Background: Patients with colonic inflammatory bowel disease (IBD) undergo endoscopic surveillance to detect and remove colorectal neoplasia (CRN). Although high-definition dye-based chromo-endoscopy (HD-CE) is the recommended modality in guidelines, high-definition white light endoscopy (HD-WLE) with matched procedural time may yield similar neoplasia detection rates. We designed a non-blinded, randomized controlled trial in four academic hospitals aiming to compare three endoscopic surveillance techniques in IBD patients for the outcome of CRN detection rate.

Methods: Eligible patients were aged ≥18 years and scheduled for colitis-associated CRN surveillance according to Dutch IBD surveillance guidelines. Patients were excluded if surveillance was insufficient (i.e. insufficient bowel cleansing or active colitis), if >50% of the colon was resected or if patients did not receive the intervention. Patients were randomly assigned (2:2:1) to undergo HD-WLE with segmental re-inspection (SR), HD-CE, or single pass HD-WLE. The primary outcome was CRN detection rate, defined as the proportion of procedures in which macroscopic CRN was detected. Secondary outcomes included the number of CRN and withdrawal time. To demonstrate non-inferiority of HD-WLE with SR compared to HD-CE (one-sided testing, α=0.05, 1−β=0.8, non-inferiority (NI) margin -10%), and superiority compared to single-pass HD-WLE (two-sided testing, estimated CRN detection rate of 24% and 12%, respectively¹) with Mantel-Haenszel analyses, a total of 566 patients were needed.

Results: In total, 666 patients were randomized, 563 fulfilled all study criteria and were analysed per protocol (table 1). Of these, 51.8% were male, with a median age of 51 years (interquartile range 35-63). CRN detection rates were 9.8% for HD-WLE with SR, 13.1% for HD-CE, and 6.1% for single pass HD-WLE. HD-WLE with SR was non-inferior compared to HD-CE (Δ-3.1%, lower boundary of the 95% confidence interval (CI) -8.1 not exceeding the NI margin of -10%, p<0.5). HD-WLE with SR was not superior compared to single-pass HD-WLE (Δ3.7%, 95% CI -2.5:9.1%). A significant difference was found between arms for the number of detected CRN (n=29 vs n=36 vs n=8 for HD-WLE with SR, HD-CE and single-pass HD-WLE, respectively (p=.04)) and withdrawal time (p=.03, table 1). One advanced lesion was detected (high-grade dysplasia in the HD-CE arm).

Conclusion: In this large, randomized controlled trial, HD-WLE with SR was non-inferior compared to HD-CE for CRN detection in IBD patients. HD-WLE with SR was not superior to single-pass HD-WLE, although this may have resulted from lower than expected neoplasia yields and subsequent lower power. These results suggest that HD-WLE with SR may provide a feasible alternative to HD-CE in clinical practice (NCT04291976).

Reference: 1: Imperatore et al. JCC 2019;13(6):714-24

<b>Table 1: Characteristics and outcomes of 563 IBD patients included in the HELIOS study (per protocol analysis)</b><br /> 1: Extensive disease based on endoscopic information and classified according to the Montreal classification for UC/IBD-U or expressed in percentages for CD. 2: Risk categories were based on the Dutch surveillance guidelines. 3: 9 missing values. Abbreviations: CRC: colorectal cancer; HD-CE: high-definition dye-based chromoendoscopy; HD-WLE: high-definition white light endoscopy; HGD: high-grade dysplasia; IND: indefinite for dysplasia; IQR=interquartile range; LGD: low-grade dysplasia; n=number; SR: segmental re-inspection; y=years.

Table 1: Characteristics and outcomes of 563 IBD patients included in the HELIOS study (per protocol analysis)
1: Extensive disease based on endoscopic information and classified according to the Montreal classification for UC/IBD-U or expressed in percentages for CD. 2: Risk categories were based on the Dutch surveillance guidelines. 3: 9 missing values. Abbreviations: CRC: colorectal cancer; HD-CE: high-definition dye-based chromoendoscopy; HD-WLE: high-definition white light endoscopy; HGD: high-grade dysplasia; IND: indefinite for dysplasia; IQR=interquartile range; LGD: low-grade dysplasia; n=number; SR: segmental re-inspection; y=years.

Dirk De Jong

Appearances