Establishment and validation of a computer-assisted colonic polyp localization system based on deep learning

AI support for colonoscopy quality control using CNN and transformer architectures

BACKGROUND

Construct deep learning models for colonoscopy quality control using different architectures and explore their decision-making mechanisms.

METHODS

A total of 4,189 colonoscopy images were collected from two medical centers, covering different levels of bowel cleanliness, the presence of polyps, and the cecum. Using these data, eight pre-trained models based on CNN and Transformer architectures underwent transfer learning and fine-tuning. The models' performance was evaluated using metrics such as AUC, Precision, and F1 score. Perceptual hash functions were employed to detect image changes, enabling real-time monitoring of colonoscopy withdrawal speed. Model interpretability was analyzed using techniques such as Grad-CAM and SHAP. Finally, the best-performing model was converted to ONNX format and deployed on device terminals.

RESULTS

The EfficientNetB2 model outperformed other architectures on the validation set, achieving an accuracy of 0.992. It surpassed models based on other CNN and Transformer architectures. The model's precision, recall, and F1 score were 0.991, 0.989, and 0.990, respectively. On the test set, the EfficientNetB2 model achieved an average AUC of 0.996, with a precision of 0.948 and a recall of 0.952. Interpretability analysis showed the specific image regions the model used for decision-making. The model was converted to ONNX format and deployed on device terminals, achieving an average inference speed of over 60 frames per second.

CONCLUSIONS

The AI-assisted quality system, based on the EfficientNetB2 model, integrates four key quality control indicators for colonoscopy. This integration enables medical institutions to comprehensively manage and enhance these indicators using a single model, showcasing promising potential for clinical applications.

Non-polypoid Colorectal Lesions Detection and False Positive Detection by Artificial Intelligence under Blue Laser Imaging and Linked Color Imaging

Objectives

Artificial intelligence (AI) with white light imaging (WLI) is not enough for detecting non-polypoid colorectal polyps and it still has high false positive rate (FPR). We developed AIs using blue laser imaging (BLI) and linked color imaging (LCI) to detect them with specific learning sets (LS).

Methods

The contents of LS were as follows, LS (WLI): 1991 WLI images of lesion of 2-10 mm, LS (IEE): 5920 WLI, BLI, and LCI images of non-polypoid and small lesions of 2-20 mm. LS (IEE) was extracted from videos and included both in-focus and out-of-focus images. We designed three AIs as follows: AI (WLI) finetuned by LS (WLI), AI (IEE) finetuned by LS (WLI)+LS (IEE), and AI (HQ) finetuned by LS (WLI)+LS (IEE) only with images in focus. Polyp detection using a test set of WLI, BLI, and LCI videos of 100 non-polypoid or non-reddish lesions of 2-20 mm and FPR using movies of 15 total colonoscopy were analyzed, compared to 2 experts and 2 trainees.

Results

The sensitivity for LCI in AI (IEE) (83%) was compared to that for WLI in AI (IEE) (76%: p=0.02), WLI in AI (WLI) (57%: p<0.01), BLI in AI (IEE) (78%: p=0.14), and LCI in trainees (74%: p<0.01). The sensitivity for LCI in AI (IEE) (83%) was significantly higher than that in AI (HQ) (78%: p<0.01). The FPR for LCI (6.5%) in AI (IEE) were significantly lower than that in AI (HQ) (17.3%: p<0.01).

Conclusions

AI finetuned by appropriate LS detected non-reddish and non-polypoid polyps under LCI.

Lack of Effectiveness of Computer Aided Detection for Colorectal Neoplasia: A Systematic Review and Meta-Analysis of Nonrandomized Studies

BACKGROUND AND AIMS

Benefits of computer-aided detection (CADe) in detecting colorectal neoplasia were shown in many randomized trials in which endoscopists' behavior was strictly controlled. However, the effect of CADe on endoscopists' performance in less-controlled setting is unclear. This systematic review and meta-analyses were aimed at clarifying benefits and harms of using CADe in real-world colonoscopy.

METHODS

We searched MEDLINE, EMBASE, Cochrane, and Google Scholar from inception to August 20, 2023. We included nonrandomized studies that compared the effectiveness between CADe-assisted and standard colonoscopy. Two investigators independently extracted study data and quality. Pairwise meta-analysis was performed utilizing risk ratio for dichotomous variables and mean difference (MD) for continuous variables with a 95% confidence interval (CI).

RESULTS

Eight studies were included, comprising 9782 patients (4569 with CADe and 5213 without CADe). Regarding benefits, there was a difference in neither adenoma detection rate (44% vs 38%; risk ratio, 1.11; 95% CI, 0.97 to 1.28) nor mean adenomas per colonoscopy (0.93 vs 0.79; MD, 0.14; 95% CI, -0.04 to 0.32) between CADe-assisted and standard colonoscopy, respectively. Regarding harms, there was no difference in the mean non-neoplastic lesions per colonoscopy (8 studies included for analysis; 0.52 vs 0.47; MD, 0.14; 95% CI, -0.07 to 0.34) and withdrawal time (6 studies included for analysis; 14.3 vs 13.4 minutes; MD, 0.8 minutes; 95% CI, -0.18 to 1.90). There was a substantial heterogeneity, and all outcomes were graded with a very low certainty of evidence.

CONCLUSION

CADe in colonoscopies neither improves the detection of colorectal neoplasia nor increases burden of colonoscopy in real-world, nonrandomized studies, questioning the generalizability of the results of randomized trials.

Real-time artificial intelligence (AI)-aided endoscopy improves adenoma detection rates even in experienced endoscopists: a cohort study in Singapore

BACKGROUND

Colonoscopy is a mainstay to detect premalignant neoplastic lesions in the colon. Real-time Artificial Intelligence (AI)-aided colonoscopy purportedly improves the polyp detection rate, especially for small flat lesions. The aim of this study is to evaluate the performance of real-time AI-aided colonoscopy in the detection of colonic polyps.

METHODS

A prospective single institution cohort study was conducted in Singapore. All real-time AI-aided colonoscopies, regardless of indication, performed by specialist-grade endoscopists were anonymously recorded from July to September 2021 and reviewed by 2 independent authors (FHK, JL). Sustained detection of an area by the program was regarded as a "hit". Histology for the polypectomies were reviewed to determine adenoma detection rate (ADR). Individual endoscopist's performance with AI were compared against their baseline performance without AI endoscopy.

RESULTS

A total of 24 (82.8%) endoscopists participated with 18 (62.1%) performing ≥ 5 AI-aided colonoscopies. Of the 18, 72.2% (n = 13) were general surgeons. During that 3-months period, 487 "hits" encountered in 298 colonoscopies. Polypectomies were performed for 51.3% and 68.4% of these polypectomies were adenomas on histology. The post-intervention median ADR was 30.4% was higher than the median baseline polypectomy rate of 24.3% (p = 0.02). Of the adenomas excised, 14 (5.6%) were sessile serrated adenomas. Of those who performed ≥ 5 AI-aided colonoscopies, 13 (72.2%) had an improvement of ADR compared to their polypectomy rate before the introduction of AI, of which 2 of them had significant improvement.

CONCLUSIONS

Real-time AI-aided colonoscopy have the potential to improved ADR even for experienced endoscopists and would therefore, improve the quality of colonoscopy.

Deep learning-assisted diagnosis of chronic atrophic gastritis in endoscopy

Background

Chronic atrophic gastritis (CAG) is a precancerous condition. It is not easy to detect CAG in endoscopy. Improving the detection rate of CAG under endoscopy is essential to reduce or interrupt the occurrence of gastric cancer. This study aimed to construct a deep learning (DL) model for CAG recognition based on endoscopic images to improve the CAG detection rate during endoscopy.

Methods

We collected 10,961 endoscopic images and 118 video clips from 4,050 patients. For model training and testing, we divided them into two groups based on the pathological results: CAG and chronic non-atrophic gastritis (CNAG). We compared the performance of four state-of-the-art (SOTA) DL networks for CAG recognition and selected one of them for further improvement. The improved network was called GAM-EfficientNet. Finally, we compared GAM-EfficientNet with three endoscopists and analyzed the decision basis of the network in the form of heatmaps.

Results

After fine-tuning and transfer learning, the sensitivity, specificity, and accuracy of GAM-EfficientNet reached 93%, 94%, and 93.5% in the external test set and 96.23%, 89.23%, and 92.37% in the video test set, respectively, which were higher than those of the three endoscopists.

Conclusions

The CAG recognition model based on deep learning has high sensitivity and accuracy, and its performance is higher than that of endoscopists.

Mucosal imaging in colon polyps: New advances and what the future may hold

An expanding range of advanced mucosal imaging technologies have been developed with the goal of improving the detection and characterization of lesions in the gastrointestinal tract. Many technologies have targeted colorectal neoplasia given the potential for intervention prior to the development of invasive cancer in the setting of widespread surveillance programs. Improvement in adenoma detection reduces miss rates and prevents interval cancer development. Advanced imaging technologies aim to enhance detection without significantly increasing procedural time. Accurate polyp characterisation guides resection techniques for larger polyps, as well as providing the platform for the “resect and discard” and “do not resect” strategies for small and diminutive polyps. This review aims to collate and summarise the evidence regarding these technologies to guide colonoscopic practice in both interventional and non-interventional endoscopists.