32?C-reactive protein (CRP) as a prognostic biomarker in advanced non-small cell lung cancer treated with immune checkpoint inhibitors. Results from a multi-center international observational study

摘要

Background CRP is an acute-phase protein produced primarily in response to interleukin IL-6 via transcriptional activation of the STAT3. Recent data have provided mechanistic insights into the immune suppressive role of elevated CRP by elucidating its influence on effector T-cell function and antigen presentation. 1 Furthermore, melanoma patients in Checkmate-064 and 067 with high baseline and on-treatment CRP were seen to have a lower response rate and shorter survival to immune checkpoint inhibitors (ICIs). 2 Given these observations, we sought to evaluate the role of CRP as a prognostic biomarker in advanced NSCLC treated with ICIs from a multi-center international cohort. Methods Between 2015–2019, 420 adult patients with advanced NSCLC treated with ICIs alone or with concurrent chemotherapy (Chemo-ICI) were identified at four (1 US and 3 European) academic centers. CRP level in peripheral blood samples collected up to 2 weeks before starting ICI based treatments was considered as baseline. Based on previously validated data, a CRP cutoff of 10 mg/l was used to define CRP-normal (CRP-N) and CRP-high (CRP-H). Association of baseline CRP with median progression-free survival (PFS) and overall survival (OS) were estimated using the Kaplan-Meier method and multivariate proportional hazards regression adjusted for multiple variables. Results Baseline CRP value was available in 75.5% of patients, with 66% having CRP-H. The median CRP was 21.0 mg/l. Single-agent nivolumab (44%) and Chemo-ICI (33.3%) were the two most common therapies. CRP-H showed a trend for stronger association with squamous histology (73.7% vs 63.3%; p= 0.063) and female sex (70.8 vs 60.0%; p=0.062) but did not show an association with PD-L1 status (0%, 1–49%, or ≥50%). Patients with CRP-H had a lower objective response rate compared with patients with CRP-N (26.9% vs. 47.6% PR; p=0.029). Compared to those with CRP-N (figure 1), patients with CRP-H had a significantly shorter median PFS [3.9 vs. 6.6 months, HR 1.41 95% CI: (1.07–1.86); p=0.0138] and OS (8.6 vs. 14.8 months, HR 1.55 95% CI [1.13- 2.14]; p=0.0060). In Cox regression analysis, CRP-H was again found to be independently associated with shorter median PFS and OS. Conclusions This is the largest international real-world dataset demonstrating significantly inferior outcomes associated with CRP 10 mg/l in NSCLC patients treated with ICI based therapies. The potential influence of the immune suppressive effects of elevated CRP and IL-6 on the anti-tumor efficacy of ICIs needs prospective evaluation and could potentially be exploited as a therapeutic avenue in NSCLC. Abstract 32 Figure 1 Kaplan-Meier Curves with 95% CI for PFS and OSSignificantly inferior median PFS and OS were seen for patients with CRP-H vs. CRP-N. Acknowledgements Susan Eubanks and Sue-Ann Joyner at the ECU IRB for their help and support. Ethics Approval The primary IRB approval for this study was conducted under an ECU (P-MAIT- UMCIRB-15-001400). Individual approval was also obtained from the respective IRB of each participating institution.