Evaluation of Spectrum-Aided Visual Enhancer (SAVE) in Esophageal Cancer Detection Using YOLO Frameworks.

The early detection of esophageal cancer presents a substantial difficulty, which contributes to its status as a primary cause of cancer-related fatalities. This study used You Only Look Once (YOLO) frameworks, specifically YOLOv5 and YOLOv8, to predict and detect early-stage EC by using a dataset sourced from the Division of Gastroenterology and Hepatology, Ditmanson Medical Foundation, Chia-Yi Christian Hospital. The dataset comprised 2741 white-light images (WLI) and 2741 hyperspectral narrowband images (HSI-NBI). They were divided into 60% training, 20% validation, and 20% test sets to facilitate robust detection. The images were produced using a conversion method called the spectrum-aided vision enhancer (SAVE). This algorithm can transform a WLI into an NBI without requiring a spectrometer or spectral head. The main goal was to identify dysplasia and squamous cell carcinoma (SCC). The model's performance was evaluated using five essential metrics: precision, recall, F1-score, mAP, and the confusion matrix. The experimental results demonstrated that the HSI model exhibited improved learning capabilities for SCC characteristics compared with the original RGB images. Within the YOLO framework, YOLOv5 outperformed YOLOv8, indicating that YOLOv5's design possessed superior feature-learning skills. The YOLOv5 model, when used in conjunction with HSI-NBI, demonstrated the best performance. It achieved a precision rate of 85.1% (CI95: 83.2-87.0%, p < 0.01) in diagnosing SCC and an F1-score of 52.5% (CI95: 50.1-54.9%, p < 0.01) in detecting dysplasia. The results of these figures were much better than those of YOLOv8. YOLOv8 achieved a precision rate of 81.7% (CI95: 79.6-83.8%, p < 0.01) and an F1-score of 49.4% (CI95: 47.0-51.8%, p < 0.05). The YOLOv5 model with HSI demonstrated greater performance than other models in multiple scenarios. This difference was statistically significant, suggesting that the YOLOv5 model with HSI significantly improved detection capabilities.

Multi-Instance Learning for Vocal Fold Leukoplakia Diagnosis Using White Light and Narrow-Band Imaging: A Multicenter Study.

OBJECTIVES: Vocal fold leukoplakia (VFL) is a precancerous lesion of laryngeal cancer, and its endoscopic diagnosis poses challenges. We aim to develop an artificial intelligence (AI) model using white light imaging (WLI) and narrow-band imaging (NBI) to distinguish benign from malignant VFL. METHODS: A total of 7057 images from 426 patients were used for model development and internal validation. Additionally, 1617 images from two other hospitals were used for model external validation. Modeling learning based on WLI and NBI modalities was conducted using deep learning combined with a multi-instance learning approach (MIL). Furthermore, 50 prospectively collected videos were used to evaluate real-time model performance. A human-machine comparison involving 100 patients and 12 laryngologists assessed the real-world effectiveness of the model. RESULTS: The model achieved the highest area under the receiver operating characteristic curve (AUC) values of 0.868 and 0.884 in the internal and external validation sets, respectively. AUC in the video validation set was 0.825 (95% CI: 0.704-0.946). In the human-machine comparison, AI significantly improved AUC and accuracy for all laryngologists (p < 0.05). With the assistance of AI, the diagnostic abilities and consistency of all laryngologists improved. CONCLUSIONS: Our multicenter study developed an effective AI model using MIL and fusion of WLI and NBI images for VFL diagnosis, particularly aiding junior laryngologists. However, further optimization and validation are necessary to fully assess its potential impact in clinical settings. LEVEL OF EVIDENCE: 3 Laryngoscope, 2024.

Topic highlight on texture and color enhancement imaging in gastrointestinal diseases.

Olympus Corporation developed texture and color enhancement imaging (TXI) as a novel image-enhancing endoscopic technique. This topic highlights a series of hot-topic articles that investigated the efficacy of TXI for gastrointestinal disease identification in the clinical setting. A randomized controlled trial demonstrated improvements in the colorectal adenoma detection rate (ADR) and the mean number of adenomas per procedure (MAP) of TXI compared with those of white-light imaging (WLI) observation (58.7% vs 42.7%, adjusted relative risk 1.35, 95%CI: 1.17-1.56; 1.36 vs 0.89, adjusted incident risk ratio 1.48, 95%CI: 1.22-1.80, respectively). A cross-over study also showed that the colorectal MAP and ADR in TXI were higher than those in WLI (1.5 vs 1.0, adjusted odds ratio 1.4, 95%CI: 1.2-1.6; 58.2% vs 46.8%, 1.5, 1.0-2.3, respectively). A randomized controlled trial demonstrated non-inferiority of TXI to narrow-band imaging in the colorectal mean number of adenomas and sessile serrated lesions per procedure (0.29 vs 0.30, difference for non-inferiority -0.01, 95%CI: -0.10 to 0.08). A cohort study found that scoring for ulcerative colitis severity using TXI could predict relapse of ulcerative colitis. A cross-sectional study found that TXI improved the gastric cancer detection rate compared to WLI (0.71% vs 0.29%). A cross-sectional study revealed that the sensitivity and accuracy for active Helicobacter pylori gastritis in TXI were higher than those of WLI (69.2% vs 52.5% and 85.3% vs 78.7%, respectively). In conclusion, TXI can improve gastrointestinal lesion detection and qualitative diagnosis. Therefore, further studies on the efficacy of TXI in clinical practice are required.

Narrow band imaging reveals field cancerisation undetected by conventional white light: Optical diagnosis versus histopathology.

OBJECTIVE: To assess whether narrow band imaging (NBI) detects fields of cancerisation around suspicious lesions in the upper aerodigestive tract, which were undetected by white light imaging (WLI). METHODS: In 96 patients with laryngeal and pharyngeal lesions suspicious for malignancy, 206 biopsies were taken during laryngoscopy: 96 biopsies of suspicious lesions detected by both WLI and NBI (WLI+/NBI+), 60 biopsies adjacent mucosa only suspicious with NBI (WLI-/NBI+), and 46 biopsies of NBI and WLI unsuspicious mucosa (WLI-/NBI-) as negative controls. Optical diagnosis according to the Ni-classification was compared with histopathology. RESULTS: Signs of (pre)malignancy were found in 88% of WLI+/NBI+ biopsies, 32% of WLI-/NBI+ biopsies and 0% in WLI-/NBI- (p < .001). In 58% of the WLI-/NBI+ mucosa any form of dysplasia or carcinoma was detected. CONCLUSION: The use of additional NBI led to the detection of (pre)malignancy in 32% of the cases, that would have otherwise remained undetected with WLI alone. This highlights the potential of NBI as a valuable adjunct to WLI in the identification of suspicious lesions in the upper aerodigestive tract.

Real-Time Laryngeal Cancer Boundaries Delineation on White Light and Narrow-Band Imaging Laryngoscopy with Deep Learning.

OBJECTIVE: To investigate the potential of deep learning for automatically delineating (segmenting) laryngeal cancer superficial extent on endoscopic images and videos. METHODS: A retrospective study was conducted extracting and annotating white light (WL) and Narrow-Band Imaging (NBI) frames to train a segmentation model (SegMENT-Plus). Two external datasets were used for validation. The model's performances were compared with those of two otolaryngology residents. In addition, the model was tested on real intraoperative laryngoscopy videos. RESULTS: A total of 3933 images of laryngeal cancer from 557 patients were used. The model achieved the following median values (interquartile range): Dice Similarity Coefficient (DSC) = 0.83 (0.70-0.90), Intersection over Union (IoU) = 0.83 (0.73-0.90), Accuracy = 0.97 (0.95-0.99), Inference Speed = 25.6 (25.1-26.1) frames per second. The external testing cohorts comprised 156 and 200 images. SegMENT-Plus performed similarly on all three datasets for DSC (p = 0.05) and IoU (p = 0.07). No significant differences were noticed when separately analyzing WL and NBI test images on DSC (p = 0.06) and IoU (p = 0.78) and when analyzing the model versus the two residents on DSC (p = 0.06) and IoU (Senior vs. SegMENT-Plus, p = 0.13; Junior vs. SegMENT-Plus, p = 1.00). The model was then tested on real intraoperative laryngoscopy videos. CONCLUSION: SegMENT-Plus can accurately delineate laryngeal cancer boundaries in endoscopic images, with performances equal to those of two otolaryngology residents. The results on the two external datasets demonstrate excellent generalization capabilities. The computation speed of the model allowed its application on videolaryngoscopies simulating real-time use. Clinical trials are needed to evaluate the role of this technology in surgical practice and resection margin improvement. LEVEL OF EVIDENCE: III Laryngoscope, 134:2826-2834, 2024.

The Detection of Nasopharyngeal Carcinomas Using a Neural Network Based on Nasopharyngoscopic Images.

OBJECTIVE: To construct and validate a deep convolutional neural network (DCNN)-based artificial intelligence (AI) system for the detection of nasopharyngeal carcinoma (NPC) using archived nasopharyngoscopic images. METHODS: We retrospectively collected 14107 nasopharyngoscopic images (7108 NPCs and 6999 noncancers) to construct a DCNN model and prepared a validation dataset containing 3501 images (1744 NPCs and 1757 noncancers) from a single center between January 2009 and December 2020. The DCNN model was established using the You Only Look Once (YOLOv5) architecture. Four otolaryngologists were asked to review the images of the validation set to benchmark the DCNN model performance. RESULTS: The DCNN model analyzed the 3501 images in 69.35 s. For the validation dataset, the precision, recall, accuracy, and F1 score of the DCNN model in the detection of NPCs on white light imaging (WLI) and narrow band imaging (NBI) were 0.845 ± 0.038, 0.942 ± 0.021, 0.920 ± 0.024, and 0.890 ± 0.045, and 0.895 ± 0.045, 0.941 ± 0.018, and 0.975 ± 0.013, 0.918 ± 0.036, respectively. The diagnostic outcome of the DCNN model on WLI and NBI images was significantly higher than that of two junior otolaryngologists (p < 0.05). CONCLUSION: The DCNN model showed better diagnostic outcomes for NPCs than those of junior otolaryngologists. Therefore, it could assist them in improving their diagnostic level and reducing missed diagnoses. LEVEL OF EVIDENCE: 3 Laryngoscope, 134:127-135, 2024.

Optical imaging for early detection of cervical cancer: state of the art and perspectives.

Significance: Cervical cancer is one of the major causes of death in females worldwide. HPV infection is the key cause of uncontrolled cell growth leading to cervical cancer. About 90% of cervical cancer is preventable because of the slow progression of the disease, giving a window of about 10 years for the precancerous lesion to be recognized and treated. Aim: The present challenges for cervical cancer diagnosis are interobserver variation in clinicians' interpretation of visual inspection with acetic acid/visual inspection with Lugol's iodine, cost of cytology-based screening, and lack of skilled clinicians. The optical modalities can assist in qualitatively and quantitatively analyzing the tissue to differentiate between cancerous and surrounding normal tissues. Approach: This work is on the recent advances in optical techniques for cervical cancer diagnosis, which promise to overcome the above-listed challenges faced by present screening techniques. Results: The optical modalities provide substantial measurable information in addition to the conventional colposcopy and Pap smear test to clinically aid the diagnosis. Conclusions: Recent optical modalities on fluorescence, multispectral imaging, polarization-sensitive imaging, microendoscopy, Raman spectroscopy, especially with the portable design and assisted by artificial intelligence, have a significant scope in the diagnosis of premalignant cervical cancer in future.

Efficacy of image-enhanced endoscopy for colorectal adenoma detection: A multicenter, randomized trial.

BACKGROUND: Improved adenoma detection at colonoscopy has decreased the risk of developing colorectal cancer. However, whether image-enhanced endoscopy (IEE) further improves the adenoma detection rate (ADR) is controversial. AIM: To compare IEE with white-light imaging (WLI) endoscopy for the detection and identification of colorectal adenoma. METHODS: This was a multicenter, randomized, controlled trial. Participants were enrolled between September 2019 to April 2021 from 4 hospital in China. Patients were randomly assigned to an IEE group with WLI on entry and IEE on withdrawal (n = 2113) or a WLI group with WLI on both entry and withdrawal (n = 2098). The primary outcome was the ADR. The secondary endpoints were the polyp detection rate (PDR), adenomas per colonoscopy, adenomas per positive colonoscopy, and factors related to adenoma detection. RESULTS: A total of 4211 patients (966 adenomas) were included in the analysis (mean age, 56.7 years, 47.1% male). There were 2113 patients (508 adenomas) in the IEE group and 2098 patients (458 adenomas) in the WLI group. The ADR in two group were not significantly different [24.0% vs 21.8%, 1.10, 95% confidence interval (CI): 0.99-1.23, P = 0.09]. The PDR was higher with IEE group (41.7%) than with WLI group (36.1%, 1.16, 95%CI: 1.07-1.25, P = 0.01). Differences in mean withdrawal time (7.90 ± 3.42 min vs 7.85 ± 3.47 min, P = 0.30) and adenomas per colonoscopy (0.33 ± 0.68 vs 0.28 ± 0.62, P = 0.06) were not significant. Subgroup analysis found that with narrow-band imaging (NBI), between-group differences in the ADR, were not significant (23.7% vs 21.8%, 1.09, 95%CI: 0.97-1.22, P = 0.15), but were greater with linked color imaging (30.9% vs 21.8%, 1.42, 95%CI: 1.04-1.93, P = 0.04). the second-generation NBI (2G-NBI) had an advantage of ADR than both WLI and the first-generation NBI (27.0% vs 21.8%, P = 0.01; 27.0% vs 21.2.0%, P = 0.01). CONCLUSION: This prospective study confirmed that, among Chinese, IEE didn't increase the ADR compared with WLI, but 2G-NBI increase the ADR.

Analysis of white-light imaging-based features predictive for determination of lesion depths of superficial flat esophageal squamous cell carcinoma: a retrospective multicenter study from China.

OBJECTIVES: Endoscopic diagnosis of invasion depth of superficial esophageal squamous cell carcinoma (SESCC) by white-light imaging (WLI) modality remains difficult. This study aims to clarify WLI-based features which are predictive for invasion depth of SESCC. METHODS: A two-phase study was performed by enrolling 1288 patients with 1396 SESCC lesions. Endoscopic appearances, clinical characteristics and post-operative pathological outcomes were collected and reviewed. The association between lesion features and invasion depth were analyzed. A predictive nomogram was constructed for prediction of invasion depth. RESULTS: Among 1396 lesions in derivation and validation cohort, 1139 (81.6%), 194 (13.9%) and 63 (4.5%) lesions were diagnosed as lesions confined into the intraepithelium or the lamina propria mucosa (T1a-EP/LPM), lesions invading the muscularis mucosa (T1a-MM) or superficial submucosa (T1b-SM1) and tumor with moderate invasion into the submucosa or deeper submucosal invasion (≥ T1b-SM2), respectively. Lesion length > 2 cm (p < 0.001), wider circumferential extension (p < 0.001, 0.002 and 0.048 for > 3/4, 1/2-3/4 and 1/4-1/2 circumferential extension, respectively), surface unevenness (p < 0.001 for both type 0-IIa/0-IIc lesions and mixed type lesions), spontaneous bleeding (p < 0.001), granularity (p < 0.001) and nodules (p < 0.001) were identified as significant factors predictive for lesion depth. A nomogram based on these factors was constructed and the values of area under the Receiver Operating Characteristics curve were 0.89 and 0.90 in the internal and external patient cohort. CONCLUSIONS: Our study provides six WLI-based morphological features predicting for lesion depth of SESCC. Our findings will make endoscopic evaluation of invasion depth for SESCC more convenient by assessing these profiles.

A novel intelligent chromo capsule endoscope for the diagnosis of neoplastic lesions in the gastrointestinal tract.

Background: Chromoendoscopy has not been fully integrated into capsule endoscopy. This study aimded to develop and validate a novel intelligent chromo capsule endoscope (ICCE). Methods: The ICCE has two modes: a white-light imaging (WLI) mode and an intelligent chromo imaging (ICI) mode. The performance of the ICCE in observing colors, animal tissues, and early gastrointestinal (GI) neoplastic lesions in humans was evaluated. Images captured by the ICCE were analysed using variance of Laplacian (VoL) values or image contrast evaluation. Results: For color observation, conventional narrow-band imaging endoscopes and the ICI mode of the ICCE have similar spectral distributions. Compared with the WLI mode, the ICI mode had significantly higher VoL values for animal tissues (2.154 ± 1.044 vs 3.800 ± 1.491, P = 0.003), gastric precancerous lesions and early gastric cancers (2.242 ± 0.162 vs 6.642 ± 0.919, P < 0.001), and colon tumors (3.896 ± 1.430 vs 11.882 ± 7.663, P < 0.001), and significantly higher contrast for differentiating tumor and non-tumor areas (0.069 ± 0.046 vs 0.144 ± 0.076, P = 0.005). More importantly, the sensitivity, specificity, and accuracy of the ICI mode for early GI tumors were 95.83%, 91.67%, and 94.64%, respectively, which were significantly higher than the values of the WLI mode (78.33% [P < 0.001], 77.08% [P = 0.01], and 77.98% [P < 0.001], respectively). Conclusions: We successfully integrated ICI into the capsule endoscope. The ICCE is an innovative and useful tool for differential diagnosis based on contrast-enhanced images and thus has great potential as a superior diagnostic tool for early GI tumor detection.

Improvement of detection sensitivity of upper gastrointestinal epithelial neoplasia in linked color imaging based on data of eye tracking.

BACKGROUND AND AIM: Linked color imaging (LCI) is useful for screening in the gastrointestinal tract; however, its true clinical benefit has not been determined. The aim of this study was to determine the objective advantage of LCI for detection of upper gastrointestinal neoplasms. METHODS: Nine endoscopists, including three novices, three trainees, and three experts, prospectively performed eye tracking. From 30 cases of esophageal or gastric neoplasm and 30 normal cases without neoplasms, a total of 120 images, including 60 pair images of white light imaging (WLI) and LCI taken at the same positions and angles, were randomly shown for 10 s. The sensitivity of tumor detection as a primary endpoint was evaluated and sensitivities by organ, size, and visual gaze pattern were also assessed. Color differences (ΔE using CIE1976 [L*a*b*]) between lesions and surrounding mucosa were measured and compared with detectability. RESULTS: A total of 1080 experiments were completed. The sensitivities of tumor detection in WLI and LCI were 53.7% (50.1-56.8%) and 68.1% (64.8-70.8%), respectively (P = 0.002). LCI provided higher sensitivity than WLI for the novice and trainee groups (novice: 42.2% [WLI] vs 65.6% [LCI], P = 0.003; trainee: 54.4% vs 70.0%, P = 0.045). No significant correlations were found between sensitivity and visual gaze patterns. LCI significantly increased ΔE, and the diagnostic accuracy with WLI depended on ΔE. CONCLUSIONS: In conclusion, LCI significantly improved sensitivity in the detection of epithelial neoplasia and enabled epithelial neoplasia detection that is not possible with the small color difference in WLI. (UMIN000047944).

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).

Efficacy of using red dichromatic imaging throughout endoscopic submucosal dissection procedure.

BACKGROUND: Management of bleeding during endoscopic submucosal dissection (ESD) is critical. Red Dichromatic Imaging (RDI), a novel image-enhanced endoscopy technology, has been reported to improve the visibility of deep vessels and bleeding source compared to white light imaging (WLI). We hypothesized that using RDI during the entire cutting process (full time RDI ESD: FTR-ESD), higher R0 resection rate, shorter procedure time and fewer complications could be achieved. Therefore, the aims of the present study were to investigate the efficacy and safety of FTR-ESD. METHODS: This retrospective observational study included a total of 82 consecutive patients who underwent ESD by a single expert endoscopist for 40 esophageal, 17 gastric and 25 colorectal cancers at our institution from January 2018 to March 2021. The clinicopathological data were collected from patients' medical records and the treatment outcomes were analyzed according to the treatment phase (early; 57 WLI-ESD and late; 25 FTR-ESD). RESULTS: The median of the greatest diameter of resected specimen was 40.0 mm. The median procedure time was relatively shorter in the FTR-ESD group (35 min) than in the WLI-ESD group (40 min), but the difference was not statistically significant (p = 0.34). The median dissection speed in the FTR-ESD group (27.23 mm2/min) was significantly faster than that in the WLI-ESD group (20.94 mm2/min) (p = 0.025). The dissection speed was not different among different organs. A multivariate analysis revealed that tumor size (more than 30 mm) and FTR-ESD were significant independent factors contributing to faster dissection speed (p < 0.05). There were no significant differences in the rates of en bloc resection, HM0, VM0 or occurrence of adverse events between WLI-ESD and FTR-ESD. CONCLUSIONS: FTR-ESD significantly increases the dissection speed compared to WLI-ESD. FTR-ESD can be performed safely and therapeutic outcomes in FTR-ESD are comparable with WLI-ESD. A further multicenter prospective study is warranted to confirm our results.

Distinguishing the papilla of Vater during biliary cannulation using texture and color enhancement imaging: A pilot study.

Objectives: Understanding the exact morphology of the bile duct opening is important for determining the success of bile duct cannulation. Texture and color enhancement imaging (TXI) has been reported to enhance slight changes in color tone and structure that are difficult to see with white light imaging. This study investigated whether TXI mode1 could improve papillary recognition by trainees inexperienced in endoscopic retrograde cholangiopancreatography. Methods: We included 31 patients with naive papilla of Vater at a single institution in the study. Trainee endoscopists (n = 4) evaluated and identified the papilla according to the Inomata classification using white light imaging and TXI. The degree of agreement with the evaluation of supervising physicians (n = 4) was examined using the McNemar test. Results: In the trainee group, the kappa coefficient agreements were κ = 0.346 and κ = 0.754 for white light imaging and TXI, respectively. When further evaluated, the separate and septal types of papilla groups showed an increased concordance rate in one of the four trainees (76.67%-96.67%, p = 0.031, respectively). Moreover, comparison for two-group evaluation showed an increased kappa coefficient in two of four trainees (0.34-0.92, p = 0.010, 0.45-0.92, p = 0.024). Conclusions: Observation of the duodenal papilla using TXI improved papillary differentiation and suggested the potential of TXI as a clinical tool. Further study of this method is necessary; it is expected to help reduce cannulation time and the incidence of pancreatitis.

Automatic detection of early gastric cancer in endoscopy based on Mask region-based convolutional neural networks (Mask R-CNN)(with video).

The artificial intelligence (AI)-assisted endoscopic detection of early gastric cancer (EGC) has been preliminarily developed. The currently used algorithms still exhibit limitations of large calculation and low-precision expression. The present study aimed to develop an endoscopic automatic detection system in EGC based on a mask region-based convolutional neural network (Mask R-CNN) and to evaluate the performance in controlled trials. For this purpose, a total of 4,471 white light images (WLIs) and 2,662 narrow band images (NBIs) of EGC were obtained for training and testing. In total, 10 of the WLIs (videos) were obtained prospectively to examine the performance of the RCNN system. Furthermore, 400 WLIs were randomly selected for comparison between the Mask R-CNN system and doctors. The evaluation criteria included accuracy, sensitivity, specificity, positive predictive value and negative predictive value. The results revealed that there were no significant differences between the pathological diagnosis with the Mask R-CNN system in the WLI test (χ2 = 0.189, P=0.664; accuracy, 90.25%; sensitivity, 91.06%; specificity, 89.01%) and in the NBI test (χ2 = 0.063, P=0.802; accuracy, 95.12%; sensitivity, 97.59%). Among 10 WLI real-time videos, the speed of the test videos was up to 35 frames/sec, with an accuracy of 90.27%. In a controlled experiment of 400 WLIs, the sensitivity of the Mask R-CNN system was significantly higher than that of experts (χ2 = 7.059, P=0.000; 93.00% VS 80.20%), and the specificity was higher than that of the juniors (χ2 = 9.955, P=0.000, 82.67% VS 71.87%), and the overall accuracy rate was higher than that of the seniors (χ2 = 7.009, P=0.000, 85.25% VS 78.00%). On the whole, the present study demonstrates that the Mask R-CNN system exhibited an excellent performance status for the detection of EGC, particularly for the real-time analysis of WLIs. It may thus be effectively applied to clinical settings.

Diagnostic accuracy of linked colour imaging versus white light imaging for early gastric cancers: a prospective, multicentre, randomized controlled trial study.

BACKGROUND: Linked colour imaging (LCI) is a novel new image-enhanced endoscopy (IEE) technology that produces bright and vivid images. The aim of this study was to assess the ability of LCI to improve the diagnostic accuracy of early gastric cancer (EGC) relative to white light imaging (WLI). MATERIALS AND METHODS: We performed this study on patients undergoing screening endoscopy from 12 medical institutions in China. Patients were randomly assigned to receive WLI followed by LCI or LCI followed by WLI. The primary outcome was to compared the diagnostic accuracy between LCI and WLI for EGC/high-grade intraepithelial neoplasms. Secondary outcomes included the numbers of suspicious lesions, neoplastic lesions and examination time by using LCI detected versus using WLI. RESULTS: A total of 1924 patients were randomly selected, and 1828 were included in the analysis. The diagnostic accuracy for EGC, which was 78.8% by using LCI and 68.4% by using WLI (p < .0001). More suspicious lesions were detected by LCI than by WLI (n = 1235 vs. 1036, p = .031), especially among differentiated EGC (p = .013). LCI greatly shortened the examination time compared with WLI (p = .019). CONCLUSIONS: LCI has better accuracy and shorter examination time in diagnosing EGC than WLI (Clinical trial registration: NCT03092414).Key messagesCompared with white light imaging (WLI), the diagnostic accuracy, sensitivity and specificity increased by using LCI.More lesions were detected by LCI alone than by WLI alone, especially among differentiated EGC.LCI may be used as a screening tool for routine clinical observation.

Using texture and colour enhancement imaging to evaluate gastrointestinal diseases in clinical practice: a review.

White light imaging (WLI) is the most common endoscopic technique used for screening of gastrointestinal diseases. However, despite the advent of a new processor that offers sufficient clear illumination and other advanced developments in endoscopic instrumentation, WLI alone is inadequate for detecting all gastrointestinal diseases with abnormalities in mucosal discoloration and morphological changes to the mucosal surface. The recent development of image-enhanced endoscopy (IEE) has dramatically improved the detection of gastrointestinal diseases. Texture and colour enhancement imaging (TXI) is a new type of IEE that enhances brightness, surface irregularities, such as elevations or depressions, and subtle colour changes. TXI with two modes, namely modes 1 and 2, can selectively enhance brightness in dark areas of an endoscopic image and subtle tissue differences such as slight morphological or colour changes while simultaneously preventing over-enhancement. Several clinical studies have investigated the efficacy of TXI for detecting and visualizing gastrointestinal diseases, including oesophageal squamous cell carcinoma (ESCC), Barret's epithelium, gastric cancer, gastric mucosal atrophy and intestinal metaplasia. Although TXI is often more useful for detecting and visualizing gastrointestinal diseases than WLI, it remains unclear whether TXI outperforms other IEEs, such as narrow-band imaging (NBI), in similar functions, and whether the performance of TXI modes 1 and 2 are comparable. Therefore, large-scale prospective studies are needed to compare the efficacy of TXI to WLI and other IEEs for endoscopic evaluation of patients undergoing screening endoscopy. Here, we review the characteristics and efficacy of TXI for the detection and visualization of gastrointestinal diseases.Key MessagesTXI mode 1 can improve the visibility of gastrointestinal diseases and qualitative diagnosis, especially for diseases associated with colour changes.The enhancement of texture and brightness with TXI mode 2 enables the detection of diseases, and is ideal for use in the first screening of gastrointestinal tract.

Low-magnification narrow-band imaging for small gastric neoplasm detection on screening endoscopy.

Background and Aims: Microsurface patterns of the gastric mucosa can be observed using magnifying narrow-band imaging (M-NBI). However, the efficacy of M-NBI at low-magnification (LM-NBI) screening for detecting small gastric neoplasms is unclear. Methods: This prospective study was conducted at a single institution. LM-NBI, defined as minimal magnification that could reveal the microsurface pattern of the gastric mucosa, was performed after routine white-light imaging (WLI) observation of the stomach. Depending on the phase in which the neoplastic lesions were initially found, they were divided into the WLI group and the LM-NBI group, and the characteristics of these neoplastic lesions were investigated accordingly. Results: Sixty-five epithelial lesions (adenomas or noninvasive carcinomas) of 20 mm or less in diameter were identified in this study. Sixteen lesions were detected only with LM-NBI. Smaller lesions were detected using LM-NBI (P = .01). WLI took about 160 to 260 seconds, while LM-NBI required about 70 to 80 seconds. All lesions in the LM-NBI group had a background of map-like redness (n = 5) or atrophic/metaplastic mucosa (n = 11). Conclusions: LM-NBI was able to detect lesions overlooked by WLI, especially those in areas of map-like redness or atrophic/metaplastic mucosa of the stomach. Approximately one-quarter of newly diagnosed neoplasms were retrieved on routine examination during an extra 1.5 minutes.

Endoluminal larynx anatomy model - towards facilitating deep learning and defining standards for medical images evaluation with artificial intelligence algorithms.

The pioneering nature of this work covers the answers to two questions: (1) Is an up-to-date anatomical model of the larynx needed for modern endoscopic diagnostics, and (2) can such a digital segmentation model be utilized for deep learning purposes. The idea presented in this article has never been proposed before, and this is a breakthrough in numerical approaches to aerodigestive videoendoscopy imaging. The approach described in this article assumes defining a process for data acquisition, integration, and segmentation (labeling), for the needs of a new branch of knowledge: digital medicine and digital diagnosis support expert systems. The first and crucial step of such a process is creating a digital model of the larynx, which has to be then validated utilizing multiple clinical, as well as technical metrics. The model will form the basis for further artificial intelligence (AI) requirements, and it may also contribute to the development of translational medicine.

Texture and color enhancement imaging for detecting colorectal adenomas: Good, but not good enough.

Texture and color enhancement imaging (TXI) has been developed as a novel image-enhancing endoscopy. However, the effectiveness of TXI detecting adenomas is inferior to narrow band imaging. Thus, future studies will need to focus on investigating the feasibility of such combination in clinical settings in order to provide patients with more accurate diagnoses.