Diagnostic performance of endocytoscopy with normal pit-like structure sign for colorectal low-grade adenoma compared with conventional modalities

Objectives

A "resect-and-discard" strategy has been proposed for diminutive adenomas in the colorectum. However, this strategy is sometimes difficult to implement because of the lack of confidence in differentiating low-grade adenoma (LGA) from advanced lesions such as high-grade adenoma or carcinoma. To perform real-time precise diagnosis of LGA with high confidence, we assessed whether endocytoscopy (EC) diagnosis, considering normal pit-like structure (NP-sign), an excellent indicator of LGA, could have additional diagnostic potential compared with conventional modalities.

Methods

All the neoplastic lesions that were observed by non-magnifying narrow-band imaging (NBI), magnifying NBI (M-NBI), magnifying pit pattern, and EC prior to pathological examination between 2005 and 2018 were retrospectively investigated. The neoplastic lesions were classified into two categories: LGA and other neoplastic lesions. We assessed the differential diagnostic ability of EC with NP-sign between LGA and other neoplastic lesions compared with that of NBI, M-NBI, pit pattern, and conventional EC in terms of sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC).

Results

A total of 1376 lesions from 1097 patients were eligible. The specificity (94.9%), accuracy (91.5%), and area under the receiver operating characteristic curve (0.95) of EC with NP-sign were significantly higher than those of NBI, M-NBI, pit pattern, and conventional EC.

Conclusions

EC diagnosis with NP-sign has significantly higher diagnostic performance for predicting colorectal LGA compared with the conventional modalities and enables stratification of neoplastic lesions for "resect-and-discard" with higher confidence.

Advances in the removal of diminutive colorectal polyps

Diminutive polyps (<5 mm in diameter) represent the majority of polyps found during colonoscopy; about a half of them are adenomatous, with low risk of advanced neoplasia. Recent studies have demonstrated that cold polypectomy should be considered the recommended approach for resecting diminutive polyps and that cold snaring may be superior to cold forceps biopsy, at least for polyps of 4-5 mm. Recently, electronic chromoendoscopy has been applied to characterization of diminutive polyps to discriminate adenomatous from nonadenomatous lesions. Optical diagnosis of polyp histology could potentially exert huge cost savings by the 'resect and discard' strategy for diminutive polyps and 'leaving-in' for diminutive hyperplastic polyps in the recto-sigmoid colon. These policies represent the mainstay for adopting endoscopy-directed post-polypectomy surveillance strategies, endorsed by both American and European Endoscopy Societies. Accuracy of both histology and surveillance intervals predictions from academic centers have been encouraging, although the same performance has not been replicated in community practices.

Proposal of a new 'resect and discard' strategy using magnifying narrow band imaging: pilot study of diagnostic accuracy

BACKGROUND AND AIM

A 'resect and discard' strategy using non-magnifying narrow band imaging (N-NBI) has been proposed for reducing screening colonoscopy costs, but it does not take into consideration advanced histology and magnifying NBI (M-NBI) that can potentially further improve the 'resect and discard' strategy. We conducted a pilot study to investigate the feasibility of M-NBI for the new 'resect and discard' strategy with consideration for advanced histology.

METHODS

The present study involved 63 patients. For each polyp, optical diagnosis was independently made using N-NBI and M-NBI, a decision (discard or send for pathology) was made based on the M-NBI findings, and histological and optical diagnosis results were compared.

RESULTS

Sensitivity, specificity and accuracy (95% confidence interval) of M-NBI for lesions with advanced histology were 1.00 (0.36-1.00), 0.92 (0.91-0.92) and 0.92 (0.90-0.92), respectively. Whereas the management of 86% of small polyps could be decided without formal histopathology, lesions with advanced histology were not discarded using M-NBI. The diagnostic accuracy of M-NBI in distinguishing neoplastic from non-neoplastic lesions was higher than that of N-NBI.

CONCLUSIONS

We propose a new resect and discard strategy with advanced histology using M-NBI that is a promising strategy to reduce the costs of histopathology and to minimize the risk of discarding important lesions.