Microsatellite Instability, Tumor Mutational Burden, and Response to Immune Checkpoint Blockade in Patients with Prostate Cancer

Authors

Lenis AT

Ravichandran V

Brown S

Alam SM

Katims A

Truong H

Reisz PA

Vasselman S

Nweji B

Autio KA

Morris MJ

Slovin SF

Rathkopf D

Danila D

Scher HI

Woo S

Vargas HA

Laudone VP

Ehdaie B

Reuter V

Arcila M

Berger MF

Viale A

Schultz N

Gopalan A

Donoghue MTA

Ostrovnaya I

Stopsack KH

Solit DB

Abida W

Doi

10.1158/1078-0432.CCR-23-3403

PMID: 38949888 · DOI: 10.1158/1078-0432.CCR-23-3403 · Journal: Clinical Cancer Research (2024)

TL;DR

This study compared the genomic features, clinical course, and response to immune checkpoint blockade (ICB) in prostate cancer patients with microsatellite instability high/mismatch repair deficient (MSI-H/dMMR) versus tumor mutational burden high/microsatellite stable (TMB-H/MSS) tumors. Among 2,257 prostate cancer patients sequenced with MSK-IMPACT, 63 (2.8%) had MSI-H/dMMR and 33 (1.5%) had TMB-H/MSS tumors. MSI-H/dMMR patients treated with ICB showed a 45% RECIST response rate and 65% PSA50 response, whereas no TMB-H/MSS patient achieved a RECIST response. MSI-H/dMMR tumors had significantly higher TMB, indel burden, and neoantigen burden compared with TMB-H/MSS tumors, potentially explaining the differential ICB efficacy.

Cohort & data

  • 2,257 prostate cancer (PRAD) patients treated at Memorial Sloan Kettering Cancer Center from April 2015 to February 2021.
  • 3,244 tumors sequenced using MSK-IMPACT targeted next-generation sequencing panel.
  • Dataset deposited in cBioPortal as prostate_msk_2024.
  • Whole exome sequencing (WES) performed on 48 tumor samples (20 MSI-H/dMMR, 28 TMB-H/MSS) via recapture of MSK-IMPACT libraries.
  • Copy number analysis performed using FACETS (v0.5.6).
  • MSI status determined by MSIsensor (v0.5); TMB defined as non-synonymous mutations per megabase in canonical exons.

Key findings

  • 63 (2.8%) patients had MSI-H/dMMR tumors, 33 (1.5%) had TMB-H/MSS tumors, and 2,161 (95.7%) had TMB-L/MSS tumors (PMID:38949888).
  • MSI-H/dMMR tumors had significantly higher TMB than TMB-H/MSS (median 41 vs. 15 mutations/Mb, p<0.01).
  • Both MSI-H/dMMR (62%) and TMB-H/MSS (59%) tumors were more commonly Gleason grade group 5 compared with TMB-L/MSS (40%, p<0.001).
  • Among 27 MSI-H/dMMR patients treated with ICB: 45% RECIST overall response rate (95% CI 26-66%), including 2 complete responses and 7 partial responses; 65% PSA50 response rate (15/23 evaluable).
  • Among 8 TMB-H/MSS patients treated with ICB: 0% RECIST response (all 5 evaluable had stable disease); 50% PSA50 response rate (3/6 evaluable).
  • Median rPFS tended to be longer for MSI-H/dMMR vs. TMB-H/MSS patients on ICB (38 vs. 7 months, HR 2.13, 95% CI 0.79-5.73, p=0.14).
  • OS from metastatic disease was not significantly different across subgroups (MSI-H/dMMR median 45 months, TMB-H/MSS 39 months, TMB-L/MSS 48 months; log-rank p=0.36).
  • WES analysis: MSI-H/dMMR tumors had significantly higher neoantigen burden (median 11.1 vs. 6.8, p=0.011), indel burden (8.7 vs. 2.1, p<0.001), and TMB (27.39 vs. 25.02, p=0.014) compared with TMB-H/MSS.
  • No statistically significant genomic differences identified between MSI-H/dMMR responders and non-responders to ICB.

Genes & alterations

  • MSH2, MSH6, MLH1, PMS2: Deleterious MMR gene alterations identified in 75% of MSI-H/dMMR patients; rare in TMB-H/MSS (MSH2 6%, MSH6 9%, MLH1 3%, PMS2 0%) (PMID:38949888).
  • JAK1: Higher mutation rate in MSI-H/dMMR (56%) vs. TMB-H/MSS (15%) and TMB-L/MSS (32%, p<0.001), likely due to a microsatellite tract within JAK1 (PMID:38949888).
  • BRCA2/BRCA1: Higher mutation rate in MSI-H/dMMR (17.5%) and TMB-H/MSS (21%) vs. TMB-L/MSS (5.2%); these may represent subclonal or passenger events rather than actionable alterations (PMID:38949888).
  • ERG/TMPRSS2: TMPRSS2-ERG fusions depleted in MSI-H/dMMR (3%) vs. TMB-H/MSS (12%) and TMB-L/MSS (p=0.015), suggesting distinct oncogenic pathways (PMID:38949888).
  • POLE: Three patients had oncogenic POLE mutations (two P286R, one V411M; TMBs of 183, 34, 169), all with microsatellite stable tumors. None of the ICB-treated patients harbored POLE mutations (PMID:38949888).
  • ASXL1, MSH3: Identified as shared putative oncogenic mutations in the exceptional responder case highlighting clonal relatedness of serial tumors (PMID:38949888).

Clinical implications

  • MSI-H/dMMR prostate cancer patients can achieve profound and durable responses to ICB, supporting NCCN recommendations for somatic tumor testing in locally advanced and metastatic prostate cancer (PMID:38949888).
  • TMB-H/MSS status alone may be an insufficient biomarker for ICB response in prostate cancer, as no RECIST responses were observed despite FDA approval of pembrolizumab for TMB-H tumors (PMID:38949888).
  • Higher indel and neoantigen burden in MSI-H/dMMR tumors may explain the differential response to ICB compared with TMB-H/MSS tumors, suggesting these could serve as more refined biomarkers (PMID:38949888).
  • Elevated BRCA2/BRCA1 mutation rates in MSI-H/dMMR and TMB-H/MSS tumors likely represent passenger events rather than actionable targets, consistent with lower responses to PARP inhibitors reported previously (PMID:38949888).

Limitations & open questions

  • Small cohort sizes for MSI-H/dMMR (n=63) and TMB-H/MSS (n=33) subgroups limited statistical power for genomic comparisons between ICB responders and non-responders.
  • Retrospective, real-world study design with variable imaging intervals and non-standardized response evaluation timing.
  • Response assessment did not incorporate iRECIST criteria, developed specifically for immunotherapy-treated tumors.
  • IHC for MMR protein expression was not performed; MSI-H/dMMR classification relied on MSIsensor score and genomic MMR alterations.
  • Tumors were sequenced from both primary and metastatic disease sites, introducing potential heterogeneity.
  • TMB-H/MSS tumors displayed heterogeneous mutational signatures (MMR, aging, APOBEC, POLE, tobacco), suggesting variable molecular pathogenesis; larger studies may identify subsets (e.g., POLE-mutant) with higher ICB response probability.
  • Complementary molecular profiling (e.g., RNA sequencing, immune infiltrate characterization) may be required to identify additional predictors of ICB response within MSI-H/dMMR tumors.

Citations from this paper used in the wiki

  • “63 (2.8%) men had MSI-H/dMMR and 33 (1.5%) had TMB-H/MSS prostate cancers.” (Results)
  • “MSI-H/dMMR patients had significantly higher TMB than TMB-H/MSS patients (median 41, interquartile range 26-57 vs. median 15, interquartile range 11-27, p<0.01)” (Results)
  • “an overall response rate of 45% (95% confidence interval 26-66%) was observed, with two (10%) complete responses (CRs), seven (35%) partial responses (PRs)” (Results)
  • “No TMB-H/MSS patient had a RECIST response and 50% had a PSA50 response.” (Abstract)
  • “neoantigen burden (11.1 vs. 6.8, p=0.011), and indel burden (8.7 vs. 2.1, p<0.001) were significantly greater in MSI-H/dMMR tumors compared to TMB-H/MSS tumors” (Results)
  • “we also found that MSI-H/dMMR prostate cancers had lower rates of TMPRSS2-ERG rearrangements compared with TMB-H/MSS and TMB-L/MSS tumors” (Discussion)
  • “MSI-H/dMMR prostate cancers are enriched for JAK1 mutations, which may be due to a microsatellite tract within JAK1” (Discussion)

This page was processed by crosslinker on 2026-05-04.