Advances in image enhancement in colonoscopy for detection of adenomas

The influence of color vision deficiency on vessel visibility during colorectal endoscopic submucosal dissection and the potential advantage of red dichromatic imaging to achieve color vision barrier-free

Objectives

Although color information is important in gastrointestinal endoscopy, there are limited studies on how endoscopic images are viewed by people with color vision deficiency. We aimed to investigate the differences in the visibility of blood vessels during endoscopic submucosal dissection (ESD) among people with different color vision characteristics and to examine the effect of red dichromatic imaging (RDI) on blood vessel visibility.

Methods

Seventy-seven pairs of endoscopic images of white light imaging (WLI) and RDI of the same site were obtained during colorectal ESD. The original images were set as type C (WLI-C and RDI-C), a common color vision. These images were computationally converted to simulate images perceived by people with color vision deficiency protanope (Type P) or deutanope (Type D) and denoted as WLI-P and RDI-P or WLI-D and RDI-D. Blood vessels and background submucosa that needed to be identified during ESD were selected in each image, and the color differences between these two objects were measured using the color difference (ΔE 00) to assess the visibility of blood vessels.

Results

ΔE 00 between a blood vessel and the submucosa was greater under RDI (RDI-C/P/D: 24.05 ± 0.64/22.85 ± 0.66/22.61 ± 0.64) than under WLI (WLI-C/P/D: 22.26 ± 0.60/5.19 ± 0.30/8.62 ± 0.42), regardless of color vision characteristics. This improvement was more pronounced in Type P and Type D and approached Type C in RDI.

Conclusions

Color vision characteristics affect the visibility of blood vessels during ESD, and RDI improves blood vessel visibility regardless of color vision characteristics.

Risk-stratified screening and colorectal cancer incidence and mortality: A retrospective study from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial

OBJECTIVE

To determine whether risk stratification can optimize the benefits of flexible sigmoidoscopy (FSG) screening.

METHODS

The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial was conducted from 1993 to 2001 in the United States. A colorectal cancer (CRC) risk stratification tool was developed in the control arm (n = 64,207) from the PLCO cohort and validated in the UK Biobank (n = 270,726). PLCO participants (n = 130,021) were classified into low-, medium-, and high-risk groups. Cumulative incidence and mortality were estimated using the Kaplan-Meier method. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations between screening and CRC incidence and mortality.

RESULTS

The CRC risk stratification tool was based on age, gender, body mass index, smoking status, family history of CRC, diabetes, regular use of aspirin, and CRC screening history. Compared with the control arm, FSG screening was significantly associated with a reduction in mortality in both the medium-risk (HR = 0.76, 95% CI = 0.63-0.92) and high-risk groups (0.58, 0.46-0.73), but not in the low-risk group (0.85, 0.61-1.19). FSG screening also reduced distal CRC incidence and mortality in the medium-risk and high-risk groups. Furthermore, it was associated with a reduction in incidence (0.74, 0.59-0.92) and mortality (0.59, 0.40-0.87) of proximal colon cancer in the high-risk group.

CONCLUSIONS

FSG screening yielded more benefits for the high-risk group than for the low-risk and medium-risk groups, supporting the development of a risk-stratified CRC screening strategy.

Blue Laser Triggered Hemostatic Peptide Hydrogel for Gastrointestinal Bleeding Treatment

In an emergency, nonvariceal upper gastrointestinal bleeding (NVUGIB), endoscopic hemostasis is considered the gold standard intervention. However, current endoscopic hemostasis is very challenging to manage bleeding in large-diameter or deep lesions highly prone to rebleeding risk. Herein, a novel hemostatic peptide hydrogel (HPH) is reported, consisting of a self-assembly peptide sequence CFLIVIGSIIVPGDGVPGDG (PFV) and gelatin methacryloyl (GelMA), which can be triggered by blue laser endoscopy (BLE) for nonvariceal upper gastrointestinal bleeding treatment without recurring bleeding concerns. Upon contact with GelMA solution, PFV immediately fibrillates into β-sheet nanofiber and solvent-induced self-assembly to form HPH gel. HPH nanofiber networks induced ultrafast coagulation by enveloping blood cells and activating platelets and coagulation factors even to the blood with coagulopathy. Besides its remarkable hemostatic performance in artery and liver injury models, HPH achieves instant bleeding management in porcine NVUGIB models within 60 s by preventing the rebleeding risk. This work demonstrates an extraordinary hemostatic agent for NVUGIB intervention by BLE for the first time, broadening potential application scenarios, including patients with coagulopathy and promising clinical prospects.

The diagnostic yield of colonoscopic surveillance following resection of early age onset colorectal cancer

BACKGROUND

The primary benefit of post-colorectal cancer (CRC) colonoscopic surveillance is to detect and remove premalignant lesions to prevent metachronous CRC. Current guidelines for long-term colonoscopic surveillance post early age onset CRC (EOCRC) resection are based on limited evidence. The aims of this study were to assess the diagnostic yield of colonoscopic surveillance post-EOCRC resection and identify molecular and clinicopathological risk factors associated with advanced neoplasia.

METHODOLOGY

A retrospective cohort study of prospectively collected data was conducted at St Mark's hospital, London, United Kingdom, for patients diagnosed with EOCRC who underwent at least one episode of post-CRC colonoscopic surveillance between 1978 and 2022. We collected clinicopathological data including tumour molecular status and neoplasia detection rates.

RESULTS

In total, 908 colonoscopic surveillance procedures were performed in 195 patients over 2581.3 person-years of follow-up. The diagnostic yields of metachronous CRC, advanced adenomas and non-advanced adenomas were 1.76%, 3.41% and 22.69% respectively. Sixteen patients (8.21%) developed metachronous CRC, and the majority (87.5%) were detected more than 3 years post index EOCRC diagnosis. Detection of advanced neoplasia was significantly higher in EOCRC patients with Lynch syndrome (26.15%) compared with those in whom Lynch syndrome was excluded (13.13%) (OR, 2.343; 95% CI, 1.014-5.256; p = 0.0349).

CONCLUSIONS

During colonoscopic surveillance post-EOCRC resection, the long-term risk of developing metachronous advanced neoplasia remains high in the context of Lynch syndrome, but this trend is not as clearly evident when Lynch syndrome has been excluded.

Artificial intelligence-aided diagnostic imaging: A state-of-the-art technique in precancerous screening

BACKGROUND AND AIM

Chromoendoscopy with the use of indigo carmine (IC) dye is a crucial endoscopic technique to identify gastrointestinal neoplasms. However, its performance is limited by the endoscopist's skill, and no standards are available for lesion identification. Thus, we developed an artificial intelligence (AI) model to replace chromoendoscopy.

METHODS

This pilot study assessed the feasibility of our novel AI model in the conversion of white-light images (WLI) into virtual IC-dyed images based on a generative adversarial network. The predictions of our AI model were evaluated against the assessments of five endoscopic experts who were blinded to the purpose of this study with a staining quality rating from 1 (unacceptable) to 4 (excellent).

RESULTS

The AI model successfully transformed the WLI of polyps with different morphologies and different types of lesions in the gastrointestinal tract into virtual IC-dyed images. The quality ratings of the real IC-dyed and AI images did not significantly differ concerning surface structure (AI vs IC: 3.08 vs 3.00), lesion border (3.04 vs 2.98), and overall contrast (3.14 vs 3.02) from 10 sets of images (10 AI images and 10 real IC-dyed images). Although the score depended significantly on the evaluator, the staining methods (AI or real IC) and evaluators had no significant interaction (P > 0.05) with each other.

CONCLUSION

Our results demonstrated the feasibility of employing AI model's virtual IC staining, increasing the possibility of being employed in daily practice. This novel technology may facilitate gastrointestinal lesion identification in the future.

Optimization of colonoscopy quality: Comprehensive review of the literature and future perspectives

Colonoscopy is crucial in preventing colorectal cancer (CRC) and reducing associated mortality. This comprehensive review examines the importance of high-quality colonoscopy and associated quality indicators, including bowel preparation, cecal intubation rate, withdrawal time, adenoma detection rate (ADR), complete resection, specimen retrieval, complication rates, and patient satisfaction, while also discussing other ADR-related metrics. Additionally, the review draws attention to often overlooked quality aspects, such as nonpolypoid lesion detection, as well as insertion and withdrawal skills. Moreover, it explores the potential of artificial intelligence in enhancing colonoscopy quality and highlights specific considerations for organized screening programs. The review also emphasizes the implications of organized screening programs and the need for continuous quality improvement. A high-quality colonoscopy is crucial for preventing postcolonoscopy CRC- and CRC-related deaths. Health-care professionals must develop a thorough understanding of colonoscopy quality components, including technical quality, patient safety, and patient experience. By prioritizing ongoing evaluation and refinement of these quality indicators, health-care providers can contribute to improved patient outcomes and develop more effective CRC screening programs.

Usefulness of AI-Equipped Endoscopy for Detecting Colorectal Adenoma during Colonoscopy Screening: Confirm That Colon Neoplasm Finely Can Be Identified by AI without Overlooking Study (Confidential Study)

In the present prospective case series study, we investigated the lesion-detection ability of an AI-equipped colonoscopy as an addition to colonoscopy (CS) screening. Participants were 100 patients aged ≥20 years who had not undergone CS at the study site in the last 3 years and passed the exclusion criteria. CS procedures were conducted using conventional white light imaging and computer-aided detection (CADe). Adenoma detection rate (ADR; number of individuals with at least one adenoma detected) was compared between the conventional group and the CADe group. Of the 170 lesions identified, the ADR of the CADe group was significantly higher than the ADR of the conventional group (69% vs. 61%, p = 0.008). For the expert endoscopists, although ADR did not differ significantly, the mean number of detected adenomas per procedure (MAP) was significantly higher in the CADe group than in the conventional group (1.7 vs. 1.45, p = 0.034). For non-expert endoscopists, ADR and MAP were significantly higher in the CADe group than in the conventional group (ADR 69.5% vs. 56.6%, p = 0.016; MAP 1.66 vs. 1.11, p < 0.001). These results indicate that the CADe function in CS screening has a positive effect on adenoma detection, especially for non-experts.

Colorectal sessile serrated lesion detection using linked-color imaging versus narrow-band imaging: a parallel randomized controlled trial

BACKGROUND

 Previous studies have reported the effectiveness of narrow-band imaging (NBI) and linked-color imaging (LCI) in improving the detection of colorectal neoplasms. There has however been no direct comparison between LCI and NBI in the detection of colorectal sessile serrated lesions (SSLs). The present study aimed to compare the effectiveness of LCI and NBI in detecting colorectal SSLs.

METHODS

 A prospective, parallel, randomized controlled trial was conducted. The participants were randomly assigned to the LCI or NBI arms. The primary end point was the SSL detection rate (SDR).

RESULTS

 406 patients were involved; 204 in the LCI arm and 202 in the NBI arm. The total polyp detection rate, adenoma detection rate, and SDR were 54.2 %, 38.7 %, and 10.8%, respectively. The SDR was not significantly different between the LCI and NBI arms (12.3 % vs. 9.4 %; P = 0.36). The differences in the detection rate and the per-patient number of polyps, adenomas, diminutive lesions, and flat lesions between LCI and NBI also were not statistically significant. Multivariate analysis showed that LCI and NBI were not independent factors associated with SDR, whereas Boston Bowel Preparation Scale score (odds ratio [OR] 1.35, 95 %CI 1.03-1.76; P = 0.03), withdrawal time (OR 1.13, 95 %CI 1.00-1.26; P = 0.04), and operator experience (OR 3.73, 95 %CI 1.67-8.32; P = 0.001) were independent factors associated with SDR.

CONCLUSIONS

 LCI and NBI are comparable for SSL detection, as well as for the detection of polyps and adenomas.

Colorectal Sessile Serrated Lesion Detection Using Linked Color Imaging: A Multicenter, Parallel Randomized Controlled Trial

BACKGROUND & AIMS

Linked color imaging (LCI) is a novel technology that improves the color differences between colorectal lesions and the surrounding mucosa. The present study aims to compare the detection of colorectal sessile serrated lesions (SSL) using LCI with white light imaging (WLI).

METHOD

A large-scale, multicenter, parallel prospective randomized controlled trial was conducted in 4 hospitals in China. The participants were randomly assigned to the LCI group and WLI group. The primary endpoint was the SSL detection rate (SDR).

RESULTS

A total of 884 patients were involved in the intention-to-treat analysis, with 441 patients in the LCI group and 443 patients in the WLI group. The total polyp detection rate, adenoma detection rate, and SDR were 51.8%, 35.7%, and 8.6%, respectively. The SDR was significantly higher in the LCI group than in the WLI group (11.3% vs 5.9%, P = .004). Furthermore, LCI significantly increased the number of polyps and adenomas detected per patient, when compared with WLI (P < .05). In addition, there was higher detection rate of diminutive and flat lesions in the LCI group (P < .05). Multivariate analysis revealed that LCI is an independent factor associated with SDR (hazard ratio, 1.990; 95% confidence interval, 1.203-3.293; P = .007), along with withdrawal time (hazard ratio, 1.157; 95% confidence interval, 1.060-1.263; P = .001) and operator experience (hazard ratio, 1.850; 95% confidence interval, 1.045-3.273; P = .035).

CONCLUSIONS

LCI is significantly superior to WLI for SSL detection, and may improve polyp and adenoma detection. LCI can be recommended as an appropriate method for routine inspection during colonoscopy (http://www.chictr.org.cn number, ChiCTR2000035705).

Segmentation of polyps based on pyramid vision transformers and residual block for real-time endoscopy imaging

Polyp segmentation is an important task in early identification of colon polyps for prevention of colorectal cancer. Numerous methods of machine learning have been utilized in an attempt to solve this task with varying levels of success. A successful polyp segmentation method which is both accurate and fast could make a huge impact on colonoscopy exams, aiding in real-time detection, as well as enabling faster and cheaper offline analysis. Thus, recent studies have worked to produce networks that are more accurate and faster than the previous generation of networks (e.g., NanoNet). Here, we propose ResPVT architecture for polyp segmentation. This platform uses transformers as a backbone and far surpasses all previous networks not only in accuracy but also with a much higher frame rate which may drastically reduce costs in both real time and offline analysis and enable the widespread application of this technology.

Potential Factors Predicting Histopathologically Upgrade Discrepancies between Endoscopic Forceps Biopsy of the Colorectal Low-Grade Intraepithelial Neoplasia and Endoscopic Resection Specimens

Background

It was gradually accepted that endoscopic fragment biopsy (EFB) diagnosis cannot accurately guarantee the absence of higher-grade neoplasms within the lesion of the digestive tract. There are no well-established predictors for histopathologically upgrade discrepancies between EFB diagnosing colorectal low-grade intraepithelial neoplasia (LGIN) and endoscopic resection (ER) specimens.

Methods

A total of 918 colorectal LGINs was histopathologically diagnosed by EFB, including 162 cases with upgrade discrepancy and 756 concordant cases. We compared clinicopathological data of EFB and ER specimens between these two groups. Multivariate analysis was performed to identify predictors for this upgrade histopathology.

Results

The predominant upgrade discrepancy of LGINs diagnosed by EFB was upgrades to high-grade dysplasia (114/918, 12.4%), followed by upgrades to intramucosal carcinoma (33/918, 3.6%), submucosal adenocarcinoma (10/918, 1.1%), and advanced adenocarcinoma (5/918, 0.5%). NSAID history (OR 4.83; 95% CI, 2.27-10.27; p < 0.001), insufficient EFB number (OR 2.99; 95% CI, 1.91-4.68; p < 0.001), maximum diameter ≥ 1.0 cm (OR 6.18; 95% CI, 1.32-28.99; p = 0.021), lobulated shape (OR 2.68; 95% CI, 1.65-4.36; p < 0.001), erythema (OR 2.42; 95% CI, 1.50-3.91; p < 0.001), erosion (OR 7.12; 95% CI, 3.91-12.94; p < 0.001), surface unevenness (OR 2.31; 95% CI, 1.33-4.01; p = 0.003), and distal location of the target adenoma (OR 3.29; 95% CI, 1.68-6.41; p < 0.001) were associated with the histologically upgrade discrepancies.

Conclusion

NSAID history, insufficient EFB number, adenoma size and location, and abnormal macroscopic patterns are potential predictors for upgrade histopathology of LGINs diagnosed by EFBs. The standardization of EFB number and advanced imaging techniques could minimize the risk of neglecting the potential of this upgrade histopathology.

Endoscopy image enhancement method by generalized imaging defect models based adversarial training

Objective. Smoke, uneven lighting, and color deviation are common issues in endoscopic surgery, which have increased the risk of surgery and even lead to failure. Approach. In this study, we present a new physics model driven semi-supervised learning framework for high-quality pixel-wise endoscopic image enhancement, which is generalizable for smoke removal, light adjustment, and color correction. To improve the authenticity of the generated images, and thereby improve the network performance, we integrated specific physical imaging defect models with the CycleGAN framework. No ground-truth data in pairs are required. In addition, we propose a transfer learning framework to address the data scarcity in several endoscope enhancement tasks and improve the network performance. Main results. Qualitative and quantitative studies reveal that the proposed network outperforms the state-of-the-art image enhancement methods. In particular, the proposed method performs much better than the original CycleGAN, for example, the structural similarity improved from 0.7925 to 0.8648, feature similarity for color images from 0.8917 to 0.9283, and quaternion structural similarity from 0.8097 to 0.8800 in the smoke removal task. Experimental results of the proposed transfer learning method also reveal its superior performance when trained with small datasets of target tasks. Significance. Experimental results on endoscopic images prove the effectiveness of the proposed network in smoke removal, light adjustment, and color correction, showing excellent clinical usefulness.

Colorectal Cancer Surveillance in Patients with Inflammatory Bowel Diseases: Chromoendoscopy or Non-Chromoendoscopy, That Is the Question

Subjects affected by ulcerative colitis and Crohn’s disease with colonic localization have an increased risk of colorectal cancer (CRC). Surveillance colonoscopy is recommended by international guidelines as it can detect early-stage CRC. Based on previous evidence, in 2015 the Surveillance for Colorectal Endoscopic Neoplasia Detection and Management in Inflammatory Bowel Disease Patients International Consensus indicated dye chromoendoscopy (DCE) as the most effective technique for detecting dysplasia. However, advances in endoscopic technology such as high-definition colonoscopes and dye-less virtual chromoendoscopy (VCE) may change future practice. In this review, we summarize the available evidence on CRC surveillance in IBD, focusing on the emerging role of high-definition white light endoscopy (HD-WLE) and VCE over the standard DCE, and the current role of random biopsies.

Current status of diagnostic and therapeutic colonoscopy in Japan: The Japan Endoscopic Database Project

OBJECTIVES

The Japan Endoscopy Database Project was initiated to develop the world's largest endoscopy data repository. This study describes the first phase of the colonoscopy project in Japan.

METHODS

Data were aggregated offline by integrating information from the endoscopy database software from January 2015 through March 2017. The study population included all patients who underwent colonoscopy at eight centers.

RESULTS

A total of 31,395 patients who underwent 38,497 colonoscopy procedures were registered. The majority of procedures were performed for screening (n = 14,156), followed by fecal immunochemical test positivity (n = 3960), abdominal symptoms (n = 3864), post-colorectal surgery surveillance (n = 3431), post-endoscopic treatment surveillance (n = 3757), thorough pre-treatment examination (n = 2822), and therapeutic purposes (n = 6507). In the screening group, advanced cancers, early cancers, and adenomas were diagnosed endoscopically in 2.1%, 1.3%, and 28.7% of cases, respectively, while in the fecal immunochemical test-positive group, they were diagnosed in 2.5%, 1.9%, and 41.6% of cases, respectively. The incidence of complications was 0.177% and 0.152% in the screening and fecal immunochemical test-positive groups, respectively. The therapeutic procedures included 1446 cold forceps polypectomy procedures, 4770 cold snare polypectomy procedures, 368 hot biopsies, 2998 hot snare polypectomy procedures, 9775 endoscopic or piecemeal endoscopic mucosal resections, and 1660 endoscopic submucosal dissections. A total of 173 procedure-related complications (0.82%) occurred in 21,017 therapeutic procedures performed in 15,744 patients.

CONCLUSIONS

The first phase of the Japan Endoscopy Database Project established the proportions of the diagnostic and therapeutic colonoscopy procedures, and complication rates in real-world settings.

Thoughts on factors related to colonoscopy quality

With the increase in health consciousness and the prevalence of early gastrointestinal cancer screening, the number of diagnosed cases of colorectal polyps and adenomas in the Chinese general population has substantially increased. However, the overall adenoma detection rate (ADR) in China remains lower than the global standard of 30%. High quality colonoscopy is pivotal for increasing the ADR and reducing the risk of interval colon cancer. A few factors are related to the quality of colonoscopy, including preoperative evaluation of examinees, bowel preparation, techniques for intubation, use of image enhancement endoscopy, withdraw time, and prevention of adverse events.

A Comprehensive Study on Colorectal Polyp Segmentation With ResUNet++, Conditional Random Field and Test-Time Augmentation

Colonoscopy is considered the gold standard for detection of colorectal cancer and its precursors. Existing examination methods are, however, hampered by high overall miss-rate, and many abnormalities are left undetected. Computer-Aided Diagnosis systems based on advanced machine learning algorithms are touted as a game-changer that can identify regions in the colon overlooked by the physicians during endoscopic examinations, and help detect and characterize lesions. In previous work, we have proposed the ResUNet++ architecture and demonstrated that it produces more efficient results compared with its counterparts U-Net and ResUNet. In this paper, we demonstrate that further improvements to the overall prediction performance of the ResUNet++ architecture can be achieved by using Conditional Random Field (CRF) and Test-Time Augmentation (TTA). We have performed extensive evaluations and validated the improvements using six publicly available datasets: Kvasir-SEG, CVC-ClinicDB, CVC-ColonDB, ETIS-Larib Polyp DB, ASU-Mayo Clinic Colonoscopy Video Database, and CVC-VideoClinicDB. Moreover, we compare our proposed architecture and resulting model with other state-of-the-art methods. To explore the generalization capability of ResUNet++ on different publicly available polyp datasets, so that it could be used in a real-world setting, we performed an extensive cross-dataset evaluation. The experimental results show that applying CRF and TTA improves the performance on various polyp segmentation datasets both on the same dataset and cross-dataset. To check the model's performance on difficult to detect polyps, we selected, with the help of an expert gastroenterologist, 196 sessile or flat polyps that are less than ten millimeters in size. This additional data has been made available as a subset of Kvasir-SEG. Our approaches showed good results for flat or sessile and smaller polyps, which are known to be one of the major reasons for high polyp miss-rates. This is one of the significant strengths of our work and indicates that our methods should be investigated further for use in clinical practice.

Randomised comparison of postpolypectomy surveillance intervals following a two-round baseline colonoscopy: the Japan Polyp Study Workgroup

OBJECTIVE

To assess whether follow-up colonoscopy after polypectomy at 3 years only, or at 1 and 3 years would effectively detect advanced neoplasia (AN), including nonpolypoid colorectal neoplasms (NP-CRNs).

DESIGN

A prospective multicentre randomised controlled trial was conducted in 11 Japanese institutions. The enrolled participants underwent a two-round baseline colonoscopy (interval: 1 year) to remove all neoplastic lesions. Subsequently, they were randomly assigned to undergo follow-up colonoscopy at 1 and 3 years (2-examination group) or at 3 years only (1-examination group). The incidence of AN, defined as lesions with low-grade dysplasia ≥10 mm, high-grade dysplasia or invasive cancer, at follow-up colonoscopy was evaluated.

RESULTS

A total of 3926 patients were enrolled in this study. The mean age was 57.3 (range: 40-69) years, and 2440 (62%) were male. Of these, 2166 patients were assigned to two groups (2-examination: 1087, 1-examination: 1079). Overall, we detected 29 AN in 28 patients at follow-up colonoscopy in both groups. On per-protocol analysis (701 in 2-examination vs 763 in 1-examination group), the incidence of AN was similar between the two groups (1.7% vs 2.1%, p=0.599). The results of the non-inferiority test were significant (p=0.017 in per-protocol, p=0.001 in intention-to-treat analysis). NP-CRNs composed of dominantly of the detected AN (62%, 18/29), and most of them were classified into laterally spreading tumour non-granular type (83%, 15/18).

CONCLUSION

After a two-round baseline colonoscopy, follow-up colonoscopy at 3 years detected AN, including NP-CRNs, as effectively as follow-up colonoscopies performed after 1 and 3 years.

Will purposely seeking detect more colorectal polyps than routine performing during colonoscopy?

BACKGROUND & GOALS

We observed that the number of colorectal polyps found intraoperatively was often higher than that encountered preoperatively during elective colonoscopic polypectomy. To evaluate whether more polyps can be detected when they are purposely sought than when they are routinely examined during colonoscopy.

MATERIALS AND METHODS

Patients undergoing colonoscopy were randomized into groups A and B. Before colonoscopy was performed, endoscopists were instructed to seek polyps for group A purposely but not for group B. Polypectomy was electively completed. In groups A and B, the cases of elective polypectomy were named groups AR and BR, including groups AR-1 and BR-1, during the first colonoscopy and groups AR-2 and BR-2 during the second colonoscopy for polypectomy, respectively. The following data were calculated: the number of polyps detected (NPD) and the polyp detection rate (PDR) in all cases and the number of polyps missed (NPM) and partial polyp miss rate (PPMR) in the cases of colorectal polyps.

RESULTS

A total of 419 cases were included in group A, 421 in group B, 43 in group AR, and 35 in group BR. No significant differences in PDR were found between groups A and B and in PPMR between groups AR-1 and BR-1 (P > .05), although PPMR in group AR-1 was higher than in group AR-2 (P < .05), similar results were found in PPMR between groups BR-1 and BR-2 (P < .05).

CONCLUSION

Purposely seeking for colorectal polyps did not result in more polyps detected compared with routine colonoscopy.

Impact of new techniques on adenoma detection rate based on meta-analysis data

The high incidence of colorectal cancer and the occurrence of interval cancers after screening colonoscopy support the need to develop methods to increase adenoma detection rate (ADR). This review focuses on the importance of ADR and the impact of new techniques on ADR based on meta-analysis data. The low-cost interventions (such as water-aided colonoscopy, second observation, and dynamic position change) were effective in increasing ADR. So were enhanced imaging techniques and add-on devices. Increase with higher cost interventions such as newer scopes is uncertain. Water exchange (WE) has the highest ADR compared with water immersion, air insufflation, and carbon dioxide insufflation. Second observation with forward or retroflexed views improved the right colon ADR. Add-on devices result in only modest improvement in ADR, of particular help in low performing endoscopists. The second-generation narrow-band imaging (NBI) provided a two-fold brighter image than the previous system. The improvement in ADR with NBI required the “best” quality bowel preparation. New endoscopic techniques incur various additional costs, nil for WE, small for tip attachments but large for the newer scopes. In conclusion, one or more of the above methods to improve ADR may be applicable in Taiwan. A comparison of these approaches to determine which is the most cost-effective is warranted.

Compared Abilities of Endoscopic Techniques to Increase Colon Adenoma Detection Rates: A Network Meta-analysis

BACKGROUND AND AIMS

Adenoma detection rate (ADR) is a quality metric for colorectal cancer screening. We performed a systematic review and network meta-analysis to assess the overall and comparative efficacies of different endoscopic techniques in adenoma detection.

METHODS

We performed a systematic review of published articles and abstracts, through March 15, 2018, to identify randomized controlled trials of adults undergoing colonoscopy that compared the efficacy of different devices in detection of adenomas. Our final analysis included 74 2-arm trials that comprised 44948 patients. These studies compared efficacies of add-on devices (cap, endocuff, endo-rings, G-EYE), enhanced imaging techniques (chromoendoscopy, narrow-band imaging, flexible spectral imaging color enhancement, blue laser imaging), new scopes (full-spectrum endoscopy, extra-wide-angle-view colonoscopy, dual focus), and low-cost optimizing existing resources (water-aided colonoscopy, second observer, dynamic position change), alone or in combination with high-definition colonoscopy or each other. Primary outcome was increase in ADR. We performed pairwise and network meta-analyses, and appraised quality of evidence using GRADE.

RESULTS

Low-cost optimizing existing resources (odds ratio [OR], 1.29; 95% CI,1.17-1.43), enhanced imaging techniques (OR,1.21; 95% CI, 1.09-1.35), and add-on devices (OR,1.18; 95% CI, 1.07-1.29) were associated with a moderate increase in ADR compared with high-definition colonoscopy; there was low to moderate confidence in estimates. Use of newer scopes was not associated with significant increases in ADR compared with high-definition colonoscopy (OR, 0.98; 95% CI, 0.79-1.21). In our comparative efficacy analysis, no specific technology for increasing ADR was superior to others. We did not find significant differences between technologies in detection of advanced ADR, polyp detection rate, or mean number of adenomas/patient.

CONCLUSIONS

In a network meta-analysis of published trials, we found that low-cost optimization of existing resources to be as effective as enhanced endoscopic imaging, or add-on devices, in increasing ADR during high-definition colonoscopy.

Diagnostic utility of linked color imaging in the evaluation of colonic mucosal inflammation in ulcerative colitis: a pilot study

Background and aims  Recent studies recommend histological mucosal healing of intestinal tissue as a treatment goal in ulcerative colitis (UC). Linked-color imaging (LCI) is a new endoscopy system that enhances the color differences of the gastrointestinal mucosa. We investigated the efficacy of LCI in the evaluation of intestinal activity, including the histological activity in UC. Method  A total of 21 UC patients who were evaluated by the LASEREO system (FUJIFILM Co., Tokyo, Japan) were enrolled from August to December in 2016. All of the target points were observed by conventional white-light imaging (WLI) and LCI and biopsied from the region of interest in each view. We quantified the color tones of 73 biopsied points on WLI and LCI using the L*a*b* color value (WLI-L, WLI-a, WLI-b, and LCI-L, LCI-a, LCI-b). We then investigated the relationships among the L*a*b* color values, endoscopic findings, and histological healing. Results  The average LCI-a and LCI-b values of patients with high mucosal activity disease were significantly higher than in those with mucosal healing ( P  < 0.01), and only LCI-a was significantly correlated with the score for histological healing. With regard to the other color values (LCI-L and all WLI values), there were no statistically significant differences in terms of average color value and correlation between patients with high mucosal activity and those with mucosal healing. In addition, the difference in the average color values with LCI and WLI was 7.1 and 3.1, respectively. Conclusion  LCI is more useful than WLI for the visualization and evaluation of mucosal inflammation in UC.

Automated Counting of Rice Panicle by Applying Deep Learning Model to Images from Unmanned Aerial Vehicle Platform

The number of panicles per unit area is a common indicator of rice yield and is of great significance to yield estimation, breeding, and phenotype analysis. Traditional counting methods have various drawbacks, such as long delay times and high subjectivity, and they are easily perturbed by noise. To improve the accuracy of rice detection and counting in the field, we developed and implemented a panicle detection and counting system that is based on improved region-based fully convolutional networks, and we use the system to automate rice-phenotype measurements. The field experiments were conducted in target areas to train and test the system and used a rotor light unmanned aerial vehicle equipped with a high-definition RGB camera to collect images. The trained model achieved a precision of 0.868 on a held-out test set, which demonstrates the feasibility of this approach. The algorithm can deal with the irregular edge of the rice panicle, the significantly different appearance between the different varieties and growing periods, the interference due to color overlapping between panicle and leaves, and the variations in illumination intensity and shading effects in the field. The result is more accurate and efficient recognition of rice-panicles, which facilitates rice breeding. Overall, the approach of training deep learning models on increasingly large and publicly available image datasets presents a clear path toward smartphone-assisted crop disease diagnosis on a global scale.

Comparison of the detection of colorectal lesions in different endoscopic modalities: A network meta-analysis and systematic review

A colonoscopy is considered to be the standard diagnostic test used to detect early colorectal lesions. Detection rates are expected to improve with optimised visualisation. A systematic review and network meta-analysis was conducted to evaluate detection efficiency in several colonoscopic modalities. Relevant articles were identified in searches of the PubMed, EMBASE and Cochrane Library databases. The modalities, comprising of standard-definition white light (SDWL), high-definition white light (HDWL), narrow-band imaging (NBI), autofluorescence imaging (AFI), PENTAX image enhanced technology (i-SCAN), Fuji Intelligent Color Enhancement (FICE), dye-based chromoendoscopy and novel image enhanced systems, including blue laser imaging (BLI) and linked color imaging (LCI), were compared to identify the most efficient modalities that could be used to detect colorectal lesions. Odds ratios (ORs) and mean differences (MDs) with 95% confidence intervals (CIs) were calculated. As a result, 40 studies fulfilled the inclusion criteria. Overall, in the network meta-analyses, NBI (OR, 1.29; 95% CI, 1.04-1.58), FICE (OR, 1.39; 95% CI, 1.11-1.77), chromoendoscopy (OR, 1.53; 95% CI, 1.22-1.93) and AFI (OR, 1.81; 95% CI, 1.07-2.87) were significantly better compared with SDWL at identifying adenoma in patients, and chromoendoscopy also proved significantly superior to HDWL (OR, 1.30; 95% CI, 1.06-1.60). In pairwise analyses, it was demonstrated that chromoendoscopy was significantly superior to HDWL at detecting the number of polyps (MD, -1.11; 95% CI, -1.46, -0.76) and flat lesions (MD, -0.30; 95% CI, -0.49, -0.10) per subject. Additionally, FICE detected a significantly greater number of subjects with polyps (OR, 0.78; 95% CI, 0.64-0.96) and NBI was significantly better at detecting the number of subjects with flat lesions (OR, 0.77; 95% CI, 0.60-0.99) compared with HDWL. Based on the meta-analysis, NBI, FICE and AFI were significantly better compared with SDWL at detecting patients with adenoma. Additionally, chromoendoscopy was significantly better than SDWL and HDWL at detecting the number of colorectal adenoma, however additional studies are needed to confirm these findings.

How I do it: Flexible 3-D endoscope for endoscopic submucosal dissection

Endoscopic submucosal dissection (ESD) is technically challenging as a result of a lack of depth perception. The present article investigated the 3-D endoscope for carrying out ESD and translated the technique from bench to clinical use. In a preclinical porcine experiment, ESD using a 3-D endoscope was compared between an experienced and a novice endoscopist. All ESD were completed without perforation. Median operative time per surface area was significantly lower for the experienced endoscopist than for the novice (197.9 s/cm² vs 434.7 s/cm² ; P = 0.05). The second part was a prospective clinical experience to evaluate use of the 3-D endoscope for carrying out ESD. Ten patients received ESD using the 3-D endoscope. Four patients had gastric ESD, two had duodenal ESD and four had sigmoid and rectal ESD. There were no complications, whereas ESD failed in one patient who had gastric neoplasia at anastomosis. Mean operative time was 99.4 min, and operative time per surface area resection was 391 s/cm² . The operating endoscopist did not complain of motion sickness, whereas the assistants had some dizziness upon prolonged ESD procedure. This study showed that carrying out ESD was safe and effective using a 3-D endoscope with an excellent 3-D view enhancing depth perception. Future study should be conducted to compare 3-D against 2-D endoscopes for ESD.

An Additional 30-s Observation of the Right-Sided Colon with Narrow Band Imaging Decreases Missed Polyps: A Pilot Study

INTRODUCTION

Previous narrow-band imaging (NBI) was dark and reported not to be useful for polyp detection. In this study, we analyzed the efficacy of an additional 30-s observation of the right-sided colon with the recent bright high-resolution NBI.

METHODS

We enrolled patients undergoing colonoscopy from February 2015 to May 2017 in two institutions. All procedures were performed with the latest system (EVIS LUCERA ELITE, Olympus). The cecum and ascending colon were first observed with white light imaging (WLI) in both the NBI and WLI group. Then, the colonoscope was re-inserted, and the cecum and ascending colon were observed for an additional 30 s. In this second observation, NBI was performed for the first 130 patients in the NBI group and WLI for the next 130 in the WLI group. The number of adenoma and sessile serrated polyps (ASPs) in the second observation were examined in both groups. According to our initial pilot study, the sample size was estimated at 126.

RESULTS

In the first observation, the number of ASPs was 72 in the NBI group and 72 in the WLI group (p = 1.0). In the second observation, the number of ASPs was 23 in the NBI group and 10 in the WLI group (p = 0.02). The polyp and adenoma detection rates in the second observation were 16.2% and 12.3% in the NBI group and 7.7% (p = 0.03) and 6.2% (p = 0.09) in the WLI group.

CONCLUSIONS

The additional 30-s observation with recent NBI decreased missed polyps in the right-sided colon.

Number of Adenomas Removed and Colorectal Cancers Prevented in Randomized Trials of Flexible Sigmoidoscopy Screening

BACKGROUND & AIMS

Screening for colorectal cancer (CRC) with sigmoidoscopy reduces CRC incidence by detecting and removing adenomas. The number needed to screen is a measure of screening efficiency, but is not directly associated with adenoma removal. We propose the following 2 new metrics for quantifying the relationship between adenoma removal and CRC prevented: number of adenomas needed to remove (NNR) and adenoma dwell time avoided (DTA).

METHODS

We collected data from 4 randomized trials of sigmoidoscopy screening (1 in the United States and 3 in Europe) to assess NNR and DTA. For each trial, NNR was computed as the number of adenomas removed from subjects in the intervention group, divided by the number of CRCs prevented. DTA was computed similarly but taking into account the timing of adenoma removal. Combined results across trials were assessed using standard meta-analytic techniques.

RESULTS

The estimated NNR for the PLCO (Prostate, Lung, Colorectal and Ovarian) trial was 74 (95% confidence interval [CI], 56-110), for the NORCCAP (Norwegian Colorectal Cancer Prevention) trial was 71 (95% CI, 44-174), for the SCORE (Screening for Colon Rectum) trial was 27 (95% CI, 14-135), and for the UKFSST (UK Flexible Sigmoidoscopy Screening Trial) was 36 (95% CI, 28-52). The combined estimate (meta-analysis) of NNR was 52 (95% CI, 36-93) assuming heterogeneity (P for heterogeneity = .014). DTA estimates among trials ranged from 278 to 730 years, with a combined estimate of 500 (95% CI, 344-833) years assuming heterogeneity (P for heterogeneity = .035), or 2 CRC cases prevented per 1000 adenoma dwell years avoided. The combined estimates of NNR and DTA restricted to advanced adenomas were 13 (95% CI, 9-22) and 122 (95% CI, 90-190) years, respectively.

CONCLUSIONS

We collected data from 4 randomized trials of sigmoidoscopy screening for CRC to develop metrics of endoscopic efficiency, NNR and DTA, which are directly linked to adenoma detection and removal. They can be used to compare screening among endoscopic modalities and to more precisely measure adenoma to carcinoma transition rates.

New technologies improve adenoma detection rate, adenoma miss rate, and polyp detection rate: a systematic review and meta-analysis

BACKGROUND AND AIMS

The need to increase the adenoma detection rate (ADR) for colorectal cancer screening has ushered in devices that mechanically or optically improve conventional colonoscopy. Recently, new technology devices (NTDs) have become available. We aimed to compare the ADR, polyp detection rate (PDR), and adenoma miss rate (AMR) between NTDs and conventional colonoscopy and between mechanical and optical NTDs.

METHODS

MEDLINE and Embase databases were searched from inception through September 2017 for articles or abstracts reporting ADR, PDR, and AMR with NTDs. Randomized controlled trials and case-control studies with >10 subjects were included. Primary outcomes included ADR, PDR, and AMR odds ratio (OR) between conventional colonoscopy and NTDs. Secondary outcomes included cecal intubation rates, adverse events, cecal intubation time, and total colonoscopy time.

RESULTS

From 141 citations, 45 studies with 20,887 subjects were eligible for ≥1 analyses. Overall, the ORs for ADR (1.35; 95% confidence interval [CI] 1.24-1.47; P < .01) and PDR (1.51; 95% CI, 1.37-1.67; P < .01) were higher with NTDs. Higher ADR (OR, 1.52 vs 1.25; P = .035) and PDR (OR, 1.63 vs 1.10; P ≤ .01) were observed with mechanical NTDs. The overall AMR with NTDs was lower compared with conventional colonoscopy (OR, .19; 95% CI, .14-.26; P < .01). Mechanical NTDs had lower AMRs compared with optical NTDs (OR, .10 vs .33; P < .01). No differences in cecal intubation rates, cecal intubation time, or total colonoscopy time were found.

CONCLUSIONS

Newer endoscopic technologies are an effective option to improve ADR and PDR and decrease AMR, particularly with mechanical NTDs. No differences in operability and safety were found.

Villous Gastrointestinal Tumors: Multimodality Imaging with Histopathologic Correlation

Villous lesions are advanced adenomas that manifest most commonly in the colon; however, they can develop throughout the gastrointestinal tract. The duodenum is the most common small-bowel site of these lesions. Although in most cases these are isolated lesions that occur sporadically, patients with certain specific colorectal cancer syndromes, including familial adenomatous polyposis and hereditary nonpolyposis colorectal cancer, may develop multiple advanced adenomas. Villous lesions are important because although they are histologically benign, they may harbor dysplasia and have potential for malignancy. These characteristics make them a primary target for colorectal cancer screening with optical and virtual colonoscopy. However, these lesions can also be symptomatic and detected at diagnostic imaging when patients present for examination. They have characteristic features at a variety of imaging examinations, including barium fluoroscopy, CT, MRI, and endoscopic US. It is important for radiologists to be aware of these lesions, their potential morphologies, and their typical appearances at multimodality imaging. Although villous tumors can be detected at imaging and confirmed with biopsy, owing to limitations in identifying dysplasia and foci of malignancy with the above modalities alone and the potential for malignancy, referral for surgical resection of these lesions ultimately is required. ^©RSNA, 2018.

Dual-modality endoscopic probe for tissue surface shape reconstruction and hyperspectral imaging enabled by deep neural networks

Surgical guidance and decision making could be improved with accurate and real-time measurement of intra-operative data including shape and spectral information of the tissue surface. In this work, a dual-modality endoscopic system has been proposed to enable tissue surface shape reconstruction and hyperspectral imaging (HSI). This system centers around a probe comprised of an incoherent fiber bundle, whose fiber arrangement is different at the two ends, and miniature imaging optics. For 3D reconstruction with structured light (SL), a light pattern formed of randomly distributed spots with different colors is projected onto the tissue surface, creating artificial texture. Pattern decoding with a Convolutional Neural Network (CNN) model and a customized feature descriptor enables real-time 3D surface reconstruction at approximately 12 frames per second (FPS). In HSI mode, spatially sparse hyperspectral signals from the tissue surface can be captured with a slit hyperspectral imager in a single snapshot. A CNN based super-resolution model, namely "super-spectral-resolution" network (SSRNet), has also been developed to estimate pixel-level dense hypercubes from the endoscope cameras standard RGB images and the sparse hyperspectral signals, at approximately 2 FPS. The probe, with a 2.1 mm diameter, enables the system to be used with endoscope working channels. Furthermore, since data acquisition in both modes can be accomplished in one snapshot, operation of this system in clinical applications is minimally affected by tissue surface movement and deformation. The whole apparatus has been validated on phantoms and tissue (ex vivo and in vivo), while initial measurements on patients during laryngeal surgery show its potential in real-world clinical applications.

Advanced diagnostic endoscopy in the lower gastrointestinal tract: A review of JGES core sessions

At each of the 89th to the 92nd congresses of the Japan Gastroenterological Endoscopy Society, a series of featured discussion sessions concerning advanced diagnostic endoscopy in the lower gastrointestinal tract were presented. In total, 45 lectures were presented in this subject area. It was shown that, in recent years, several convenient and less invasive colonoscopic modalities have been developed. This review article summarizes these core sessions and the efficacy of the techniques discussed.