Magnifying chromoendoscopy does not seem helpful in avoiding unnecessary surgery once deep submucosal invasion is suspected

We read with interest the article by Zhang et al. on the diagnostic accuracy of magnifying narrow-band imaging (M-NBI) for the optical diagnosis of T1 colorectal cancer (CRC) with deep submucosal invasion [1]. The authors claim that, although the specificity of M-NBI was comparable to magnifying chromoendoscopy (M-CE), the sensitivity was significantly lower (0.74, 95 % confidence interval [CI] 0.66 – 0.81 for M-NBI vs. 0.84, 95 %CI 0.76 – 0.89 for M-CE). This contradicts the conclusion of a recently published meta-analysis by our study group [2], in which no difference was observed. In this letter, we would like to discuss some concerns regarding the conclusion and recommendations proposed by Zhang et al.

The authors analyzed only studies in which M-NBI was conducted, with or without subsequent M-CE. By selecting these studies, the authors actually compared M-NBI with M-NBI + M-CE, and not M-NBI with M-CE as they erroneously state. This nonblinded unidirectional order of M-CE following M-NBI does not introduce “some bias” as the authors discuss, but actually hampers them addressing their study aim.

Furthermore, the conclusion of superior sensitivity for M-CE seems too firm when looking in detail at the performance of M-NBI for different subgroups. The sensitivity of M-NBI in real-time diagnosis studies (0.84, 95 %CI 0.76 – 0.89) is similar to the reported sensitivity of M-CE (0.84, 95 %CI 0.76 – 0.89), and comparable to the sensitivity of M-CE (0.81, 95 %CI 0.75 – 0.87) reported in our meta-analysis in which studies that evaluated M-CE without NBI were also included [2]. In addition, the sensitivity of M-NBI in the current review was higher when excluding abstracts that were not peer reviewed (0.76, 95 %CI 0.67 – 0.83). This obviously questions the robustness of evidence supporting the authors’ conclusion.

Importantly, the authors recommend to use M-NBI first for optical diagnosis to make an endoscopic therapeutic decision, and to use M-CE when a suspected T1 CRC with deep submucosal invasion is detected, in order to re-evaluate invasion depth to avoid unnecessary surgery. This recommendation is not, however, supported by their findings, as this would require a superior specificity of M-CE (i. e. reducing false positives), which was not observed. Superior sensitivity, if present, would lower the number of false negatives. In other words, the current results do not suggest that unnecessary surgery can be avoided once deep submucosal invasion is suspected, but instead suggest that M-CE could be of value in situations where no deep submucosal invasion is suspected with M-NBI, in order to avoid unnecessary endoscopic resection of lesions with deep invasion.

Taken together, both reviews demonstrate the usefulness of advanced imaging techniques, such as M-NBI and M-CE, for the optical diagnosis of deep submucosal invasion. In our opinion, the evidence is lacking that one of the techniques is superior to the other. At this moment, it therefore seems more reasonable to base the choice for a specific advanced imaging modality on local expertise, training opportunities, and available technical resources.

• References

• 1 Zhang QW, Teng LM, Zhang XT. et al. Narrow-band imaging in the diagnosis of deep submucosal colorectal cancers: a systematic review and meta-analysis. Endoscopy 2017; 49: 564-580
• 2 Backes Y, Moss A, Reitsma JB. et al. Narrow band imaging, magnifying chromoendoscopy, and gross morphological features for the optical diagnosis of T1 colorectal cancer and deep submucosal invasion: a systematic review and meta-analysis. Am J Gastroenterol 2017; 112: 54-64