18F-FDG PET/CT for prediction of PD-L1 expression and tumor mutational burden in non-small cell lung cancer: biomarkers prediction models development and immunotherapy responses verification

Abstract

Background: Whether metabolic parameters on PET/CT are noninvasive alternatives to programmed cell death ligand 1 (PD-L1) expression and tumor mutational burden (TMB) in adenocarcinoma (ADC) and squamous cell carcinoma (SCC) remains unknown. We identified predictors and developed PET/CT-based nomogram models for predicting PD-L1 expression and TMB, and tested the usefulness of models for stratifying immunotherapy responses. Methods: We enrolled 305 patients with non-small cell lung cancer from January 2017 to April 2024 as the primary cohort to investigate independent predictors of PD-L1 expression and TMB for ADC (n = 183) and SCC (n = 122). Clinicopathological characteristics, semantic CT features, and PET metabolic parameters were reviewed and analyzed. Significant predictors of ADC biomarkers were identified, and ADC were randomly assigned 7:3 to the training (n = 128) and validation (n = 55) cohorts to develop and validate predictive models for PD-L1 expression, TMB, and the combination of both. Clinical, SULmax, and clinical-SULmax combined models were constructed using logistic regression analysis. Model discrimination, calibration, and clinical usefulness were assessed. An independent test cohort (n = 29) with ADC receiving neoadjuvant immunotherapy was used to validate the models in stratifying pathologic responses. Results: PD-L1 expression did not differ significantly between SCC and ADC; however, the TMB was significantly higher in SCC (10 vs. 5 mutations/Mb, p < 0.001). In the analysis of ADC, PD-L1 expression was associated with clinical stage, differentiation, and EGFR mutation status (p < 0.05). TMB was associated with age, gender, smoking history, and EGFR mutation status (p < 0.05). SULmax was an independent predictor of both PD-L1-Pos and TMB-High (p < 0.001). The clinical-SULmax combined models for predicting both PD-L1 expression and TMB demonstrated great performance (AUC = 0.805 for PD-L1-Pos and TMB-High; AUC = 0.798 for PD-L1-Neg and TMB-Low). The usefulness of models in stratifying pathologic responses was confirmed on the ADC test cohort receiving neoadjuvant immunotherapy (p = 0.035 for PD-L1-Pos and TMB-High model; p = 0.001 for PD-L1-Neg and TMB-Low model). However, no significant predictors were identified to develop models in the analysis of SCC. Conclusion: SULmax is an independent predictor of both PD-L1-Pos and TMB-High in ADC. The clinical-SULmax combined models effectively predicted PD-L1-Pos and TMB-High status, as well as PD-L1-Neg and TMB-Low status in ADC, indicating the clinical utility in patient selection and immunotherapy response stratification.