Circulating tumor DNA is emerging as a promising biomarker for risk stratification in muscle-invasive bladder cancer, but its prognostic performance in tumors with histological subtypes has remained less clear.
On August 12, 2026, European Urology Focus published the study titled “Prognostic Value of Perioperative Circulating Tumor DNA in Pure Urothelial Carcinoma and Histological Subtypes of Bladder Cancer.”
Authors: Pierre-Etienne Gabriel, Alec Zhu, Christopher Karim, Ahmed Eraky, Raghav Gupta, Reuben Ben-David, Peter Wiklund, Reza Mehrazin, and John P. Sfakianos.
Evaluating ctDNA Across Different Bladder Cancer Histologies
The retrospective study analyzed a prospectively maintained database of consecutive patients with nonmetastatic muscle-invasive bladder cancer who underwent robot-assisted radical cystectomy at Mount Sinai Hospital between November 2021 and May 2025. All patients had at least one perioperative tumor-informed ctDNA assessment performed as part of routine clinical practice using the Signatera 16-plex multiplex PCR next-generation sequencing assay.
Histology was classified as either pure urothelial carcinoma or a histological subtype. A histological subtype was defined by the presence of a nonurothelial component representing at least 5% of the tumor in the cystectomy specimen. Overall, 159 patients were included, of whom 92 (58%) had pure urothelial carcinoma and 67 (42%) had histological subtypes.
Preoperative ctDNA data were available for 138 patients and postoperative ctDNA data for 144 patients. Preoperative samples were obtained within 1 month before or on the day of radical cystectomy, at a median of 1 day before surgery. Postoperative ctDNA was assessed using the first sample collected within 90 days after surgery, at a median of 32 days. Patients were evaluated according to both histology and ctDNA detectability, with disease-free survival as the outcome.
Histological Subtypes Were Associated With More Aggressive Tumor Features
Patients with histological subtypes were more likely to have detectable preoperative ctDNA than those with pure urothelial carcinoma, at 61% versus 40%, respectively.
They also had more advanced pathological features. Pathological stage ≥pT2 was observed in 90% of patients with histological subtypes compared with 57% of those with pure urothelial carcinoma, while lymph node involvement was present in 39% versus 23%, respectively.
Adjuvant therapy was also used more frequently in patients with histological subtypes, at 33% compared with 19% in the pure urothelial carcinoma group. The histological subtype cohort included micropapillary, squamous, glandular, plasmacytoid, sarcomatoid, nested, small cell, and other less common tumor subtypes.
Preoperative ctDNA and Disease-Free Survival
Among the 138 patients with available preoperative ctDNA results, 45 had undetectable ctDNA and pure urothelial carcinoma, 37 had detectable ctDNA and pure urothelial carcinoma, 15 had undetectable ctDNA and histological subtypes, and 41 had detectable ctDNA and histological subtypes.
After a median follow-up of 21 months, 46 patients experienced disease recurrence or death. In patients with pure urothelial carcinoma, 24-month disease-free survival was 89% among those with undetectable preoperative ctDNA compared with 49% among those with detectable ctDNA (p<0.001).
A similar pattern was observed in patients with histological subtypes. Twenty-four-month disease-free survival was 79% with undetectable ctDNA versus 43% with detectable ctDNA (p=0.012). After multivariable adjustment, detectable preoperative ctDNA remained independently associated with recurrence or death in both histological groups.
For patients with pure urothelial carcinoma, the hazard ratio was 4.07 (95% CI 1.29–12.81; p=0.020). For patients with histological subtypes, the hazard ratio was 3.99 (95% CI 1.24–12.82; p=0.023). There were no significant differences in disease-free survival between pure urothelial carcinoma and histological subtypes among either the ctDNA-detectable or ctDNA-undetectable populations.
Postoperative ctDNA Showed a Similar Prognostic Pattern
Postoperative ctDNA results were available for 144 patients. Among them, 66 had undetectable ctDNA and pure urothelial carcinoma, 15 had detectable ctDNA and pure urothelial carcinoma, 43 had undetectable ctDNA and histological subtypes, and 20 had detectable ctDNA and histological subtypes.
After a median follow-up of 22 months, 49 patients had experienced disease recurrence or death. In the pure urothelial carcinoma group, 24-month disease-free survival was 81% in patients with undetectable postoperative ctDNA compared with 30% in those with detectable ctDNA (p<0.001). Among patients with histological subtypes, the corresponding rates were 66% and 29% (p<0.001).
Detectable postoperative ctDNA remained independently associated with disease recurrence or death after multivariable adjustment. The hazard ratio was 4.38 (95% CI 1.66–11.52; p<0.01) in patients with pure urothelial carcinoma and 4.52 (95% CI 1.86–11.03; p<0.001) in those with histological subtypes.
Again, disease-free survival did not differ significantly between pure urothelial carcinoma and histological subtypes within either the ctDNA-detectable or ctDNA-undetectable populations.
What Do These Findings Mean?
The results suggest that the prognostic value of perioperative tumor-informed ctDNA is preserved across the histological groups evaluated in this study. Both before and after radical cystectomy, detectable ctDNA identified patients with a substantially higher risk of recurrence or death, with adjusted hazard ratios of approximately four in both pure urothelial carcinoma and tumors with histological subtypes.
The authors noted that this may be particularly relevant for patients with histological subtypes, which can have distinct biological characteristics and are often associated with more aggressive pathological features. The findings also add to growing evidence supporting postoperative ctDNA as a marker of minimal residual disease and a potential tool for identifying patients at particularly high risk after radical cystectomy.
However, the authors emphasized that further prospective validation is required to determine how ctDNA-guided strategies can be optimally implemented in clinical practice and whether they can improve clinical outcomes.
Study Limitations
Several limitations should be considered. The analysis was retrospective and observational, creating the possibility of selection bias. Perioperative ctDNA testing was performed at nonstandardized time points, which may have introduced measurement variability.
Histological subtype classification was based on surgical specimens reviewed by different pathologists, and biologically distinct histological subtypes were grouped together for analysis. The limited number of patients with histological subtypes in certain subgroups (<20 patients) also reduced the precision of hazard ratio estimates and limited the interpretation and generalizability of the results.
The authors additionally noted the possibility of bias related to postoperative ctDNA assessment because samples were collected after radical cystectomy, although no events occurred between surgery and the first postoperative ctDNA measurement. Follow-up was also relatively short.
Conclusion
In patients undergoing radical cystectomy for nonmetastatic muscle-invasive bladder cancer, detectable tumor-informed ctDNA before or after surgery was independently associated with a higher risk of disease recurrence or death in both pure urothelial carcinoma and tumors with histological subtypes.
The consistency of these associations across histological groups supports the potential role of perioperative ctDNA in refining perioperative risk stratification. Further prospective validation will be needed to establish how ctDNA-guided management should be implemented and whether it can improve clinical outcomes.
The full article is available in European Urology Focus.
Find more videos on OncoDaily YouTube TV.