Urine tumour DNA found BCG failures a median four months before cystoscopy did

Design

multicentre validation cohort with Kaplan-Meier and Cox analyses

Population

57 patients receiving intravesical BCG for high-risk non-muscle-invasive bladder cancer, with pre- and post-treatment urine

Primary outcome

12-month recurrence-free survival by urine tumour DNA status

Effect

pre-BCG positive 61% against 85% (HR 3.2, P=0.04); post-BCG positive despite clinical negativity 25% against 91% (HR 10.0, P<0.001); median lead time 4.1 months

High-risk non-muscle-invasive bladder cancer recurs within the first year in a substantial minority despite adequate BCG, and cystoscopic surveillance detects that only once it is visible. This multicentre study analysed urine from 57 patients before and after BCG using a tumour-naïve urine tumour DNA assay.

Before the first instillation, 40% (23 of 57) were positive. Their 12-month recurrence-free survival was 61% against 85% in negative patients (hazard ratio 3.2, P=0.04). The post-treatment result is stronger: among 43 patients who were clinically negative after BCG, 19% (eight patients) were urine tumour DNA positive, and their 12-month recurrence-free survival was 25% against 91% for the negative patients (hazard ratio 10.0, P<0.001). Of all future recurrences, a third were flagged by urine tumour DNA while clinical surveillance was still negative, with a median lead time of 4.1 months. Two-thirds of those lead-time cases went on to radical cystectomy and a third progressed to muscle-invasive disease.

Fifty-seven patients is small, and eight positives after BCG is a very thin basis for a hazard ratio of 10. There is also no evidence yet that acting earlier on a molecular signal improves anything - the lead-time cases progressed, which shows the test detects real disease but not that detecting it sooner changed the outcome. That is the trial that needs to happen. Until it does, this is a promising assay, not a surveillance strategy, and in Indian practice the cost question would arrive long before the evidence question.

• Continue cystoscopic surveillance as the standard after BCG - urine tumour DNA has not been shown to improve outcomes
• Note the group with the strongest signal: clinically negative after BCG but urine tumour DNA positive, where three-quarters recurred within a year
• Treat a positive result obtained elsewhere as a reason for closer surveillance, not as an indication for cystectomy
• Bear in mind the assay is tumour-naïve, so it does not require the original tumour tissue - that is its practical advantage
• Watch for a trial testing whether earlier intervention on a molecular signal changes progression

Why it matters

It identifies the BCG failures while cystoscopy still looks clear, which is the point at which intensification would have to happen to be worth anything.

Don't overread it

Fifty-seven patients, eight events in the key subgroup, and no evidence that earlier detection changed any outcome.

The statistics, in plain English

A hazard ratio of 10.0 sounds definitive and rests on eight patients in the positive group - with numbers that small, the point estimate is unstable even when the P value is convincing, and no confidence interval is reported to bound it. The pre-treatment result (hazard ratio 3.2, P=0.04) is the more modest and more trustworthy finding. The 4.1-month lead time is a genuine measurement, but lead time on its own can create the appearance of benefit without any: knowing sooner only helps if something useful can be done in the interval.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.