A combined immune prognostic index in esophageal squamous cell carcinoma patients treated with anti-PD-1 therapy

Only a fraction of patients with esophageal squamous cell carcinoma (ESCC) show tumor responses to anti-programmed cell death protein 1 (PD-1) therapy. The predictive value of single biomarkers for prognosis is limited, and a more comprehensive approach that incorporates multiple factors may improve the prognostic prediction. Here, we conducted a retrospective study to develop a combined immune prognostic index (CIPI) for predicting clinical outcomes of ESCC patients treated with anti-PD-1 therapy.

We performed a pooled analysis of two multicenter clinical trials comparing immunotherapy versus chemotherapy as second-line treatment in ESCC patients. The discovery cohort comprised patients who received anti-PD-1 inhibitors (N = 322) and the control cohort comprised patients who received chemotherapy (N = 307). The validation cohort included patients with pan-cancers treated with PD-1/programmed cell death ligand-1 inhibitors, except for ESCC (N = 110). Multivariable Cox proportional hazard regression was used to assess the prediction value of variables on survival.

In the discovery cohort, neutrophil-to-lymphocyte ratio, serum albumin, and liver metastasis were independently associated with overall survival (OS) and progression-free survival (PFS). We integrated the three variables into CIPI and found that CIPI could categorize patients into four subgroups (CIPI 0 to CIPI 3) with distinct OS, PFS, and tumor responses. The CIPI was also predictive of clinical outcomes in the validation cohort, but not in the control cohort. Furthermore, patients with CIPI 0, CIPI 1, and CIPI 2 were more likely to benefit from anti-PD-1 monotherapy than chemotherapy, while patients with CIPI 3 did not benefit from anti-PD-1 monotherapy over chemotherapy.

The CIPI score was a robust biomarker for prognostic prediction in ESCC patients treated with anti-PD-1 therapy and was immunotherapy specific. The CIPI score may also be applicable for prognostic prediction in pan-cancers.