An analysis based on a review of the recent US Food and Drug Administration (FDA) approval of pembrolizumab in patients with treatment-refractory cancers and a tumour mutational burden (TMB) greater than 10 mutations per megabase, found that the overall survival (OS) benefit was decreased in several tumour types and after re-evaluation according to stratification by other criteria.
In correspondence published on 25 March 2021 in The New England Journal of Medicine, first author Benoit Rousseau and investigators from the Memorial Sloan Kettering Cancer Centre in New York, NY and Johns Hopkins Medical Institutes in Baltimore, MD, US questioned whether this FDA approval, which was based solely on a TMB cut-off value, could be too broad.
They explained that the FDA approval of the programmed death 1 (PD-1) blocking antibody, pembrolizumab, was for treatment of any type of solid tumour. They further noted that approval was based on retrospective evidence from a study indicating that high TMB was predictive of a favourable radiographic response to treatment with immune checkpoint inhibitors (ICIs) in patients with 10 different rare cancers.
However, they pointed out that, in this study, high TMB was not predictive of improved OS after treatment with ICIs, and other major tumour types as well as the cause of the high-mutation phenotype were not evaluated.
Patient outcome following ICI treatment was evaluated in patients with colorectal and other types of cancer
According to the authors, the FDA approval would affect approximately 18% of patients with advanced colorectal cancer, prompting them to retrospectively evaluate the outcome of ICI treatment in 137 patients with this cancer.
The investigators found that median OS was longer in patients with advanced colorectal cancer and high TMB (TMB ≥10 mutations per megabase) than in similar patients with low TMB following ICI therapy (hazard ratio [HR] for death, 0.40; 95% confidence interval [CI] 0.24-0.65).
Importantly, the OS benefit was no longer evident after the cohort was stratified according to mismatch repair–deficiency status or with respect to the presence of pathogenic mutations in polymerase ε (POLE) or polymerase δ1 (POLD1), which are collectively referred to as pol-d (HR for death, 1.17; 95% CI 0.59-2.32).
The investigators broadened this analysis to include 1661 patients treated with ICIs for various tumour types; in this cohort, TMB of 10 or more mutations per megabase was associated with improved OS, but only in a limited subgroup of patients that also had mismatch repair–proficient tumours. Following stratification by tumour type, OS was improved only in patients with metastatic head and neck cancer, non–small cell lung cancer, and melanoma (HR for death, 0.52; 95% CI 0.31-0.64).
Kaplan–Meier plots provided by the authors showed no benefit in patients with 10 other tumour types that had intact mismatch repair or without pol-d mutations, including kidney, breast, and neuroendocrine tumours, uveal melanoma, and mucosal melanoma (HR for death, 0.84; 95% CI 0.63-1.11).
They noted that mismatch-repair deficiency is a well-established biomarker of improved OS after ICI treatment, and suggested that pol-d status may also indicate benefit with ICIs.
They found that, in addition to patients with mismatch-repair deficiency and pol-d mutations, the only patients with hypermutated tumours showing benefit from ICI treatment had cancers that were strongly associated with environmental carcinogens, such as chronic exposure to ultraviolet radiation or tobacco.
Conclusions
The authors cautioned that the current FDA approval granted on the basis of TMB may be too broad. This approval, which was granted solely on the basis of response rate, neglects more meaningful clinical end points, including survival and quality of life, which might slow the development of more effective therapies for patients with hypermutated tumours.
The investigators propose that ICIs should be considered in the context of the cause of the high TMB and not based solely on an absolute threshold.
No external funding was disclosed.
Reference
Rousseau B, Foote MB, Maron SB, et al. The Spectrum of Benefit from Checkpoint Blockade in Hypermutated Tumors. N Engl J Med 2021;384:1168-1170. DOI: 10.1056/NEJMc2031965.