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

Multicentre study to assess the performance of an artificial intelligence instrument to support qualitative diagnosis of colorectal polyps

OBJECTIVE

Computer-aided diagnosis (CAD) using artificial intelligence (AI) is expected to support the characterisation of colorectal lesions, which is clinically relevant for efficient colorectal cancer prevention. We conducted this study to assess the diagnostic performance of commercially available CAD systems.

METHODS

This was a multicentre, prospective performance evaluation study. The endoscopist diagnosed polyps using white light imaging, followed by non-magnified blue light imaging (non-mBLI) and mBLI. AI subsequently assessed the lesions using non-mBLI (non-mAI), followed by mBLI (mAI). Eventually, endoscopists made the final diagnosis by integrating the AI diagnosis (AI+endoscopist). The primary endpoint was the accuracy of the AI diagnosis of neoplastic lesions. The diagnostic performance of each modality (sensitivity, specificity and accuracy) and confidence levels were also assessed.

RESULTS

Overall, 380 lesions from 139 patients were included in the analysis. The accuracy of non-mAI was 83%, 95% CI (79% to 87%), which was inferior to that of mBLI (89%, 95% CI (85% to 92%)) and mAI (89%, 95% CI (85% to 92%)). The accuracy (95% CI) of diagnosis by expert endoscopists using mAI (91%, 95% CI (87% to 94%)) was comparable to that of expert endoscopists using mBLI (91%, 95% CI (87% to 94%)) but better than that of non-expert endoscopists using mAI (83%, 95% CI (75% to 90%)). The level of confidence in making a correct diagnosis was increased when using magnification and AI.

CONCLUSIONS

The diagnostic performance of mAI for differentiating colonic lesions is comparable to that of endoscopists, regardless of their experience. However, it can be affected by the use of magnification as well as the endoscopists' level of experience.

Accuracy of artificial intelligence on histology prediction and detection of colorectal polyps: a systematic review and meta-analysis

BACKGROUND AND AIMS

We performed a meta-analysis of all published studies to determine the diagnostic accuracy of artificial intelligence (AI) on histology prediction and detection of colorectal polyps.

METHOD

We searched Embase, PubMed, Medline, Web of Science, and Cochrane library databases to identify studies using AI for colorectal polyp histology prediction and detection. The quality of included studies was measured by the Quality Assessment of Diagnostic Accuracy Studies tool. We used a bivariate meta-analysis following a random-effects model to summarize the data and plotted hierarchical summary receiver operating characteristic curves. The area under the hierarchical summary receiver operating characteristic curve (AUC) served as an indicator of the diagnostic accuracy and during head-to-head comparisons.

RESULTS

A total of 7680 images of colorectal polyps from 18 studies were included in the analysis of histology prediction. The accuracy of the AI (AUC) was .96 (95% confidence interval [CI], .95-.98), with a corresponding pooled sensitivity of 92.3% (95% CI, 88.8%-94.9%) and specificity of 89.8% (95% CI, 85.3%-93.0%). The AUC of AI using narrow-band imaging (NBI) was significantly higher than the AUC using non-NBI (.98 vs .84, P < .01). The performance of AI was superior to nonexpert endoscopists (.97 vs .90, P < .01). For characterization of diminutive polyps using a deep learning model with nonmagnifying NBI, the pooled negative predictive value was 95.1% (95% CI, 87.7%-98.1%). For polyp detection, the pooled AUC was .90 (95% CI, .67-1.00) with a sensitivity of 95.0% (95% CI, 91.0%-97.0%) and a specificity of 88.0% (95% CI, 58.0%-99.0%).

CONCLUSIONS

AI was accurate in histology prediction and detection of colorectal polyps, including diminutive polyps. The performance of AI was better under NBI and was superior to nonexpert endoscopists. Despite the difference in AI models and study designs, AI performances are rather consistent, which could serve as a reference for future AI studies.

Clinical exploration of marking targeting biopsy in the intraoperative localization value of colon polypectomy

Objective

To evaluate the feasibility and safety of marking targeting biopsy (MTB) in the intraoperative localization value of colon polypectomy.

Methods

The clinical data from patients with polyp of colon discovered under colonoscopy from January 2014 to January 2016 were retrospectively analyzed. A total of 87 patients conformed to the inclusion criteria, among them, 43 received colonoscopic polypectomy one week after MTB (MTB group), while 44 underwent colonoscopic polypectomy one week after conventional biopsy (conventional group). The time consumption in colonoscopic treatment, polypectomy rate and postoperative complications between two groups were compared.

Results

The time consumed in operation in the MTB group was 25.5 (±8.6) minutes, while that in conventional group was 42.0 (±20.5) minutes, and the difference was statistically significant (P<0.01). There were a total of 86 polyps in the MTB group, among which 83 were removed, yielding the removal rate of 96.5%. There were altogether 88 polyps in the conventional group, among which 54 were removed, resulting in the removal rate of 61.4%, and the difference was statistically significant (P<0.05). three polyps in the MTB group were detached after MTB, or the wound surface became flat after gross polyp removal, and no polypectomy was required, but the marking targeting solution was clearly visible. two respective polyps in 12 cases in conventional group could not be found in colonoscopic treatment, and 10 of them had respective one polyp that could not be found again. 12 cases in MTB group suffered from abdominal pain after surgery, and no hemorrhage was seen intraoperatively and postoperatively. 10 cases in the conventional group had abdominal pain after surgery, and one case had delayed hemorrhage after surgery. The results between two groups displayed no statistical significance (P>0.05).

Conclusions

The localization value of MTB in colon polypectomy is definitely feasible, safe and effective, which can greatly shorten the time of endoscopic colon polypectomy, mitigate patient sufferings, and reduce the incidence of false negative rate of polyp. It displays favorable clinical application value and is worthy of being promoted in clinic.

Optical Technologies for Endoscopic Real-Time Histologic Assessment of Colorectal Polyps: A Meta-Analysis

OBJECTIVES

Accurate, real-time, endoscopic risk stratification of colorectal polyps would improve decision-making and optimize clinical efficiency. Technologies to manipulate endoscopic optical outputs can be used to predict polyp histology in vivo; however, it remains unclear how accuracy has progressed and whether it is sufficient for routine clinical implementation.

METHODS

A meta-analysis was conducted by searching MEDLINE, Embase, and the Cochrane Library. Studies were included if they prospectively deployed an endoscopic optical technology for real-time in vivo prediction of adenomatous colorectal polyps. Polyposis and inflammatory bowel diseases were excluded. Bayesian bivariate meta-analysis was performed, presenting 95% confidence intervals (CI).

RESULTS

One hundred two studies using optical technologies on 33,123 colorectal polyps were included. Digital chromoendoscopy differentiated neoplasia (adenoma and adenocarcinoma) from benign polyps with sensitivity of 92.2% (90.6%-93.9% CI) and specificity of 84.0% (81.5%-86.3% CI), with no difference between constituent technologies (narrow-band imaging, Fuji intelligent Chromo Endoscopy, iSCAN) or with only diminutive polyps. Dye chromoendoscopy had sensitivity of 92.7% (90.1%-94.9% CI) and specificity of 86.6% (82.9%-89.9% CI), similarly unchanged for diminutive polyps. Spectral analysis of autofluorescence had sensitivity of 94.4% (84.0%-99.1% CI) and specificity of 50.9% (13.2%-88.8% CI). Endomicroscopy had sensitivity of 93.6% (85.3%-98.3% CI) and specificity of 92.5% (81.8%-98.1% CI). Computer-aided diagnosis had sensitivity of 88.9% (74.2%-96.7% CI) and specificity of 80.4% (52.6%-95.7% CI). Prediction confidence and endoscopist experience alone did not significantly improve any technology. The only subgroup to demonstrate a negative predictive value for adenoma above 90% was digital chromoendoscopy, making high confidence predictions of diminutive recto-sigmoid polyps. Chronologic meta-analyses show a falling negative predictive value over time. A significant publication bias exists.

DISCUSSION

This novel approach to meta-analysis demonstrates that existing optical technologies are increasingly unlikely to allow safe "resect and discard" strategies and that step-change innovation may be required. A "diagnose and leave" strategy may be supported for diminutive recto-sigmoid polyps diagnosed with high confidence; however, limitations exist in the evidence base for this cohort.

Endoscopic prediction of advanced histology in diminutive and small colorectal polyps

BACKGROUND AND AIM

Most polyps detected during colonoscopies are diminutive or small, and they rarely have advanced histology. Real-time prediction of advanced histology would help clinicians to assess the need for pathological evaluation. Here, we investigated endoscopic predictors of advanced histology in diminutive and small polyps.

METHODS

Consecutive patients with adenomatous polyps (<10 mm) removed endoscopically from January 2013 to December 2014 at a single tertiary cancer center were eligible for inclusion. Two endoscopists reviewed all endoscopic images to identify significant findings associated with advanced histology using multivariate models. The sensitivity, specificity, and negative predictive value of the identified endoscopic predictors for advanced histology were calculated.

RESULTS

Of 6170 polyps (4746 diminutive) removed from 2611 patients, 320 (5.2%) showed advanced histology, including five submucosal invasive cancers. In multivariate analysis, advanced histology was significantly associated with the following: loss of lobulation (odds ratio [OR] 61.7; 95% confidence interval [95% CI]: 19.1-199.0); heterogeneity in mucosal patterns (OR 29.0; 95% CI: 14.6-57.3); non-polypoid growth (OR 15.7; 95% CI: 4.4-55.5); white spots (OR 13.5; 95% CI: 7.8-23.5); and surface redness (OR 6.6; 95% CI: 3.0-14.5); and irregular capillary pattern (OR 4.8; 95% CI: 2.5-9.1). These significant predictors successfully predicted all submucosal invasive cancers as advanced histology. The sensitivity, specificity, and negative predictive values were 37.2%, 97.8%, and 96.6%.

CONCLUSIONS

We identified six endoscopic predictors for advanced histology in diminutive or small colon polyps. Diminutive and small polyps lacking these predictors would not be considered to have advanced histology.

Investigation of the Local Recurrence Rate after Colorectal Endoscopic Mucosal Resection: Is Incomplete Polyp Resection Really a Clinically Important Problem? Analysis of the Rationale for the "Resect and Discard" Strategy

BACKGROUND/AIMS

The "Resect and Discard" strategy is a potentially useful strategy. At present, only the lesion size and accuracy of diagnosis are cited as considerations for clinical adoption of this strategy. On the other hand, histopathology of the resected specimens after Endoscopic Mucosal Resection (EMR) reveals often an unclear or positive-margin status, implying Incomplete Polyp Resection (IPR). If IPR indeed increased the risk of local recurrence, histopathological evaluation of the margin would be indispensable and clinical adoption of this strategy is difficult. The aim of this study is to verify the association between IPR and the risk of local recurrence.

METHODS

The 1872 polyps and 603 EMR cases in 597 patients who had EMR between May 2013 and May 2014 were enrolled. The local recurrence rate until 3 years after the EMR in cases with the target lesions of the "Resect and Discard" strategy was determined in the negative-margin and IPR groups.

RESULTS

The final analysis was performed using the data of 1092 polyps, and 222 were categorized into the IPR group. There were no cases of recurrence in either of the groups.

CONCLUSION

This is the world's first report conducted to examine the correlation of IPR and the local recurrence rate for clinical practice of "Resect and Discard" strategy. There is the possibility that pathological evaluation of the margins after EMR in patients with small polyps can be skipped.

Management and risk factors for incomplete resection associated with jumbo forceps polypectomy for diminutive colorectal polyps: a single-institution retrospective study

BACKGROUND

Cold forceps polypectomy is simple and widely used in clinical practice. However, there are concerns about the risk of incomplete resection using this technique. In recent years, it has been reported that polypectomy with jumbo forceps (JF) is an effective treatment modality for diminutive polyps (DPs) because JF are able to remove large tissue samples with the combined advantage of a higher complete histological resection rate for DPs than standard forceps. To our knowledge, no studies have evaluated the risk factors for incomplete resection when polypectomy with JF is performed for DPs.

METHODS

From among 1129 DPs resected using JF at Hiroshima City Asa Citizens Hospital between November 2015 and December 2016, we retrospectively evaluated the clinical outcomes of 999 tumors with known histopathology and investigated the relationship between incomplete resection and clinicopathological factors.

RESULTS

Most lesions [985 (87%)] were low-grade dysplasia and 14 (1%) were high-grade dysplasia. The en bloc resection rate was 92% (918/999) and the histological en bloc resection rate was 78% (777/999). Multivariate analysis showed that the significant independent predictors of incomplete resection were tumor size ≥ 4 mm [odds ratio (OR) 3.8; 95% confidence interval (CI) 2.65-5.37; p < 0.01], non-tangential direction of forceps in relation to the tumor (OR 1.73; 95% CI 1.21-2.45; p < 0.01), and lack of muscularis mucosae in the pathological specimen (OR 15.7; 95% CI 9.16-27.7; p < 0.01).

CONCLUSIONS

This study identified significant independent predictors of incomplete resection of DPs which may be helpful when planning polypectomy with JF.

Clinical outcomes of the "resect and discard" strategy using magnifying narrow-band imaging for small (< 10 mm) colorectal polyps

Background and study aim  The "resect and discard" strategy is a new paradigm for the management of small colorectal polyps that reduces the cost and effort related to pathological diagnosis after polypectomy. This retrospective study aimed to clarify the clinical outcome of the "resect and discard" strategy for small colorectal polyps. Patients and methods  The clinical records were reviewed from 501 consecutive patients who underwent the "resect and discard" strategy for colorectal polyps smaller than 10 mm at our hospital between January 2008 and December 2010. All colorectal lesions were evaluated onsite under magnifying narrow-band imaging after careful conventional white-light imaging. In cases of low grade adenoma predicted with high confidence, colonoscopists selected the "resect and discard" option without formal histopathology. The mid-term outcomes were evaluated to validate the curability of the "resect and discard" strategy. Results  The present study included 501 consecutive patients with 816 lesions. The mid-term outcomes were examined for 476 (95 %) patients who received follow-up for at least 1 year after undergoing the "resect and discard" strategy. The median observation period was 83 months (range 12 - 117 months). No patient died from colorectal cancer related to the procedure, resulting in a disease-specific survival rate of 100 %. There were no local and/or distant recurrences detected during follow-up. Conclusions  The "resect and discard" strategy for small colorectal polyps under strict preoperative diagnosis achieves excellent mid-term outcome.

Clinical impact of surveillance colonoscopy using magnification without diminutive polyp removal

BACKGROUND AND AIM

In Western countries, endoscopic removal of all adenomas during colonoscopy is recommended. The present study evaluates the usefulness of magnifying colonoscopy without removal of diminutive (≤5 mm) colorectal polyps.

METHODS

Patients with diminutive polyps who were observed for over 5 years using magnification at Hiroshima University Hospital were selected retrospectively. Lesions ≥6 mm in size, depressed lesions, and lesions with type V pit pattern were indications for endoscopic resection. We investigated the characteristics of lesions indicated for endoscopic resection detected on surveillance colonoscopy and the risk factors for the incidence of lesions indicated for endoscopic resection.

RESULTS

A total of 706 consecutive patients were enrolled. Sixty-eight lesions indicated for endoscopic resection were detected, averaging 9.0 ± 4.8 mm, and 33 (49%) lesions were located in the right colon. Pathological diagnoses were adenoma, Tis carcinoma, and T1 carcinoma in 58 (85%), eight (12%), and two (3%) lesions, respectively. Five lesions were considered to grow from previously detected diminutive polyps. Relative risks for the incidence of a lesion indicated for endoscopic resection were 1.76 (95% confidence interval [CI], 1.004-3.23) for males compared with females, 3.76 (95% CI, 2.03-7.50) for more than three polyps at initial colonoscopy compared with one polyp, and 2.84 (95% CI, 1.43-5.24) for patients with carcinoma at initial colonoscopy compared with patients with no lesion indicated for endoscopic resection. Nine carcinomas were resected endoscopically.

CONCLUSION

Diminutive low-grade adenomas detected by using magnifying colonoscopy may not necessarily require removal.

Clinical, endoscopic and pathological characteristics of colorectal polyps in elderly patients: Single-center experience

Increasing age is a risk factor for the development of colorectal adenomas and advanced adenomas. However, few studies have been published on the features of colorectal polyps in the elderly. The present study aimed to investigate the clinical, enteroscopic and pathological characteristics of colorectal polyps in Chinese elderly patients in a single center (The Central Hospital of Wuhan, Hubei, China). The endoscopic and pathological reports of colonoscopies performed in our center were retrospectively analyzed. A total of 7,795 consecutive patients referred for colonoscopy were evaluated between January 2013 and December 2014. Of the 297 who met the inclusion criteria, 279 polyps were observed in men and 230 in women. Of all the polyps, 263 were non-adenomatous polyps, 104 were non-advanced adenomas and 142 were advanced adenomas. 336 polyps were left-sided and 173 were right-sided. Polyps ≥10 mm were more likely to exhibit an adenomatous component and advanced features, and these findings continued to hold true when the size cut-off was set at 5 mm. The data shown in the present study have revealed that a significant number of polyps lie proximal to the splenic flexure. Thus, evaluation of the whole bowel is particularly important in elderly patients who are undergoing colonoscopy. In addition, the polyp size was associated with the presence of adenoma, and advanced component, diminutive and small polyps should not be ignored in elderly patients.

What Can We Do to Optimize Colonoscopy and How Effective Can We Be?

In the USA, colorectal cancer is the third most common cancer and third leading cause of cancer death among both men and women. Declining rates of colon cancer in the past decade have been attributed in part to screening and removal of precancerous polyps via colonoscopy. Recent emphasis has been placed on measures to increase the quality and effectiveness of colonoscopy. These have been divided into pre-procedure quality metrics (bowel preparation), procedural quality metrics (cecal intubation, withdrawal time, and adenoma detection rate), post-procedure metrics (surveillance interval), and other quality metrics (patient satisfaction and willingness to repeat the procedure). The purpose of this article is to review the data and controversies surrounding each of these and identify ways to optimize the performance of colonoscopy.

Feasibility of cold snare polypectomy in Japan: A pilot study

AIM: To investigate the feasibility of cold snare polypectomy (CSP) in Japan.

METHODS: The outcomes of 234 non-pedunculated polyps smaller than 10 mm in 61 patients who underwent CSP in a Japanese referral center were retrospectively analyzed. The cold snare polypectomies were performed by nine endoscopists with no prior experience in CSP using an electrosurgical snare without electrocautery.

RESULTS: CSPs were completed for 232 of the 234 polyps. Two (0.9%) polyps could not be removed without electrocautery. Immediate postpolypectomy bleeding requiring endoscopic hemostasis occurred in eight lesions (3.4%; 95%CI: 1.1%-5.8%), but all were easily managed. The incidence of immediate bleeding after CSP for small polyps (6-9 mm) was significantly higher than that of diminutive polyps (≤ 5 mm; 15% vs 1%, respectively). Three (5%) patients complained of minor bleeding after the procedure but required no intervention. The incidence of delayed bleeding requiring endoscopic intervention was 0.0% (95%CI: 0.0%-1.7%). In total, 12% of the resected lesions could not be retrieved for pathological examination. Tumor involvement in the lateral margin could not be histologically assessed in 70 (40%) lesions.

CONCLUSION: CSP is feasible in Japan. However, immediate bleeding, retrieval failure and uncertain assessment of the lateral tumor margin should not be underestimated. Careful endoscopic diagnosis before and evaluation of the tumor residue after CSP are recommended when implementing CSP in Japan.

An alternative option for "resect and discard" strategy, using magnifying narrow-band imaging: a prospective "proof-of-principle" study

BACKGROUND

The "resect and discard" strategy is beneficial for cost savings on screening and surveillance colonoscopy, but it has the risk to discard lesions with advanced histology or small invasive cancer (small advanced lesion; SALs). The aim of this study was to prove the principle of new "resect and discard" strategy with consideration for SALs using magnifying narrow-band imaging (M-NBI).

METHODS

Patients undergoing colonoscopy at a tertiary center were involved in this prospective trial. For each detected polyp <10 mm, optical diagnosis (OD) and virtual management ("leave in situ", "discard" or "send for pathology") were independently made using non-magnifying NBI (N-NBI) and M-NBI, and next surveillance interval were predicted. Histological and optical diagnosis results of all polyps were compared.

RESULTS

While the management could be decided in 82% of polyps smaller than 10 mm, 24/31 (77%) SALs including two small invasive cancers were not discarded based on OD using M-NBI. The sensitivity [90% confidence interval (CI)] of M-NBI for SALs was 0.77 (0.61-0.89). The risk for discarding SALs using N-NBI was significantly higher than that using M-NBI (53 vs. 23%, p = 0.02). The diagnostic accuracy (95% CI) of M-NBI in distinguishing neoplastic from non-neoplastic lesions [0.88 (0.86-0.90)] was significantly better than that of N-NBI [0.84 (0.82-0.87)] (p = 0.005).

CONCLUSIONS

The results of our study indicated that our "resect and discard" strategy using M-NBI could work to reduce the risk for discarding SALs including small invasive cancer (UMIN-CTR, UMIN000003740).

Diagnostic performance of magnifying narrow-band imaging for early gastric cancer: A meta-analysis

AIM: To investigate the performance of magnifying endoscopy with narrow-band imaging (ME-NBI) in the diagnosis of early gastric cancer (EGC).

METHODS: Systematic literature searches were conducted until February 2014 in PubMed, EMBASE, Web of Science, Ovid, Scopus and the Cochrane Library databases by two independent reviewers. Meta-analysis was performed to calculate the pooled sensitivity, specificity and diagnostic odds ratio and to construct a summary receiver operating characteristic (ROC) curve. Subgroup analyses were performed based on the morphology type of lesions, diagnostic standard, the size of lesions, type of assessment, country and sample size to explore possible sources of heterogeneity. A Deeks’ asymmetry test was used to evaluate the publication bias.

RESULTS: Fourteen studies enrolling 2171 patients were included. The pooled sensitivity, specificity and diagnostic odds ratio for ME-NBI diagnosis of EGC were 0.86 (95%CI: 0.83-0.89), 0.96 (95%CI: 0.95-0.97) and 102.75 (95%CI: 48.14-219.32), respectively, with the area under ROC curve being 0.9623. Among the 14 studies, six also evaluated the diagnostic value of conventional white-light imaging, with a sensitivity of 0.57 (95%CI: 0.50-0.64) and a specificity of 0.79 (95%CI: 0.76-0.81). When using “VS” (vessel plus surface) ME-NBI diagnostic systems in gastric lesions of depressed macroscopic type, the pooled sensitivity and specificity were 0.64 (95%CI: 0.52-0.75) and 0.96 (95%CI: 0.95-0.98). For the lesions with a diameter less than 10 mm, the sensitivity and specificity were 0.74 (95%CI: 0.65-0.82) and 0.98 (95%CI: 0.97-0.98).

CONCLUSION: ME-NBI is a promising endoscopic tool in the diagnosis of early gastric cancer and might be helpful in further target biopsy.