Tissue TMB Shows Potential as Biomarker of Response to Pembrolizumab in Thyroid Cancer

Patients with tissue tumor mutational burden (tTMB)-high solid tumors represent a subset of patients who could have a robust tumor response to the immune checkpoint inhibitor pembrolizumab (Keytruda) as monotherapy, suggesting tTMB may be a novel and useful predictor of response in patients with previously treated recurrent or metastatic advanced solid tumors, including thyroid cancer, according to findings from the phase 2 KEYNOTE-158 clinical trial.

The multicohort, single-arm, open-label study assessed pembrolizumab monotherapy in patients with less frequently occurring solid tumor types, which was conducted across 81 sites in 21 countries. Patients were enrolled to 1 of 10 tumor type-specific cohorts, and cohort I, specifically, included patients with papillary or follicular thyroid carcinoma.

Other cohorts included anal squamous cell carcinoma (cohort A); biliary adenocarcinoma except for ampulla of Vater cancers (cohort B); well and moderately differentiated neuroendocrine tumors of the lung, appendix, small intestine, colon, rectum, or pancreas (cohort C); endometrial carcinoma except sarcomas and mesenchymal tumors (cohort D); cervical squamous cell carcinoma (cohort E); vulvar squamous cell carcinoma (cohort F); small-cell lung carcinoma (cohort G); malignant pleural mesothelioma (cohort H); and salivary gland carcinoma except for sarcomas and mesenchymal tumors (cohort J). Any other advanced solid tumor, with the exception of colorectal cancer, with microsatellite instability (MSI)-High (MSI-H) were also enrolled and included in cohort K.

The primary end point of the study was the proportion of patients with an objective response per RECIST version 1.1, and secondary end points included duration of response (DOR), progression-free survival (PFS), overall survival (OS), and safety and tolerability.

Patients received 200 mg pembrolizumab once every 3 weeks intravenously, which was continued until completion of 35 cycles or until disease progression, unacceptable toxicity, intercurrent illness preventing further treatment, or investigator decision. A total of 1073 patients were enrolled, and 1066 were treated as of the data cut-off of June 27, 2019. Among these patients, 805 (76%) were evaluable for TMB, and 105 of those (13%) had tTMB-high and were assessed for safety.

Additionally, 1050 of 1066 (98%) enrolled by at least 26 weeks before the data cutoff date, and 790 of these patients (75%) were evaluable for TMB and were included in the efficacy analyses. A total of 102 (13%) of the 790 patients had tTMB-high status with ≥10 mutations per megabase, and 688 (87%) had non-tTMB-high status. Among patients with thyroid cancer, 2 (2%) in the efficacy population fell under the tTMB-high group, 78 (11%) in the non-tTMB-high group, and 2 (2%) were included in the safety population.

Objective responses were observed in 30 (29%) of the 102 patients in the tTMB-high group (95% CI, 21-39) and 43 (6%) of the 688 patients in the non-tTMB-high group. Complete responses occurred in 4 (4%) of the tTMB-high group and 11 (2%) of the non-tTMB-high group, while partial responses were observed in 26 (25%) and 32 (5%), respectively.

Fifty-three (58%) of the 91 patients in the tTMB-high group who were evaluable for post-baseline image assessments and 214 (35%) of 619 in the non-tTMB-high group had a reduction from baseline in tumor size. The median DOR had not been reached in the tTMB-high group (range, 2.2+ to 34.8+) and 33.1 months (range, 4.0 to 35.7+) in the non-tTMB-high group.

In the cohort of patients with thyroid cancer, 2 of 2 patients in the tTMB-high group had objective responses, as well as 3 of 78 patients in the non-tTMB-high group.

Across the majority of tumor type cohorts, the proportion of patients with an objective response was higher among patients without tTMB-high status, including the cohort of patients with thyroid cancer.

An association between TMB score and PD-L1 combined positive score (CPS) was not observed, either in the overall efficacy population (P =.18) or in patients who had a response (P =.07) or did not have a response (P=.15). Among those with both tTMB-high status and PD-L1 positivity, objective responses were observed in 24 (35%) of 68 patients (95% CI, 24-48).

Objective responses also occurred in 6 (21%) of 29 patients with tTMB-high and PD-L1 negativity with CPS < 1 (95% CI, 8-40). Among patients in the non-tTMB-high group, objective responses were seen in 33 (9%) of 383 with PD-L1-positive tumors (95% CI, 6-12) and 9 (3%) of 274 patients with PD-L1-negative tumors (95% CI, 2-6).

Disease progression or death occurred in 78 (76%) of 102 patients in the tTMB-high group and 645 (94%) of 688 in the non-tTMB-high group. The median PFS was 2.1 months in the tTMB-high group (95% CI, 2.1-4.1) and 2.1 months in the non-tTMB-high group (95% CI, 2.1-2.2). Sixty-nine (68%) in the tTMB-high group and 534 (78%) in the non-tTMB-high group had died as of the data cutoff.

Overall, 67 (64%) of patients had at least 1 treatment-related adverse event (AE), and the most common were fatigue, asthenia, and hypothyroidism. Treatment-related serious AEs were observed in 11 (10%) patients, which included comprised acute respiratory distress syndrome, blood creatinine increased, chest pain, colitis, fulminant type 1 diabetes, hypotension, lower abdominal pain, pneumonia, pneumonitis, primary adrenal insufficiency, secondary adrenocortical insufficiency, and sepsis (n = 1 each). One patient had both lower abdominal pain and chest pain.

Sixteen (15%) patients had a grade 3 to 5 treatment-related AE, but the only event to occur in more than 1 patient was colitis, which was observed in 2 (2%) patients, both cases resolved. Immune-mediated AEs and infusion reactions occurred in 26 (25%) patients, and the most frequently occurring events included hypothyroidism, hyperthyroidism, colitis, and pneumonitis. These events were mostly mild to moderate in severity, and grade 3 events occurred in 9 (9%) of patients. No deaths due to immune-mediated events and infusion reactions occurred in the study.

_Reference

_{Marabelle A, Fakih M, Lopez J, et al. Association of tumour mutational burden with outcomes in patients with advanced solid tumours treated with pembrolizumab: prospective biomarker analysis of the multicohort, open-label, phase 2 KEYNOTE-158 study. Lancet Oncol. Published Online September 10, 2020. doi: 10.1016/S1470-2045(20)30445-9.}