Organ-specific metastatic landscape dissects PD-(L)1 blockade efficacy in advanced non-small cell lung cancer: applicability from clinical trials to real-world practice
Si-Cong Ma,Xue Bai,Xue-Jun Guo,Li Liu,Lu-Shan Xiao,Yan Lin,Jia-Le Tan,Xiao-Ting Cai,Yu-Xiang Wen,Hu Ma,Q. John Fu,Meng-Xin Leng,Yan-Pei Zhang,Li-Li Long,Ze-Qin Guo,De-Hua Wu,Jian-Guo Zhou,Zhong-Yi Dong
DOI: https://doi.org/10.1186/s12916-022-02315-2
IF: 9.3
2022-04-12
BMC Medicine
Abstract:Abstract Background Organ-specific metastatic context has not been incorporated into the clinical practice of guiding programmed death-(ligand) 1 [PD-(L)1] blockade, due to a lack of understanding of its predictive versus prognostic value. We aim at delineating and then incorporating both the predictive and prognostic effects of the metastatic-organ landscape to dissect PD-(L)1 blockade efficacy in non-small cell lung cancer (NSCLC). Methods A total of 2062 NSCLC patients from a double-arm randomized trial (OAK), two immunotherapy trials (FIR, BIRCH), and a real-world cohort (NFyy) were included. The metastatic organs were stratified into two categories based on their treatment-dependent predictive significance versus treatment-independent prognosis. A metastasis-based scoring system (METscore) was developed and validated for guiding PD-(L)1 blockade in clinical trials and real-world practice. Results Patients harboring various organ-specific metastases presented significantly different responses to immunotherapy, and those with brain and adrenal gland metastases survived longer than others [overall survival (OS), p = 0.0105; progression-free survival (PFS), p = 0.0167]. In contrast, survival outcomes were similar in chemotherapy-treated patients regardless of metastatic sites (OS, p = 0.3742; PFS, p = 0.8242). Intriguingly, the immunotherapeutic predictive significance of the metastatic-organ landscape was specifically presented in PD-L1-positive populations (PD-L1 > 1%). Among them, a paradoxical coexistence of a favorable predictive effect coupled with an unfavorable prognostic effect was observed in metastases to adrenal glands, brain, and liver (category I organs), whereas metastases to bone, pleura, pleural effusion, and mediastinum yielded consistent unfavorable predictive and prognostic effects (category II organs). METscore was capable of integrating both predictive and prognostic effects of the entire landscape and dissected OS outcome of NSCLC patients received PD-(L)1 blockade ( p < 0.0001) but not chemotherapy ( p = 0.0805) in the OAK training cohort. Meanwhile, general performance of METscore was first validated in FIR ( p = 0.0350) and BIRCH ( p < 0.0001), and then in the real-world NFyy cohort ( p = 0.0181). Notably, METscore was also applicable to patients received PD-(L)1 blockade as first-line treatment both in the clinical trials (OS, p = 0.0087; PFS, p = 0.0290) and in the real-world practice (OS, p = 0.0182; PFS, p = 0.0045). Conclusions Organ-specific metastatic landscape served as a potential predictor of immunotherapy, and METscore might enable noninvasive forecast of PD-(L)1 blockade efficacy using baseline radiologic assessments in advanced NSCLC.
medicine, general & internal