What does this research mean for the field?

A multimodal prognostic model integrating PET/CT radiomics, deep learning features, and clinical characteristics effectively predicts overall and progression-free survival in locally advanced non-small cell lung cancer, with radiomics features being the strongest predictors. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To develop and evaluate a multimodal prognostic model combining clinical characteristics, tumor burden, PET/CT radiomics, and deep learning features for locally advanced non-small cell lung cancer.

August 27, 2026BMC CancerOpen Access

PET/CT radiomics-based prognostic model with exploratory deep learning features in locally advanced non-small cell lung cancer

View Full Paper

Ask AI

Bookmark

Authors

JZJingjing Zhang DPDa PanMinistry of Education KCKun Chen

Discussion

Member takes

Implication

Retrospective cohort study reveals multimodal PET/CT radiomics and deep learning predict survival in locally advanced non-small cell lung cancer, indicating potential for personalized prognosis.

Key Points

To develop and evaluate a multimodal prognostic model combining clinical characteristics, tumor burden, PET/CT radiomics, and deep learning features for locally advanced non-small cell lung cancer.
Retrospective single-center study evaluating 221 patients with locally advanced non-small cell lung cancer split into a training cohort (n = 176) and an internal test cohort (n = 45) with pretreatment 18F-FDG PET/CT scans.
Extracted PET/CT radiomic features using elastic-net Cox regression and generated deep-learning risk scores via a dual-branch 3D CNN-Transformer architecture to construct overall survival (OS) and progression-free survival (PFS) models.
Model discrimination was evaluated using the concordance index (C-index) and time-dependent AUCs, with feature contributions interpreted using SHAP analysis.
The integrated multimodal model achieved overall survival C-indices of 0.731 in the training cohort and 0.743 in the test cohort, yielding 1-, 2-, and 3-year overall survival AUCs of 0.658, 0.820, and 0.874 in the test cohort.
For progression-free survival, the standalone deep-learning model demonstrated a higher test C-index than the integrated model, while SHAP analysis identified the radiomics risk score as the primary contributor to model predictions.

Cite This Study

Zhang et al. (2026) studied this question.

synapsesocial.com/papers/6a8fe9ad10c91c1e926217dd https://doi.org/10.1186/s12885-026-16816-3

View Full Paper

Ask AI

Bookmark

Implication

Retrospective cohort study reveals multimodal PET/CT radiomics and deep learning predict survival in locally advanced non-small cell lung cancer, indicating potential for personalized prognosis.

PET/CT radiomics-based prognostic model with exploratory deep learning features in locally advanced non-small cell lung cancer

Discussion

Member takes

Implication

Key Points

Cite This Study

PET/CT radiomics-based prognostic model with exploratory deep learning features in locally advanced non-small cell lung cancer

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider