Hyperprogressive disease during PD-1/PD-L1 blockade in patients with non-small-cell lung cancer

Immune checkpoint blockade with Programmed cell death 1 (PD-1)/PD-L1 inhibitors has been effective in various malignancies and is considered as a standard treatment modality for patients with non-small-cell lung cancer (NSCLC). However, emerging evidence show that PD-1/PD-L1 blockade can lead to hyperprogressive disease (HPD), a flair-up of tumor growth linked to dismal prognosis. This study aimed to evaluate the incidence of HPD and identify the determinants associated with HPD in patients with NSCLC treated with PD-1/PD-L1 blockade. We enrolled patients with recurrent and/or metastatic NSCLC treated with PD-1/PD-L1 inhibitors between April 2014 and November 2018. Clinicopathologic variables, dynamics of tumor growth, and treatment outcomes were analyzed in patients with NSCLC who received PD-1/PD-L1 blockade. HPD was defined according to tumor growth kinetics (TGK), tumor growth rate (TGR), and time to treatment failure (TTF). Immunophenotyping of peripheral blood CD8+ T lymphocytes was conducted to explore the potential predictive biomarkers of HPD. A total of 263 patients were analyzed. HPD was observed in 55 (20.9%), 54 (20.5%), and 98 (37.3%) patients according to the TGK, TGR, and TTF. HPD meeting both TGK and TGR criteria was associated with worse progression-free survival [hazard ratio (HR) 4.619; 95% confidence interval (CI) 2.868–7.440] and overall survival (HR, 5.079; 95% CI, 3.136–8.226) than progressive disease without HPD. There were no clinicopathologic variables specific for HPD. In the exploratory biomarker analysis with peripheral blood CD8+ T lymphocytes, a lower frequency of effector/memory subsets (CCR7−CD45RA− T cells among the total CD8+ T cells) and a higher frequency of severely exhausted populations (TIGIT+ T cells among PD-1+CD8+ T cells) were associated with HPD and inferior survival rate. HPD is common in NSCLC patients treated with PD-1/PD-L1 inhibitors. Biomarkers derived from rationally designed analysis may successfully predict HPD and worse outcomes, meriting further investigation of HPD.