Patients with predominant variant histology (VH) of bladder tumors, defined as involving >50 % of the tumor specimens, are typically excluded from clinical trials, and for these patients, the efficacy of standard chemotherapy is limited.

To evaluate the activity of preoperative pembrolizumab in patients with muscle-invasive bladder carcinoma (MIBC) and VH, enrolled in PURE-01 study (NCT02736266).

In the open-label, single-arm, phase 2 PURE-01 study, three courses of 200 mg pembrolizumab preceding radical cystectomy (RC) were administered in T2-4aN0M0 MIBC patients. The amended study design included patients with predominant VH.

Neoadjuvant pembrolizumab and RC.

Pathological complete response (pT0) in intention-to-treat population was the primary endpoint. Biomarker analyses included programmed cell-death ligand-1 (PD-L1) expression using the combined positive score (CPS; Dako 22C3 antibody) and comprehensive genomic profiling (FoundationOne assay). Multivariable logistic regression analyses (MVAs) evaluated the histological category (predominant VH vs nonpredominant VH vs pure urothelial carcinoma), tumor mutational burden (TMB) and CPS in association with the pathological response.

From February 2017 to June 2019, 114 patients were enrolled; 34 (30%) of them presented with VH, including 19 (17%) with predominant VH. In total, the pT0 rate was 37% (95% confidence interval [CI]: 28-46) and the pT ≤ 1 rate was 55% (95% CI: 46-65). The majority of predominant VH patients presented with squamous-cell carcinoma (SCC; N = 7), and six of seven (86%) had downstaging to pT ≤ 1, with one pT0; two of three lymphoepithelioma-like (LEL) variants had a pT0 response. None of the remaining nine predominant VHs had a response. On MVA, TMB and CPS were associated with both the pT0 and the pT ≤ 1 response, regardless of tumor histology.

The updated PURE-01 results confirm the activity of neoadjuvant pembrolizumab in MIBC. Patients with SCC and LEL features may be suitable for neoadjuvant immunotherapy trials. CPS and TMB are the key response predictors irrespective of the histological subtypes.

In the PURE-01 study, we have preliminarily evaluated the activity of neoadjuvant pembrolizumab in patients with predominant variant histology (VH). Of these patients, those harboring squamous-cell carcinoma or a lymphoepithelioma-like variant feature had major, although preliminary, pathological responses compared with those with other predominant VHs. Expression of programmed cell-death ligand-1 and tumor mutational burden may predict the pathological response to pembrolizumab, and provide a rationale for selecting patients according to these features instead of the histological bladder cancer subtypes.