Molecular subtyping reveals luminal bladder tumors have lower rates of pathological upstaging at radical cystectomy

Lotan Y. ¹,
Boorjian S.A. ²,
Zhang J. ³,
Bivalacqua T.J. ⁴,
Porten S.P. ⁵,
Lerner S.P. ⁶,
Wheeler T.M. ⁷,
Hutchinson R. ¹,
Francis F. ⁸,
Du Plessis M. ³,
Davicioni E. ¹²,
Svatek R. ¹⁰,
Black P. ¹¹,
Gibb E.A. ⁹

¹ University of Texas Southwestern Medical Center, Dept. of Urology, Dallas, United States of America ² Mayo Clinic, Dept. of Urology, Rochester, United States of America ³ GenomeDx Inc., Clinical Development, Vancouver, Canada ⁴ John Hopkins Medical Institute, Dept. of Urology, Baltimore, United States of America ⁵ University of California San Francisco, Dept. of Urology, San Francisco, United States of America ⁶ Baylor College of Medicine Medical Center, Dept. of Urology, Houston, United States of America ⁷ Baylor College of Medicine Medical Center, Dept. of Pathology and Immunology, Houston, United States of America ⁸ University of Texas Southwestern Medical Center, Dept. of Pathology, Dallas, United States of America ⁹ GenomeDX, R&D, San Diego, United States of America ¹⁰ University of Texas Health San Antonio, Dept. of Urology, San Antonio, United States of America ¹¹ University of British Columbia, Dept. of Urological Sciences, Vancouver, Canada ¹² GenomeDx Inc., R&D, Vancouver, Canada

Abstract

Publication: March 2019

Introduction & Objectives

Upstaging of bladder cancer from precystectomy clinical stage ≤T2 to non-organ confined pathologic stage ≥T3 or N+ at radical cystectomy (RC) is common and can significantly increase the risk of cancer-specific mortality (CSM). We assessed the ability of a genomic subtyping classifier (GSC) to predict pathological upstaging in a multi-institutional cohort of patients with clinical T1-T2 bladder cancer treated with RC.

Materials & Methods

A cohort of 206 patients with clinical high-grade, organ-confined (cT1-T2, N0M0) treated with RC and bilateral PLND without neoadjuvant chemotherapy (NAC) was selected from the registries of 7 academic hospitals. Precystectomy bladder tumor transurethral resection FFPE specimens were submitted for genomic testing with the Decipher Bladder assay (GenomeDx, San Diego, CA). Uni- and multi-variable logistic regression analysis (UVA/MVA) were used to evaluate a locked GSC for upstaging, defined as pT3/T4 and/or pTanyN1-3 disease at RC. Cumulative incidence of CSM was calculated using Fine-Gray competing risks.

Results

Upstaging to non-organ confined disease (≥T3 and/or ≥N1) occurred in 23% of cT1 and 57% of cT2 cases, respectively. Stratifying for clinical stage, fewer patients with luminal tumors were upstaged to ≥pT3 compared to non-luminal tumors (M-H p-value = 0.002; cT1: 13% vs 34%, cT2: 34% vs 58%, respectively). On UVA and MVA, non-luminal patients were significantly more likely to upstage (≥pT3) at RC compared to luminal patients (p<0.001 for both). Conversely, rates of upstaging to ≥N1 disease were similar when considering clinical stage and subtype. Rates of upstaging to ≥N1 at cT1 were 13% for both luminal and non-luminal patients (p>0.9). Similarly, for cT2, upstaging to ≥N1 was 15% for luminal and 23% for non-luminal patients (p=0.56). Luminal patients also trended toward having better prognosis as determined by lower CSM rates compared to non-luminal patients (p=0.061).

Conclusions

Molecular subtyping revealed luminal tumors have lower rates of upstaging compared to non-luminal tumors. Combined with previous reports that the benefit of NAC may be greatest in basal tumors, these data suggest that subtyping may have utility in identifying higher risk patients who could be prioritized for NAC. Nevertheless, approximately one-third of luminal tumors with cT2 tumors are upstaged to non-organ confined disease. These patients should continue to receive NAC until further studies evaluate the implications of immediate cystectomy.