Detectable circulating tumour DNA in non-muscle-invasive bladder cancer marks disease that is not really superficial

Design

Prospective cohort with serial tumour-informed circulating tumour DNA testing and tumour-normal whole-exome sequencing in a subset

Population

52 patients with high-risk non-muscle-invasive bladder cancer; 29 BCG-naive, 23 BCG-exposed or unresponsive; 21 underwent cystectomy

Primary outcome

Clinical and pathological upstaging and recurrence by ctDNA status

Effect

ctDNA detectable in 17/52 (33%); clinical upstaging 59% vs 14% (p=0.002); pathological upstaging at cystectomy 78% vs 8% (p=0.002); distant recurrence 3/17 vs 0/35

Non-muscle-invasive bladder cancer is defined by what the cystoscope and the pathologist can see, and a proportion of it behaves as though it is already systemic. This cohort tested whether tumour-informed circulating tumour DNA identifies that group, following 52 patients with high-risk disease using a personalised assay at presentation and at about three-monthly intervals through routine treatment.

Circulating tumour DNA was detectable in 17 of 52 (33%) - 28% of BCG-naive and 39% of BCG-exposed or unresponsive patients. Clinical upstaging occurred in 10 of 17 positive patients against 5 of 35 negative (59% vs 14%, p=0.002). Among the 21 who went to cystectomy, pathological upstaging occurred in 7 of 9 positive against 1 of 12 negative (78% vs 8%, p=0.002). Distant recurrence occurred in three positive patients and none of the negative. TP53 alterations were enriched in positive tumours and FGFR3 alterations in negative ones - a genomic pattern consistent with the clinical one.

The striking part is that this often happened without cystoscopic or radiographic evidence, which is exactly the failure mode of surveillance based on looking. The limitation is the cohort: 52 patients, 17 positive, and the cystectomy comparison rests on 21 people, so the 78% versus 8% figure is built on single-digit counts.

The clinical hypothesis this generates is worth stating because it is testable and consequential: a ctDNA-positive patient may be one who will not benefit from continued intravesical salvage therapy and should be considered for cystectomy sooner. Nobody should act on that yet - the assay is tumour-informed, expensive and not available in routine practice - but it is the question the prospective trial needs to ask.

• Do not order tumour-informed ctDNA in bladder cancer outside a study - this is 52 patients
• Recognise that cystoscopic surveillance misses biologically advanced disease; keep imaging thresholds low in high-risk patients
• TP53 alteration on tumour sequencing is the genomic correlate worth noting where sequencing is done
• Reconsider prolonged intravesical salvage in a patient whose disease keeps recurring after BCG failure
• The cystectomy comparison rests on single-digit patient counts in each cell

The statistics, in plain English

59% versus 14% with p=0.002 is a large difference, but it comes from 10 events among 17 patients and 5 among 35 - small enough that a handful of reclassifications would change it. The cystectomy figures, 7 of 9 against 1 of 12, are more striking and rest on even fewer patients, and they are also subject to selection: who went to cystectomy was decided clinically, possibly influenced by the same features that predict upstaging. Distant recurrence in 3 of 17 versus 0 of 35 has no p value reported and should be read as a signal rather than a result.

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.