Advice

Clinical and technical evidence

Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement for medtech innovation briefings. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting mibs@nice.org.uk.

Published evidence

Five studies are summarised in this briefing, including 1 systematic review and network meta-analysis, 3 prospective studies and 1 retrospective cohort study. The evidence base for Bladder EpiCheck presented in this briefing included 3,064 people having surveillance after treatment for non‑muscle‑invasive bladder cancer (NMIBC).

Another study was identified in which Bladder EpiCheck was used to detect disease in people with upper urinary tract carcinoma before having a radical nephroureterectomy (Pierconti et al. 2021a). The study is not presented in detail here because the population was out of scope for this briefing.

The clinical evidence and its strengths and limitations are summarised in the overall assessment of the evidence.

Overall assessment of the evidence

Studies were done in appropriate populations and the sensitivity and specificity of the test were compared to an appropriate reference standard (cystoscopy, cytology and histology).

Across the studies, the overall sensitivity of the test ranged from 62% to 90% and its specificity ranged from and 82% to 88%. The high-grade sensitivity of the test ranged from 79% to 100%, and the high-grade specificity ranged from 85% to 91%.

Bladder EpiCheck's pooled high-grade negative predictive value (NPV) was reported as 98%, and the overall pooled NPV was reported as 94% (Laukhtina et al. 2021), which is higher than the standard care methods of cystoscopy and cytology. Also, based on the pooled specificity reported, using Bladder EpiCheck could result in fewer false-positive results. It may be possible that the test is detecting early epigenetic changes in pre-cancerous cells, but future studies with longer-term follow up would be needed to correlate false-positive results with later recurrence. Many of the studies were limited by their single-visit design or did not clearly report the length of follow up. The longest duration of follow up reported was a median of 12 months (ranging between 9 and 15 months; Pierconti et al. 2021b).

Many of the studies did not clearly report whether the urologists and pathologists were blinded to Bladder EpiCheck results; not blinding would increase the risk of bias. There is currently no published evidence on the technology being used in the UK or the NHS. Most of the studies evaluated the diagnostic accuracy of the test only. Larger studies with longer-term follow up would be helpful, preferably done in an NHS setting. In addition to diagnostic accuracy outcomes, studies could also evaluate the downstream consequences of using the test, such as the impact on clinical outcomes and resource use, including costs associated with false test results and inappropriate treatment.

Laukhtina et al. (2021)

Interventions and comparator

The systematic review and meta-analysis included studies that assessed the diagnostic accuracy of urinary biomarker tests (Xpert Bladder Cancer, Bladder EpiCheck, ADXBLADDER, Uromonitor, Cxbladder monitor) compared with a reference standard of cystoscopy or histopathology.

Key outcomes

The meta-analysis included 21 studies, with a reported total of 7,330 people. This included 10 studies for Xpert Bladder (2,806 people), 5 for Bladder EpiCheck (1,684 people), 3 for ADXBLADDER (2,053 people) and 2 each for Uromonitor (262 people) and Cxbladder Monitor (1,112 people). Overall, the tests showed sensitivities of up to 93%, specificities of up to 84%, positive predictive values (PPVs) of up to 67%, and NPVs of up to 99%. The detection of high-grade recurrence showed similar diagnostic accuracy compared with that of any-grade recurrence for the tests evaluated (Expert bladder, Bladder EpiCheck and ADXBLADDER). The pooled results for Bladder EpiCheck specifically were as follows: sensitivity 74%, specificity 84%, PPV 48% and NPV 94%. For high-grade recurrence, the pooled results for Bladder EpiCheck were: sensitivity 91%, specificity 81%, PPV 43% and NPV 98%. The network meta-analysis (based on 13 of the studies) showed that most of the diagnostic values of the tests (except for specificity) were significantly higher than those of cytology for detecting recurrence. The authors concluded that the high diagnostic accuracy of the studied novel urinary biomarkers supports their utility in the NMIBC surveillance setting. They noted that all of these have the potential to help prevent unnecessary cystoscopies safely in the NMIBC surveillance population.

Strengths and limitations

Strengths were as follows. The study was a meta-analysis including a range of studies of urine biomarker tests. Study selection and data extraction were done by 3 reviewers independently. Discrepancies were resolved by referring to the senior author (study selection) or by consensus with the co-authors (data extraction). Risk of bias for included studies was evaluated using a validated tool (the Quality Assessment of Diagnostic Accuracy Studies [QUADAS-2] tool).

Weaknesses were as follows. The included studies were heterogeneous in terms of patient population, reference standards used and the prevalence of recurrence rate. But the study used a random-effect model to account for heterogeneity across the studies. Some of the studies did not report data on blinding and most of the studies did not include cut-off values for the tests; however, this is not relevant for Bladder EpiCheck because it has a set cut-off. The protocol used for cystoscopy follow up was not reported in most of the included studies. Not all biomarkers in the meta-analysis had multicentre prospective studies, but Bladder EpiCheck did.

Cochetti et al. (2021)

Intervention and comparator

The diagnostic performance of Bladder EpiCheck, photodynamic diagnosis (PDD)-guided cystoscopy and urinary cytology were compared with a histological diagnosis.

Key outcomes

Bladder EpiCheck had an area under the receiver operator curve (AUROC) of 0.95. The high-grade results for detecting recurrence were as follows: sensitivity 100%, specificity 90.9%, PPV 90% and NPV 100%. There was good agreement between diagnosis using Bladder EpiCheck and histological findings (Cohen's kappa test score of 0.9, p<0.001). Test performance was not affected by haematuria or turbid urine (2 and 5 people, respectively). The type of instillation therapy received did not significantly impact the diagnostic performance of the test. The sensitivity of the test was 100% for people who had either Bacillus Calmette–Guérin (BCG) or mitomycin C instillation, but its specificity was slightly lower in people who had BCG therapy than in people who had mitomycin C (92.9% compared with 87.5%). PDD had an area under the curve of 0.51. For detecting recurrence, overall sensitivity was 61%, and high-grade results were as follows: sensitivity 29%, specificity 41%, PPV 46% and NPV 56%.

Strengths and limitations

Strengths: The study had a prospective design and enrolled people consecutively. The Bladder EpiCheck results were not used in clinical practice, so the investigators were blinded to the findings. Bladder EpiCheck was tested in an important subset of people with NMIBC, those who had intravesical instillations, which might affect the sensitivity and specificity of a urine biomarker. Bladder EpiCheck was compared with another established diagnostic test, PDD-guided cystoscopy.

Weaknesses: The study was done in a single centre and had a small sample size. It was also limited by the single-visit design. The sub-analysis of intravesical treatments may have been underpowered.

Pierconti et al. (2021b)

Study size, design and location

Prospective single-centre cohort study in 205 people with high-grade NMIBC who had treatment with intravesical BCG or mitomycin C therapy, in Italy. During follow up, 151 people had a recurrence, which were all high-grade; 54 had negative cytology, cystoscopy and histology results and were considered to not have a recurrence.

Intervention and comparator

Bladder EpiCheck tests results were compared with urine cytology using histological diagnosis as a reference standard. Results were compared at early follow up (within the 3 months after intravesical therapy) or later (more than 3 months after intravesical therapy).

Key outcomes

In early follow up, the high-grade results for Bladder EpiCheck were as follows: sensitivity 92.1%, specificity 85.1%, PPV 77.8% and NPV 95.0%. For cytology, the respective results were 85.0%, 86.3%, 72.3% and 93.2%. In later follow up, the high-grade results for Bladder EpiCheck were as follows: sensitivity 96.8%, specificity 78.0%, PPV 60.0% and NPV 98.6%. For cytology, the respective results were 77.4%, 85.7%, 64.9% and 91.8%. The sensitivity and specificity of Bladder EpiCheck in people with papillary disease at early follow up were as follows: sensitivity 76.9% and specificity 96.3%. For cytology, the respective results were 73.3% and 90.4%. In people with carcinoma in situ the sensitivity and specificity of Bladder EpiCheck at early follow up were as follows: sensitivity 100% and specificity 80.9%. For cytology, the respective results were 92.0% and 81.4%, respectively. The AUROC was 99.5% for Bladder EpiCheck and 85.5% for cytology. The sensitivity of Bladder EpiCheck was always higher than that of cytology during the whole follow-up period both for papillary NMIBC and carcinoma in situ.

Strengths and limitations

Strengths: The study was prospective in design and enrolled people consecutively. People in the study had BCG or mitomycin C, and results were compared with cytology. Samples were evaluated by 2 expert cytopathologists with more than 10 years' experience. When a consensus could not be reached, a third uropathologist expert was consulted. The median follow up was 12 months (range 9 to 15 months).

Weaknesses: The study included a relatively small number of people from a single centre. It did not include a multivariate analysis or a comparison with low-grade cancer. The study did not compare Bladder EpiCheck with cystoscopy.

Pierconti et al. (2021c)

Intervention and comparator

Bladder EpiCheck test results were compared with urine cytology using histological diagnosis as a reference standard. Bladder EpiCheck test results were assessed for correlation with the different categories of The Paris System for Reporting Urinary Cytology (TPS).

Key outcomes

Results showed that EpiScore increased in TPS categories from negative for high-grade urothelial carcinoma (NHGUC) to high-grade urothelial carcinoma (HGUC) TPS categories. When cytological categories of NHGUC and atypical urothelial cells were compared, an EpiScore of less than 60 correlated with NHGUC (p=0.0003, odds ratio 3.925, 95% confidence interval [CI] 1.907 to 8.081, Fisher's exact test). When atypical urothelial cell and suspicious for high-grade urothelial carcinoma (SHGUC) or SHGUC and HGUC categories were compared, an EpiScore of 60 or more correlated with SHGUC (p=0.0031, OR 3.791, 95% CI 1.612 to 8.915) and with HGUC (p=0.0027, OR 3.957, 95% CI 1.639 to 9.550, Fisher's exact test). For NHGUC, the sensitivity of Bladder EpiCheck was 100%, specificity was 89.9%, PPV was 100% and NPV was 5%. For atypical urothelial cells the results were 81.8%, 52.3%, 42.8% and 84.6%, respectively. For SHGUC they were 86.6%, 52.3%, 56.5% and 84.6%, respectively. For HGUC they were 98.8%, 100%, 100% and 85.7%, respectively. The company stated that when comparing Bladder EpiCheck and cytology with the histological reference standard, the high-grade sensitivity for Bladder EpiCheck was 95%, and specificity was 82%, and the high-grade sensitivity for cytology was 90%, and specificity was 84%. The company provided the data comparing Bladder EpiCheck and cytology with the histological reference standard and stated that, although not reported in the study, values could be calculated from the publication.

Strengths and limitations

Strengths: All results were reviewed by at least 2 experienced uropathologists. The cohort focused on the subpopulation of people with a history of high‑grade disease having treatment. Follow up was for 1 year.

Weaknesses: The study was retrospective in design and included people undergoing surveillance at a single centre in Italy. The study did not compare Bladder EpiCheck with cystoscopy.

Witjes et al. (2018)

Intervention and comparator

Bladder EpiCheck tests results were compared with cytology, and cystoscopy results were confirmed by histology.

Key outcomes

Of the 440 people enrolled in the study, 353 were included in the diagnostic performance analysis. Results from 87 people were excluded because of inconclusive diagnosis according to the reference standard (50 people), no Bladder EpiCheck results (30 people), or both (7 people). For Bladder EpiCheck, the overall results were as follows: sensitivity 68.2%, specificity 88.0%, PPV 95.1%, NPV 44.8% and AUROC 82%. When excluding low‑grade recurrence, the high-grade sensitivity was 91.7%, NPV was 99.3% and AUROC was 94%. Recent intravesical instillations did not impact the test's performance.

Strengths and limitations

Strengths: The study was prospective in design and included a relatively good number of people having surveillance at multiple centres in Europe. The investigators were blinded to Bladder EpiCheck test results.

Weaknesses: The study population included only people of European family background, so the results may not be generalisable to a wider, more diverse population. No follow-up data was collected, so false positives could not be correlated with later recurrences. No data was available about the presence of urinary tract infections.

Sustainability

The company did not make any sustainability claims for the technology.

Recent and ongoing studies