Sarah K. McGill

Narrow band imaging to differentiate neoplastic and non-neoplastic colorectal polyps in real time: a meta-analysis of diagnostic operating characteristics

Purpose Many studies have reported on the use of narrow band imaging (NBI) colonoscopy to differentiate neoplastic from non-neoplastic colorectal polyps. It has potential to replace pathological diagnosis of diminutive polyps. We aimed to perform a systematic review and meta-analysis on the real-time diagnostic operating characteristics of NBI colonoscopy. Methods We searched PubMed, SCOPUS and Cochrane databases and abstracts. We used a two-level bivariate meta-analysis following a random effects model to summarise the data and fit hierarchical summary receiver-operating characteristic (HSROC) curves. The area under the HSROC curve serves as an indicator of the diagnostic test strength. We calculated summary sensitivity, specificity and negative predictive value (NPV). We assessed agreement of surveillance interval recommendations based on endoscopic diagnosis compared to pathology. Results For NBI diagnosis of colorectal polyps, the area under the HSROC curve was 0.92 (95% CI 0.90 to 0.94), based on 28 studies involving 6280 polyps in 4053 patients. The overall sensitivity was 91.0% (95% CI 87.6% to 93.5%) and specificity was 82.6% (95% CI 79.0% to 85.7%). In eight studies (n=2146 polyps) that used high-confidence diagnostic predictions, sensitivity was 93.8% and specificity was 83.3%. The NPVs exceeded 90% when 60% or less of all polyps were neoplastic. Surveillance intervals based on endoscopic diagnosis agreed with those based on pathology in 92.6% of patients (95% CI 87.9% to 96.3%). Conclusions NBI diagnosis of colorectal polyps is highly accurate—the area under the HSROC curve exceeds 0.90. High-confidence predictions provide >90% sensitivity and NPV. It shows high potential for real-time endoscopic diagnosis.

Endoscopists can sustain high performance for the optical diagnosis of colorectal polyps following standardized and continued training.

BACKGROUND AND STUDY AIMS The learning curve for optical diagnosis of colorectal polyps with the narrow-band imaging (NBI) is unknown. To forego histological analysis of diminutive polyps diagnosed optically with high confidence, guidelines recommend ≥ 90 % negative predictive value (NPV) and concordance of ≥ 90 % for surveillance intervals predicted optically and histologically. We aimed to study the learning of optical diagnosis for colorectal polyps. PATIENTS AND METHODS We studied five endoscopists as part of a randomized multisite trial comparing near-focus and standard-focus views for optical diagnosis. They trained using a computer-based module, followed by 10 real-time colonoscopies with pathology correlation. Endoscopists then optically diagnosed and resected all the polyps found during 558 consecutive colonoscopies, and diagnoses were compared with pathology. Endoscopists repeated the training module at the study midpoint. NPV and concordance of surveillance intervals for diminutive polyps diagnosed optically with high confidence were measured over time. RESULTS Endoscopists showed high diagnostic performance, with a nonsignificant trend toward higher NPV in the second half of the study. For the 445 polyps in the standard-view arm, the NPV was 88.0 % (95 %CI 75.7 % - 95.5 %) in the first half and 95.8 % (88.3 % - 99.1 %) in the second; P = 0.7. Three endoscopists in the first half and four in the second achieved > 90 % NPV. Concordance of surveillance intervals was identical in the first and second halves at 98.1 % (95 %CI 93.3 % - 99.8 %). CONCLUSIONS High NPV for the prediction of non-neoplasms with NBI was achieved and maintained in this group of endoscopists who participated in standardized and continued training. Both NPV and surveillance interval agreement indicated high performance in the optical diagnosis of colorectal polyps and exceeded thresholds.

The learning curve for detection of non-polypoid (flat and depressed) colorectal neoplasms

We read with interest the two excellent studies on interval colorectal cancers (CRC) following colonoscopy and wish to present important data relevant to the prevention of interval CRC. In a pooled multicohort analysis, Robertson and coauthors found that invasive cancer was diagnosed in 0.6% of patients after clearing colonoscopy among eight large North American trials, and concluded that half of these were likely missed lesions.1 In a population-based study, le Clercq and coauthors reported that CRC found following colonoscopy were more likely to have non-polypoid morphology than prevalent cancers.2 We previously described that non-polypoid colorectal neoplasms (NP-CRN) harbour more advanced pathology than polypoid neoplasms.3 Indeed, due to their subtle appearance, NP-CRN may be a key contributor to interval CRC. The learning curve in their detection is largely unknown. Using available data, we summarised the learning curve (see figure 1) for the detection …

Image-enhanced endoscopy in practice

The detection, diagnosis and treatment of early cancers offers the best hope for the prevention and cure of gastrointestinal cancers – one of the leading causes of death worldwide (1). The detection of pre- or early cancer using white light endoscopy can be challenging because their morphology can be inconspicuous (ie, nonpolypoid; slightly elevated, flat, or slightly depressed]) and their colour can be minimally altered. Indeed, nonpolypoid neoplasms have been shown to be common and important in the esophagus, Barrett’s mucosa and stomach (Figure 1). Our recent prevalence study (2), highlighted their importance in the colon. We showed that nonpolypoid colorectal neoplasms (NP-CRNs) are relatively common and are potentially more dangerous than polypoid neoplasms of similar size because they have a higher risk of containing in situ or submucosal invasive carcinoma. Some nonpolypoid gastrointestinal neoplasms are fairly easy to detect and diagnose, whereas others can be quite difficult to visualize using white light illumination. The current technique and technology of image-enhanced endoscopy (IEE) is available to augment the detection, diagnosis and treatment of these subtle lesions. Figure 1) Distribution of superficial lesions according to morphological classification showing the relevance and importance of the nonpolypoid types in our endoscopic practice. Data regarding lesions of the upper gastrointestinal tract were derived from the Paris ... Dr Tonya Kaltenbach is a gastroenterologist with the VA Healthcare System in Palo Alto, California, USA There are two methods of IEE: dye-based and equipment-based (3). The objective of these two methods is to increase the contrast of structures, thus making the mucosal topography, morphology and borders of lesions viewable in finer detail. Used alone or in tandem, they may complement the white light examination as well each other (Figure 2). Detailed examination of the mucosa provides a cross-sectional view of the underlying pathology and facilitates discrimination between normal, non-neoplastic and neoplastic tissue. Taken together with size and morphology, and observation during submucosal injection, important information regarding the likelihood of submucosal invasion and whether the patient can undergo a safe and curative endoscopic resection is obtained. Figure 2) The benefit of the detection and diagnosis of pre- and early cancer of the gastrointestinal tract is significantly curtailed if safe and efficacious endoscopic treatment is unavailable. Image-enhanced endoscopy is an integral component of endoscopic diagnosis, ... The present article describes the techniques and applications of IEE and provides readers of the Journal with a resource to begin or to potentially improve on their use of IEE. For critical analysis of the literature, readers are directed toward a comprehensive review of the subject (3,4). We will present a description of the various techniques of IEE, and their specific preparations and properties. We will also include an outline of the supporting data regarding the use of IEE for the detection, diagnosis and therapy of a variety of nonpolypoid gastrointestinal neoplasms.

Tu1157 Gastroenterology Trainees Can Perform Real Time Optical Diagnosis of Diminutive Colorectal Polyps Using Narrow Band Imaging

BACKGROUND Proposed clinical thresholds for optical diagnosis of colorectal polyps recommend ≥90% agreement between opticaland histopathology-based surveillance intervals, and ≥90% negative predictive value in the rectosigmoid colon. Experienced gastroenterologists have exceeded these thresholds in studies, but the performance of trainees is unknown. We aimed to prospectively assess real-time trainee performance of optical diagnosis of diminutive colorectal polyps. METHODS We studied the optical diagnoses of colorectal polyps using narrow band imaging (NBI) in gastroenterology trainees over a 9-month period during a prospective, randomized polyp detection tandem colonoscopy trial. Trainees viewed a computer-based learning module on NBI International Colorectal Endoscopic (NICE) classification and took a test that required minimum 90% accuracy prior to the study. For each polyp found at colonoscopy, trainees stated the diagnosis (neoplastic vs. non-neoplastic) and confidence level in the diagnosis (high vs. low), then removed and submitted it for histopathologic diagnosis. Central, blinded pathology was the reference standard. An optical diagnosis was high confidence if the polyp had at least one of the features of neoplasm or non-neoplasm and none of the features of the opposing diagnosis. Otherwise, low confidence