Are ¹⁸F-FDG PET/CT based radiomics features useful for prediction of PD-L1 expression in non-small cell lung cancer?

This study investigated the diagnostic test accuracy of fluorine-18-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) based radiomics features for prediction of programmed cell death protein 1 and its ligand (PD-L1) expression in non-small cell lung cancer (NSCLC).

A systematic search was performed in PubMed and EMBASE (last updated in 31 August 2024). Studies evaluating diagnostic performance of ¹⁸F-FDG PET/CT based radiomics features for prediction of PD-L1 expression in NSCLC. The sensitivities, specificities, positive and negative likelihood ratios (LR+ and LR-), and pooled area under curve (AUC) were estimated.

The pooled sensitivity of ¹⁸F-FDG PET/CT was 0.75 (95% CI; 0.64-0.83) and a pooled specificity of 0.66 (95% CI; 0.52-0.78) for prediction of >1% expression of PD-L1. For prediction of >50% expression of PD-L1, the pooled sensitivity of ¹⁸F-FDG PET/CT was 0.77 (95% CI; 0.67-0.85) and a pooled specificity of 0.61 (95% CI; 0.55-0.66). For >1% expression of PD-L1, the pooled AUC of fixed effects was 0.791 (95% CI; 0.771-0.811) and of random effects was 0.783 (95% CI; 0.722-0.845). For >50% expression of PD-L1, the pooled AUC of fixed effects was 0.735 (95% CI; 0.718-0.751) and of random effects was 0.766 (95% CI; 0.706-0.825).

Analysis of the available studies indicated that ¹⁸F-FDG PET/CT based radiomics features showed a moderate diagnostic performance for prediction of PD-L1 expression in NSCLC. However, future studies would be necessary for standardization of the method for prediction of PD-L1 expression in NSCLC using ¹⁸F-FDG PET/CT based radiomics features.