Multimodal deep learning for survival prediction and biomarker discovery in non-small cell lung cancer.
1/5 보강
[BACKGROUND] Deep learning (DL) is revolutionizing high-dimensional analysis for clinical applications, with multimodal approaches showing promise in integrating diverse datasets and improving predict
APA
Wang Y, Xia P, et al. (2026). Multimodal deep learning for survival prediction and biomarker discovery in non-small cell lung cancer.. BioData mining, 19(1). https://doi.org/10.1186/s13040-026-00522-8
MLA
Wang Y, et al.. "Multimodal deep learning for survival prediction and biomarker discovery in non-small cell lung cancer.." BioData mining, vol. 19, no. 1, 2026.
PMID
41703583 ↗
Abstract 한글 요약
[BACKGROUND] Deep learning (DL) is revolutionizing high-dimensional analysis for clinical applications, with multimodal approaches showing promise in integrating diverse datasets and improving predictive accuracy. However, its use in prognostic assessment, particularly for non-small-cell lung cancer (NSCLC), remains limited.
[METHODS] Whole-slide imaging (WSI), next-generation sequencing (NGS), and clinical features from 2,366 NSCLC patients sourced from various databases were analyzed. A Sequential-Adapted Attention (SeAttn) model was developed to process WSI, while multiple DL models analyzed genetic and clinical data. Multimodal features were combined using a COX-based DL approach to predict long-term prognosis. Validation ( = 100) and test ( = 101) cohorts were used to confirm model performance.
[RESULTS] The SeAttn model achieved an area under the curve (AUC) of 0.98 for NSCLC histologic subtype prediction in the test cohort ( < 0.0001). Individually, WSI, clinical, and genetic features predicted prognosis with concordance indices (C-indices) of 0.55–0.67 using the test cohort. Combined multimodal features improved the test cohort C-index to 0.71 ( = 0.002). The model accurately predicted prognosis up to 5 years post-diagnosis ( < 0.05 for all time-dependent AUCs) and stratified patients by overall survival ( = 0.012). SeAttn demonstrated attention shifts when comparing subtyping and prognosis models and revealed immuno-hot and immuno-cold stroma enrichment underlying survival differences. Novel prognostic markers, including mutation, microRNA cluster amplification, and - co-deletion, were identified and validated.
[CONCLUSIONS] This multimodal DL framework demonstrates robust prognostic capabilities for NSCLC, integrating clinical, genetic, and imaging data to yield accurate survival estimates and identify biomarkers. These findings highlight the scalability of multimodal DL for broader cancer applications.
[SUPPLEMENTARY INFORMATION] The online version contains supplementary material available at 10.1186/s13040-026-00522-8.
[METHODS] Whole-slide imaging (WSI), next-generation sequencing (NGS), and clinical features from 2,366 NSCLC patients sourced from various databases were analyzed. A Sequential-Adapted Attention (SeAttn) model was developed to process WSI, while multiple DL models analyzed genetic and clinical data. Multimodal features were combined using a COX-based DL approach to predict long-term prognosis. Validation ( = 100) and test ( = 101) cohorts were used to confirm model performance.
[RESULTS] The SeAttn model achieved an area under the curve (AUC) of 0.98 for NSCLC histologic subtype prediction in the test cohort ( < 0.0001). Individually, WSI, clinical, and genetic features predicted prognosis with concordance indices (C-indices) of 0.55–0.67 using the test cohort. Combined multimodal features improved the test cohort C-index to 0.71 ( = 0.002). The model accurately predicted prognosis up to 5 years post-diagnosis ( < 0.05 for all time-dependent AUCs) and stratified patients by overall survival ( = 0.012). SeAttn demonstrated attention shifts when comparing subtyping and prognosis models and revealed immuno-hot and immuno-cold stroma enrichment underlying survival differences. Novel prognostic markers, including mutation, microRNA cluster amplification, and - co-deletion, were identified and validated.
[CONCLUSIONS] This multimodal DL framework demonstrates robust prognostic capabilities for NSCLC, integrating clinical, genetic, and imaging data to yield accurate survival estimates and identify biomarkers. These findings highlight the scalability of multimodal DL for broader cancer applications.
[SUPPLEMENTARY INFORMATION] The online version contains supplementary material available at 10.1186/s13040-026-00522-8.
🏷️ 키워드 / MeSH 📖 같은 키워드 OA만
같은 제1저자의 인용 많은 논문 (5)
- "I wanna look like the person in that picture": Linking selfies on social media to cosmetic surgery consideration based on the tripartite influence model.
- ZmSKIP enhances drought tolerance by reducing stomatal aperture in maize.
- c.7374_7375insAlu is a French-Canadian founder pathogenic variant associated with predisposition to pancreatic and breast cancer.
- Enhancing Node-RADS for preoperative assessment of cervical lymph node metastases in papillary thyroid carcinoma: validation and modification.
- Aging modulation of the immune system and immunotherapy efficacy in cancer.
🏷️ 같은 키워드 · 무료전문 — 이 논문 MeSH/keyword 기반
- Hidden findings: How often does Holmium laser enucleation of the prostate (HoLEP) uncover prostate cancer?
- Outcomes of point-of-care manufactured CAR T cell therapy for B cell acute lymphoblastic leukemia and non-Hodgkin lymphoma in Vietnam.
- Clinicopathological characteristics and therapeutic outcomes in patients with non-small cell lung cancer harboring mutations.
- Cell-Based Immunization Combined with Single-Round Cell Panning Enables Discovery of PSMA-Targeting Nanobodies from Phage Display Libraries.
- Incidental pulmonary findings on CT in daily practice: the nodule and the interstitial lung abnormalities - what's old, what's new.
- An Algorithmic Approach for Quantitative Determination of Microsatellite Status in NGS-Based Cancer Diagnostics.