งค๊ฐœ์ฒด๋ฅผ ์ œ๊ฑฐํ•จ์œผ๋กœ์จ ์—ผ์ฆ ๊ท ํ˜•์„ ์กฐ์ ˆํ•œ๋‹ค [26]. ๊ทธ๋Ÿฌ๋‚˜ ํ˜„์žฌ ์ž„์ƒ ์‹œํ—˜์—์„œ๋Š” ์‚ฌ๋ง๋ฅ  ๊ฐœ์„ ์— ์žˆ์–ด ์œ ์˜๋ฏธํ•œ ์ฐจ์ด๋ฅผ ๋ณด์ด์ง€ ์•Š์•˜์œผ๋ฉฐ, ์ด๋Š” ํ™˜์ž ์ด์งˆ์„ฑ(patient heterogeneity)๊ณผ phenotype-based selection์˜ ๋ถ€์žฌ๋กœ ์„ค๋ช…๋œ๋‹ค [26]. ECCOโ‚‚R์€ ๊ณ ํƒ„์‚ฐํ˜ˆ์ฆ(hypercapnia)์ด ์ฃผ๋œ ํŠน์ง•์ธ ARDS phenotype์—์„œ, CytoSorb๋Š” ๊ณผ์—ผ์ฆ ์ƒํƒœ(hyperinflammatory state)๋ฅผ ๋ณด์ด๋Š” ํ™˜์ž์—์„œ ๋” ํšจ๊ณผ์ ์ผ ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฐ€์„ค์ด ์ œ์‹œ๋˜๊ณ  ์žˆ๋‹ค [26]. ์ด๋Ÿฌํ•œ ์ •๋ฐ€ ์˜ํ•™ ์ ‘๊ทผ๋ฒ•์€ ILD๋กœ ์ธํ•œ ํ˜ธํก๋ถ€์ „์ด๋‚˜ ์ „์‹  ์—ผ์ฆ ๋ฐ˜์‘์ด ์‹ฌํ•œ ํ™˜์ž ๊ด€๋ฆฌ์—๋„ ์ ์šฉ๋  ๊ฐ€๋Šฅ์„ฑ์ด ์žˆ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” NSCLC ์น˜๋ฃŒ์—์„œ durvalizumab์˜ ํšจ๋Šฅ๊ณผ immune-related adverse events (irAEs)์˜ ์—ฐ๊ด€์„ฑ์„ ๋‹ค๋ฃฌ ๋…ผ๋ฌธ์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด, ICIs ๊ด€๋ จ ๋ถ€์ž‘์šฉ์— ๋Œ€ํ•œ ์ผ๋ถ€ ์ดํ•ด๊ฐ€ ์กด์žฌํ•œ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ILD ํŠน์ด์ ์ธ ์‹ ์žฅ ๋…์„ฑ(renal toxicity) ๊ด€๋ฆฌ๋‚˜, Sepsis/ARDS์—์„œ์˜ extracorporeal purification ๊ธฐ์ˆ  ์ ์šฉ์— ๋Œ€ํ•œ ๋ฌธํ—Œ์€ ๋ถ€์กฑํ•˜๋ฏ€๋กœ, ์ „์‹  ์น˜๋ฃŒ์˜ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ ๋ฐ ์ค‘์ฆ ํ˜ธํก๊ธฐ ํ•ฉ๋ณ‘์ฆ ๋Œ€์‘ ์ „๋žต์„ ๋ณด๊ฐ•ํ•  ํ•„์š”๊ฐ€ ์žˆ๋‹ค.

์„ฌ์œ ํ™” ๊ธฐ์ „ ๋ฐ ๋ฏธ์„ธํ™˜๊ฒฝ์˜ ๋ณ‘๋ฆฌ์ƒ๋ฆฌํ•™์  ์ดํ•ด

ILD์˜ ํ•ต์‹ฌ ๋ณ‘๋ฆฌ ๊ณผ์ •์ธ ํ ์„ฌ์œ ํ™”(pulmonary fibrosis)๋Š” ์‹ ์žฅ ์„ฌ์œ ํ™”(renal fibrosis)์™€ ์œ ์‚ฌํ•œ ์„ธํฌ ๋ฐ ๋ถ„์ž ๊ธฐ์ „์„ ๊ณต์œ ํ•œ๋‹ค. Tubulointerstitial fibrosis๋Š” ๋…ธํ™”์™€ ๋งŒ์„ฑ ์‹ ์žฅ ์งˆํ™˜(CKD)์—์„œ ์ง„ํ–‰์„ฑ์œผ๋กœ ๋ฐœ์ƒํ•˜๋ฉฐ, 70์„ธ ์ด์ƒ ์„ฑ์ธ์˜ ์ ˆ๋ฐ˜๊ณผ ์„ธ๊ณ„ ์ธ๊ตฌ์˜ 10%๋ฅผ ์˜ํ–ฅ์„ ๋ฏธ์นœ๋‹ค [14]. ์„ฌ์œ ํ™” ๊ณผ์ •์€ ์ƒํ”ผ ์„ธํฌ ์†์ƒ ํ›„ ๋ถ„๋น„๋˜๋Š” pro-fibrotic ๋ฐ pro-inflammatory paracrine signals์— ์˜ํ•ด ๊ฐ„์งˆ myofibroblast์˜ ํ™œ์„ฑํ™”๊ฐ€ ์กฐ์ ˆ๋œ๋‹ค [14]. ํ˜„์žฌ ์‹ ์žฅ ์„ฌ์œ ํ™”๋ฅผ ๋Šฆ์ถ”๋Š” ํ‘œ์  ์น˜๋ฃŒ์ œ๋Š” ์•„์ง ์—†์œผ๋‚˜, ์ด๋Ÿฌํ•œ ๊ธฐ์ „ ์ดํ•ด๋ฅผ ๋ฐ”ํƒ•์œผ๋กœ ์ƒˆ๋กœ์šด ์น˜๋ฃŒ ํ‘œ์  ๊ฐœ๋ฐœ์ด ์ง„ํ–‰ ์ค‘์ด๋‹ค [14]. ํ์—์„œ๋„ ์œ ์‚ฌํ•˜๊ฒŒ, Cancer-associated fibroblasts (CAFs)๋Š” ์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(tumor microenvironment)์˜ ํ•ต์‹ฌ ๊ตฌ์„ฑ ์š”์†Œ๋กœ, ๋งคํŠธ๋ฆญ์Šค ์นจ์ฐฉ ๋ฐ ์žฌ๊ตฌ์„ฑ, ์•” ์„ธํฌ์™€์˜ ์‹ ํ˜ธ ๊ตํ™˜, ์นจ์œค ๋ฐฑํ˜ˆ๊ตฌ์™€์˜ ์ƒํ˜ธ์ž‘์šฉ์„ ํ†ตํ•ด ์งˆ๋ณ‘ ์ง„ํ–‰์— ๊ธฐ์—ฌํ•œ๋‹ค [4]. CAFs์˜ ๊ธฐ์›๊ณผ ๊ธฐ๋Šฅ ์ด์งˆ์„ฑ(heterogeneity)์€ ์น˜๋ฃŒ์  ์กฐ์ ˆ์˜ ์ฃผ์š” ์žฅ์• ๋ฌผ์ด์ง€๋งŒ, ์ผ๋ถ€ ํ•ญ์ข…์–‘ ๊ธฐ๋Šฅ์„ ์œ ์ง€ํ•˜๋ฉด์„œ ์„ฌ์œ ํ™” ๊ณผ์ •์„ ์–ต์ œํ•˜๋Š” ์ „๋žต์ด ์—ฐ๊ตฌ๋˜๊ณ  ์žˆ๋‹ค [4].

Epithelialโ€“mesenchymal transition (EMT)๋Š” ์ƒํ”ผ ์„ธํฌ๊ฐ€ ์ค‘๊ฐ„์—ฝ ํ‘œํ˜„ํ˜•์œผ๋กœ ์ „ํ™˜๋˜๋Š” ๋™์  ๊ณผ์ •์œผ๋กœ, ์„ธํฌ ์ด๋™๊ณผ ์นจ์Šต์„ฑ ๋ณ€ํ™”๋ฅผ ์œ ๋ฐœํ•˜๋ฉฐ ILD ๋ฐ ์•” ์ง„ํ–‰์—์„œ ์ค‘์š”ํ•œ ์—ญํ• ์„ ํ•œ๋‹ค [7]. EMT๋Š” ๋ณด์กด๋œ ์œ ๋„ ์‹ ํ˜ธ(inducing signals), ์ „์‚ฌ ์กฐ์ ˆ์ž(transcriptional regulators), ํ•˜์œ„ ํšจ๊ณผ๋ฌผ(downstream effectors)์— ์˜ํ•ด ๊ตฌ๋™๋˜๋ฉฐ, TEMTIA (The EMT International Association)์˜ ํ•ฉ์˜ ์„ฑ๋ช…์„œ๋Š” ์ด ๋ถ„์•ผ์˜ ์šฉ์–ด ํ†ต์ผ๊ณผ ์—ฐ๊ตฌ ๊ฐ€์ด๋“œ๋ผ์ธ์„ ์ œ์‹œํ•˜์—ฌ ๋ฐ์ดํ„ฐ ํ•ด์„์˜ ์˜คํ•ด๋ฅผ ์ค„์ด๊ณ  ํ•™์ œ ๊ฐ„ ํ˜‘๋ ฅ์„ ์ด‰์ง„ํ•˜๊ณ ์ž ํ•œ๋‹ค [7]. Notch signaling pathway ์—ญ์‹œ ์กฐ์ง ํ•ญ์ƒ์„ฑ(tissue homeostasis)๊ณผ ์งˆ๋ณ‘์—์„œ ์ค‘์š”ํ•œ ์—ญํ• ์„ ํ•˜๋ฉฐ, ์„ธํฌ-์„ธํฌ ๊ฐ„ ํ†ต์‹ (cell-cell communication)์„ ํ†ตํ•ด ์„ธํฌ ์šด๋ช…์„ ์กฐ์ ˆํ•œ๋‹ค [21]. Notch ์‹ ํ˜ธ ์ „๋‹ฌ์˜ ์ด์ƒ์€ ํ”ผ๋ถ€, ๊ฐ„, ํ, ์žฅ, ํ˜ˆ๊ด€ ๋“ฑ ๋‹ค์–‘ํ•œ ์žฅ๊ธฐ์˜ ๋ฐœ๋‹ฌ ๋ฐ ์งˆํ™˜(ํŠนํžˆ ์•”)๊ณผ ์—ฐ๊ด€๋˜์–ด ์žˆ์œผ๋ฉฐ, ILD์—์„œ์˜ ์„ฌ์œ ํ™” ์ง„ํ–‰์—๋„ ๊ด€์—ฌํ•  ๊ฐ€๋Šฅ์„ฑ์ด ์žˆ๋‹ค [21].

์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ์˜ ์‚ฐ์„ฑ๋„(acidity)๋Š” ์„ธํฌ ๋Œ€์‚ฌ, ์ฆ์‹, ์ƒ์กด, ๊ทธ๋ฆฌ๊ณ  ๋ฉด์—ญ ๋ฐ˜์‘์— ๋ณตํ•ฉ์ ์ธ ์˜ํ–ฅ์„ ๋ฏธ์ณ ์งˆ๋ณ‘ ์ง„ํ–‰์„ ์ด‰์ง„ํ•œ๋‹ค [15]. ๊ณ ๋Œ€์‚ฌ ํ™œ๋™๊ณผ ๋ถˆ์ถฉ๋ถ„ํ•œ ๊ด€๋ฅ˜(perfusion)๋กœ ์ธํ•ด ์ข…์–‘ ๊ฐ„์งˆ ๊ณต๊ฐ„์˜ pH๊ฐ€ ๋‚ฎ์•„์ง€๋ฉด, ์ด๋Š” ์œ ์ „์  ๋ถˆ์•ˆ์ •์„ฑ, ํ›„์ƒ์œ ์ „ํ•™์  ๋ณ€ํ™”, ์„ธํฌ ๋Œ€์‚ฌ ์กฐ์ ˆ ๋“ฑ์„ ํ†ตํ•ด ์•…์„ฑ ํ‘œํ˜„ํ˜•์„ ๊ฐ•ํ™”ํ•œ๋‹ค [15]. ๋˜ํ•œ, ์‚ฐ์„ฑ ํ™˜๊ฒฝ์€ ์„ธํฌ ์ด๋™์„ฑ ์ฆ๊ฐ€, ์„ธํฌ์™ธ ๊ธฐ์งˆ ๋ถ„ํ•ด, ๋ฉด์—ญ ๋ฐ˜์‘ ๊ฐ์†Œ๋ฅผ ์œ ๋ฐœํ•˜์—ฌ ์นจ์Šต์„ฑ์„ ๋†’์ธ๋‹ค [15]. ์ด๋Ÿฌํ•œ ๋ฏธ์„ธํ™˜๊ฒฝ์˜ ๋ฌผ๋ฆฌํ™”ํ•™์  ํŠน์„ฑ์€ ILD์—์„œ์˜ ์„ฌ์œ ํ™” ์กฐ์ง ํ˜•์„ฑ๊ณผ๋„ ์œ ์‚ฌํ•œ ๋งฅ๋ฝ์—์„œ ๊ณ ๋ ค๋  ์ˆ˜ ์žˆ์œผ๋ฉฐ, TRP (Transient Receptor Potential) cation channels์™€ ๊ฐ™์€ ์ด์˜จ ์ฑ„๋„์€ ์—ผ์ฆ ์ƒ์„ฑ๋ฌผ ๋ฐ ์ž๊ทน์ œ์— ๋ฐ˜์‘ํ•˜์—ฌ ๋‹ค์–‘ํ•œ ์งˆํ™˜์˜ ๋ณ‘์ธ ๋ฐœ์ƒ(pathogenesis)์— ๊ธฐ์—ฌํ•œ๋‹ค [13].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ILD ๋ถ„๋ฅ˜(Ers/ATS statement)์™€ ์ง„ํ–‰์„ฑ ํ์„ฌ์œ ์ฆ์˜ ์˜์ƒํ•™์  ๋ถ„์„์— ์ดˆ์ ์„ ๋งž์ถ”๊ณ  ์žˆ์œผ๋‚˜, EMT, Notch signaling, CAFs ๋“ฑ ์„ธํฌ ์ˆ˜์ค€์˜ ์„ฌ์œ ํ™” ๊ธฐ์ „๊ณผ ๋ฏธ์„ธํ™˜๊ฒฝ ์กฐ์ ˆ์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ๋ฌธํ—Œ์€ ๋ถ€์กฑํ•˜๋‹ค. ํŠนํžˆ "Update of the international multidisciplinary classification of the interstitial pneumonias"๋Š” ๋ถ„๋ฅ˜ ์ฒด๊ณ„์— ์ค‘์ ์„ ๋‘๊ณ  ์žˆ์œผ๋ฏ€๋กœ, ๋ณ‘๋ฆฌ์ƒ๋ฆฌํ•™์  ๋ฉ”์ปค๋‹ˆ์ฆ˜(EMT, Notch, CAFs)์„ ๋‹ค๋ฃจ๋Š” ๋ฆฌ๋ทฐ ๋…ผ๋ฌธ๋“ค์„ ์ถ”๊ฐ€ํ•˜๋ฉด ILD์˜ ๊ทผ๋ณธ์ ์ธ ์›์ธ์— ๋Œ€ํ•œ ์ดํ•ด๊ฐ€ ๊นŠ์–ด์งˆ ๊ฒƒ์ด๋‹ค.

ํ™˜๊ฒฝ์  ์š”์ธ ๋ฐ ์žฅ๊ธฐ ํ›„์œ ์ฆ ๊ด€๋ฆฌ

ILD์™€ ํ˜ธํก๊ธฐ ์งˆํ™˜์€ ํ™˜๊ฒฝ์  ๋…ธ์ถœ๊ณผ ๋ฐ€์ ‘ํ•œ ๊ด€๋ จ์ด ์žˆ์œผ๋ฉฐ, ์ž์—ฐ์žฌํ•ด๋‚˜ ์ธ์œ„์  ์žฌํ•ด(disasters) ์ƒํ™ฉ์—์„œ๋Š” ํ˜ธํก๊ธฐ ์‘๊ธ‰ ์ƒํ™ฉ์ด ๋นˆ๋ฒˆํ•˜๊ฒŒ ๋ฐœ์ƒํ•œ๋‹ค. ์ง€์ง„ ํŒŒํŽธ์˜ ์‹ค๋ฆฌ์นด/์„๋ฉด ๋จผ์ง€, ์‚ฐ๋ถˆ ์—ฐ๊ธฐ์˜ ๋ฏธ์„ธ ์ž…์ž, ํ™์ˆ˜ ํ”ผํ•ด ๊ฑด๋ฌผ์˜ ๊ณฐํŒก์ด ํฌ์ž, ํ™”์‚ฐ ๋ถ„ํ™”์˜ ์žฌ ๋ฐ ์‚ฐ์„ฑ ๊ฐ€์Šค, ํ™”ํ•™ ์‚ฌ๊ณ ์˜ ์œ ๋… ๋ฌผ์งˆ ๋“ฑ์€ ๊ฐ๊ฐ ๋…ํŠนํ•œ ๊ธฐ์ „์œผ๋กœ ๊ธ‰์„ฑ ๋ฐ ๋งŒ์„ฑ ํ˜ธํก๊ธฐ ์งˆํ™˜์„ ์œ ๋ฐœํ•œ๋‹ค [30]. ํŠนํžˆ ๋งŒ์„ฑ ํ˜ธํก๊ธฐ ์งˆํ™˜ ํ™˜์ž๋Š” ์žฌ๋ฐœ(exacerbations) ์œ„ํ—˜์ด ๋†’์œผ๋ฉฐ, ์ž์›์ด ์ œํ•œ๋œ ํ™˜๊ฒฝ์—์„œ Acute respiratory distress syndrome (ARDS) ๊ด€๋ฆฌ๊ฐ€ ํฐ ๋„์ „ ๊ณผ์ œ๋กœ ๋‚จ์•„ ์žˆ๋‹ค [30]. ๊ธฐํ›„ ๋ณ€ํ™”๋กœ ์ธํ•ด ์žฌํ•ด ๊ด€๋ จ ํ˜ธํก๊ธฐ ์งˆํ™˜์˜ ๋ถ€๋‹ด์ด ์ฆ๊ฐ€ํ•˜๊ณ  ์žˆ์–ด, ์ทจ์•ฝ ๊ณ„์ธต์„ ์œ„ํ•œ ๋Œ€์‘ ํ”„๋กœํ† ์ฝœ ๊ฐ•ํ™”๊ฐ€ ์‹œ๊ธ‰ํ•˜๋‹ค [30].

COVID-19 ํŒฌ๋ฐ๋ฏน ์ดํ›„ Long COVID (post-viral sequelae)์— ๋Œ€ํ•œ ๊ด€์‹ฌ์ด ๋†’์•„์กŒ์œผ๋ฉฐ, ์ด๋Š” ILD ํ™˜์ž๋‚˜ ๋งŒ์„ฑ ํ˜ธํก๊ธฐ ์งˆํ™˜์ž์—๊ฒŒ ์ถ”๊ฐ€์ ์ธ ๋ถ€๋‹ด์œผ๋กœ ์ž‘์šฉํ•  ์ˆ˜ ์žˆ๋‹ค. ์ฒด๊ณ„์  ๋ฌธํ—Œ๊ณ ์ฐฐ ๋ฐ ๋ฉ”ํƒ€๋ถ„์„ ๊ฒฐ๊ณผ, SARS-CoV-2 ๊ฐ์—ผ ํ™˜์ž์˜ 80%๊ฐ€ ํ•˜๋‚˜ ์ด์ƒ์˜ ์žฅ๊ธฐ ์ฆ์ƒ(long-term symptoms)์„ ๊ฒฝํ—˜ํ•˜๋Š” ๊ฒƒ์œผ๋กœ ์ถ”์ •๋œ๋‹ค [5]. ๊ฐ€์žฅ ํ”ํ•œ ์ฆ์ƒ์€ ํ”ผ๋กœ(fatigue, 58%), ๋‘ํ†ต(headache, 44%), ์ฃผ์˜๋ ฅ ์žฅ์• (attention disorder, 27%), ํƒˆ๋ชจ(hair loss, 25%), ํ˜ธํก๊ณค๋ž€(dyspnea, 24%)์ด๋‹ค [5]. ๋˜ ๋‹ค๋ฅธ living systematic review์—์„œ๋Š” ์ฆ์ƒ ๋ฐœํ˜„ 12์ฃผ ์ดํ›„์˜ Long COVID ํŠน์ง•์„ ๋ถ„์„ํ•œ ๊ฒฐ๊ณผ, ์•ฝ์ (weakness, 41%), ์ „๋ฐ˜์ ์ธ ๋ถˆ์พŒ๊ฐ(general malaise, 33%), ํ”ผ๋กœ(fatigue, 31%), ์ง‘์ค‘๋ ฅ ์ €ํ•˜(concentration impairment, 26%), ํ˜ธํก๊ณค๋ž€(breathlessness, 25%)์ด ์ฃผ์š” ์ฆ์ƒ์ด๋ฉฐ, ํ™˜์ž์˜ 37%๊ฐ€ ์‚ถ์˜ ์งˆ ๊ฐ์†Œ(reduced quality of life)๋ฅผ ๋ณด๊ณ ํ•˜๊ณ  26%์˜ ์—ฐ๊ตฌ์—์„œ ํ ๊ธฐ๋Šฅ ์ €ํ•˜(evidence of reduced pulmonary function)๊ฐ€ ๊ด€์ฐฐ๋˜์—ˆ๋‹ค [24]. ์ด๋Ÿฌํ•œ ์žฅ๊ธฐ ํ›„์œ ์ฆ์€ ๋‹คํ•™์ œ ํŒ€(multi-disciplinary teams)์„ ํ†ตํ•œ ์˜ˆ๋ฐฉ, ์žฌํ™œ, ์ž„์ƒ ๊ด€๋ฆฌ ์ „๋žต์ด ํ•„์š”ํ•จ์„ ์‹œ์‚ฌํ•œ๋‹ค [5].

์žฌํ•ด ์ƒํ™ฉ์—์„œ์˜ ํ˜ธํก๊ธฐ ์‘๊ธ‰ ๊ด€๋ฆฌ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ, ์ผ์ƒ์ ์ธ ํ™˜๊ฒฝ ๋…ธ์ถœ(์˜ˆ: ๋‚˜๋…ธ๋ฌผ์งˆ inhalation)๋„ ํ˜ธํก๊ธฐ ๊ฑด๊ฐ•์— ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ๋‹ค. Nanomaterials๋Š” ์ „์ž ๋ถ€ํ’ˆ, ์Šคํฌ์ธ  ์žฅ๋น„, ์˜๋ฃŒ ์ œํ’ˆ ๋“ฑ ๋‹ค์–‘ํ•œ ๋ถ„์•ผ์— ์‚ฌ์šฉ๋˜์ง€๋งŒ, ํƒ„์†Œ ๊ธฐ๋ฐ˜ ๋‚˜๋…ธํŠœ๋ธŒ(carbon-based nanotubes)์™€ ๊ฐ™์€ ๋ฌผ์งˆ์ด ์„๋ฉด(asbestos)๊ณผ ์œ ์‚ฌํ•œ ๋ฐœ์•” ์œ„ํ—˜์„ ๊ฐ€์งˆ ์ˆ˜ ์žˆ๋‹ค๋Š” ์šฐ๋ ค๊ฐ€ ์ œ๊ธฐ๋˜๊ณ  ์žˆ๋‹ค [12]. ํก์ž… ๊ฒฝ๋กœ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ํ”ผ๋ถ€ ๋ฐ ์ฃผ์‚ฌ ๊ฒฝ๋กœ ๋“ฑ์„ ํ†ตํ•œ ๋‚˜๋…ธ๋ฌผ์งˆ์˜ ์ฒด๋‚ด ์ถ•์ ๊ณผ ์žฅ๊ธฐ์  ๋…์„ฑ(long-term consequences)์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๋Š” ์•„์ง ๋ถ€์กฑํ•˜๋ฉฐ, ์ž‘์—…์ž์™€ ์ตœ์ข… ์‚ฌ์šฉ์ž๋ฅผ ๋ณดํ˜ธํ•˜๊ธฐ ์œ„ํ•œ ์•ˆ์ „ ๊ธฐ์ค€ ๋งˆ๋ จ์ด ํ•„์š”ํ•˜๋‹ค [12]. ์ด๋Ÿฌํ•œ ํ™˜๊ฒฝ์  ์š”์ธ๋“ค์€ ILD์˜ ๋ฐœ๋ณ‘์ด๋‚˜ ์•…ํ™”์— ๊ฐ„์ ‘์ ์œผ๋กœ ๊ธฐ์—ฌํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ, ํ™˜์ž ๊ต์œก ๋ฐ ์˜ˆ๋ฐฉ ์ „๋žต ์ˆ˜๋ฆฝ ์‹œ ๊ณ ๋ คํ•ด์•ผ ํ•  ์š”์†Œ์ด๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋Š” ์ฃผ๋กœ ์•” ๋‹จ๊ณ„ํ™”(staging)์™€ ICIs ์น˜๋ฃŒ์— ์ง‘์ค‘๋˜์–ด ์žˆ์–ด, ํ™˜๊ฒฝ์  ์š”์ธ(์žฌํ•ด, ๋‚˜๋…ธ๋ฌผ์งˆ)์ด๋‚˜ Long COVID์™€ ๊ฐ™์€ ๊ฐ์—ผ ํ›„ ์žฅ๊ธฐ ํ˜ธํก๊ธฐ ํ›„์œ ์ฆ์— ๋Œ€ํ•œ ๋ฌธํ—Œ์ด ์ „๋ฌดํ•˜๋‹ค. ILD ํ™˜์ž์˜ ์ „๋ฐ˜์ ์ธ ๊ฑด๊ฐ• ๊ด€๋ฆฌ์™€ ์˜ˆ๋ฐฉ ์˜ํ•™ ๊ด€์ ์—์„œ, ์žฌํ•ด ๊ด€๋ จ ํ˜ธํก๊ธฐ ์‘๊ธ‰ ๊ด€๋ฆฌ [30] ๋ฐ Long COVID์˜ ํ ๊ธฐ๋Šฅ ์˜ํ–ฅ [24, 5]์— ๋Œ€ํ•œ ๋ฌธํ—Œ์„ ์ถ”๊ฐ€ํ•˜๋ฉด, ํ™˜์ž์—๊ฒŒ ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š” ์™ธ๋ถ€ ์š”์ธ์— ๋Œ€ํ•œ ํฌ๊ด„์ ์ธ ์ดํ•ด๊ฐ€ ๊ฐ€๋Šฅํ•ด์งˆ ๊ฒƒ์ด๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1Adverse Renal Effects of Immune Checkpoint Inhibitors: A Narrative ReviewAmerican Journal of Nephrology ยท 201715291PubMed ยท DOI
2Characteristics of SARS-CoV-2 and COVID-19Nature Reviews Microbiology ยท 20205506PubMed ยท DOI
3Deep Learning in Medical Image AnalysisAnnual Review of Biomedical Engineering ยท 20174929PubMed ยท DOI
4A framework for advancing our understanding of cancer-associated fibroblastsNature reviews. Cancer ยท 20203939PubMed ยท DOI
5More than 50 long-term effects of COVID-19: a systematic review and meta-analysisScientific Reports ยท 20212519PubMed ยท DOI
6The Next Generation of Platinum Drugs: Targeted Pt(II) Agents, Nanoparticle Delivery, and Pt(IV) ProdrugsChemical Reviews ยท 20162482PubMed ยท DOI
7Guidelines and definitions for research on epithelialโ€“mesenchymal transitionNature Reviews Molecular Cell Biology ยท 20202367PubMed ยท DOI
8Chemokines and Chemokine Receptors: Positioning Cells for Host Defense and ImmunityAnnual Review of Immunology ยท 20142133PubMed ยท DOI
9Acute kidney disease and renal recovery: consensus report of the Acute Disease Quality Initiative (ADQI) 16 WorkgroupNature Reviews Nephrology ยท 20171564PubMed ยท DOI
10Human organs-on-chips for disease modelling, drug development and personalized medicineNature Reviews Genetics ยท 20221526PubMed ยท DOI
11The Sleep-Immune Crosstalk in Health and DiseasePhysiological Reviews ยท 20191486PubMed ยท DOI
12Toxicity of nanomaterialsChemical Society Reviews ยท 20111444PubMed ยท DOI
13Transient Receptor Potential Cation Channels in DiseasePhysiological Reviews ยท 20071396PubMed ยท DOI
14Mechanisms of Renal FibrosisAnnual Review of Physiology ยท 20171279PubMed ยท DOI
15The Acidic Tumor Microenvironment as a Driver of CancerAnnual Review of Physiology ยท 20191263PubMed ยท DOI
16Challenges and strategies in anti-cancer nanomedicine development: An industry perspectiveAdvanced Drug Delivery Reviews ยท 20161199PubMed ยท DOI
17Trophic macrophages in development and diseaseNature reviews. Immunology ยท 20091178PubMed ยท DOI
18Toward a Rational and Mechanistic Account of Mental EffortAnnual Review of Neuroscience ยท 20171175PubMed ยท DOI
19A guide to the organ-on-a-chipNature Reviews Methods Primers ยท 20221079DOI
20Pulmonary arterial pressure during rest and exercise in healthy subjects: a systematic reviewEuropean Respiratory Journal ยท 20091066PubMed ยท DOI
21Notch Signaling in Development, Tissue Homeostasis, and DiseasePhysiological Reviews ยท 20171023PubMed ยท DOI
22Anti-obesity drug discovery: advances and challengesNature Reviews Drug Discovery ยท 20211001PubMed ยท DOI
23The global burden of multiple chronic conditions: A narrative reviewPreventive Medicine Reports ยท 2018997PubMed ยท DOI
24Characterising long COVID: a living systematic reviewBMJ Global Health ยท 2021989PubMed ยท DOI
25Diagnostic accuracy of point-of-care lung ultrasound compared to chest radiography for identifying community-acquired pneumonia in children: a systematic review and meta-analysisEgyptian Pediatric Association Gazette ยท 20260DOI
26Extracorporeal Blood Purification Techniques in Sepsis and ARDS: Evaluation of ECCOโ‚‚R and CytoSorbยฎ Hemoadsorption โ€“ A Narrative ReviewJournal of Education Health and Sport ยท 20260DOI
27Diagnostic imaging and emerging image-guided neurointervention in brain tumors: a narrative reviewFrontiers in Radiology ยท 20260DOI
28Early detection, treatment initiation, and long-term health outcomes in patients with connective tissue disease-associated pulmonary arterial hypertension: A systematic literature review and expert consensusModern Rheumatology ยท 20260PubMed ยท DOI
29Safety and efficacy of Meridian sinew tuina (MST) for post-surgical upper limb lymphedema: a systematic review and meta-analysisJournal of Obstetrics and Gynaecology ยท 20260PubMed ยท DOI
30Respiratory Emergencies in Disasters: A Narrative ReviewThoracic research and practice ยท 20260PubMed ยท DOI
31Diagnostic yield and safety of small versus standard cryoprobes in transbronchial lung cryobiopsy for interstitial lung disease: a systematic review and meta-analysisEgyptian Journal of Bronchology ยท 20260DOI
32Exercise as a modulator of purinergic signaling and inflammation in cancer-associated sarcopenia: a narrative reviewInflammation Research ยท 20260PubMed ยท DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1A Review of Antisynthetase Syndrome-Associated Interstitial Lung Disease.International journal of molecular sciences ยท 2024-Apr-18PubMed ยท DOI
2Smoking-Related Interstitial Lung Disease and Emphysema.Clinics in chest medicine ยท 2024-JunPubMed ยท DOI
3Approach to diagnosing and managing granulomatous-lymphocytic interstitial lung disease.EClinicalMedicine ยท 2024-SepPubMed ยท DOI
4Refining dermatomyositis.Journal of the American Academy of Dermatology ยท 2024-DecPubMed ยท DOI
5Drug-induced interstitial lung disease: a narrative review of a clinical conundrum.Expert review of respiratory medicine ยท 2024PubMed ยท DOI
6Transbronchial lung cryo-rebiopsy for progressive interstitial lung disease.Respiratory investigation ยท 2025-SepPubMed ยท DOI
7Pathogenesis of interstitial lung disease in systemic sclerosis.Rheumatology and immunology research ยท 2024-SepPubMed ยท DOI
8What Is Interstitial Lung Disease?JAMA ยท 2025-Aug-26PubMed ยท DOI
9Smoking-Related Interstitial Fibrosis and Smoker's Macrophages.Journal of Nippon Medical School = Nippon Ika Daigaku zasshi ยท 2024-Mar-09PubMed ยท DOI
10Oxaliplatin-Induced Pulmonary Fibrosis: A Rare but Fatal Reality.Cureus ยท 2023-DecPubMed ยท DOI
SCLC โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 24ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
SCLC์˜ ํ‘œ์ค€ ์น˜๋ฃŒ ์ „๋žต ๋ฐ ๋ฐฉ์‚ฌ์„  ์š”๋ฒ•์˜ ์ตœ์ ํ™”

์†Œ์„ธํฌํ์•”(SCLC)์€ ๊ณต๊ฒฉ์ ์ธ ์ž„์ƒ ์–‘์ƒ์„ ๋ณด์ด๋ฉฐ, ํ™”ํ•™์š”๋ฒ•์ด ์น˜๋ฃŒ์˜ ํ•ต์‹ฌ ์ถ•์„ ์ฐจ์ง€ํ•ฉ๋‹ˆ๋‹ค. ํŠนํžˆ 1์ฐจ ์น˜๋ฃŒ์—์„œ Cisplatin๊ณผ Carboplatin ๊ธฐ๋ฐ˜ ์š”๋ฒ•์˜ ์„ ํƒ์€ ์˜ค๋žซ๋™์•ˆ ๋…ผ์Ÿ์˜ ๋Œ€์ƒ์ด์—ˆ์œผ๋‚˜, COCIS ๋ฉ”ํƒ€๋ถ„์„[9]์€ ๋‘ ์•ฝ๋ฌผ ๊ฐ„์— Overall Survival(OS)์ด๋‚˜ Progression-Free Survival(PFS), Objective Response Rate(ORR)์—์„œ ํ†ต๊ณ„์ ์œผ๋กœ ์œ ์˜๋ฏธํ•œ ์ฐจ์ด๊ฐ€ ์—†์Œ์„ ์ž…์ฆํ–ˆ์Šต๋‹ˆ๋‹ค. Median OS๋Š” Cisplatin๊ตฐ 9.6๊ฐœ์›”, Carboplatin๊ตฐ 9.4๊ฐœ์›”๋กœ ์œ ์‚ฌํ–ˆ์œผ๋ฉฐ(HR 1.08, P=.37), ๋…์„ฑ ํ”„๋กœํŒŒ์ผ๋งŒ ์ƒ์ดํ•˜์—ฌ(Carboplatin์€ ํ˜ˆ์•กํ•™์  ๋…์„ฑ์ด ๋†’๊ณ , Cisplatin์€ ๋น„ํ˜ˆ์•กํ•™์  ๋…์„ฑ์ด ๋†’์Œ) ์ž„์ƒ์  ํŽธ์˜์„ฑ๊ณผ ํ™˜์ž ์ƒํƒœ์— ๋”ฐ๋ผ ์„ ํƒ๋  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ Cisplatin์˜ ์ถ”๊ฐ€์  ์ด์ ์„ ๊ทœ๋ช…ํ•˜๊ธฐ ์œ„ํ•œ ๋ฉ”ํƒ€๋ถ„์„[21]์—์„œ๋Š” Cisplatin ํฌํ•จ ์š”๋ฒ•์ด ๋ฐ˜์‘๋ฅ (OR 1.35, P<10^-5)์„ ๋†’์ด์ง€๋งŒ, ๋…์„ฑ ๊ด€๋ จ ์‚ฌ๋ง ์œ„ํ—˜๊นŒ์ง€ ๊ณ ๋ คํ–ˆ์„ ๋•Œ ์ƒ์กด์œจ์—์„œ ์œ ์˜ํ•œ ์šฐ์œ„๋ฅผ ๋ณด์ด์ง€ ์•Š๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ์Šต๋‹ˆ๋‹ค. ์ด๋Š” SCLC ์น˜๋ฃŒ์—์„œ Platinum ๊ธฐ๋ฐ˜ ์š”๋ฒ•์˜ ํ•„์ˆ˜์„ฑ์€ ์ธ์ •๋˜์ง€๋งŒ, ํŠน์ • ์•ฝ๋ฌผ์˜ ์ ˆ๋Œ€์  ์šฐ์œ„๋ณด๋‹ค๋Š” ํ™˜์ž์˜ ์ „์‹  ์ƒํƒœ์™€ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ๊ฐ€ ๋” ์ค‘์š”ํ•จ์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค.

๋ฐฉ์‚ฌ์„  ์š”๋ฒ•(Radiotherapy, RT)์˜ ์‹œ๊ธฐ ๋ฐ ๋ฐฉ๋ฒ•์€ ๊ตญ์†Œ๊ธฐ(Limited-Stage, LS-SCLC)์—์„œ ์ƒ์กด์œจ์— ๊ฒฐ์ •์ ์ธ ์˜ํ–ฅ์„ ๋ฏธ์นฉ๋‹ˆ๋‹ค. Early Thoracic Radiation Therapy (ERT)์™€ Late Thoracic Radiation Therapy (LRT)๋ฅผ ๋น„๊ตํ•œ ๋ฉ”ํƒ€๋ถ„์„[14]์€ ERT๊ฐ€ 2๋…„ OS์—์„œ ์œ ์˜ํ•œ ์ด์ ์„ ์ œ๊ณตํ•จ์„ ๋ณด์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค(RR 1.17, P=.03). ํŠนํžˆ Hyperfractionated RT์™€ Platinum ๊ธฐ๋ฐ˜ ํ™”ํ•™์š”๋ฒ•์„ ๋ณ‘์šฉํ•  ๋•Œ ์ด๋Ÿฌํ•œ ์ด์ ์€ ๋”์šฑ ๋šœ๋ ทํ•˜๊ฒŒ ๋‚˜ํƒ€๋‚ฌ์Šต๋‹ˆ๋‹ค(Hyperfractionation: RR 1.44, P=.001; Platinum-based: RR 1.30, P=.002). ์ด๋Š” SCLC์˜ ๋น ๋ฅธ ์„ฑ์žฅ ์†๋„๋ฅผ ๊ณ ๋ คํ•  ๋•Œ, ๊ฐ€๋Šฅํ•œ ํ•œ ์ดˆ๊ธฐ์— ๊ตญ์†Œ ๋ณ‘๋ณ€์„ ํ†ต์ œํ•˜๋Š” ๊ฒƒ์ด ์ „์‹  ์น˜๋ฃŒ์™€ ์‹œ๋„ˆ์ง€ ํšจ๊ณผ๋ฅผ ๋‚ด์–ด ์ƒ์กด์„ ์—ฐ์žฅ์‹œํ‚ฌ ์ˆ˜ ์žˆ์Œ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.

๋‡Œ์ „์ด ์˜ˆ๋ฐฉ์„ ์œ„ํ•œ Prophylactic Cranial Irradiation (PCI)์˜ ์—ญํ•  ๋˜ํ•œ SCLC ๊ด€๋ฆฌ์—์„œ ์ค‘์š”ํ•œ ๋ถ€๋ถ„์ž…๋‹ˆ๋‹ค. ๋ฉ”ํƒ€๋ถ„์„[23]์€ ํ™”ํ•™์š”๋ฒ• ํ›„ ์™„์ „ ๋ฐ˜์‘(CR)์„ ๋ณด์ธ SCLC ํ™˜์ž์—์„œ PCI๊ฐ€ ๋‡Œ์ „์ด ๋ฐœ์ƒ๋ฅ ์„ ํ˜„์ €ํžˆ ๊ฐ์†Œ์‹œํ‚ค๋ฉฐ(HR 0.48, 95% CI: 0.39-0.60), ์ƒ์กด์œจ๋„ ๊ฐœ์„ ์‹œํ‚จ๋‹ค๋Š” ๊ฒƒ์„ ํ™•์ธํ–ˆ์Šต๋‹ˆ๋‹ค(HR 0.82, 95% CI: 0.71-0.96). ๊ทธ๋Ÿฌ๋‚˜ ์žฅ๊ธฐ์ ์ธ ์‹ ๊ฒฝ๋…์„ฑ์— ๋Œ€ํ•œ ๋ฐ์ดํ„ฐ๊ฐ€ ๋ถ€์กฑํ•˜๋‹ค๋Š” ํ•œ๊ณ„๊ฐ€ ์ง€์ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ๋ฐ˜๋ฉด, ์ด๋ฏธ ๋‡Œ์ „์ด๊ฐ€ ๋ฐœ์ƒํ•œ ๊ฒฝ์šฐ์˜ ์น˜๋ฃŒ ์ ‘๊ทผ๋ฒ•์€ SCLC์˜ ๋ฐฉ์‚ฌ์„  ๊ฐ์ˆ˜์„ฑ(Radiosensitivity)์„ ๊ณ ๋ คํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค. Whole Brain Radiation Therapy (WBRT)์— ๋Œ€ํ•œ ๊ฐ€์ด๋“œ๋ผ์ธ[20]์€ SCLC์™€ ๊ฐ™์ด radiosensitiveํ•œ ์กฐ์งํ•™์—์„œ๋Š” WBRT์˜ ์šฉ๋Ÿ‰/๋ถ„ํ•  ์Šค์ผ€์ค„ ๋ณ€๊ฒฝ์ด ์ƒ์กด์ด๋‚˜ ์‹ ๊ฒฝ์ธ์ง€ ๊ฒฐ๊ณผ์— ์œ ์˜ํ•œ ์ฐจ์ด๋ฅผ ๋งŒ๋“ค์ง€ ์•Š๋Š”๋‹ค๊ณ  ๊ฒฐ๋ก ์ง€์—ˆ์œผ๋ฉฐ, ์ด๋Š” SCLC๊ฐ€ ๋‹ค๋ฅธ ๊ณ ํ˜•์•”๊ณผ ๋‹ฌ๋ฆฌ ๋ฐฉ์‚ฌ์„ ์— ๋งค์šฐ ๋ฏผ๊ฐํ•˜๋ฏ€๋กœ ํ‘œ์ค€ WBRT๋กœ๋„ ์ถฉ๋ถ„ํ•œ ํšจ๊ณผ๋ฅผ ๊ธฐ๋Œ€ํ•  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ์‚ฌ์šฉ์ž์˜ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” NSCLC ์ค‘์‹ฌ์˜ Neoadjuvant Nivolumab ์—ฐ๊ตฌ๋‚˜ Stage III NSCLC์—์„œ์˜ Durvalumab ์œ ์ง€์š”๋ฒ•(PACIFIC trial ๊ด€๋ จ)์— ๋Œ€ํ•œ ๋…ผ๋ฌธ๋“ค์ด ๋‹ค์ˆ˜ ํฌํ•จ๋˜์–ด ์žˆ์œผ๋‚˜, SCLC ํŠน์ด์ ์ธ ๋ฐฉ์‚ฌ์„  ์š”๋ฒ• ํƒ€์ด๋ฐ(ERT vs LRT)์ด๋‚˜ PCI์˜ ์ตœ์‹  ์‹ ๊ฒฝ๋…์„ฑ ๊ด€๋ฆฌ ๋ฐ์ดํ„ฐ๋Š” ๋ช…์‹œ์ ์œผ๋กœ ๋‹ค๋ฃจ์–ด์ง€์ง€ ์•Š์€ ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. ํŠนํžˆ SCLC์—์„œ์˜ Immunotherapy ๊ฒฐํ•ฉ ์š”๋ฒ•์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๋ฉ”ํƒ€๋ถ„์„ ๋ฐ์ดํ„ฐ๋Š” ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ถ€์žฌํ•ฉ๋‹ˆ๋‹ค.

๋ฉด์—ญ๊ด€๋ฌธ์–ต์ œ์ œ(ICI) ๊ธฐ๋ฐ˜ ์น˜๋ฃŒ ๋ฐ ๋‚ด์„ฑ ๊ธฐ์ „

Immune Checkpoint Inhibitors (ICIs)๋Š” ์•” ๋ฉด์—ญ์น˜๋ฃŒ์˜ ํŒจ๋Ÿฌ๋‹ค์ž„์„ ๋ฐ”๊พธ์—ˆ์œผ๋‚˜, SCLC๋ฅผ ํฌํ•จํ•œ ๋‹ค์–‘ํ•œ ์ข…์–‘์—์„œ ๋‚ด์„ฑ ๋ฌธ์ œ๋Š” ์—ฌ์ „ํžˆ ํ•ด๊ฒฐํ•ด์•ผ ํ•  ๊ณผ์ œ์ž…๋‹ˆ๋‹ค[1]. ICIs์˜ ํšจ๋Šฅ์€ ์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(TME) ๋‚ด ๋ฉด์—ญ ์„ธํฌ์˜ ์žฌํ”„๋กœ๊ทธ๋ž˜๋ฐ์— ์˜์กดํ•˜๋ฉฐ, PD-1/PD-L1 ๊ฒฝ๋กœ๋ฅผ ํ‘œ์ ์œผ๋กœ ํ•˜๋Š” ํ•ญ์ฒด๋“ค์ด FDA ์Šน์ธ์„ ๋ฐ›์œผ๋ฉฐ ์ƒ์กด์œจ์„ ์—ฐ์žฅ์‹œ์ผฐ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ๋ชจ๋“  ํ™˜์ž๊ฐ€ ์ง€์† ๊ฐ€๋Šฅํ•œ ๋ฐ˜์‘์„ ๋ณด์ด๋Š” ๊ฒƒ์€ ์•„๋‹ˆ๋ฉฐ, Primary Resistance์™€ Acquired Resistance๊ฐ€ ์น˜๋ฃŒ ๊ฒฐ๊ณผ๋ฅผ ์ œํ•œํ•ฉ๋‹ˆ๋‹ค[1]. ์ด๋Ÿฌํ•œ ๋‚ด์„ฑ ๊ธฐ์ „์„ ์ดํ•ดํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” ์ข…์–‘ ์„ธํฌ ์ž์ฒด์˜ ๋ณ€ํ™”๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์ „์‹  ๋ฉด์—ญ ์ฒด๊ณ„์˜ ๋ณ€ํ™”๋ฅผ ํ•จ๊ป˜ ๊ณ ๋ คํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.

SCLC ํ™˜์ž์—์„œ์˜ ICI ๋ฐ˜์‘์—๋Š” ์„ฑ๋ณ„์— ๋”ฐ๋ฅธ ์ด์งˆ์„ฑ์ด ์กด์žฌํ•ฉ๋‹ˆ๋‹ค. Meta-analysis[22]๋Š” ๋‚จ์„ฑ ํ™˜์ž๊ฐ€ ์—ฌ์„ฑ ํ™˜์ž๋ณด๋‹ค Anti-PD-1/PD-L1 ์š”๋ฒ•์— ๋” ํฐ ํ˜œํƒ์„ ๋ฐ›๋Š” ๊ฒฝํ–ฅ์ด ์žˆ์Œ์„ ๋ณด์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค(Pooled OS-HR ratio < 1 indicating greater effect in men). ์ด๋Š” ๋ฉด์—ญ ์ฒด๊ณ„์˜ ์„ฑ๋ณ„ ๋‹คํ˜•์„ฑ(Sex-dimorphism)์ด ์น˜๋ฃŒ ๋ฐ˜์‘์— ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•˜๋ฉฐ, ํ–ฅํ›„ ๊ฐœ์ธํ™”๋œ ์น˜๋ฃŒ ์ „๋žต ์ˆ˜๋ฆฝ ์‹œ ์„ฑ๋ณ„์„ ์ค‘์š”ํ•œ ๋ณ€์ˆ˜๋กœ ๊ณ ๋ คํ•ด์•ผ ํ•จ์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค.

ICI ์น˜๋ฃŒ์˜ ๋ถ€์ž‘์šฉ ๊ด€๋ฆฌ, ํŠนํžˆ Immune-related Adverse Events (irAEs)๋Š” ์น˜๋ฃŒ ์ง€์†์„ฑ๊ณผ ์ง๊ฒฐ๋ฉ๋‹ˆ๋‹ค. ICIs ์œ ๋„ ๊ด€์ ˆ์—ผ(ICIs-IA)์€ ICI ์น˜๋ฃŒ ํ™˜์ž์˜ ์•ฝ 7%์—์„œ ๋ฐœ์ƒํ•˜๋ฉฐ, ํ์•” ํ™˜์ž์—์„œ๋„ ์ƒ๋‹น ๋น„์œจ(27.7%)์„ ์ฐจ์ง€ํ•ฉ๋‹ˆ๋‹ค[26]. ์ด๋Ÿฌํ•œ Rh-irAEs์˜ ์กฐ๊ธฐ ์ง„๋‹จ๊ณผ ์ ์ ˆํ•œ ์Šคํ…Œ๋กœ์ด๋“œ ์‚ฌ์šฉ์€ ์น˜๋ฃŒ ์ค‘๋‹จ ์—†์ด ๋ฉด์—ญ์š”๋ฒ•์˜ ์ด์ ์„ ์œ ์ง€ํ•˜๋Š” ๋ฐ ์ค‘์š”ํ•ฉ๋‹ˆ๋‹ค. ๋‡Œ์ „์ด ํ™˜์ž์—๊ฒŒ ์Šคํ…Œ๋กœ์ด๋“œ๋ฅผ ์‚ฌ์šฉํ•  ๊ฒฝ์šฐ, ์ฆ์ƒ ์ •๋„์— ๋”ฐ๋ผ Dexamethasone 4-8mg/์ผ(๊ฒฝ์ฆ) ๋˜๋Š” 16mg/์ผ ์ด์ƒ(์ค‘์ฆ)์„ ์‹œ์ž‘ํ•˜๊ณ , 2์ฃผ ์ด์ƒ ์„œ์„œํžˆ ๊ฐ๋Ÿ‰ํ•˜๋Š” ๊ฒƒ์ด ๊ถŒ์žฅ๋ฉ๋‹ˆ๋‹ค[15]. ์ด๋Š” SCLC์—์„œ ๋นˆ๋ฒˆํ•˜๊ฒŒ ๋ฐœ์ƒํ•˜๋Š” ๋‡Œ์ „์ด ๊ด€๋ฆฌ ์‹œ ICI์™€ ๋ณ‘์šฉ ์น˜๋ฃŒ ์ค‘ ์‹ ๊ฒฝํ•™์  ์ฆ์ƒ ์กฐ์ ˆ์— ํ•„์ˆ˜์ ์ธ ๊ฐ€์ด๋“œ๋ผ์ธ์ž…๋‹ˆ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” NSCLC์—์„œ์˜ PD-L1 ์˜ˆ์ธก ์ธ์ž ์—ฐ๊ตฌ๋‚˜ Nivolumab/Pembrolizumab์˜ ํšจ๋Šฅ ํ‰๊ฐ€ ๋…ผ๋ฌธ๋“ค์ด ์กด์žฌํ•˜์ง€๋งŒ, SCLC ํŠน์ด์ ์ธ ICI ๋ฐ˜์‘์˜ ์„ฑ๋ณ„ ์ฐจ์ด[22]๋‚˜ ICI ์œ ๋„ ๊ด€์ ˆ์—ผ๊ณผ ๊ฐ™์€ ํŠน์ • irAEs์— ๋Œ€ํ•œ ์ฒด๊ณ„์  ๊ฒ€ํ† ๋Š” ํฌํ•จ๋˜์–ด ์žˆ์ง€ ์•Š์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ SCLC์—์„œ์˜ ICI ๋‚ด์„ฑ ๊ธฐ์ „์— ๋Œ€ํ•œ ๋ถ„์ž ์ˆ˜์ค€์˜ ๋ถ„์„์€ ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—์„œ ํ™•์ธ๋˜์ง€ ์•Š์Šต๋‹ˆ๋‹ค.

์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(TME) ์กฐ์ ˆ ๋ฐ ํ‘œ์  ์น˜๋ฃŒ์˜ ์ƒˆ๋กœ์šด ์ง€ํ‰

์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(TME)์€ SCLC์˜ ์ง„ํ–‰๊ณผ ์ „์ด์— ํ•ต์‹ฌ์ ์ธ ์—ญํ• ์„ ํ•˜๋ฉฐ, ์ตœ๊ทผ ์—ฐ๊ตฌ๋“ค์€ TME ๋‚ด ๋‹ค์–‘ํ•œ ์„ธํฌ ๊ตฌ์„ฑ ์š”์†Œ์™€ ์‹ ํ˜ธ ์ „๋‹ฌ ๊ฒฝ๋กœ๋ฅผ ํ‘œ์ ์œผ๋กœ ํ•˜๋Š” ์ƒˆ๋กœ์šด ์น˜๋ฃŒ ์ „๋žต์„ ๋ชจ์ƒ‰ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. B7-H3๋Š” ์ข…์–‘ ์„ธํฌ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ํ˜ˆ๊ด€, ๊ฐ„์งˆ, ๋ฉด์—ญ ์„ธํฌ์—์„œ ๋ฐœํ˜„๋˜๋ฉฐ, ํŠนํžˆ Myeloid-derived suppressor cells (MDSCs)์™€ ๋Œ€์‹์„ธํฌ์—์„œ์˜ ๊ณ ๋ฐœํ˜„์€ ๋ฉด์–ต ํšŒํ”ผ์™€ ์—ฐ๊ด€๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค[31]. B7-H3์˜ TME ๋‚ด ๋ถ„ํฌ๋ฅผ ์ดํ•ดํ•˜๋Š” ๊ฒƒ์€ ์ฐจ์„ธ๋Œ€ ํ•ญ์ฒด-์•ฝ๋ฌผ ์ ‘ํ•ฉ์ฒด(ADC)๋‚˜ ๋ฉด์—ญ์น˜๋ฃŒ์ œ ๊ฐœ๋ฐœ์— ์ค‘์š”ํ•œ ํƒ€๊ฒŸ์ด ๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์‹ค์ œ๋กœ ADCs๋Š” ์ข…์–‘ ํŠน์ด์  ํ•ญ์›์„ ํ‘œ์ ์œผ๋กœ ํ•˜์—ฌ ์„ธํฌ๋…์„ฑ ์•ฝ๋ฌผ์„ ์ „๋‹ฌํ•จ์œผ๋กœ์จ ์น˜๋ฃŒ ์ง€์ˆ˜๋ฅผ ๋†’์ด๋Š” ์œ ๋งํ•œ ์ „๋žต์œผ๋กœ ๋ถ€์ƒํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค[18].

TME์˜ ์„ฌ์œ ์•„์„ธํฌ(Cancer-associated fibroblasts, CAFs)์™€ ๊ด€๋ จ๋œ Renin-Angiotensin System (RAS) ์‹ ํ˜ธ ์ „๋‹ฌ๋„ ๋ฉด์—ญ์น˜๋ฃŒ ๋ฐ˜์‘์— ์˜ํ–ฅ์„ ๋ฏธ์นฉ๋‹ˆ๋‹ค. Meta-analysis[30]๋Š” RAS ์–ต์ œ์ œ(RAS inhibitors)๋ฅผ ๋ณ‘์šฉ ์‚ฌ์šฉํ–ˆ์„ ๋•Œ ํ์•” ํ™˜์ž์˜ OS(HR 0.74)์™€ PFS(HR 0.81)๊ฐ€ ๊ฐœ์„ ๋˜๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ์œผ๋‚˜, ์ถœํŒ ํŽธํ–ฅ ๊ฐ€๋Šฅ์„ฑ๋„ ์ง€์ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” TME์˜ ๊ตฌ์กฐ์  ์žฌ๊ตฌ์„ฑ์„ ์กฐ์ ˆํ•จ์œผ๋กœ์จ ICI์˜ ํšจ๋Šฅ์„ ์ฆ์ง„์‹œํ‚ฌ ์ˆ˜ ์žˆ๋Š” ๊ฐ€๋Šฅ์„ฑ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ Statins์˜ ์‚ฌ์šฉ ์—ญ์‹œ ํ์•” ํ™˜์ž์˜ ICI ์น˜๋ฃŒ ๊ฒฐ๊ณผ์— ๊ธ์ •์ ์ธ ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค[27]. Meta-analysis ๊ฒฐ๊ณผ, Statin ์‚ฌ์šฉ๊ตฐ์€ OS(HR 0.76)์™€ PFS(HR 0.82)๊ฐ€ ์œ ์˜ํ•˜๊ฒŒ ๊ฐœ์„ ๋˜์—ˆ์œผ๋ฉฐ, ์ด๋Š” Statins์˜ ๋ฉด์—ญ์กฐ์ ˆ ๋ฐ ํ•ญ์—ผ์ฆ ํŠน์„ฑ์ด ์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ์„ ์œ ๋ฆฌํ•˜๊ฒŒ ๋ณ€ํ™”์‹œ์ผœ ํ•ญ์ข…์–‘ ๋ฉด์—ญ ๋ฐ˜์‘์„ ๊ฐ•ํ™”ํ•˜๊ธฐ ๋•Œ๋ฌธ์œผ๋กœ ํ•ด์„๋ฉ๋‹ˆ๋‹ค.

Notch ์‹ ํ˜ธ ์ „๋‹ฌ ๊ฒฝ๋กœ ๋˜ํ•œ SCLC๋ฅผ ํฌํ•จํ•œ ์•”์—์„œ ์ด์ค‘์ ์ธ ์—ญํ• ์„ ํ•ฉ๋‹ˆ๋‹ค[3, 5]. Notch๋Š” ์„ธํฌ-์„ธํฌ ๊ฐ„ ํ†ต์‹ ์„ ํ†ตํ•ด ์„ธํฌ ์šด๋ช…์„ ๊ฒฐ์ •ํ•˜๋ฉฐ, ๊ทธ ๊ธฐ๋Šฅ์€ ๋ฌธ๋งฅ(Context)์— ๋”ฐ๋ผ ์ข…์–‘ ์ด‰์ง„ ๋˜๋Š” ์ข…์–‘ ์–ต์ œ ์—ญํ• ๋กœ ๋‹ฌ๋ผ์งˆ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. SCLC์—์„œ Notch ๊ฒฝ๋กœ์˜ ์ด์ƒ ์กฐ์ ˆ์€ ์น˜๋ฃŒ ๋‚ด์„ฑ ๋ฐ ์žฌ๋ฐœ๊ณผ ์—ฐ๊ด€๋  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ, ์ด๋ฅผ ํ‘œ์ ์œผ๋กœ ํ•˜๋Š” ์ „๋žต์ด ์—ฐ๊ตฌ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ CXCL8-CXCR1/2 ๊ฒฝ๋กœ[6]๋Š” ์ข…์–‘ ์„ฑ์žฅ, ์นจ์Šต, ์ „์ด๋ฅผ ์ด‰์ง„ํ•˜๋ฉฐ ํ˜ˆ๊ด€ ์‹ ์ƒ์„ ์œ ๋„ํ•˜์—ฌ TME๋ฅผ ํ˜•์„ฑํ•˜๋Š” ๋ฐ ๊ธฐ์—ฌํ•ฉ๋‹ˆ๋‹ค. ์ด๋Ÿฌํ•œ ์‚ฌ์ดํ† ์นด์ธ ๊ฒฝ๋กœ๋ฅผ ์ฐจ๋‹จํ•˜๋Š” ๊ฒƒ์€ SCLC์˜ ์ง„ํ–‰์„ ๋Šฆ์ถ”๋Š” ์ž ์žฌ์  ์น˜๋ฃŒ ํ‘œ์ ์ด ๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” Hypoxia-inducible miR-210๊ณผ ๊ฐ™์€ ๋ฏธ์„ธํ™˜๊ฒฝ ๊ด€๋ จ ์—ฐ๊ตฌ๊ฐ€ ์ผ๋ถ€ ํฌํ•จ๋˜์–ด ์žˆ์œผ๋‚˜, B7-H3์˜ TME ๋‚ด ๋ถ„ํฌ[31]๋‚˜ RAS ์–ต์ œ์ œ/Statins์™€ ICI์˜ ๋ณ‘์šฉ ํšจ๊ณผ[27, 30]์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๋ฉ”ํƒ€๋ถ„์„ ๋ฐ์ดํ„ฐ๋Š” ๋ˆ„๋ฝ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. SCLC ํŠน์ด์ ์ธ TME ์กฐ์ ˆ ํ‘œ์ (B7-H3, Notch, CXCL8)์— ๋Œ€ํ•œ ์ข…ํ•ฉ์ ์ธ ๊ฒ€ํ† ๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค.

์ง„๋‹จ biomarker ๋ฐ ์•ก์ฒด ์ƒ๊ฒ€(Liquid Biopsy)์˜ ์ž„์ƒ ์ ์šฉ

SCLC์˜ ์กฐ๊ธฐ ์ง„๋‹จ๊ณผ ์น˜๋ฃŒ ๋ฐ˜์‘ ๋ชจ๋‹ˆํ„ฐ๋ง์„ ์œ„ํ•ด ๋‹ค์–‘ํ•œ Biomarker์™€ Liquid Biopsy ๊ธฐ์ˆ ์ด ๊ฐœ๋ฐœ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. SHOX2์™€ RASSF1A์˜ ๋ฉ”ํ‹ธํ™” ํŒจ๋„์€ ํ์•” ์ง„๋‹จ์—์„œ ๋†’์€ ๋ฏผ๊ฐ๋„(77.8%)์™€ ํŠน์ด๋„(89.0%)๋ฅผ ๋ณด์˜€์œผ๋ฉฐ, AUC๋Š” 0.916์œผ๋กœ ์šฐ์ˆ˜ํ•œ ์ง„๋‹จ ์„ฑ๋Šฅ์„ ์ž…์ฆํ–ˆ์Šต๋‹ˆ๋‹ค[29]. ์ด๋Š” ์กฐ์ง ์ƒ๊ฒ€์˜ ํ•œ๊ณ„๋ฅผ ๋ณด์™„ํ•  ์ˆ˜ ์žˆ๋Š” ๋น„์นจ์Šต์  ์ง„๋‹จ ๋„๊ตฌ๋กœ์„œ์˜ ๊ฐ€๋Šฅ์„ฑ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. ํŠนํžˆ SCLC์™€ NSCLC ๋ชจ๋‘์—์„œ ์ ์šฉ ๊ฐ€๋Šฅํ•œ Liquid Biopsy ํ”Œ๋žซํผ์€ ์‹ค์‹œ๊ฐ„ ๋ถ„์ž ์ •๋ณด๋ฅผ ์ œ๊ณตํ•˜์—ฌ ์น˜๋ฃŒ ์ „๋žต ์ˆ˜๋ฆฝ์— ๊ธฐ์—ฌํ•ฉ๋‹ˆ๋‹ค[25].

์ „์‹  ๋ฉด์—ญ-์—ผ์ฆ ์ง€์ˆ˜(Systemic Immune-Inflammation Index, SII)๋Š” ๊ณ ํ˜•์•”์˜ ์˜ˆํ›„ ์˜ˆ์ธก ์ธ์ž๋กœ ์ฃผ๋ชฉ๋ฐ›๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. Meta-analysis[19]๋Š” ๋†’์€ SII๊ฐ€ SCLC๋ฅผ ํฌํ•จํ•œ ์—ฌ๋Ÿฌ ์•”์—์„œ ๋ถˆ๋Ÿ‰ํ•œ OS(HR 1.55)์™€ ์—ฐ๊ด€๋˜์–ด ์žˆ์Œ์„ ๋ณด์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ํ˜ˆ์•ก ๊ฒ€์‚ฌ๋งŒ์œผ๋กœ ํ™˜์ž์˜ ๋ฉด์—ญ-์—ผ์ฆ ์ƒํƒœ๋ฅผ ํ‰๊ฐ€ํ•˜๊ณ  ์˜ˆํ›„๋ฅผ ์ถ”์ •ํ•  ์ˆ˜ ์žˆ๋Š” ๊ฐ„๋‹จํ•œ biomarker๋กœ์„œ์˜ ๊ฐ€์น˜๋ฅผ ์ž…์ฆํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ Ki-67 ๋ฐœํ˜„ ์ˆ˜์ค€์€ ์„ธํฌ ์ฆ์‹๋ฅ ์„ ๋ฐ˜์˜ํ•˜๋ฉฐ, NSCLC์—์„œ๋Š” ๋†’์€ Ki-67์ด ๋ถˆ๋Ÿ‰ํ•œ ์˜ˆํ›„(HR 1.56)์™€ ์—ฐ๊ด€๋˜์ง€๋งŒ[24], SCLC์—์„œ์˜ ๊ตฌ์ฒด์ ์ธ prognostic value๋Š” ์ถ”๊ฐ€ ์—ฐ๊ตฌ๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. RAS ๋Œ์—ฐ๋ณ€์ด๋Š” ์ฃผ๋กœ NSCLC์—์„œ ์—ฐ๊ตฌ๋˜์—ˆ์œผ๋‚˜[7], SCLC์—์„œ์˜ ์—ญํ• ์€ ์ œํ•œ์ ์œผ๋กœ ๋ณด๊ณ ๋˜์–ด ์žˆ์œผ๋ฉฐ, p53 ์ด์ƒ ์—ญ์‹œ ํ์•” ์ „๋ฐ˜์—์„œ ๋ถˆ๋Ÿ‰ํ•œ ์˜ˆํ›„(HR 1.44-2.24)์™€ ์—ฐ๊ด€๋ฉ๋‹ˆ๋‹ค[12].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” Radiomics ๊ธฐ๋ฐ˜์˜ ์˜ˆ์ธก ๋ชจ๋ธ์ด๋‚˜ PD-L1 IHC assay ๋น„๊ต ์—ฐ๊ตฌ๊ฐ€ ํฌํ•จ๋˜์–ด ์žˆ์œผ๋‚˜, SHOX2/RASSF1A ๋ฉ”ํ‹ธํ™” ํŒจ๋„[29]์ด๋‚˜ SII[19]์™€ ๊ฐ™์€ ํ˜ˆ์•ก ๊ธฐ๋ฐ˜ biomarker์— ๋Œ€ํ•œ ์ฒด๊ณ„์  ๊ฒ€ํ† ๋Š” ์—†์Šต๋‹ˆ๋‹ค. SCLC ํŠน์ด์ ์ธ Liquid Biopsy ๋งˆ์ปค(์˜ˆ: ctDNA fragmentomics)์˜ ์ž„์ƒ ์œ ํšจ์„ฑ์— ๋Œ€ํ•œ ๋ฐ์ดํ„ฐ๊ฐ€ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ถ€์žฌํ•ฉ๋‹ˆ๋‹ค.

์น˜๋ฃŒ ์ง€์—ฐ, ์ ‘๊ทผ์„ฑ ๋ฐ ์ž„์ƒ ์‹œํ—˜ ์„ค๊ณ„์˜ ์˜ํ–ฅ

์น˜๋ฃŒ ์‹œ์ž‘ ์‹œ๊ธฐ์™€ ์˜๋ฃŒ ์ ‘๊ทผ์„ฑ์€ SCLC ํ™˜์ž์˜ ์ƒ์กด์— ์ง์ ‘์ ์ธ ์˜ํ–ฅ์„ ๋ฏธ์นฉ๋‹ˆ๋‹ค. Meta-analysis[2]๋Š” ์•” ์น˜๋ฃŒ ์ง€์—ฐ์ด ์‚ฌ๋ง๋ฅ  ์ฆ๊ฐ€์™€ ์œ ์˜๋ฏธํ•˜๊ฒŒ ์—ฐ๊ด€๋˜์–ด ์žˆ์Œ์„ ๋ณด์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค. ์ˆ˜์ˆ ์ด๋‚˜ ์ „์‹  ์น˜๋ฃŒ์˜ ๊ฒฝ์šฐ, 4์ฃผ๋งˆ๋‹ค ์ง€์—ฐ๋  ๋•Œ๋งˆ๋‹ค Hazard Ratio์ด 1.01-1.28 ๋ฒ”์œ„์—์„œ ์ฆ๊ฐ€ํ•˜๋ฉฐ, ์ด๋Š” SCLC์ฒ˜๋Ÿผ ์ง„ํ–‰์ด ๋น ๋ฅธ ์ข…์–‘์—์„œ๋Š” ์น˜๋ฃŒ ์‹œ์ž‘์˜ ์‹ ์†์„ฑ์ด ํŠนํžˆ ์ค‘์š”ํ•จ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ ์ง€๋ฆฌ์  ์š”์ธ์— ๋”ฐ๋ฅธ ํ™”ํ•™์š”๋ฒ• ์ ‘๊ทผ์„ฑ์˜ ์ฐจ์ด[13]๋Š” 'postcode prescribing' ํ˜„์ƒ์„ ์ดˆ๋ž˜ํ•˜์—ฌ, ์ผ๋ถ€ ์ง€์—ญ์—์„œ๋Š” ์น˜๋ฃŒ ์ด์šฉ๋ฅ ์ด 4-5๋ฐฐ๊นŒ์ง€ ์ฐจ์ด๊ฐ€ ๋‚  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” SCLC์™€ ๊ฐ™์ด ๊ธด๊ธ‰ํ•œ ์น˜๋ฃŒ๊ฐ€ ํ•„์š”ํ•œ ํ™˜์ž๊ตฐ์—์„œ ๊ฑด๊ฐ• ๋ถˆํ‰๋“ฑ์„ ์‹ฌํ™”์‹œํ‚ฌ ์ˆ˜ ์žˆ๋Š” ์‹ฌ๊ฐํ•œ ๋ฌธ์ œ์ž…๋‹ˆ๋‹ค.

์ƒˆ๋กœ์šด ์น˜๋ฃŒ๋ฒ•์˜ ๊ฐœ๋ฐœ์„ ์œ„ํ•ด Master Protocols (Basket, Umbrella, Platform trials)๊ฐ€ ์ ์  ๋” ๋งŽ์ด ํ™œ์šฉ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค[11]. ์ด๋Ÿฌํ•œ ์„ค๊ณ„๋Š” ์—ฌ๋Ÿฌ ๊ฐ€์„ค์„ ๋™์‹œ์— ๊ฒ€์ฆํ•˜๊ฑฐ๋‚˜ ์น˜๋ฃŒ๊ตฐ์„ ๋™์ ์œผ๋กœ ์ถ”๊ฐ€/์ œ๊ฑฐํ•  ์ˆ˜ ์žˆ์–ด ํšจ์œจ์„ฑ์„ ๋†’์ž…๋‹ˆ๋‹ค. SCLC์™€ ๊ฐ™์€ ํฌ๊ท€ ๋˜๋Š” ๊ณต๊ฒฉ์ ์ธ ์•”์ข…์—์„œ ์ƒˆ๋กœ์šด ํ‘œ์  ์น˜๋ฃŒ์ œ๋‚˜ ๋ฉด์—ญ์น˜๋ฃŒ ์กฐํ•ฉ์„ ํ‰๊ฐ€ํ•  ๋•Œ, ์ด๋Ÿฌํ•œ ์œ ์—ฐํ•œ ์ž„์ƒ ์‹œํ—˜ ์„ค๊ณ„๋Š” ํ™˜์ž ๋ชจ์ง‘์˜ ์–ด๋ ค์›€๊ณผ ๋น ๋ฅธ ์งˆ๋ณ‘ ์ง„ํ–‰ ์†๋„๋ฅผ ๊ณ ๋ คํ•  ๋•Œ ์œ ์šฉํ•œ ๋„๊ตฌ๊ฐ€ ๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ํŠนํžˆ SCLC์—์„œ์˜ ICI์™€ ๋ฐฉ์‚ฌ์„  ์š”๋ฒ•์˜ ์กฐํ•ฉ[32]์€ OS(HR 0.865)์™€ PFS(HR 0.799)๋ฅผ ๊ฐœ์„ ์‹œํ‚ค๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ์œผ๋ฉฐ, ์ด๋Š” ๊ธฐ์กด ํ‘œ์ค€ ์น˜๋ฃŒ๋ฅผ ๋„˜์–ด์„œ๋Š” ์ƒˆ๋กœ์šด ์น˜๋ฃŒ ํŒจ๋Ÿฌ๋‹ค์ž„์„ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ ํ˜„ํ™ฉ ๋ฐ Gap: ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” Neoadjuvant ์š”๋ฒ•์ด๋‚˜ Real-world data ๋ถ„์„์ด ํฌํ•จ๋˜์–ด ์žˆ์œผ๋‚˜, ์น˜๋ฃŒ ์ง€์—ฐ์˜ ์ •๋Ÿ‰์  ์˜ํ–ฅ[2]์ด๋‚˜ ์ง€๋ฆฌ์  ์ ‘๊ทผ์„ฑ ๊ฒฉ์ฐจ[13]์— ๋Œ€ํ•œ ๋…ผ์˜๋Š” ์—†์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ Master Protocols๋ฅผ ํ†ตํ•œ SCLC ์‹ ์•ฝ ๊ฐœ๋ฐœ ์ „๋žต[11]์ด๋‚˜ ICI+Radiotherapy ์กฐํ•ฉ์˜ ๋ฉ”ํƒ€๋ถ„์„[32] ๋ฐ์ดํ„ฐ๋Š” ํ˜„์žฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ช…์‹œ์ ์œผ๋กœ ํฌํ•จ๋˜์–ด ์žˆ์ง€ ์•Š์Šต๋‹ˆ๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1Immune Checkpoint Inhibitors for the Treatment of Cancer: Clinical Impact and Mechanisms of Response and ResistanceAnnual Review of Pathology Mechanisms of Disease ยท 20202263PubMed ยท DOI
2Mortality due to cancer treatment delay: systematic review and meta-analysisBMJ ยท 20201742PubMed ยท DOI
3Notch Signaling in Development, Tissue Homeostasis, and DiseasePhysiological Reviews ยท 20171023PubMed ยท DOI
4Critical review of the molecular design progress in non-fullerene electron acceptors towards commercially viable organic solar cellsChemical Society Reviews ยท 2018991PubMed ยท DOI
5The Varied Roles of Notch in CancerAnnual Review of Pathology Mechanisms of Disease ยท 2016745PubMed ยท DOI
6The CXCL8-CXCR1/2 pathways in cancerCytokine & Growth Factor Reviews ยท 2016687PubMed ยท DOI
7The role of RAS oncogene in survival of patients with lung cancer: a systematic review of the literature with meta-analysisBritish Journal of Cancer ยท 2004612PubMed ยท DOI
8Progression and metastasis of lung cancerCancer and Metastasis Reviews ยท 2016585PubMed ยท DOI
9Carboplatin- or Cisplatin-Based Chemotherapy in First-Line Treatment of Small-Cell Lung Cancer: The COCIS Meta-Analysis of Individual Patient DataJournal of Clinical Oncology ยท 2012575PubMed ยท DOI
10The role of stereotactic radiosurgery in the management of patients with newly diagnosed brain metastases: a systematic review and evidence-based clinical practice guidelineJournal of Neuro-Oncology ยท 2009518PubMed ยท DOI
11Systematic review of basket trials, umbrella trials, and platform trials: a landscape analysis of master protocolsTrials ยท 2019450PubMed ยท DOI
12Role of p53 as a prognostic factor for survival in lung cancer: a systematic review of the literature with a meta-analysisEuropean Respiratory Journal ยท 2001418PubMed ยท DOI
13A systematic review of geographical variation in access to chemotherapyBMC Cancer ยท 2015405PubMed ยท DOI
14Systematic Review Evaluating the Timing of Thoracic Radiation Therapy in Combined Modality Therapy for Limited-Stage Small-Cell Lung CancerJournal of Clinical Oncology ยท 2004391PubMed ยท DOI
15The role of steroids in the management of brain metastases: a systematic review and evidence-based clinical practice guidelineJournal of Neuro-Oncology ยท 2009361PubMed ยท DOI
16The biology and treatment of Merkel cell carcinoma: current understanding and research prioritiesNature Reviews Clinical Oncology ยท 2018322PubMed ยท DOI
17Clinical development of new drugโ€“radiotherapy combinationsNature Reviews Clinical Oncology ยท 2016317PubMed ยท DOI
18Antibodyโ€“Drug Conjugates for Cancer TreatmentAnnual Review of Medicine ยท 2018311PubMed ยท DOI
19Prognostic role of systemic immune-inflammation index in solid tumors: a systematic review and meta-analysisOncotarget ยท 2017308PubMed ยท DOI
20The role of whole brain radiation therapy in the management of newly diagnosed brain metastases: a systematic review and evidence-based clinical practice guidelineJournal of Neuro-Oncology ยท 2009303PubMed ยท DOI
21Is there a case for cisplatin in the treatment of small-cell lung cancer? A meta-analysis of randomized trials of a cisplatin-containing regimen versus a regimen without this alkylating agentBritish Journal of Cancer ยท 2000293PubMed ยท DOI
22Sex-Based Heterogeneity in Response to Lung Cancer Immunotherapy: A Systematic Review and Meta-AnalysisJNCI Journal of the National Cancer Institute ยท 2019288PubMed ยท DOI
23Prophylactic cranial irradiation in small cell lung cancer: a systematic review of the literature with meta-analysisDOAJ (DOAJ: Directory of Open Access Journals) ยท 2001282DOI
24Ki-67 expression and patients survival in lung cancer: systematic review of the literature with meta-analysisBritish Journal of Cancer ยท 2004281PubMed ยท DOI
25Consolidated Evidence and New Frontiers of Liquid Biopsy in Lung Cancer: A Narrative ReviewCells ยท 20260DOI
26Profiling immune check point inhibitors โ€“ Induced arthritis: A systematic review and pooled analyses of cohort studiesJournal of Autoimmunity ยท 20260PubMed ยท DOI
27Statins use and prognosis in lung cancer patients treated with immune checkpoint inhibitors: evidence from a meta-analysisFrontiers in Immunology ยท 20260DOI
28Exploring the Regenerative Role of Exosomes as a Potential Therapy for Peripheral Nerve Injury: A Systematic ReviewBiomolecules & Therapeutics ยท 20260PubMed ยท DOI
29Clinical value of combined SHOX2 and RASSF1A methylation in lung cancer diagnosis across tissue and liquid biopsy samples: a systematic review and meta-analysisOncology Reviews ยท 20260DOI
30Reninโ€“angiotensin system inhibitors and immunotherapy outcomes in lung cancer: a systematic review and meta-analysis with complementary transcriptomic analysesFrontiers in Immunology ยท 20260PubMed ยท DOI
31Mapping B7-H3 in the tumour microenvironment: a systematic review of stromal, vascular and immune expressionBritish Journal of Cancer ยท 20260PubMed ยท DOI
32The clinical value of adding immune checkpoint inhibitors to radiotherapy for cancer: a systematic review and meta-analysisAnnals of Medicine ยท 20260PubMed ยท DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1Small Cell Lung Cancer: A Review.JAMA ยท 2025-Jun-03PubMed ยท DOI
2Clinical insights into small cell lung cancer: Tumor heterogeneity, diagnosis, therapy, and future directions.CA: a cancer journal for clinicians ยท 2023PubMed ยท DOI
3Tumor- and circulating-free DNA methylation identifies clinically relevant small cell lung cancer subtypes.Cancer cell ยท 2024-Feb-12PubMed ยท DOI
4B7-H3/CD276 and small-cell lung cancer: What's new?Translational oncology ยท 2024-JanPubMed ยท DOI
5Emerging advances in defining the molecular and therapeutic landscape of small-cell lung cancer.Nature reviews. Clinical oncology ยท 2024-AugPubMed ยท DOI
6YAP silencing by RB1 mutation is essential for small-cell lung cancer metastasis.Nature communications ยท 2023-Sep-22PubMed ยท DOI
7New developments in immunotherapy for SCLC.Journal for immunotherapy of cancer ยท 2025-Jan-06PubMed ยท DOI
8Cell-free and extrachromosomal DNA profiling of small cell lung cancer.Trends in molecular medicine ยท 2025-JanPubMed ยท DOI
9Crosstalk between small-cell lung cancer cells and astrocytes mimics brain development to promote brain metastasis.Nature cell biology ยท 2023-OctPubMed ยท DOI
10Jumonji histone demethylases are therapeutic targets in small cell lung cancer.Oncogene ยท 2024-SepPubMed ยท DOI
Radiomics โ€” review 32ํŽธ ยท ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ 14ํŽธ ยท ์—…๋ฐ์ดํŠธ 2026-09-12 ๐Ÿ†• ์ตœ๊ทผ ์—…๋ฐ์ดํŠธ
์ง„๋‹จ ๋ฐ ์˜์ƒ ๋ถ„์„์˜ ์ •ํ™•์„ฑ๊ณผ ํ‘œ์ค€ํ™”

Radiomics์™€ ์ธ๊ณต์ง€๋Šฅ(AI) ๊ธฐ๋ฐ˜ ์˜๋ฃŒ ์˜์ƒ ๋ถ„์„์€ ์งˆ๋ณ‘์˜ ์กฐ๊ธฐ ๋ฐœ๊ฒฌ๊ณผ ์ •ํ™•ํ•œ ์ง„๋‹จ์—์„œ ํ˜์‹ ์ ์ธ ๊ฐ€๋Šฅ์„ฑ์„ ์ œ์‹œํ•˜๊ณ  ์žˆ์œผ๋‚˜, ๊ทธ ์ž„์ƒ์  ์œ ์šฉ์„ฑ์€ ๋ชจ๋ธ์˜ ์žฌํ˜„์„ฑ(reproducibility)๊ณผ ์™ธ๋ถ€ ๊ฒ€์ฆ(external validation)์— ํฌ๊ฒŒ ์ขŒ์šฐ๋œ๋‹ค. ํŠนํžˆ ์ข…์–‘ํ•™ ๋ถ„์•ผ์—์„œ๋Š” CT, MRI, PET ๋“ฑ ๋‹ค์–‘ํ•œ ์˜์ƒ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ์—์„œ ์ถ”์ถœ๋œ radiomic features๊ฐ€ ๋ณ‘๋ฆฌํ•™์  ์ƒํƒœ์™€ ์ƒ๊ด€๊ด€๊ณ„๊ฐ€ ์žˆ์Œ์ด ์ž…์ฆ๋˜์—ˆ์œผ๋‚˜, ์ด๋Ÿฌํ•œ ํŠน์ง•๋“ค์˜ ์•ˆ์ •์„ฑ์€ ์ด๋ฏธ์ง€ ํš๋“ ์„ค์ •, ์žฌ๊ตฌ์„ฑ ์•Œ๊ณ ๋ฆฌ์ฆ˜, ์ „์ฒ˜๋ฆฌ ๊ณผ์ • ๋ฐ ์ถ”์ถœ ์†Œํ”„ํŠธ์›จ์–ด์— ๋”ฐ๋ผ ํฌ๊ฒŒ ๋‹ฌ๋ผ์งˆ ์ˆ˜ ์žˆ๋‹ค [5]. ์˜ˆ๋ฅผ ๋“ค์–ด, 1์ฐจ ํ†ต๊ณ„์  ํŠน์ง•(first-order features)์€ ํ˜•ํƒœํ•™์ (shape) ๋ฐ ์งˆ๊ฐ(textural) ํŠน์ง•๋ณด๋‹ค ์ผ๋ฐ˜์ ์œผ๋กœ ๋” ๋†’์€ ์žฌํ˜„์„ฑ์„ ๋ณด์ด์ง€๋งŒ, entropy์™€ ๊ฐ™์€ ํŠน์ • ์งˆ๊ฐ ํŠน์ง•์€ ์—ฌ์ „ํžˆ ๋ณ€๋™์„ฑ์ด ํฐ ๊ฒƒ์œผ๋กœ ๋ณด๊ณ ๋œ๋‹ค [5]. ์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์  ๋ฏผ๊ฐ์„ฑ์€ ๋ชจ๋ธ์˜ ์ผ๋ฐ˜ํ™” ๊ฐ€๋Šฅ์„ฑ์„ ์ €ํ•ดํ•˜๋Š” ์ฃผ์š” ์š”์ธ์œผ๋กœ ์ž‘์šฉํ•˜๋ฉฐ, ํŠนํžˆ ๋‹ค๊ธฐ๊ด€ ์—ฐ๊ตฌ์—์„œ ํ‘œ์ค€ํ™”๋œ ํ”„๋กœํ† ์ฝœ ๋ถ€์žฌ๋Š” ๊ฒฐ๊ณผ์˜ ๋น„๊ต๋ฅผ ์–ด๋ ต๊ฒŒ ๋งŒ๋“ ๋‹ค.

๋”ฅ๋Ÿฌ๋‹(DL) ์•Œ๊ณ ๋ฆฌ์ฆ˜์˜ ์ง„๋‹จ ์ •ํ™•๋„์— ๋Œ€ํ•œ ์ฒด๊ณ„์  ๊ฒ€ํ† ๋Š” ๋ถ„์•ผ๋ณ„ ์ฐจ์ด๋ฅผ ๋ณด์—ฌ์ค€๋‹ค. ์•ˆ๊ณผ ์˜์ƒ์—์„œ๋Š” ๋‹น๋‡จ๋ง๋ง‰๋ณ‘์ฆ์ด๋‚˜ ํ™ฉ๋ฐ˜๋ณ€์„ฑ ์ง„๋‹จ์—์„œ AUC๊ฐ€ 0.933~1.0์— ๋‹ฌํ•˜๋Š” ๋†’์€ ์„ฑ๋Šฅ์„ ๋ณด์˜€์œผ๋‚˜, ํ˜ธํก๊ธฐ ์˜์ƒ์˜ ํ๊ฒฐ์ ˆ ์ง„๋‹จ(AUC 0.864~0.937) ๋ฐ ์œ ๋ฐฉ ์˜์ƒ์˜ ์œ ๋ฐฉ์•” ์ง„๋‹จ(AUC 0.868~0.909)์—์„œ๋„ ์šฐ์ˆ˜ํ•œ ๊ฒฐ๊ณผ๋ฅผ ๊ธฐ๋กํ•˜๋ฉด์„œ๋„ ์—ฐ๊ตฌ ๊ฐ„ ์ด์งˆ์„ฑ(heterogeneity)์ด ๋งค์šฐ ๋†’์•˜๋‹ค [4]. ์ด๋Š” ๋ฐฉ๋ฒ•๋ก , ์šฉ์–ด ๋ฐ ๊ฒฐ๊ณผ ์ธก์ •์น˜์˜ ๋ถˆ์ผ์น˜๋กœ ์ธํ•ด DL ์•Œ๊ณ ๋ฆฌ์ฆ˜์˜ ์‹ค์ œ ์ง„๋‹จ ์ •ํ™•์„ฑ์ด ๊ณผ๋Œ€ํ‰๊ฐ€๋  ์ˆ˜ ์žˆ์Œ์„ ์‹œ์‚ฌํ•œ๋‹ค. ๋˜ํ•œ, ๋ฐฉ์‚ฌ์„ ํ•™ ๋ถ„์•ผ์—์„œ ์ˆ˜ํ–‰๋œ ๊ด‘๋ฒ”์œ„ํ•œ ๊ฒ€ํ† (RAISE)์— ๋”ฐ๋ฅด๋ฉด, ๋Œ€๋ถ€๋ถ„์˜ ์—ฐ๊ตฌ๊ฐ€ ํ›„ํ–ฅ์  ์ฝ”ํ˜ธํŠธ ์—ฐ๊ตฌ์ด๋ฉฐ, ์™ธ๋ถ€ ๊ฒ€์ฆ ์‹œ ์„ฑ๋Šฅ์ด ํ‰๊ท  6% ๊ฐ์†Œํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋‚˜ํƒ€๋‚ฌ๋‹ค [18]. ์ด๋Š” ๋‚ด๋ถ€ ๊ฒ€์ฆ์—์„œ์˜ ๋†’์€ Dice ๊ณ„์ˆ˜(0.89)๋‚˜ AUC(0.903)๊ฐ€ ์‹ค์ œ ์ž„์ƒ ํ™˜๊ฒฝ์—์„œ์˜ ์„ฑ๋Šฅ์„ ์™„์ „ํžˆ ๋ฐ˜์˜ํ•˜์ง€ ๋ชปํ•จ์„ ์˜๋ฏธํ•œ๋‹ค.

ํŠน์ • ์งˆํ™˜๊ตฐ์— ์ดˆ์ ์„ ๋งž์ถ˜ ์—ฐ๊ตฌ๋“ค๋„ ์œ ์‚ฌํ•œ ํŒจํ„ด์„ ๋ณด์ธ๋‹ค. ์œ ๋ฐฉ์•” ์ง„๋‹จ์—์„œ CNN(Convolutional Neural Network)์ด ๊ฐ€์žฅ ์ •ํ™•ํ•˜๊ณ  ๋„๋ฆฌ ์‚ฌ์šฉ๋˜๋Š” ๋ชจ๋ธ๋กœ ํ™•์ธ๋˜์—ˆ์œผ๋ฉฐ, ์œ ๋ฐฉ์•” ๊ฒ€์ถœ์˜ ์ •ํ™•์„ฑ ํ‰๊ฐ€์—๋Š” Accuracy ์ง€ํ‘œ๊ฐ€ ์ฃผ๋กœ ์‚ฌ์šฉ๋˜์—ˆ๋‹ค [22]. ๋ฐ˜๋ฉด, ๋‹ด๋‚ญ์•”(Gallbladder Cancer, GBC)์˜ ๊ฒฝ์šฐ AI ๊ธฐ๋ฐ˜ ์˜์ƒ ๋ชจ๋ธ์€ ๋ฏผ๊ฐ๋„ 0.888, ํŠน์ด๋„ 0.838, AUC 0.921์„ ๊ธฐ๋กํ•˜์—ฌ ๋†’์€ ์ง„๋‹จ ์ •ํ™•์„ฑ์„ ๋ณด์˜€์œผ๋‚˜, ์ž„์ƒ ๊ฒ€์ฆ ์ฝ”ํ˜ธํŠธ์™€ ๊ธฐ์ˆ ์  ๋ฒค์น˜๋งˆํฌ ์—ฐ๊ตฌ๊ฐ€ ํ˜ผ์žฌ๋˜์–ด ์žˆ์–ด ํ•ด์„์— ์ฃผ์˜๊ฐ€ ํ•„์š”ํ•˜๋‹ค [31]. ๋˜ํ•œ, ๋Œ€์กฐ์ œ ๊ฐ•ํ™” ์œ ๋ฐฉ์ดฌ์˜์ˆ (CEM)์—์„œ AI์˜ ์ ์šฉ์€ ๋ณ‘๋ณ€์˜ ์„ ๋ช…๋„๋ฅผ ๋†’์ด๊ณ  ์ž๋™ ๊ฒ€์ถœ์„ ์ง€์›ํ•˜์ง€๋งŒ, ์†Œ๊ทœ๋ชจ ์ƒ˜ํ”Œ ํฌ๊ธฐ, ํš๋“ ํ”„๋กœํ† ์ฝœ์˜ ์ด์งˆ์„ฑ, ๊ทธ๋ฆฌ๊ณ  ์ž„์ƒ์ ยทํ˜ธ๋ฅด๋ชฌ์  ์ •๋ณด ํ†ตํ•ฉ์˜ ๋ถ€์žฌ๋กœ ์ธํ•ด ์ž„์ƒ ๊ตฌํ˜„์ด ์ œํ•œ์ ์ด๋‹ค [30]. ์ด๋Ÿฌํ•œ ์ง„๋‹จ ๋ชจ๋ธ๋“ค์˜ ์„ฑ์ˆ™๋„๋Š” ์•ฝ๋ฌผ ๊ฐœ๋ฐœ ๋‹จ๊ณ„์— ๋น„์œ ํ•  ๋•Œ, ๋Œ€๋ถ€๋ถ„์˜ radiomics ์—ฐ๊ตฌ๊ฐ€ Phase 0~II ๋‹จ๊ณ„์— ๋จธ๋ฌผ๋Ÿฌ ์žˆ์œผ๋ฉฐ, ์‹ค์ œ ์ž„์ƒ ์ ์šฉ์„ ์œ„ํ•œ Phase III ์ˆ˜์ค€์˜ ๊ฒ€์ฆ์€ ์•„์ง ๋ถ€์กฑํ•˜๋‹ค [17].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” ์ด๋ฏธ 'Radiomics Quality Score 2.0' ๋ฐ 'feature normalization methods'์™€ ๊ฐ™์€ ๋ฐฉ๋ฒ•๋ก ์  ํ‘œ์ค€ํ™” ๊ด€๋ จ ๋…ผ๋ฌธ์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด, ๊ธฐ์ˆ ์  ์žฌํ˜„์„ฑ ๋ฌธ์ œ์— ๋Œ€ํ•œ ๊ธฐ์ดˆ๊ฐ€ ๋งˆ๋ จ๋˜์–ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ๋‹ค์–‘ํ•œ ์˜์ƒ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ(CT, MRI, PET) ๊ฐ„์˜ ๊ต์ฐจ ๊ฒ€์ฆ ์—ฐ๊ตฌ๋‚˜, ์‹ค์ œ ์ž„์ƒ ์›Œํฌํ”Œ๋กœ์šฐ์—์„œ์˜ ํ†ตํ•ฉ ๊ฐ€๋Šฅ์„ฑ์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๊ฐ€์ด๋“œ๋ผ์ธ์€ ์—ฌ์ „ํžˆ ๋ถ€์กฑํ•˜๋‹ค. ํŠนํžˆ, ๋‹จ์ผ ๊ธฐ๊ด€ ๋ฐ์ดํ„ฐ๋กœ ๊ฐœ๋ฐœ๋œ ๋ชจ๋ธ์ด ๋‹ค๊ธฐ๊ด€ ํ™˜๊ฒฝ์—์„œ ์–ผ๋งˆ๋‚˜ ๊ฒฌ๊ณ ํ•˜๊ฒŒ ์ž‘๋™ํ•˜๋Š”์ง€์— ๋Œ€ํ•œ ์™ธ๋ถ€ ๊ฒ€์ฆ ์—ฐ๊ตฌ์˜ ๋ถ€์žฌ๋Š” ์ฃผ์š” gap์œผ๋กœ ์ง€์ ๋œ๋‹ค.

๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ๋ฐ์ดํ„ฐ ์œตํ•ฉ ๋ฐ ์ž๊ธฐ๊ฐ๋… ํ•™์Šต(Self-Supervised Learning)

๋‹จ์ผ ์˜์ƒ ๋ฐ์ดํ„ฐ๋งŒ์œผ๋กœ๋Š” ํ™˜์ž์˜ ์ „์ฒด์ ์ธ ์ž„์ƒ์  ๋งฅ๋ฝ์„ ํฌ์ฐฉํ•˜๊ธฐ ์–ด๋ ต๊ธฐ ๋•Œ๋ฌธ์—, ์˜๋ฃŒ ์˜์ƒ๊ณผ ์ „์ž๊ฑด๊ฐ•๊ธฐ๋ก(EHR), ์œ ์ „์ฒด ๋ฐ์ดํ„ฐ ๋“ฑ ๋‹ค์–‘ํ•œ ์†Œ์Šค์˜ ๋ฐ์ดํ„ฐ๋ฅผ ์œตํ•ฉํ•˜๋Š” ์ ‘๊ทผ๋ฒ•์ด ์ค‘์š”ํ•ด์ง€๊ณ  ์žˆ๋‹ค. ๋”ฅ๋Ÿฌ๋‹ ๋ชจ๋ธ์€ ์ฃผ๋กœ ํ”ฝ์…€ ๊ฐ’ ์ •๋ณด๋งŒ์„ ์ฒ˜๋ฆฌํ•˜์ง€๋งŒ, ์ž„์ƒ๋ ฅ ๋ฐ ์‹คํ—˜์‹ค ๋ฐ์ดํ„ฐ์™€ ๊ฐ™์€ ๋น„์˜์ƒ ์ •๋ณด๋ฅผ ํ†ตํ•ฉํ•จ์œผ๋กœ์จ ์ง„๋‹จ ์ •ํ™•๋„์™€ ์ž„์ƒ์  ์˜์‚ฌ๊ฒฐ์ •์˜ ์งˆ์„ ํ–ฅ์ƒ์‹œํ‚ฌ ์ˆ˜ ์žˆ๋‹ค [6]. ์ด๋Ÿฌํ•œ ๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ์œตํ•ฉ(multimodal data fusion)์€ ํŠนํžˆ ๋‡Œ์กธ์ค‘(Stroke)๊ณผ ๊ฐ™์€ ์‹ ๊ฒฝํ•™์  ์งˆํ™˜์—์„œ ๋‘๋“œ๋Ÿฌ์ง„๋‹ค. EEG์˜ ๋ฐ€๋ฆฌ์ดˆ ๋‹จ์œ„ ์‹ ๊ฒฝ ํ™œ๋™ ์ •๋ณด์™€ MRI์˜ ๊ณ ํ•ด์ƒ๋„ ๊ตฌ์กฐ/๊ธฐ๋Šฅ ์˜์ƒ์„ ๊ฒฐํ•ฉํ•˜๋ฉด, ๊ธ‰์„ฑ๊ธฐ ์น˜๋ฃŒ ๊ฒฐ์ •๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์˜ˆํ›„ ํ‰๊ฐ€ ๋ฐ ์žฌํ™œ ์ „๋žต ์ˆ˜๋ฆฝ์— ๋” ์œ ์šฉํ•œ ์ •๋ณด๋ฅผ ์ œ๊ณตํ•  ์ˆ˜ ์žˆ๋‹ค [29]. ๊ทธ๋Ÿฌ๋‚˜ ์ด๋Ÿฌํ•œ ์œตํ•ฉ ๊ธฐ์ˆ ์˜ ์ž„์ƒ์  ์ „ํ™˜์€ ๊ณต๊ฐœ๋œ ๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ๋ฐ์ดํ„ฐ์˜ ๋ถ€์กฑ, ๊ธ‰์„ฑ๊ธฐ์—์„œ ๋งŒ์„ฑ๊ธฐ๊นŒ์ง€๋ฅผ ์•„์šฐ๋ฅด๋Š” ์ข…๋‹จ ์—ฐ๊ตฌ(longitudinal studies)์˜ ๋ถ€์žฌ, ๊ทธ๋ฆฌ๊ณ  ํš๋“ ๋ฐ ์ „์ฒ˜๋ฆฌ ํ”„๋กœํ† ์ฝœ์˜ ํ‘œ์ค€ํ™” ๊ฒฐ์—ฌ๋กœ ์ธํ•ด ์—ฌ์ „ํžˆ ๋„์ „์ ์ด๋‹ค [29].

๋ฐ์ดํ„ฐ ๋ผ๋ฒจ๋ง์˜ ๋น„์šฉ๊ณผ ์‹œ๊ฐ„์  ์ œ์•ฝ์€ ์˜๋ฃŒ ์˜์ƒ AI ๊ฐœ๋ฐœ์˜ ์ฃผ์š” ์žฅ๋ฒฝ์ด๋‹ค. ์ด์— ์ž๊ธฐ๊ฐ๋… ํ•™์Šต(Self-supervised learning)์ด ๋Œ€์•ˆ์œผ๋กœ ๋ถ€์ƒํ•˜๊ณ  ์žˆ๋‹ค. ์ด ๋ฐฉ๋ฒ•์€ ๋ฐฉ๋Œ€ํ•œ unlabeled medical datasets์—์„œ ์œ ์šฉํ•œ ํ†ต์ฐฐ์„ ํ•™์Šตํ•˜์—ฌ, ๋ผ๋ฒจ๋œ ๋ฐ์ดํ„ฐ๊ฐ€ ์ ์€ ์ƒํ™ฉ์—์„œ๋„ ๊ฐ•๊ฑดํ•œ ๋ชจ๋ธ์„ ๊ตฌ์ถ•ํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•œ๋‹ค [10]. 2012๋…„๋ถ€ํ„ฐ 2022๋…„๊นŒ์ง€์˜ ์—ฐ๊ตฌ๋ฅผ ๊ฒ€ํ† ํ•œ ๊ฒฐ๊ณผ, ์ž๊ธฐ๊ฐ๋… ํ•™์Šต ์ „๋žต์€ ์˜๋ฃŒ ์˜์ƒ ๋ถ„๋ฅ˜ ๋ชจ๋ธ ๊ฐœ๋ฐœ์— ์žˆ์–ด ์ค‘์š”ํ•œ ๊ธฐ์—ฌ๋ฅผ ํ•  ์ˆ˜ ์žˆ์œผ๋ฉฐ, ํ–ฅํ›„ ์—ฐ๊ตฌ์ž๋“ค์—๊ฒŒ ๊ตฌํ˜„ ๊ฐ€์ด๋“œ๋ผ์ธ์„ ์ œ๊ณตํ•ด์•ผ ํ•จ์ด ๊ฐ•์กฐ๋˜์—ˆ๋‹ค [10]. ๋˜ํ•œ, Vision Transformers(ViTs)์™€ Convolutional Neural Networks(CNNs)์˜ ๋น„๊ต ์—ฐ๊ตฌ์—์„œ๋Š” ViTs๊ฐ€ ๋ณต์žกํ•œ ์˜๋ฃŒ ๋ฐ์ดํ„ฐ์…‹ ์ฒ˜๋ฆฌ์—์„œ CNN ๋Œ€๋น„ ์šฐ์ˆ˜ํ•œ ์„ฑ๋Šฅ๊ณผ ํ™•์žฅ์„ฑ์„ ๋ณด์ด๋ฉฐ, ์‚ฌ์ „ ํ•™์Šต(pre-training)์˜ ์ค‘์š”์„ฑ์ด ๋ถ€๊ฐ๋˜๊ณ  ์žˆ๋‹ค [14]. YOLO(You Only Look Once) ์•Œ๊ณ ๋ฆฌ์ฆ˜๋„ ๊ฐ์ฒด ํƒ์ง€ ๋ถ„์•ผ์—์„œ ๋†’์€ ํšจ์œจ์„ฑ์„ ์ž…์ฆํ–ˆ์œผ๋‚˜, ๊ท ํ˜• ์žกํžŒ ์ฃผ์„ ๋ฐ์ดํ„ฐ์…‹์˜ ํ•„์š”์„ฑ๊ณผ ๋†’์€ ๊ณ„์‚ฐ ์š”๊ตฌ์‚ฌํ•ญ์ด ํ•œ๊ณ„๋กœ ์ง€์ ๋œ๋‹ค [15].

ํŠนํžˆ ์œ„์•” ๋ฐ ์‹๋„์œ„junction ์•”(Gastric and GEJ adenocarcinoma)์—์„œ H&E ์—ผ์ƒ‰ ์Šฌ๋ผ์ด๋“œ๋กœ๋ถ€ํ„ฐ MSI/dMMR, EBV, HER2, PD-L1, CLDN18.2์™€ ๊ฐ™์€ ์ค‘์š”ํ•œ ๋ฐ”์ด์˜ค๋งˆ์ปค ์ƒํƒœ๋ฅผ ์˜ˆ์ธกํ•˜๋Š” DL ๋ชจ๋ธ์˜ ์ •ํ™•๋„๊ฐ€ ํ‰๊ฐ€๋˜์—ˆ๋‹ค. ์ด ์—ฐ๊ตฌ๋“ค์€ ๋ถ„์ž ๋ฐ ๋ฉด์—ญ์กฐ์งํ™”ํ•™(IHC) ๊ฒ€์‚ฌ์˜ ๋น„์šฉ๊ณผ ์ ‘๊ทผ์„ฑ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•  ์ˆ˜ ์žˆ๋Š” ์ €๋น„์šฉ ์„ ๋ณ„ ๋„๊ตฌ๋กœ์„œ์˜ ์ž ์žฌ๋ ฅ์„ ๋ณด์—ฌ์ฃผ์—ˆ์œผ๋‚˜, ๋…๋ฆฝ์ ์ธ ๊ฒ€์ฆ ๋ฐ์ดํ„ฐ์…‹์˜ ๋ถ€์กฑ์œผ๋กœ ์ธํ•ด ํ†ต๊ณ„์  ํ’€๋ง์ด ์–ด๋ ค์› ๋‹ค [28]. ๋˜ํ•œ, MASLD(Metabolic dysfunction-associated steatotic liver disease) ์ง„๋‹จ์—์„œ AI๋Š” ์ดˆ์ŒํŒŒ, CT, MRI, ํƒ„์„ฑ์˜์ƒ(elastography)์„ ํ™œ์šฉํ•˜์—ฌ ์ง€๋ฐฉ๋ณ€(steatosis) ์ง„๋‹จ(AUROC 0.85-0.99) ๋ฐ ์„ฌ์œ ํ™”(fibrosis) ๋‹จ๊ณ„ ๊ตฌ๋ถ„(AUROC 0.82-0.97)์—์„œ ๋›ฐ์–ด๋‚œ ํšจ๊ณผ๋ฅผ ๋ณด์˜€๋‹ค [25, 32]. ํŠนํžˆ ๊ธฐ๊ณ„ํ•™์Šต ์•Œ๊ณ ๋ฆฌ์ฆ˜์ด ์ผ๋ฐ˜์ ์ธ ์‹คํ—˜์‹ค ์ง€ํ‘œ, ๋Œ€์‚ฌ์ฒดํ•™, ๋‹จ๋ฐฑ์งˆ์ฒดํ•™ ๋ฐ ์žฅ๋‚ด ๋ฏธ์ƒ๋ฌผ๊ตฐ ๋ฐ์ดํ„ฐ๋ฅผ ํ†ตํ•ฉํ•  ๋•Œ, FIB-4๋‚˜ NAFLD Fibrosis Score์™€ ๊ฐ™์€ ๊ธฐ์กด ๋น„์นจ์Šต์  ์•Œ๊ณ ๋ฆฌ์ฆ˜๋ณด๋‹ค ์šฐ์ˆ˜ํ•œ ์„ฑ๋Šฅ์„ ๋ฐœํœ˜ํ–ˆ๋‹ค [25].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” 'Radiogenomic-based multiomic analysis' ๋ฐ 'Imaging Phenotyping' ๊ด€๋ จ ๋…ผ๋ฌธ์ด ์žˆ์–ด, ์˜์ƒ๊ณผ ์œ ์ „์ฒด/๋ณ‘๋ฆฌ ๋ฐ์ดํ„ฐ์˜ ์—ฐ๊ด€์„ฑ์— ๋Œ€ํ•œ ์ดํ•ด๊ฐ€ ์–ด๋А ์ •๋„ ํ˜•์„ฑ๋˜์–ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ EHR๊ณผ์˜ ์‹ค์‹œ๊ฐ„ ์œตํ•ฉ์„ ํ†ตํ•œ ์ž„์ƒ ์˜์‚ฌ๊ฒฐ์ • ์ง€์› ์‹œ์Šคํ…œ(CDSS) ๊ตฌ์ถ•, ๊ทธ๋ฆฌ๊ณ  ์ž๊ธฐ๊ฐ๋… ํ•™์Šต์„ ํ™œ์šฉํ•œ ๋Œ€๊ทœ๋ชจ unlabeled ๋ฐ์ดํ„ฐ ์ฒ˜๋ฆฌ์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ๊ตฌํ˜„ ๊ฐ€์ด๋“œ๋ผ์ธ์€ ์•„์ง ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ฏธ๋น„ํ•˜๋‹ค. ํŠนํžˆ, ๋‹ค๊ธฐ๊ด€ ๋ฐ์ดํ„ฐ๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ํ•œ ์ž๊ธฐ๊ฐ๋… ํ•™์Šต ๋ชจ๋ธ์˜ ์ผ๋ฐ˜ํ™” ๊ฐ€๋Šฅ์„ฑ์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ์ถ”๊ฐ€๋กœ ํ•„์š”ํ•˜๋‹ค.

๋ฉด์—ญ ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก ๋ฐ ์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(TME) ๋ถ„์„

Radiomics๋Š” ๋‹จ์ˆœํ•œ ์ง„๋‹จ์„ ๋„˜์–ด, ์ข…์–‘ ๋ฏธ์„ธํ™˜๊ฒฝ(Tumor Microenvironment, TME)์˜ ํŠน์„ฑ์„ ๋น„์นจ์Šต์ ์œผ๋กœ ํ‰๊ฐ€ํ•˜๊ณ  ๋ฉด์—ญ ์น˜๋ฃŒ(Immunotherapy) ๋ฐ˜์‘์„ฑ์„ ์˜ˆ์ธกํ•˜๋Š” ๋ฐ ์ค‘์š”ํ•œ ์—ญํ• ์„ ํ•œ๋‹ค. CD8+ ์ข…์–‘ ์นจ์œค ๋ฆผํ”„๊ตฌ(CD8+ TILs)๋Š” ์•” ์„ธํฌ ์‚ฌ๋ฉธ์— ํ•ต์‹ฌ์ ์ธ ์—ญํ• ์„ ํ•˜๋ฉฐ, ๋†’์€ CD8+ TILs ์ˆ˜์ค€์€ ๋ฉด์—ญ ์ฒดํฌํฌ์ธํŠธ ์–ต์ œ์ œ(ICIs) ์น˜๋ฃŒ๋ฅผ ๋ฐ›๋Š” ํ™˜์ž์—์„œ ์ „์ฒด ์ƒ์กด์œจ(OS; HR 0.52), ๋ฌด์ง„ํ–‰ ์ƒ์กด์œจ(PFS; HR 0.52) ๋ฐ ๊ฐ๊ด€์  ๋ฐ˜์‘๋ฅ (ORR; OR 4.08)์˜ ์œ ์˜๋ฏธํ•œ ๊ฐœ์„ ๊ณผ ๊ด€๋ จ์ด ์žˆ๋‹ค [11]. ์ด๋Ÿฌํ•œ ์—ฐ๊ด€์„ฑ์€ NSCLC, ํ‘์ƒ‰์ข… ๋“ฑ ๋‹ค์–‘ํ•œ ์•” ์ข…๋ฅ˜์™€ ์น˜๋ฃŒ ๋ฐฉ์‹(ICI monotherapy ๋˜๋Š” combination therapy)์— ๊ฑธ์ณ ์ผ๊ด€๋˜๊ฒŒ ๊ด€์ฐฐ๋˜์—ˆ๋‹ค. ๋ฐ˜๋ฉด, ๋ง์ดˆํ˜ˆ์•ก ๋‚ด ์ˆœํ™˜ CD8+ T ์„ธํฌ๋Š” OS๋‚˜ PFS ๊ฐœ์„ ๊ณผ ์œ ์˜๋ฏธํ•œ ๊ด€๋ จ์ด ์—†์—ˆ๋‹ค [11].

Radiomic features๋Š” TME์˜ ๋ฉด์—ญ ํ”„๋กœํŒŒ์ผ๊ณผ ๋ฐ€์ ‘ํ•˜๊ฒŒ ์—ฐ๊ด€๋˜์–ด ์žˆ๋‹ค. NSCLC ํ™˜์ž์—์„œ ์˜์ƒ ๊ธฐ๋ฐ˜ ๋ฐ”์ด์˜ค๋งˆ์ปค๋Š” PD-L1 ๋ฐœํ˜„ ์ˆ˜์ค€ ๋ฐ ์ƒ์กด ๊ฒฐ๊ณผ์™€ ์ƒ๊ด€๊ด€๊ณ„๊ฐ€ ์žˆ์œผ๋ฉฐ, ์ด๋Š” Nivolumab์ด๋‚˜ Pembrolizumab๊ณผ ๊ฐ™์€ ICIs ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก์— ํ™œ์šฉ๋  ์ˆ˜ ์žˆ๋‹ค [Library: Imaging-Based Biomarkers Predict Programmed Death-Ligand 1...]. ๋˜ํ•œ, delta-radiomics features(์น˜๋ฃŒ ์ „ํ›„์˜ ์˜์ƒ ๋ณ€ํ™”)์™€ ํ˜ˆ์•กํ•™์  ์ง€ํ‘œ๋ฅผ ๊ฒฐํ•ฉํ•˜๋ฉด, ์žฌsectable NSCLC ํ™˜์ž์—์„œ neoadjuvant immunochemotherapy ํ›„ ๋ณ‘๋ฆฌ์  ์™„์ „ ๋ฐ˜์‘(pCR)์„ ์˜ˆ์ธกํ•˜๋Š” ๋ฐ ์œ ์šฉํ•˜๋‹ค [Library: Delta-radiomics features combined with haematological index...]. ์ด๋Ÿฌํ•œ ์ ‘๊ทผ๋ฒ•์€ ์ข…์–‘ ๋‚ด ์ด์งˆ์„ฑ(intratumor heterogeneity)์„ ๋ฐ˜์˜ํ•˜์—ฌ, ๋‹จ์ผ ์‹œ์ ์˜ ์˜์ƒ๋ณด๋‹ค ๋” ๋™์ ์ธ ์น˜๋ฃŒ ๋ฐ˜์‘์„ ํฌ์ฐฉํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•œ๋‹ค.

๊ทธ๋Ÿฌ๋‚˜ radiomics์™€ ์ข…์–‘ ์ƒ๋ฌผํ•™(tumor biology) ๊ฐ„์˜ ์—ฐ๊ด€์„ฑ์„ ๋‹ค๋ฃจ๋Š” ์—ฐ๊ตฌ๋“ค์˜ ๋ฐฉ๋ฒ•๋ก ์  ์งˆ์€ ์—ฌ์ „ํžˆ ๋‚ฎ๋‹ค. Radiomics Quality Score(RQS)๋ฅผ ์ ์šฉํ•œ ๊ฒ€ํ† ์— ๋”ฐ๋ฅด๋ฉด, ๋Œ€๋ถ€๋ถ„์˜ ์—ฐ๊ตฌ๊ฐ€ 50% ๋ฏธ๋งŒ์˜ ์ ์ˆ˜๋ฅผ ๋ฐ›์•˜์œผ๋ฉฐ, ์ผ๋ถ€๋Š” 0์ ์„ ๊ธฐ๋กํ–ˆ๋‹ค [19]. ์ด๋Š” ์˜์ƒ ํŠน์ง• ์ถ”์ถœ์˜ ํ‘œ์ค€ํ™” ๋ถ€์กฑ๊ณผ ํ†ต๊ณ„์  ๊ฒ€์ฆ์˜ ๋ถ€์žฌ๋ฅผ ์˜๋ฏธํ•œ๋‹ค. ํŠนํžˆ PET/CT ์œตํ•ฉ ๋ฐ์ดํ„ฐ๋ฅผ ์ด์šฉํ•œ head and neck cancer(HNC) ์—ฐ๊ตฌ์—์„œ๋Š”, ์œตํ•ฉ ๋ชจ๋ธ์ด PET-only ๋ชจ๋ธ๋ณด๋‹ค concordance index์—์„œ 0.042 ๋†’๊ฒŒ ๋‚˜ํƒ€๋‚ฌ์œผ๋‚˜(p=0.0115), CT-only ๋ชจ๋ธ ๋Œ€๋น„ ์šฐ์œ„๋Š” ํ†ต๊ณ„์  ์œ ์˜์„ฑ ๊ธฐ์ค€(Bonferroni threshold 0.025)์„ ์ถฉ์กฑํ•˜์ง€ ๋ชปํ–ˆ๋‹ค [27]. ์ด๋Š” ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ๊ฐ„ ์ •๋ณด์˜ ์ค‘๋ณต์„ฑ๊ณผ ํŽธํ–ฅ(bias) ์œ„ํ—˜์„ ์‹œ์‚ฌํ•œ๋‹ค.

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” 'Predicting response to immunotherapy in advanced non-small-cell lung cancer using tumor mutational burden radiomic biomarker' ๋ฐ 'Deciphering the tumor microenvironment through radiomics in non-small cell lung cancer'์™€ ๊ฐ™์€ ๋…ผ๋ฌธ์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด, ๋ฉด์—ญ ์น˜๋ฃŒ ๋ฐ˜์‘ ์˜ˆ์ธก์— ๋Œ€ํ•œ radiomics์˜ ์ ์šฉ ์‚ฌ๋ก€๊ฐ€ ์ž˜ ์ •๋ฆฌ๋˜์–ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ CD8+ TILs์™€ ๊ฐ™์€ ํŠน์ • ๋ฉด์—ญ ์„ธํฌ ํ•˜์œ„ ์ง‘ํ•ฉ๊ณผ์˜ ์ •๋Ÿ‰์  ์ƒ๊ด€๊ด€๊ณ„ ๋ถ„์„, ๋˜๋Š” ๋‹ค์–‘ํ•œ ์•”์ข… ๊ฐ„์˜ ๋น„๊ต ์—ฐ๊ตฌ๋Š” ์—ฌ์ „ํžˆ ๋ถ€์กฑํ•˜๋‹ค. ํŠนํžˆ, acquired resistance(ํš๋“ ๋‚ด์„ฑ) ํŒจํ„ด์„ decodํ•˜๋Š” longitudinal profiling์— ๋Œ€ํ•œ ์‹ฌ์ธต์ ์ธ ๋ฉ”ํƒ€๋ถ„์„์€ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— 'Longitudinal Blood Immune-Inflammatory and Radiomic Profiling...' ๋…ผ๋ฌธ์ด ์žˆ์œผ๋‚˜, ๋‹ค๋ฅธ ์•”์ข…์œผ๋กœ์˜ ํ™•์žฅ ๊ฐ€๋Šฅ์„ฑ์— ๋Œ€ํ•œ ๋…ผ์˜๊ฐ€ ๋” ํ•„์š”ํ•˜๋‹ค.

์„ค๋ช… ๊ฐ€๋Šฅํ•œ ์ธ๊ณต์ง€๋Šฅ(XAI) ๋ฐ ์ž„์ƒ ์˜์‚ฌ๊ฒฐ์ • ์ง€์› ์‹œ์Šคํ…œ(CDSS)

์˜๋ฃŒ AI ๋ชจ๋ธ, ํŠนํžˆ ๋”ฅ ๋‰ด๋Ÿด ๋„คํŠธ์›Œํฌ๋Š” ๋†’์€ ์„ฑ๋Šฅ์„ ๋ณด์ด์ง€๋งŒ '๋ธ”๋ž™๋ฐ•์Šค(black-box)' ํŠน์„ฑ์œผ๋กœ ์ธํ•ด ์ž„์ƒ ํ˜„์žฅ์—์„œ์˜ ์‹ ๋ขฐ์„ฑ๊ณผ ์ฑ„ํƒ๋ฅ ์ด ๋‚ฎ๋‹ค. ์„ค๋ช… ๊ฐ€๋Šฅํ•œ ์ธ๊ณต์ง€๋Šฅ(Explainable Artificial Intelligence, XAI)์€ ์ด๋Ÿฌํ•œ ๋ถˆํˆฌ๋ช…์„ฑ์„ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด ๋“ฑ์žฅํ–ˆ์œผ๋ฉฐ, LIME(Local Interpretable Model-Agnostic Explanations)๊ณผ SHAP(SHapley Additive exPlanations) ๊ฐ™์€ ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ํ†ตํ•ด ๋ชจ๋ธ์˜ ์˜ˆ์ธก ๊ทผ๊ฑฐ๋ฅผ ๋ช…์‹œ์ ์œผ๋กœ ์ œ์‹œํ•œ๋‹ค [13]. XAI๋Š” ์ž„์ƒ ์˜์‚ฌ๊ฒฐ์ • ์ง€์› ์‹œ์Šคํ…œ(CDSS)์—์„œ ํŠนํžˆ ์ค‘์š”ํ•˜๋ฉฐ, ์˜์‚ฌ๊ฐ€ AI์˜ ์ถ”์ฒœ์„ ์ดํ•ดํ•˜๊ณ  ๋งฅ๋ฝ์— ๋งž๊ฒŒ ํ•ด์„ํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•จ์œผ๋กœ์จ ๊ณผ์‹  ๋˜๋Š” ๊ณผ์†Œ์‹ ๋ขฐ ๋ฌธ์ œ๋ฅผ ๋ฐฉ์ง€ํ•œ๋‹ค [8]. ๋˜ํ•œ, ์—ญ์‚ฌ์  ๋ฐ์ดํ„ฐ๋กœ ํ›ˆ๋ จ๋œ AI๊ฐ€ ๊ธฐ์กด ํŽธํ–ฅ์„ ๊ฐ•ํ™”ํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ, XAI๋ฅผ ํ†ตํ•ด ์ด๋Ÿฌํ•œ ํŽธํ–ฅ์„ ์‹๋ณ„ํ•˜๊ณ  ์œค๋ฆฌ์ ยท๊ณต์ •ํ•œ ์˜์‚ฌ๊ฒฐ์ •์„ ๋ณด์žฅํ•˜๋Š” ๊ฒƒ์ด ํ•„์ˆ˜์ ์ด๋‹ค [8].

๊ทธ๋Ÿฌ๋‚˜ ํ˜„์žฌ ์˜๋ฃŒ ์˜์ƒ ๋ถ„์„ ๋ถ„์•ผ์˜ ํˆฌ๋ช…์„ฑ ์—ฐ๊ตฌ๋Š” ์ฃผ๋กœ ๊ณ„์‚ฐ์  ์‹คํ–‰ ๊ฐ€๋Šฅ์„ฑ(computational feasibility)์— ์ง‘์ค‘๋˜์–ด ์žˆ์œผ๋ฉฐ, ์ตœ์ข… ์‚ฌ์šฉ์ž(์ž„์ƒ ์ „๋ฌธ๊ฐ€)์˜ ์š”๊ตฌ์‚ฌํ•ญ์„ ์ถฉ๋ถ„ํžˆ ๋ฐ˜์˜ํ•˜์ง€ ๋ชปํ•˜๊ณ  ์žˆ๋‹ค [16]. ์ฒด๊ณ„์  ๊ฒ€ํ†  ๊ฒฐ๊ณผ, ํ˜•์„ฑ์  ์‚ฌ์šฉ์ž ์—ฐ๊ตฌ(formative user research)๋ฅผ ํ†ตํ•ด ์‚ฌ์šฉ์ž ํ•„์š”์™€ ๋„๋ฉ”์ธ ์š”๊ตฌ์‚ฌํ•ญ์„ ์ดํ•ดํ•˜๋Š” ๋‹จ๊ณ„๊ฐ€ ๊ฑฐ์˜ ์ˆ˜ํ–‰๋˜์ง€ ์•Š์•˜์œผ๋ฉฐ, ํˆฌ๋ช…์„ฑ ์ฃผ์žฅ์ด ๊ฒฝํ—˜์  ์‚ฌ์šฉ์ž ํ‰๊ฐ€๋ฅผ ํ†ตํ•ด ๊ฒ€์ฆ๋œ ์‚ฌ๋ก€๋„ ๋“œ๋ฌผ๋‹ค [16]. ์ด๋Š” ํ˜„์žฌ์˜ ํˆฌ๋ช… ML ๋ชจ๋ธ์ด ์‚ฌ์šฉ์ž์—๊ฒŒ incomprehensibleํ•˜์—ฌ ์ž„์ƒ์ ์œผ๋กœ ๊ด€๋ จ์„ฑ์ด ๋–จ์–ด์งˆ ์œ„ํ—˜์ด ์žˆ์Œ์„ ์‹œ์‚ฌํ•œ๋‹ค. ์ด๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด INTRPRT ๊ฐ€์ด๋“œ๋ผ์ธ๊ณผ ๊ฐ™์ด ์ธ๊ฐ„ ์ค‘์‹ฌ ์„ค๊ณ„(human-centered design) ์›์น™์„ ๋„์ž…ํ•ด์•ผ ํ•œ๋‹ค [16].

XAI์˜ ์ ์šฉ์€ ๋‹ค์ค‘ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ ๋ฐ ๋‹ค๊ธฐ๊ด€ ๋ฐ์ดํ„ฐ ์œตํ•ฉ๊ณผ ๊ฒฐํ•ฉ๋  ๋•Œ ๋”์šฑ ํšจ๊ณผ์ ์ด๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, multi-modal and multi-centre data fusion์„ ํ™œ์šฉํ•œ XAI ์†”๋ฃจ์…˜์€ ์‹ค์ œ ์ž„์ƒ ์‹œ๋‚˜๋ฆฌ์˜ค์—์„œ ๊ฒ€์ฆ๋˜์—ˆ์œผ๋ฉฐ, ์ •๋Ÿ‰์ ยท์ •์„ฑ์  ๋ถ„์„์„ ํ†ตํ•ด ๊ทธ ์œ ํšจ์„ฑ์„ ์ž…์ฆํ–ˆ๋‹ค [7]. ์ด๋Š” AI ๋ชจ๋ธ์˜ ๋‚ด๋ถ€ ๋ฉ”์ปค๋‹ˆ์ฆ˜์„ ํ•ด์„ํ•˜๋Š” ๊ฒƒ์„ ๋„˜์–ด, ๋‹ค์–‘ํ•œ ๋ฐ์ดํ„ฐ ์†Œ์Šค์˜ ์ƒํ˜ธ์ž‘์šฉ์„ ์„ค๋ช…ํ•จ์œผ๋กœ์จ ์˜์‚ฌ์˜ ์‹ ๋ขฐ๋ฅผ ๋†’์ด๋Š” ๋ฐ ๊ธฐ์—ฌํ•œ๋‹ค. ๋˜ํ•œ, CDSS์—์„œ์˜ XAI ์ ์šฉ์€ ํ‘œ๋ณธ ๋ฐ์ดํ„ฐ ์ฒ˜๋ฆฌ(tabular data)์—์„œ ๊ฐ€์žฅ ํ”ํ•˜๋ฉฐ, ํ…์ŠคํŠธ ๋ถ„์„์—์„œ๋Š” ์ƒ๋Œ€์ ์œผ๋กœ ๋“œ๋ฌผ๋‹ค [8]. ์ง€์—ญ์  ์„ค๋ช…(local explanations)์— ๋Œ€ํ•œ ๊ฐœ๋ฐœ์ž์˜ ๊ด€์‹ฌ์ด ๋†’์œผ๋ฉฐ, ์‚ฌํ›„ ์„ค๋ช…(post-hoc)๊ณผ ์‚ฌ์ „ ์„ค๋ช…(ante-hoc), ๋ชจ๋ธ ํŠน์ด์ (model-specific)๊ณผ ๋ชจ๋ธ ๋ฌด๊ด€(model-agnostic) ๊ธฐ๋ฒ• ๊ฐ„์˜ ๊ท ํ˜•์ด ์ด๋ฃจ์–ด์ง€๊ณ  ์žˆ๋‹ค [8].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” XAI์— ๋Œ€ํ•œ ์ง์ ‘์ ์ธ ๋…ผ๋ฌธ์€ ๋ช…์‹œ์ ์œผ๋กœ ๋‚˜์—ด๋˜์ง€ ์•Š์•˜์œผ๋‚˜, 'Radiomics in Oncology: A Practical Guide'์™€ ๊ฐ™์€ ์‹ค์šฉ์  ๊ฐ€์ด๋“œ๊ฐ€ ํฌํ•จ๋˜์–ด ์žˆ์–ด ๋ฐฉ๋ฒ•๋ก ์  ๊ธฐ์ดˆ๋Š” ๋งˆ๋ จ๋˜์–ด ์žˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ธ๊ฐ„ ์ค‘์‹ฌ ์„ค๊ณ„ ์›์น™์„ ์ ์šฉํ•œ XAI ๋ชจ๋ธ์˜ ์ž„์ƒ ๊ฒ€์ฆ ์—ฐ๊ตฌ, ํŠนํžˆ ๋‹ค๊ธฐ๊ด€ ํ™˜๊ฒฝ์—์„œ์˜ ์‚ฌ์šฉ์ž ์ˆ˜์šฉ๋„ ๋ฐ ํŽธํ–ฅ ๊ฐ์‹œ(bias monitoring)์— ๋Œ€ํ•œ ๊ตฌ์ฒด์ ์ธ ์‚ฌ๋ก€ ์—ฐ๊ตฌ๋Š” ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๋ถ€์กฑํ•˜๋‹ค. AI์˜ ์œค๋ฆฌ์  ๊ตฌํ˜„๊ณผ ์ฑ…์ž„(accountability)์— ๋Œ€ํ•œ ๋…ผ์˜๊ฐ€ ๋” ๊นŠ์ด ์žˆ๊ฒŒ ํ†ตํ•ฉ๋˜์–ด์•ผ ํ•œ๋‹ค.

์‹ ๊ฒฝํ‡ดํ–‰์„ฑ ์งˆํ™˜ ๋ฐ ๋Œ€์‚ฌ์„ฑ ์งˆํ™˜์—์„œ์˜ Radiomics ์ ์šฉ

Radiomics์™€ AI ๊ธฐ์ˆ ์€ ์ข…์–‘ํ•™ ์™ธ์—๋„ ์‹ ๊ฒฝํ‡ดํ–‰์„ฑ ์งˆํ™˜(Neurodegenerative diseases) ๋ฐ ๋Œ€์‚ฌ์„ฑ ์งˆํ™˜(Metabolic diseases)์˜ ์ง„๋‹จ๊ณผ ๋ชจ๋‹ˆํ„ฐ๋ง์—์„œ ์ค‘์š”ํ•œ ์—ญํ• ์„ ํ•˜๊ณ  ์žˆ๋‹ค. ๋‡Œ ์ฒ ๋ถ„(Iron) ๋ถ„ํฌ๋ฅผ ์ •๋Ÿ‰์ ์œผ๋กœ ํ‰๊ฐ€ํ•˜๋Š” ์ž๊ธฐ๊ฐ์‘ ๋งคํ•‘(Quantitative Susceptibility Mapping, QSM) ๊ธฐ์ˆ ์€ ์•Œ์ธ ํ•˜์ด๋จธ๋ณ‘(Alzheimer's disease), ํŒŒํ‚จ์Šจ๋ณ‘(Parkinson's disease), ๊ทผ์œ„์ถ•์„ฑ ์ธก์‚ญ ๊ฒฝํ™”์ฆ(ALS) ๋“ฑ ๋‹ค์–‘ํ•œ ์‹ ๊ฒฝํ‡ดํ–‰์„ฑ ์งˆํ™˜์˜ ๋ณ‘๋ฆฌํ•™์  ๋ณ€ํ™”์™€ ๊ด€๋ จ์ด ์žˆ๋‹ค [20]. QSM ์—ฐ๊ตฌ๋“ค์€ ๊ฐ ์งˆํ™˜์˜ ํŠน์ง•์ ์ธ ๋‡Œ ์˜์—ญ(์˜ˆ: ์•Œ์ธ ํ•˜์ด๋จธ์˜ ํŽธ๋„์ฒด ๋ฐ ๊ผฌ๋ฆฌํ•ต, ํŒŒํ‚จ์Šจ์˜ ํ‘์งˆ)์—์„œ ์ž๊ฐ์‘(magnetic susceptibility) ์ฆ๊ฐ€, ์ฆ‰ ์ฒ ๋ถ„ ์ถ•์ ์„ ๋ณด์—ฌ์ฃผ์—ˆ์œผ๋ฉฐ, ์ด๋Š” ์งˆ๋ณ‘ ์ง€์† ๊ธฐ๊ฐ„ ๋ฐ ์ž„์ƒ์  ์ค‘์ฆ๋„์™€ ์ƒ๊ด€๊ด€๊ณ„๊ฐ€ ์žˆ๋‹ค [20].

์•Œ์ธ ํ•˜์ด๋จธ๋ณ‘์˜ ์กฐ๊ธฐ ์ง„๋‹จ์„ ์œ„ํ•ด ๋”ฅ๋Ÿฌ๋‹(DL) ์•Œ๊ณ ๋ฆฌ์ฆ˜์ด ๋„๋ฆฌ ํ™œ์šฉ๋˜๊ณ  ์žˆ๋‹ค. DL์€ ๋‡Œ ์†์ƒ์ด ๋น„๊ฐ€์—ญ์ ์ด ๋˜๊ธฐ ์ „์— ์œ„ํ—˜ ์ธ์‹์„ ๋†’์ด๊ณ  ์˜ˆ๋ฐฉ ์กฐ์น˜๋ฅผ ์ทจํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•จ์œผ๋กœ์จ ์น˜๋ฃŒ ๊ฒฐ๊ณผ๋ฅผ ๊ฐœ์„ ํ•œ๋‹ค [21]. ๊ทธ๋Ÿฌ๋‚˜ ๋Œ€๋ถ€๋ถ„์˜ ๊ธฐ๊ณ„ ๊ฐ์ง€ ๋ฐฉ๋ฒ•์€ ์„ ์ฒœ์  ๊ด€์ฐฐ(congenital observations)์— ์ œํ•œ๋˜์–ด ์žˆ์œผ๋ฉฐ, ์˜ˆ์ธก(prediction)๋ณด๋‹ค๋Š” ์ง„๋‹จ(diagnosis)์— ๋” ์ดˆ์ ์ด ๋งž์ถฐ์ ธ ์žˆ๋‹ค [21]. ์ด๋Š” ์งˆ๋ณ‘ ๋ฐœํ˜„ ์ „ ๋‹จ๊ณ„์—์„œ์˜ ์˜ˆ์ธก ๋ชจ๋ธ ๊ฐœ๋ฐœ์ด ์—ฌ์ „ํžˆ ๊ณผ์ œ๋กœ ๋‚จ์•„ ์žˆ์Œ์„ ์˜๋ฏธํ•œ๋‹ค.

๋Œ€์‚ฌ์„ฑ ์งˆํ™˜ ์ค‘ MASLD(Metabolic dysfunction-associated steatotic liver disease, formerly NAFLD)๋Š” ์ „ ์„ธ๊ณ„ ์„ฑ์ธ 28%๋ฅผ ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š” ์ฃผ์š” ๋งŒ์„ฑ ๊ฐ„ ์งˆํ™˜์ด๋‹ค [25, 32]. ๊ฐ„ ์ƒ๊ฒ€(liver biopsy)์€ ๊ธˆํ‘œ์ค€(gold standard)์ด์ง€๋งŒ ์นจ์Šต์ ์ด๊ณ  ๋น„์šฉ์ด ๋งŽ์ด ๋“ค๋ฉฐ ์ƒ˜ํ”Œ ์˜ค๋ฅ˜ ๊ฐ€๋Šฅ์„ฑ์ด ์žˆ๋‹ค. ์ด์— AI ๊ธฐ๋ฐ˜ ์˜์ƒ ๋ชจ๋ธ(์ดˆ์ŒํŒŒ, CT, MRI, ํƒ„์„ฑ์˜์ƒ)์€ ์ง€๋ฐฉ๋ณ€ ์ง„๋‹จ(AUROC 0.85-0.99) ๋ฐ ์„ฌ์œ ํ™” ๋‹จ๊ณ„ ๊ตฌ๋ถ„(AUROC 0.82-0.97)์—์„œ ๋›ฐ์–ด๋‚œ ์ •ํ™•์„ฑ์„ ๋ณด์˜€๋‹ค [25]. ํŠนํžˆ, ๊ธฐ๊ณ„ํ•™์Šต ์•Œ๊ณ ๋ฆฌ์ฆ˜์ด ์‹คํ—˜์‹ค ์ง€ํ‘œ, ๋Œ€์‚ฌ์ฒดํ•™, ๋‹จ๋ฐฑ์งˆ์ฒดํ•™ ๋ฐ ์žฅ๋‚ด ๋ฏธ์ƒ๋ฌผ๊ตฐ ๋ฐ์ดํ„ฐ๋ฅผ ํ†ตํ•ฉํ•  ๋•Œ, ๊ธฐ์กด ๋น„์นจ์Šต์  ์•Œ๊ณ ๋ฆฌ์ฆ˜(FIB-4, NAFLD Fibrosis Score)๋ณด๋‹ค ์šฐ์ˆ˜ํ•œ ์„ฑ๋Šฅ์„ ๋ฐœํœ˜ํ–ˆ๋‹ค [25]. ๋””์ง€ํ„ธ ๋ณ‘๋ฆฌ(digital pathology)๋ฅผ ํ†ตํ•œ ๊ฐ„ ์ƒ๊ฒ€ ์ด๋ฏธ์ง€์˜ ๋”ฅ๋Ÿฌ๋‹ ๋ถ„์„์€ ๊ฐ„๋ณ‘๋ฆฌํ•™์ž(hepatopathologist) ์ „๋ฌธ๊ฐ€์™€ ๋™๋“ฑํ•œ ์ง„๋‹จ ์ •ํ™•๋„๋ฅผ ๋‹ฌ์„ฑํ•˜๊ธฐ๋„ ํ–ˆ๋‹ค [25].

์‚ฌ์šฉ์ž ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์—๋Š” MASLD๋‚˜ ์‹ ๊ฒฝํ‡ดํ–‰์„ฑ ์งˆํ™˜์— ํŠนํ™”๋œ radiomics ๋…ผ๋ฌธ์ด ๋ช…์‹œ์ ์œผ๋กœ ํฌํ•จ๋˜์–ด ์žˆ์ง€ ์•Š๋‹ค. ์ด๋Š” ํ•ด๋‹น ๋ถ„์•ผ์—์„œ์˜ radiomics ์ ์šฉ์ด ์ƒ๋Œ€์ ์œผ๋กœ ์ƒˆ๋กœ์šด ์˜์—ญ์ž„์„ ์‹œ์‚ฌํ•œ๋‹ค. ํŠนํžˆ, QSM์„ ํ™œ์šฉํ•œ ๋‡Œ ์ฒ ๋ถ„ ํ”„๋กœํŒŒ์ผ๋ง๊ณผ MASLD์˜ ๋‹ค์ค‘ ์˜ค๋ฏน์Šค(multi-omics) ํ†ตํ•ฉ ๋ถ„์„์€ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์˜ ์ฃผ์š” gap์œผ๋กœ ์ง€์ ๋œ๋‹ค. ํ–ฅํ›„ ์‹ ๊ฒฝํ•™์  ๋ฐ ๋Œ€์‚ฌ์„ฑ ์งˆํ™˜์—์„œ์˜ ๋น„์นจ์Šต์  ๋ฐ”์ด์˜ค๋งˆ์ปค ๊ฐœ๋ฐœ์„ ์œ„ํ•œ radiomics ์—ฐ๊ตฌ๊ฐ€ ํ™•๋Œ€๋˜์–ด์•ผ ํ•  ๊ฒƒ์ด๋‹ค.

์ฐธ๊ณ  ๋ฆฌ๋ทฐ

#๋ฆฌ๋ทฐ์ €๋„ ยท ์—ฐ๋„์ธ์šฉ๋งํฌ
1Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisalBMJ ยท 20203302PubMed ยท DOI
2Imaging biomarker roadmap for cancer studiesNature Reviews Clinical Oncology ยท 20161104PubMed ยท DOI
3Challenges to curing primary brain tumoursNature Reviews Clinical Oncology ยท 2019996PubMed ยท DOI
4Diagnostic accuracy of deep learning in medical imaging: a systematic review and meta-analysisnpj Digital Medicine ยท 2021941PubMed ยท DOI
5Repeatability and Reproducibility of Radiomic Features: A Systematic ReviewInternational Journal of Radiation Oncology*Biology*Physics ยท 2018913PubMed ยท DOI
6Fusion of medical imaging and electronic health records using deep learning: a systematic review and implementation guidelinesnpj Digital Medicine ยท 2020904PubMed ยท DOI
7Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: A mini-review, two showcases and beyondInformation Fusion ยท 2021724PubMed ยท DOI
8Current Challenges and Future Opportunities for XAI in Machine Learning-Based Clinical Decision Support Systems: A Systematic ReviewApplied Sciences ยท 2021617DOI
9A systematic literature review of artificial intelligence in the healthcare sector: Benefits, challenges, methodologies, and functionalitiesJournal of Innovation & Knowledge ยท 2023552DOI
10Self-supervised learning for medical image classification: a systematic review and implementation guidelinesnpj Digital Medicine ยท 2023450PubMed ยท DOI
11The association between CD8+ tumor-infiltrating lymphocytes and the clinical outcome of cancer immunotherapy: A systematic review and meta-analysisEClinicalMedicine ยท 2021374PubMed ยท DOI
12False Discovery Rates in PET and CT Studies with Texture Features: A Systematic ReviewPLoS ONE ยท 2015345PubMed ยท DOI
13The enlightening role of explainable artificial intelligence in medical & healthcare domains: A systematic literature reviewComputers in Biology and Medicine ยท 2023333PubMed ยท DOI
14Comparison of Vision Transformers and Convolutional Neural Networks in Medical Image Analysis: A Systematic ReviewJournal of Medical Systems ยท 2024318PubMed ยท DOI
15A Comprehensive Systematic Review of YOLO for Medical Object Detection (2018 to 2023)IEEE Access ยท 2024292DOI
16Explainable medical imaging AI needs human-centered design: guidelines and evidence from a systematic reviewnpj Digital Medicine ยท 2022284PubMed ยท DOI
17Towards clinical application of image mining: a systematic review on artificial intelligence and radiomicsEuropean Journal of Nuclear Medicine and Molecular Imaging ยท 2019262PubMed ยท DOI
18Radiology artificial intelligence: a systematic review and evaluation of methods (RAISE)European Radiology ยท 2022247PubMed ยท DOI
19Tracking tumor biology with radiomics: A systematic review utilizing a radiomics quality scoreRadiotherapy and Oncology ยท 2018247PubMed ยท DOI
20Systematic Review: Quantitative Susceptibility Mapping (QSM) of Brain Iron Profile in Neurodegenerative DiseasesFrontiers in Neuroscience ยท 2021235PubMed ยท DOI
21Alzheimerโ€™s Diseases Detection by Using Deep Learning Algorithms: A Mini-ReviewIEEE Access ยท 2020224DOI
22Deep Learning Based Methods for Breast Cancer Diagnosis: A Systematic Review and Future DirectionDiagnostics ยท 2023223PubMed ยท DOI
23Unveiling the Influence of AI Predictive Analytics on Patient Outcomes: A Comprehensive Narrative ReviewCureus ยท 2024200PubMed ยท DOI
24Where Is the Artificial Intelligence Applied in Dentistry? Systematic Review and Literature AnalysisHealthcare ยท 2022191PubMed ยท DOI
25AI In Diagnosing Metabolic Dysfunction- Associated Steatotic Liver Disease (Masld): A Narrative ReviewZenodo (CERN European Organization for Nuclear Research) ยท 20260DOI
26AI In Diagnosing Metabolic Dysfunction- Associated Steatotic Liver Disease (Masld): A Narrative ReviewZenodo (CERN European Organization for Nuclear Research) ยท 20260DOI
27Radiomics Using Fused PET and CT Data for Prognostic and Diagnostic Modeling in Head and Neck Cancer: A Systematic Review and Meta-AnalysisDiagnostics ยท 20260DOI
28Deep learning prediction of MSI/dMMR, EBV, HER2, PD-L1, and CLDN18.2 status directly from hematoxylin and eosin-stained slides in gastric and gastroesophageal junction adenocarcinoma: a systematic reviewFrontiers in Oncology ยท 20260DOI
29Multimodal EEGโ€“MRI signal processing approaches for stroke and rehabilitation: a narrative reviewScientific Reviews ยท 20260DOI
30Artificial intelligence in contrast-enhanced mammography: Current evidence, clinical integration, and future perspectives โ€“ a narrative reviewEurasian Journal of Medicine and Oncology ยท 20260DOI
31Artificial intelligence for imaging-based detection of gallbladder cancer: A systematic review and diagnostic test accuracy meta-analysisWorld Journal of Gastroenterology ยท 20260DOI
32Artificial intelligence in metabolic dysfunction-associated steatotic liver disease: A short narrative review of diagnostic and therapeutic implicationsArtificial Intelligence in Gastroenterology ยท 20260DOI
์ถ”์ฒœ ๋…ผ๋ฌธ โ€” ๋‹ค์Œ์— ์ฐพ์•„๋ณผ (๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ์—†๋Š” ๊ฒƒ)

์•„๋ž˜ ๋…ผ๋ฌธ ์ค‘ ํ•„์š”ํ•œ ๊ฒƒ์„ Zotero์— ๋‹ด์•„ ๋‘๋ฉด ๋‹ค์Œ ์‹คํ–‰๋ถ€ํ„ฐ ์ด ๋ชฉ๋ก์—์„œ ๋น ์ง‘๋‹ˆ๋‹ค.

#๋…ผ๋ฌธ์ €๋„ ยท ์—ฐ๋„๋งํฌ
1CT-based whole lung radiomics nomogram for identification of PRISm from non-COPD subjects.Respiratory research ยท 2024-Sep-03PubMed ยท DOI
2Non-linear Logistic Regression applied to Radiomics.Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference ยท 2024-JulPubMed ยท DOI
3CT-based delta-radiomics for predicting pathological response to neoadjuvant immunochemotherapy in esophageal squamous cell carcinoma: a multicenter study.BMC medical imaging ยท 2024-Dec-03PubMed ยท DOI
4Computed tomography radiomics to predict microsatellite instability status and immunotherapy response in gastric cancer.Insights into imaging ยท 2025-Aug-14PubMed ยท DOI
5A radiomics model based on transrectal ultrasound for predicting prostate cancer.Medical ultrasonography ยท 2024-Jun-21PubMed ยท DOI
6Muscle CT Radiomics is Feasible in the Identification of Gout.Current medical imaging ยท 2024PubMed ยท DOI
7CT-based radiomics models decode fibrosis content and molecular differences in pancreatic ductal adenocarcinoma: a multi-institutional study.Insights into imaging ยท 2025-Sep-12PubMed ยท DOI
8Postoperative Pancreatic Fistula After Pancreatoduodenectomy: Can Radiomics Improve Clinical Risk Scores?Annals of surgery ยท 2025-Nov-28PubMed ยท DOI
9Delta-radiomics features for predicting the major pathological response to neoadjuvant chemoimmunotherapy in non-small cell lung cancer.European radiology ยท 2024-AprPubMed ยท DOI
10Radiomics approach to distinguish between benign and malignant soft tissue tumors on magnetic resonance imaging.European journal of radiology open ยท 2024-JunPubMed ยท DOI