Circulating Tumor DNA and Response to Cisplatin-based Chemotherapy in Patients with Metastatic Urothelial Carcinoma Enrolled in CALGB 90601 (Alliance)

Authors

Brendan J. Guercio

Karissa Whiting

Ronak H. Shah

Karla V. Ballman

Susan Halabi

Ashley M. Regazzi

Jennifer H. Milbank

Dean F. Bajorin

Himisha Beltran

Michael J. Morris

David B. Solit

Michael F. Berger

Gopa Iyer

Venkatraman Seshan

Jonathan E. Rosenberg

Doi

10.1016/j.euros.2025.03.009

PMID: 40256659 · DOI: 10.1016/j.euros.2025.03.009 · Journal: European Urology Open Science (2025)

TL;DR

Guercio et al. analyzed pretreatment plasma cell-free DNA (cfDNA) from 201 patients with metastatic urothelial carcinoma (mUC) nested within the phase 3 CALGB 90601 (Alliance) trial of first-line gemcitabine/cisplatin with or without bevacizumab. Using the MSK-ACCESS 129-gene cfDNA panel, they tested whether DNA damage response (DDR) gene alterations, ctDNA variant allele frequency (VAF), and other genomic features predict outcomes after cisplatin-based chemotherapy. Oncogenic DDR alterations (including ERCC2) were not significantly associated with response, OS, or PFS — but the analysis was underpowered (only 13% of cases had oncogenic DDR alterations). Higher pretreatment ctDNA VAF and cfDNA alterations in the TERT promoter, PIK3CA, and ERBB2 were each independently associated with shorter overall survival.

Cohort & data

  • Cohort: 201 patients with metastatic urothelial carcinoma drawn from the 506-patient phase 3 CALGB 90601 (Alliance) trial (NCT00942331), a randomized first-line trial of gemcitabine/cisplatin with bevacizumab or placebo. Inclusion required ≥5 mL of baseline plasma; cfDNA sequencing was successful for 201/212 (95%) eligible cases.
  • Cancer type: Metastatic urothelial carcinoma / bladder cancer (BLCA). Primary tumor sites: bladder 76%, upper tract 20%, other 3.5%.
  • Dataset: blca_msk_2025 — released via cBioPortal per the data availability statement.
  • Assays:
    • Pretreatment cfDNA + matched germline DNA via MSK-ACCESS — 129-gene tumor-uninformed panel with unique molecular indexes, depth >15,000×, allele-frequency detection threshold 0.1%.
    • Matched archival tumor (107 patients, 53%) via the MSK-IMPACT panel.
  • Endpoints: Primary endpoint of the parent trial was overall survival (OS). Progression-free survival (PFS) and response were secondary endpoints. cfDNA cohort median OS 14 mo (95% CI 13–18) on chemo+bevacizumab vs 16 mo (95% CI 15–18) on chemo+placebo; PFS 7.8 vs 7.6 mo respectively. Treatment-arm comparable to overall trial cohort.
  • Statistics: Cox proportional hazards; Benjamini-Hochberg FDR (q<0.05); cube-root–transformed 75th-percentile VAF (³√VAF₇₅ₚc) used to stabilize VAF distributions. Multivariable adjustment for ECOG performance status (0 vs 1) and visceral metastases; PIK3CA/ERBB2 models additionally adjusted for ctDNA VAF.

Key findings

  • DDR-negative result (underpowered). Oncogenic DDR gene alterations were detected in pretreatment cfDNA in 27 (13%) of 201 patients (11 ATM, 5.5%; 9 ERCC2, 4.5%; 6 BRCA1, 3%; 2 BRCA2, 1%). DDR+ vs DDR− response rate was 50% (12/24) vs 41% (60/145), p = 0.4. OS HR 0.78 (95% CI 0.50–1.22, p = 0.3); PFS HR 0.77 (95% CI 0.48–1.22, p = 0.3). Authors note insufficient power because of low DDR alteration frequency (4.5% ERCC2 in this cohort vs 11% in the TCGA muscle-invasive bladder cohort).
  • High ctDNA VAF predicts shorter survival. ³√VAF₇₅ₚc was inversely associated with OS (multivariable HR 2.51, 95% CI 1.26–5.00; p = 0.009) and PFS (HR 2.18, 95% CI 1.02–4.67; p = 0.045) after adjustment for visceral metastases and performance status. Univariable VAF₇₅ₚc top tertile (>0.082) vs lowest (≤0.018): OS HR 1.81 (95% CI 1.25–2.63; p = 0.002, q = 0.003).
  • TERT promoter alterations → shorter OS. Multivariable HR 1.59 (95% CI 1.15–2.19; p = 0.005). TERT was the most frequently altered gene in the cohort (57% of patients, primarily TERT promoter mutations) and was highly correlated with ctDNA VAF (p < 0.001) — therefore not co-modeled with VAF because of collinearity.
  • PIK3CA alterations → shorter OS. Multivariable HR 1.91 (95% CI 1.20–3.04; p = 0.006), adjusted for visceral metastases, performance status, and ³√VAF₇₅ₚc. CPE 0.63 (95% CI 0.59–0.67). PIK3CA altered in 11–17% of cases (cfDNA 17% vs matched tumor 13%).
  • ERBB2 alterations → shorter OS. Multivariable HR 1.64 (95% CI 1.08–2.49; p = 0.019), adjusted as above. CPE 0.63 (95% CI 0.59–0.67). ERBB2 altered in 14% of cfDNA samples.
  • cfDNA vs tumor concordance. Among 107 paired cases, 57.1% (n=24) of DDR alterations were detected in both cfDNA and matched archival tumor; 21.4% (n=9) in tumor only; 21.4% (n=9) in cfDNA only. cfDNA detected numerically higher rates of alteration in PIK3CA (17% vs 13%), RB1 (14% vs 12%), ERCC2 (6.5% vs 4.7%), BRAF (5.6% vs 4.7%), and KRAS (5.6% vs 3.7%) than matched tumor.
  • Mutational landscape. 175/201 samples (87.06%) had ≥1 alteration in the queried genes. Top frequencies (cfDNA): TERT 57%, TP53 ~52% (per oncoprint), FGFR3 19%, KDM6A 18%, ERBB2 14%, PIK3CA 11%, RB1 11%, TSC1 7%, ATM 5%.
  • Treatment-arm robustness. Stratification by bevacizumab vs placebo did not meaningfully alter DDR or VAF results; an exploratory TERT × bevacizumab analysis showed no differential effect.

Genes & alterations

  • TERT — promoter alterations dominant; 57% of cfDNA samples; multivariable OS HR 1.59 (95% CI 1.15–2.19; p = 0.005). Highly correlated with ctDNA VAF.
  • PIK3CA — 11% in cfDNA cohort (17% vs 13% in cfDNA-vs-tumor concordance subset); multivariable OS HR 1.91 (95% CI 1.20–3.04; p = 0.006) after adjustment for visceral metastases, performance status, and ³√VAF₇₅ₚc.
  • ERBB2 — 14% of cfDNA cohort; multivariable OS HR 1.64 (95% CI 1.08–2.49; p = 0.019). Authors highlight clinical relevance given mUC response rates of 33–83% to HER2-directed antibody-drug conjugates such as trastuzumab deruxtecan and disitamab vedotin.
  • ERCC2 — oncogenic alterations in only 9 cases (4.5%); no statistically significant association with response (5/11 [45%] response in any-ERCC2-altered patients, p > 0.9), OS (HR 0.93, 95% CI 0.50–1.72; p = 0.8), or PFS (HR 0.85, 95% CI 0.46–1.58; p = 0.6). Substantially below the 11% rate seen in the TCGA muscle-invasive cohort.
  • ATM — 11 oncogenic alterations (5.5%); part of underpowered DDR pooled analysis.
  • BRCA1 — 6 alterations (3%); pooled into DDR analysis.
  • BRCA2 — 2 alterations (1%); univariable signal of shorter OS (HR 6.06, 95% CI 1.45–25.3; p = 0.014) but would not survive multiple-comparison adjustment, and n = 2 limits inference.
  • FGFR3 — alterations in 19% of cfDNA samples; consistent with prior bladder cancer studies.
  • KDM6A — 18% in cfDNA.
  • RB1 — 11% in cfDNA cohort; numerically higher in cfDNA than matched tumor in the concordance subset (14% vs 12%).
  • TSC1 — 7% in cfDNA.
  • BRAF, KRAS, MDM2 — included as biologically relevant non-TCGA-significant genes; BRAF and KRAS show numerically higher detection in cfDNA than matched tumor.

Clinical implications

  • The authors explicitly position pretreatment ctDNA features (TERT promoter status, PIK3CA, ERBB2, and ³√VAF₇₅ₚc) as candidate prognostic biomarkers for mUC patients receiving cisplatin-based chemotherapy, suitable for clinical-trial stratification rather than treatment selection.
  • Despite supplantation of platinum chemotherapy by enfortumab vedotin plus pembrolizumab as first-line standard for mUC, gemcitabine/cisplatin remains relevant as second-line therapy after EV+pembrolizumab — keeping these biomarkers clinically actionable.
  • ERBB2-altered mUC is highlighted as a candidate for HER2-directed antibody-drug conjugates; the prognostic burden seen here adds urgency to validating HER2-targeted strategies.
  • The result is null (DDR alterations) or hypothesis-generating (VAF, TERT, PIK3CA, ERBB2): the authors call for further validation in independent cohorts before clinical use.

Limitations & open questions

  • DDR analysis underpowered. Only 27 (13%) cases had oncogenic DDR alterations and only 9 (4.5%) had ERCC2 alterations — well below the 11% TCGA muscle-invasive rate. Authors flag completion of randomized trials such as NCT03609216 as crucial to definitively clarify DDR-gene predictive value.
  • Selection bias. Only 212/506 (42%) of CALGB 90601 patients consented to blood collection with sufficient plasma; authors argue baseline characteristics were similar to the parent cohort but acknowledge potential bias.
  • Mixed treatment. Half of the cfDNA cohort received bevacizumab; pooling both arms relies on the parent trial’s null bevacizumab effect. Stratified analyses were consistent.
  • Panel coverage. MSK-ACCESS interrogates 129 genes — no coverage of all DDR genes, no epigenomic or transcriptional features. Authors note panel was selected from >25,000 solid tumors to capture the most recurrent oncogenic and clinically actionable alterations.
  • TERT-VAF collinearity. TERT alterations could not be co-modeled with ³√VAF₇₅ₚc because of strong correlation (p < 0.001); the independent contribution of TERT vs overall ctDNA burden cannot be cleanly separated.
  • External validation. The KEYNOTE-361 cfDNA analysis (n = 125) did not find a statistically significant VAF–outcome association after platinum-based chemo, although the result was directionally consistent. Larger replication is needed.

Citations from this paper used in the wiki

  • “Baseline cfDNA was sequenced from 201 patients with mUC. There was no statistically significant association between alterations in DNA damage response (DDR) genes and response to cisplatin-based chemotherapy (12/24; 50% response rate in DDR+ vs 60/145; 41% response rate in DDR−; p = 0.4)…” (Abstract, p. 80)
  • “Higher variant allele frequency (VAF) in circulating tumor (ct)DNA was associated with shorter OS (HR 2.51, 95% CI 1.26–5.00; p = 0.009) and PFS (HR 2.18, 95% CI 1.02–4.67; p = 0.045).” (Abstract, p. 80–81)
  • “Shorter OS was associated with cfDNA alterations in TERT (HR 1.59, 95% CI 1.15–2.19; p = 0.005), PIK3CA (HR 1.91, 95% CI 1.20–3.04; p = 0.006), and ERBB2 (HR 1.64, 95% CI 1.08–2.49; p = 0.019).” (Abstract, p. 81)
  • “Baseline plasma samples and matched germline DNA were analyzed using MSK-ACCESS, a tumor-uninformed platform that examines mutations and copy-number alterations in 129 genes… depth of coverage of >15,000× that allows a detection threshold of 0.1% for allele frequency.” (Methods §2.2, p. 81)
  • “Oncogenic DDR gene alterations were identified in pretreatment cfDNA in 27 (13%) cases, including 11 ATM (5.5%); nine ERCC2 (4.5%); six BRCA1 (3%) and two BRCA2 (1%) alterations.” (Results §3.2, p. 82)
  • “Alterations in TERT (primarily the TERT promoter) were most frequent; FGFR3 alterations were detected in 19% of patients…” (Results §3.1, p. 82)
  • “While 57.1% (n = 24) of DDR gene alterations were detected in both cfDNA and matched tumor tissue, 21.4% (n = 9) were detected in tumor alone, and 21.4% (n = 9) in cfDNA alone.” (Results §3.1, p. 82)
  • “The study data are available via cBioPortal.” (Data availability §2.4, p. 87)

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