Predictive Multimodal Signatures Associated With Response to Treatment and Prognosis of Patients With Stage IV Non-small Cell Lung Cancer

Deep Learning-Enabled Exploration of Predictive Signatures in a Multicenter Retrospective and Prospective Observational Study Allowing the Analysis of the Aggregation of Multimodal Clinical, Biological, Genomic and Radiomics Data Associated With the Response to Treatment and Prognosis of Patients With Stage IV Non-small Cell Lung Cancer

Predicting response to therapy and disease progression in stage IV NSCLC patients treated with pembrolizumab monotherapy, chemotherapy-pembrolizumab combination therapy or chemotherapy alone in the first-line setting.

Inclusion Criteria:

• Adult ≥18 years old
• Patient diagnosed with Stage IV NSCLC (de novo or earlier stage progression to stage IV)
• Absence of oncogene activating mutations eligible patients to targeted therapy (EGFR, ALK)
• Cohort A: Received first line treatment with pembrolizumab monotherapy
• Cohort B: Received first line treatment with chemotherapy and pembrolizumab combination therapy
• Cohort C: Received first line treatment with chemotherapy doublet

Exclusion Criteria:

• Prior anti-cancer therapy for actual stage IV NSCLC
• Critical data missing (e.g., PD-L1 status, baseline millimetric imaging, first evaluation millimetric imaging)
• Patients participating in other clinical trials that modify the standard of care