Urinary Cell-Free DNA Predicts Measurable Residual Disease in High-Risk Bladder Cancer With High Accuracy

In a study presented at the 2024 American Society of Clinical Oncology Annual Meeting, Joshua Linscott, MD, of the H. Lee Moffitt Cancer Center and Research Institute in Tampa, Florida, and colleagues demonstrated that genomic alterations and copy number variations detected in urinary cell-free tumor DNA (utDNA) can accurately predict measurable residual disease (MRD) in patients with high-risk non-muscle invasive bladder cancer (HR-NMIBC). This approach has the potential to improve preoperative assessments and treatment planning.

The study involved 53 patients with HR-NMIBC, who were enrolled prior to undergoing repeat transurethral resection of bladder tumors (rTURBT). The researchers employed PredicineWES+ whole-exome sequencing, which includes enhanced coverage of over 600 oncogenes, to analyze tissue samples from the initial transurethral resection (iTURBT). Urine samples were then collected from the patients immediately before their rTURBT procedures.

To detect utDNA and estimate the copy number burden (CNB), low-pass whole-genome sequencing was used. Tumor fraction (tf) in the utDNA was measured with PredicineBEACON, which uses ultra-deep sequencing MRD probes designed from somatic variants identified in the iTURBT tissue. These MRD probes incorporate up to 60 patient-specific alterations alongside a fixed 500-gene hotspot panel.

The study’s primary objective was to evaluate the efficacy of CNB and tf in detecting residual cancer, based on the pathological findings from rTURBT. Analysis included receiver operating characteristic (ROC) curves to establish optimal thresholds for test positivity, which were then applied to assess the performance characteristics of CNB and tf both individually and in a combined model.

Pathologic evaluations of the rTURBT samples revealed that 36 of 53 patients (68%) had detectable residual disease. The CNB score from preoperative urine samples detected residual disease with an area under the curve (AUC) of 0.86. A CNB threshold greater than 6.10 yielded a positive likelihood ratio of 11.70 and a negative likelihood ratio of 0.37 for detecting MRD.

For patients with personalized urinary tf (n=34), the ROC analysis showed an AUC of 0.89. The median urinary tf was significantly higher in patients with MRD (3.50%) compared with those who were disease-free (less than 0.05%, undetectable; P<.001). The optimal thresholds for MRD detection, according to Youden’s index, were tf ≥6.8% and CNB >6.1.

When these 2 assays were combined into a step-wise model, the combined approach achieved a sensitivity of 85.7%, a specificity of 100%, and an overall accuracy of 91.4%. These results suggest that using both CNB and tf from utDNA can provide a highly accurate method for detecting MRD before rTURBT.

The study concludes that MRD can be detected using utDNA with high accuracy prior to rTURBT, presenting a promising urinary biomarker.