Clinical decision thresholds for urine‐based fibroblast growth factor receptor (FGFR) testing across the bladder cancer care continuum: From molecular classification to treatment selection

Background Fibroblast growth factor receptor (FGFR) alterations are clinically actionable in urothelial cancer, yet tissue‐based molecular testing may fail when specimens are unavailable, depleted, aged, low in tumor content, or analytically inadequate. Urine molecular testing has the potential to add value throughout the patient journey by identifying FGFR alterations when tissue testing is unsuccessful; however, published urine‐testing evidence comes from different assays, populations, and clinical settings, and therefore cannot establish the expected performance or clinical utility of a single integrated FGFR urine pathway. Methods A deterministic threshold‐and‐scenario model followed 1000 patients from tissue FGFR testing through reflex urine testing after an uninformative tissue result, treatment selection, and a 1‐year surveillance analogue. ANNAR tissue validity and positivity, and TAR‐210 urine yield and positivity, informed the reflex testing scenario as transported inputs; GALEAS Bladder sensitivity and specificity informed a separately labeled external cross‐assay calibration that does not validate the intended assay. A one‐time $2919.60 Current Procedural Terminology coding‐scale price proxy was applied after tissue failure. The model calculated the minimum discounted quality‐adjusted life‐year (QALY) gain required per additionally treated true‐positive patient at $100 000/QALY; it did not estimate comparative effectiveness or an incremental cost‐effectiveness ratio. The surveillance stage hypothetically compared three annual cystoscopy‐based encounters with one retained safety cystoscopy and two urine‐monitoring intervals, using an 18.2% confirmatory‐cystoscopy trigger rate borrowed from a non‐FGFR marker trial and a provisional $597.91 price proxy per interval. The primary perspective was the US payer; patient and limited societal consequences were reported separately. Results Under transported TAR‐210 inputs, reflex urine FGFR testing increased scenario‐projected molecular‐classification coverage from 66.0% to 85.7%, adding 197.2 informative results and 82.8 projected FGFR‐positive results per 1000 patients. A separate GALEAS‐based cross‐study calibration estimated 80.3 true‐positive classifications; this was not an observed clinical yield for the intended assay. With 80% clinical acceptance and 80% treatment uptake, 51.4 additional cross‐study‐estimated true‐positive patients would be treated, and the required gain was 0.193 QALY, approximately 70 quality‐adjusted life‐days, per treated patient; full acceptance and uptake yielded an optimistic lower‐bound threshold of 0.124 QALY. In the hypothetical surveillance analogue, the borrowed 18.2% trigger reduced cystoscopies from 3.000 to 1.364 per patient‐year, avoiding 1.636 (54.5%). Payer break‐even required a complete displaced encounter cost of $730.94 before implementation costs; this exceeded the $281.24 office cystoscopy‐plus‐cytology estimate, so the analogue remained an incremental payer investment in that setting. Conclusion The model illustrates two conditional value mechanisms for FGFR‐focused urine testing: recovery of actionable FGFR classification when tissue testing is uninformative and a separate hypothetical reduction of selected invasive surveillance encounters. Economic favorability is conditional on intended‐assay performance, clinical acceptance, treatment uptake, complete episode costs, implementation costs, and retained oncologic safety.