Sequential Monitoring of PD-L1 on Circulating Tumor Stromal Cells Predicts Survival Outcomes for Unresectable Stage 3 NSCLC Treated With Immunotherapies After Definitive Chemoradiation
التفاصيل البيبلوغرافية
العنوان:
Sequential Monitoring of PD-L1 on Circulating Tumor Stromal Cells Predicts Survival Outcomes for Unresectable Stage 3 NSCLC Treated With Immunotherapies After Definitive Chemoradiation
PURPOSE/OBJECTIVE(S) Cancer Associated Macrophage-Like cells (CAMLs) are a type of circulating stromal cell found in the blood of cancer patients (pts), described as of myeloid origin that represent the inflammatory state of the tumor microenvironment. We have previously reported that in non-small cell lung carcinoma (NSCLC), CAML PD-L1 expression is dynamic and trackable throughout chemoradiation (CRT). However, it is unknown if PD-L1 on CAMLs has clinical utility, particularly in immunotherapy (IMT) efficacy after CRT. We studied stage 3 unresectable NSCLC pts treated with (n = 80) or without (n = 55) IMT agents after CRT to evaluate CAML PD-L1 expression on progression free survival (PFS) & overall survival (OS). MATERIALS/METHODS We used 2 sets of IMT treated pts: first set from a single arm phase 2 trial treated with CRT and atezolizumab (atezo) (n = 34) and a second from a multi-year prospective study treated with standard of care CRT with consolidation durvalumab (durva) (n = 46). A control group of CRT alone pts (n = 55) were evaluated during the same period (majority before 2018). In all, 135 pts with pathologically confirmed unresectable stage 3 NSCLC were included. 15mL blood samples were taken at Baseline (BL) & ∼1 month after completion of CRT. Blood was filtered & CAMLs quantified for PD-L1 expression using a binary score (0/1 = low & 2/3 = high), to evaluate PFS & OS hazard ratios (HRs) by censored univariate & multivariate analysis at 24 months. RESULTS CAMLs were found in 92% of all samples, on average 5.2 CAMLs/15mL. From the available tumor biopsies, > 1% PD-L1 expression using DAKO 22c3 did not predict CRT+IMT (atezo/durva) response (PFS HR 1.4 P = 0.642 & OS HR 1.9 P = 0.293). At BL, high CAML PD-L1 was not associated with PFS in CRT alone (HR 1.5 P = 0.435), CRT+atezo (HR 1.0 P = 0.843), or CRT+durva (HR 1.6 P = 0.6180). After CRT, while high CAML PD-L1 did not predict PFS in CRT alone (HR 1.0 P = 0.925), it was associated with superior PFS in CRT+atezo (HR 2.8 P = 0.046), & CRT+durva (HR 5.2 P = 0.036). For OS, high CAML PD-L1 was not predictive of outcomes in CRT alone (HR 0.8 P = 0.929), but strongly predictive of superior OS for CRT+atezo (HR 4.3 P = 0.036) & CRT+durva (HR 5.9 P = 0.033) groups. Combined (atezo or durva, n = 80), 57% of pts had either maintained high PD-L1 before and after CRT (N = 25) or increased from BL (N = 21). Pts with high PD-L1 post CRT had better outcomes than pts with low expression, PFS HR = 4.1 P < 0.001 and OS HR = 4.7 P < 0.001. This pattern was not seen in pts treated with CRT alone (PFS HR = 1.2 P = 0.763 and OS HR = 0.9 P = 0.977). CONCLUSION Our data suggests that in unresectable stage 3 NSCLC, high PD-L1 expression on CAMLs after CRT completion is predictive of superior clinical outcomes in pts treated with IMT, but not in those treated with CRT alone. Prospective validation is needed to confirm these findings.