Abstract
Purpose Immunotherapy has improved outcomes in non-small cell lung cancer (NSCLC), but only a subset of patients achieves durable survival benefit. Conventional biomarkers such as PD-L1 expressions have limited predictive power. This study aimed to develop an artificial intelligence (AI)–based radiomic framework integrating deep learning derived and traditional radiomic features with clinical variables to predict survival outcomes in advanced NSCLC treated with immune checkpoint inhibitor (ICI)–containing regimens.
Methods Ninety-four patients with stage IV NSCLC treated with ICI-containing regimens between 2013 and 2022 were analyzed. Baseline CT scans were segmented and processed using LIFEx software, and deep radiomic features were extracted using a UNET convolutional encoder–decoder. Models were trained using nested five-fold cross-validation. Performance was evaluated using area under curve (AUC) and calibration assessment.
Results The ensemble model combining clinical, traditional radiomic, and deep radiomic features achieved AUCs of 0.70 and 0.84 for 12– and 36-month overall survival (OS), and 0.65 and 0.73 for 12– and 36-month progression-free survival (PFS). For 12-month OS, the model integrating clinical variables with deep radiomic features showed the highest performance (AUC 0.78). Beyond deep features, traditional radiomic features reflecting tumor heterogeneity and morphology were the most influential predictors. Patients classified as high risk by the ensemble model had significantly shorter survival (OS HR 2.10, p = 0.005; PFS HR 1.74, p = 0.015).
Conclusion Integrating deep learning–derived radiomic features with clinical data improved survival prediction in advanced NSCLC treated with ICI-containing regimens. These findings support the potential of an AI-based radiomic risk stratification model as a noninvasive tool for individualized outcome prediction, warranting prospective validation.
Question Can combining deep learning-derived radiomic features with clinical variables improve prediction of survival outcomes in patients with non–small cell lung cancer (NSCLC) treated with immune checkpoint inhibitor-containing regimens?
Findings In a retrospective cohort of 94 patients with stage IV NSCLC, an ensemble model integrating clinical, traditional radiomic, and deep learning-derived radiomic features showed strong discrimination for 12– and 36-month overall survival with an area under curve of 0.70 and 0.84, respectively. Radiomic markers of tumor heterogeneity and morphology were among the most influential factors.
Meaning Deep learning-based radiomic-clinical modeling can enhance individualized survival prediction in NSCLC.
Competing Interest Statement
J. Kim reports personal income from GeneCker Co. outside of the study. No disclosures were reported by other authors.
Funding Statement
This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
Institutional Review Board of Northwestern University gave ethical approval for this work (Protocol: STU0020711).
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Yes
Data Availability
All clinical data collected during the study is available upon request to the corresponding author.
