ASCO 2020: Association of Molecular Subtypes with Pathologic Response Interface to Studying of Coexpression Extrapolation (COXEN) with Neoadjuvant Chemotherapy for Localized, Muscle Invasive Bladder Cancer

(UroToday.com) In this poster presented at ASCO 2020, the authors used the COXEN and gene expression model with chemotherapy specific scores to identify responders to neoadjuvant chemotherapy. In a prospective randomized phase two clinical trial (SWOG S1314), the gemcitabine/cisplatin (GC) score was associated with pathology downstaging in the pooled arms. The authors investigated RNA based molecular subtypes as additional predictive biomarkers for response to neoadjuvant chemotherapy in the SWOG S1314 trial.

Eligible patients included those with clinical stage T2-T4a N0 M0, with predominant urothelial tumors larger than 5 millimeters, who were cisplatin eligible and planned for radical cystectomy and pelvic lymph node dissection.

A total of 237 patients were randomized to either 4 cycles of dose-dense combination of methotrexate, vinblastine, doxorubicin, and cisplatin (ddMVAC) vs. GC. Based on Affymetrix transcriptomic data used to assign COXEN scores, the authors determined subtypes for a subset of evaluable patients (n=161) using three classifiers TCGA (k=5), Consensus (K=6), and MD Anderson (MDA; k=3).

The primary objective of this study was to assess the subtype association with pathologic response to neoadjuvant chemotherapy in the pooled arms and determine any association with COXEN. The authors also tested whether each classifier contributed additional predictive power when added to a model based on predefined stratification factors (performance status 0 vs. 1; disease stage T2 vs. >=T3).

The study scheme and details of the classifiers used are shown in Figure 1. Table 1 depicts the collapsed classifier groups and pathologic response outcomes with pooled treatment arms. Importantly, 78% of patients had a performance status of 0; 89% of patients had T2 disease, with 51% of patients receiving ddMVAC, and 49% of patients receiving GC. A pathological complete response was seen in 33% of patients with 52% experiencing downstaging.

Figure 1 – Study scheme and classifiers used:

Table 1 – Collapsed classifier groups and pathologic response outcomes, pooled treatment arms:

Table 2 demonstrates the various modeling done for complete pathological response (pT0), with the best area under the curve seen with the GC COXEN combined with the stratification factors (performance status 0 vs. 1, T2 vs. >=T3). This was very close to the TCGA3 group predictor. In contrast, the CONSENUS, MD ANDERSON, and MVAC COXEN were like the stratification factors used alone. Lastly, the ROC curves for pathologic complete response and downstaging for each classifier are shown in Figure 2.

Table 2 – Modelling pT0:

Figure 2 – ROC curves for pT0 and downstaging:

The authors concluded that the CONSENSUS classifier, which is based in part on the TCGA and MDA classifiers, modestly improved prediction for pathologic downstaging, when added to the clinical stage and performance status of patients. The authors hope that with additional follow up, they will be able to assess the association of the COXEN scores and subtypes with overall survival in the near future.

Presented by: Seth Lerner, MD, FACS, Baylor College of Medicine, Houston, TX

Written By: Hanan Goldberg, MD, MSc., Urology Department, SUNY Upstate Medical University, Syracuse, NY, USA, Twitter: @GoldbergHanan at the 2020 ASCO Annual Meeting, Virtual Scientific Program #ASCO20, May 29-31, 2020

Related Content:

Login