Real-World and Clinical Trial Validation of a Deep Learning Radiomic Biomarker for PD-(L)1 Immune Checkpoint...

Area of Science:

• Radiomics and Artificial Intelligence in Oncology
• Medical Imaging and Diagnostics
• Immunotherapy Biomarker Development

Background:

• Predicting response to immune checkpoint inhibitors (ICIs) in advanced non-small cell lung cancer (NSCLC) is crucial for optimizing patient treatment.
• Traditional imaging biomarkers often fall short in accurately identifying patients who will benefit from ICI therapy.
• Real-world data (RWD) and clinical trial data offer valuable insights for developing robust predictive biomarkers.

Purpose of the Study:

• To develop and validate a novel deep learning radiomic biomarker for estimating ICI therapy response in advanced NSCLC.
• To utilize real-world data (RWD) and clinical trial data for comprehensive biomarker evaluation.
• To assess the performance of the developed biomarker against established imaging metrics.

Main Methods:

• A retrospective analysis of 1,829 advanced NSCLC patients treated with PD-(L)1 ICIs from academic and community institutions was conducted.
• A deep learning radiomic pipeline generated a computed tomography (CT) response score (CTRS) from pretreatment CT/PET-CT scans.
• The CTRS, and an enhanced version (eCTRS) incorporating clinical factors, were validated on independent RWD and a prospective clinical trial dataset.

Main Results:

• In RWD, the CTRS identified patients with a high probability of response, showing significant hazard ratios for progression-free survival (PFS) and overall survival (OS).
• The CTRS demonstrated predictive value in a prospective clinical trial dataset, with a significant OS hazard ratio.
• Both CTRS and eCTRS outperformed traditional lesion size biomarkers in predicting PFS and OS in RWD and OS in the clinical trial.

Conclusions:

• A deep learning radiomic biomarker (CTRS) was successfully developed and validated for predicting ICI benefit in advanced NSCLC.
• The CTRS utilizes routine pretreatment CT/PET-CT scans, offering a non-invasive method to identify likely responders to ICI therapy.
• This novel biomarker has the potential to improve treatment selection and outcomes for NSCLC patients receiving immunotherapy.