Endoscopic Video Research Articles

The influence of clinical experience on reliable evaluation of pharyngeal and laryngeal lesions: comparison of high-definition laryngoscopy using narrow band imaging with fibre-optic laryngoscopy.

Fibre-optic laryngoscopy is still widely used in daily clinical practice; however, high-definition laryngoscopy using narrow band imaging could be more reliable in characterising pharyngeal and laryngeal lesions. Endoscopic videos were assessed in a tertiary referral hospital by 12 observers with different levels of clinical experience. Thirty pairs of high-definition laryngoscopy with narrow band imaging and fibre-optic laryngoscopy videos were judged twice, with an interval of two to four weeks, in a random order. Inter- and intra-observer reliability, sensitivity and specificity were calculated in terms of detecting a malignant lesion and a specific histological entity, for beginners, trained observers and experts. Using high-definition laryngoscopy with narrow band imaging, inter-observer reliability for detecting malignant lesions increased from moderate to substantial in trained observers and experts (high-definition laryngoscopy with narrow band imaging κ = 0.66 and κ = 0.77 vs fibre-optic laryngoscopy κ = 0.51 and κ = 0.56, for trained observers and experts respectively) and sensitivity increased by 16 per cent. Inter-observer reliability increased with the level of clinical experience, especially when using high-definition laryngoscopy with narrow band imaging.

AUGMENTED REALITY-BASED NASAL ENDOSCOPE VIDEO RECONSTRUCTION AND REGISTRATION

Neuroendoscopic surgery is a minimally invasive surgical technique commonly used to remove a tumor through the patient’s mouth or nose, which requires the surgeon to avoid important neural and vascular structures. Augmented reality-based surgical navigation can provide surgeons with more information and visual aids, clarifying the structure and details of the surgical area. It requires real-time registration of intraoperative endoscopic video with pre-operative CT models. However, 3D reconstruction and camera tracking from nasal endoscopic video are challenging due to the narrow nasal cavity and the lack of texture. Besides, the nasal endoscopic datasets that can be used for deep learning-based depth estimation are scarce and hard to elaborate. To this end, a style-augmented module (SAM) is proposed in this study to minimize discrepancies between different endoscopic datasets. An unsupervised approach was trained to generate depth maps and camera motion paths, which were used to reconstruct the nasal scene’s 3D model. The obtained reconstruction was aligned with a pre-operative CT model for augmented reality-based surgical navigation. The results showed that the proposed SAM improved the generalization of the depth estimation model on nasal endoscopy data. The proposed approach to combining augmented reality technology with surgical navigation is considered instrumental in furnishing surgeons with richer information on the nasal endoscopic surgical scene.

Transformer with Hybrid Attention Mechanism for Stereo Endoscopic Video Super Resolution

With stereo cameras becoming widely used in invasive surgery systems, stereo endoscopic images provide important depth information for delicate surgical tasks. However, the small size of sensors and their limited lighting conditions lead to low-quality and low-resolution endoscopic images and videos. In this paper, we propose a stereo endoscopic video super-resolution method using transformer with a hybrid attention mechanism named HA-VSR. Stereo video SR aims to reconstruct high-resolution (HR) images from corresponding low-resolution (LR) videos. In our method, the stereo correspondence and temporal correspondence are incorporated into the HA-VSR model. Specifically, the Swin transformer architecture is utilized in proposed framework with hybrid attention mechanisms. The parallel attention mechanism is utilized by using the symmetry and consistency of left and right images, and the temporal attention mechanism is utilized by using the consistency of consecutive frames. Detailed quantitative evaluation and experiments on two datasets show the proposed model achieves advanced SR reconstruction performance, showing that the proposed stereo VSR framework outperforms alternative approaches.

Clinical and Morphological Parallels in the Etiopathogenesis of Secretory Otitis Media

Introduction. According to global statistics, secretory otitis media (SOM) ranks first among middle ear infections. This pathology is believed to be characteristic of the pediatric population; however, recent global statistics indicate an increase in the percentage of SOM cases among the adult population. Information about adult-onset SOM varies, with no consensus on the causes of its occurrence. Diverse clinical presentation of the condition, often unsatisfactory treatment outcomes, and a high percentage of recurrences have necessitated a more in-depth study of the etiological factors associated with adult-onset SOM.
 The objective of the research was to determine clinical and morphological parallels in the etiopathogenesis of adult-onset SOM as well as to establish the relationship between the clinical, otoscopic, and endoscopic presentations and pathomorphological changes in the tubal tonsils in these patients.
 Materials and Methods. Sixty-eight (38 men and 30 women) patients with SOM at the age of 18 to 65 years were examined. Fifty-three patients were diagnosed with unilateral pathological process, while in 15 individuals, bilateral SOM was observed. All the patients underwent a comprehensive assessment of their complaints, collection of their medical history, and a full ear, nose, and throat examination using conventional methods, optical techniques, video endoscopy of the nasopharynx and tubal tonsils, and otomicroscopy. In 15 patients, to exclude a malignant process, tissue samples from the tubal tonsils were taken for pathological examination.
 Results. One of the causes of adult-onset SOM could be pathomorphological changes in the tubal tonsils, specifically their inflammation, hypertrophy, and in some cases, tumours. The clinical course and stages of SOM development depend on the duration of auditory tube obstruction. Inflammation of the tubal tonsils usually results in catarrhal and serous SOM, while hypertrophy or hyperplasia of the tubal tonsils is observed in patients with mucous SOM.
 Conclusions. Pathomorphological changes in the tubal tonsils have been established as one of the reasons for developing SOM in adults. Both the clinical presentation of SOM and the stage of its progression depends on the type of pathomorphological changes, their localization (upper or lower pole), and the duration of auditory tube obstruction. A prolonged, recurrent, and treatment-resistant course of SOM may be associated with hypertrophy of the tubal tonsil or tumours. Clinical and morphological parallels identified by us will enable timely utilization of effective etiopathogenetic treatment of SOM and prevent the occurrence of complications.

Functional and organic dysphonia in voice professionals (lecture)

An urgent task of otorhinolaryngology (phoniatry) is the diagnosis and treatment of disorders of voice function in voice professionals or persons of voice-speech professions. An undiagnosed laryngeal disease worsens the quality of life of patients, creates difficulties in communication and can even lead to professional unfitness, because the patient does not withstand the necessary vocal load. The leading role in the modern diagnosis of laryngeal diseases belongs to endoscopic research methods video endoscopy and video endostroboscopy. All disorders of voice function are classified as functional and organic. Recent studies have shown that from 30.5 to 40% of voice professionals who have applied for phoniatric care suffer from functional disorders. Functional changes are understood to mean changes in which laryngoscopic examination does not always reveal any visible pathology. But with the help of special research methods, such as stroboscopy (videostroboscopy), spectral analysis of the voice, pathognomonic changes can be determined, and the patient presents characteristic complaints. More often than others, hypotonic and hypo-hypertonic dysphonia are diagnosed. Organic laryngeal diseases are characteristic of 69.5% of patients with dysphonia. They are represented by various nosological forms. Certain difficulties arise in the diagnosis of professional laryngitis and vascular pathology of the larynx. The lecture presents pathognomonic clinical changes of the larynx in voice professionals, knowledge of which is necessary for practical doctors to make a diagnosis.

Using Computer Vision to Improve Endoscopic Disease Quantification in Therapeutic Clinical Trials of Ulcerative Colitis

Endoscopic assessment of ulcerative colitis (UC) typically reports only the maximum severity observed. Computer vision methods may better quantify mucosal injury detail, which varies among patients. Endoscopic video from the UNIFI clinical trial (A Study to Evaluate the Safety and Efficacy of Ustekinumab Induction and Maintenance Therapy in Participants With Moderately to Severely Active Ulcerative Colitis) comparing ustekinumab and placebo for UC were processed in a computer vision analysis that spatially mapped Mayo Endoscopic Score (MES) to generate the Cumulative Disease Score (CDS). CDS was compared with the MES for differentiating ustekinumab vs placebo treatment response and agreement with symptomatic remission at week 44. Statistical power, effect, and estimated sample sizes for detecting endoscopic differences between treatments were calculated using both CDS and MES measures. Endoscopic video from a separate phase 2 clinical trial replication cohort was performed for validation of CDS performance. Among 748 induction and 348 maintenance patients, CDS was lower in ustekinumab vs placebo users at week 8 (141.9 vs 184.3; P < .0001) and week 44 (78.2 vs 151.5; P < .0001). CDS was correlated with the MES (P < .0001) and all clinical components of the partial Mayo score (P < .0001). Stratification by pretreatment CDS revealed ustekinumab was more effective than placebo (P < .0001) with increasing effect in severe vs mild disease (-85.0 vs -55.4; P < .0001). Compared with the MES, CDS was more sensitive to change, requiring 50% fewer participants to demonstrate endoscopic differences between ustekinumab and placebo (Hedges' g= 0.743 vs 0.460). CDS performance in the JAK-UC replication cohort was similar to UNIFI. As an automated and quantitative measure of global endoscopic disease severity, the CDS offers artificial intelligence enhancement of traditional MES capability to better evaluate UC in clinical trials and potentially practice.

SurgT challenge: Benchmark of soft-tissue trackers for robotic surgery

This paper introduces the “SurgT: Surgical Tracking” challenge which was organized in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 2022). There were two purposes for the creation of this challenge: (1) the establishment of the first standardized benchmark for the research community to assess soft-tissue trackers; and (2) to encourage the development of unsupervised deep learning methods, given the lack of annotated data in surgery. A dataset of 157 stereo endoscopic videos from 20 clinical cases, along with stereo camera calibration parameters, have been provided. Participants were assigned the task of developing algorithms to track the movement of soft tissues, represented by bounding boxes, in stereo endoscopic videos. At the end of the challenge, the developed methods were assessed on a previously hidden test subset. This assessment uses benchmarking metrics that were purposely developed for this challenge, to verify the efficacy of unsupervised deep learning algorithms in tracking soft-tissue. The metric used for ranking the methods was the Expected Average Overlap (EAO) score, which measures the average overlap between a tracker’s and the ground truth bounding boxes. Coming first in the challenge was the deep learning submission by ICVS-2Ai with a superior EAO score of 0.617. This method employs ARFlow to estimate unsupervised dense optical flow from cropped images, using photometric and regularization losses. Second, Jmees with an EAO of 0.583, uses deep learning for surgical tool segmentation on top of a non-deep learning baseline method: CSRT. CSRT by itself scores a similar EAO of 0.563. The results from this challenge show that currently, non-deep learning methods are still competitive. The dataset and benchmarking tool created for this challenge have been made publicly available at https://surgt.grand-challenge.org/. This challenge is expected to contribute to the development of autonomous robotic surgery and other digital surgical technologies.

A Temporal Learning Approach to Inpainting Endoscopic Specularities and Its Effect on Image Correspondence

Video streams are utilised to guide minimally-invasive surgery and diagnosis in a wide range of procedures, and many computer-assisted techniques have been developed to automatically analyse them. These approaches can provide additional information to the surgeon such as lesion detection, instrument navigation, or anatomy 3D shape modelling. However, the necessary image features to recognise these patterns are not always reliably detected due to the presence of irregular light patterns such as specular highlight reflections. In this paper, we aim at removing specular highlights from endoscopic videos using machine learning. We propose using a temporal generative adversarial network (GAN) to inpaint the hidden anatomy under specularities, inferring its appearance spatially and from neighbouring frames, where they are not present in the same location. This is achieved using in-vivo data from gastric endoscopy (Hyper Kvasir) in a fully unsupervised manner that relies on the automatic detection of specular highlights. System evaluations show significant improvements to other methods through direct comparison and ablation studies that depict the importance of the network’s temporal and transfer learning components. The generalisability of our system to different surgical setups and procedures was also evaluated qualitatively on in-vivo data of gastric endoscopy and ex-vivo porcine data (SERV-CT, SCARED). We also assess the effect of our method in comparison to other methods on computer vision tasks that underpin 3D reconstruction and camera motion estimation, namely stereo disparity, optical flow, and sparse point feature matching. These are evaluated quantitatively and qualitatively and results show a positive effect of our specular inpainting method on these tasks in a novel comprehensive analysis. Our code and dataset are made available at https://github.com/endomapper/Endo-STTN.

Assessment of the Sensitivity of the Vocal Tract in Parkinson Disease by Nasal Video Endoscopy.

Introduction Sensory deficits of the upper airways are common in people with Parkinson disease. Compounded by considerable deterioration of the swallowing function, these deficits may contribute to the elevated rates of morbidity and mortality among this population. Objective To examine the sensitivity of the vocal tracts of people with Parkinson disease using nasal videoendoscopy and to compare the results with paired controls. Methods The present was a prospective, observational, case-control study. Sensitivity assessments were conducted in a sample of 24 people divided into 2 groups: one group of 12 patients with Parkinson's disease, and a control group with 12 healthy subjects. The study group also underwent a fiberoptic endoscopic evaluation of swallowing and answered the Swallowing Disturbance Questionnaire (SQD) to detect dysphagia. Results There was a significant difference ( p < 0.05) in the region of the arytenoid cartilages, showing that sensitivity was better preserved among the control group, and that sensitivity deficits were present in the study group. The qualitative results showed sensorial impairment in the study group than in the control group regarding the base of the tongue, the vestibular folds, and the vocal folds. The study group showed self-awareness regarding the deficits in the swallowing function, but there was no statistically significant association between swallowing function deficits and deterioration of sensorial function among them. Conclusions Sensorial deficits were present in the study group when compared to the healthy subjects, mainly in the qualitative evaluation.

S3236 Spontaneous Elimination of a Video Endoscopy Capsule from a Small Bowel Stricture After 9 Months in a Patient With Crohn’s Disease

S3236 Spontaneous Elimination of a Video Endoscopy Capsule from a Small Bowel Stricture After 9 Months in a Patient With Crohn’s Disease

Capsule endoscopy with artificial intelligence-assisted technology: Real-world usage of a validated AI model for capsule image review.

Background and study aims Capsule endoscopy is a time-consuming procedure with a significance error rate. Artificial intelligence (AI) can potentially reduce reading time significantly by reducing the number of images that need human review. An OMOM Artificial Intelligence-enabled small bowel capsule has been recently trained and validated for small bowel capsule endoscopy video review. This study aimed to assess its performance in a real-world setting in comparison with standard reading methods. Patients and methods In this single-center retrospective study, 40 patient studies performed using the OMOM capsule were analyzed first with standard reading methods and later using AI-assisted reading. Reading time, pathology identified, intestinal landmark identification and bowel preparation assessment (Brotz Score) were compared. Results Overall diagnosis correlated 100% between the two reading methods. In a per-lesion analysis, 1293 images of significant lesions were identified combining standard and AI-assisted reading methods. AI-assisted reading captured 1268 (98.1%, 95% CI 97.15-98.7) of these findings while standard reading mode captured 1114 (86.2%, 95% confidence interval 84.2-87.9), P < 0.001. Mean reading time went from 29.7 minutes with standard reading to 2.3 minutes with AI-assisted reading ( P < 0.001), for an average time saving of 27.4 minutes per study. Time of first cecal image showed a wide discrepancy between AI and standard reading of 99.2 minutes (r = 0.085, P = 0.68). Bowel cleansing evaluation agreed in 97.4% (r = 0.805 P < 0.001). Conclusions AI-assisted reading has shown significant time savings without reducing sensitivity in this study. Limitations remain in the evaluation of other indicators.

Monocular endoscope 6-DoF tracking with constrained evolutionary stochastic filtering.

Monocular endoscopic 6-DoF camera tracking plays a vital role in surgical navigation that involves multimodal images to build augmented or virtual reality surgery. Such a 6-DoF camera tracking generally can be formulated as a nonlinear optimization problem. To resolve this nonlinear problem, this work proposes a new pipeline of constrained evolutionary stochastic filtering that originally introduces spatial constraints and evolutionary stochastic diffusion to deal with particle degeneracy and impoverishment in current stochastic filtering methods. With its application to endoscope 6-DoF tracking and validation on clinical data including more than 59,000 endoscopic video frames acquired from various surgical procedures, the experimental results demonstrate the effectiveness of the new pipeline that works much better than state-of-the-art tracking methods. In particular, it can significantly improve the accuracy of current monocular endoscope tracking approaches from (4.83 mm, 10.2∘) to (2.78 mm, 7.44∘).

Dynamic surface reconstruction in robot-assisted minimally invasive surgery based on neural radiance fields.

The purpose of this study was to improve surgical scene perception by addressing the challenge of reconstructing highly dynamic surgical scenes. We proposed a novel depth estimation network and a reconstruction framework that combines neural radiance fields to provide more accurate scene information for surgical task automation and AR navigation. We added a spatial pyramid pooling module and a Swin-Transformer module to enhance the robustness of stereo depth estimation. We also improved depth accuracy by adding unique matching constraints from optimal transport. To avoid deformation distortion in highly dynamic scenes, we used neural radiance fields to implicitly represent scenes in the time dimension and optimized them with depth and color information in a learning-based manner. Our experiments on the KITTI and SCARED datasets show that the proposed depth estimation network performs close to the state-of-the-art method on natural images and surpasses the SOTA method on medical images with 1.12% in 3 px Error and 0.45 px in EPE. The proposed dynamic reconstruction framework successfully reconstructed the dynamic cardiac surface on a totally endoscopic coronary artery bypass video, achieving SOTA performance with 27.983dB in PSNR, 0.812 in SSIM, and 0.189 in LPIPS. Our proposed depth estimation network and reconstruction framework provide a significant contribution to the field of surgical scene perception. The framework achieves better results than SOTA methods on medical datasets, reducing mismatches on depth maps and resulting in more accurate depth maps with clearer edges. The proposed ER framework is verified on a series of dynamic cardiac surgical images. Future efforts will focus on improving the training speed and solving the problem of limited field of view.

Urologist Workforce and Services in Ethiopia.

Urological conditions are a cause of diminishing quality of life, hence affecting productivity. Despite the need for urological treatment, it was excluded from receiving priority in both United Nation and Lancet commission. Most of the surgeries in sub-Saharan Africa are open surgeries. The lack of basic endourology equipment and a shortage of experts have limited Africans from receiving the privileges of minimally invasive surgeries, especially in urology. This study describes the socio demographics of the urologists in Ethiopia, the field of services they provide and their access to endourology equipment. This study presents data from a survey of urologists in Ethiopia who are members of Urology Society of Ethiopia (USE). Thirty-three urologists of the 43 responded, making the response rate 76.5%. Qualification by urology residency in Ethiopia accounted for 66.7% of participants, followed by 21.2% by fellowship training abroad after general surgery training. All respondents practice open surgeries and 75.8% perform endourology. Video endoscope and cystoscopy sets were available to all those practicing endourology, with Direct Visual Internal Urethrotomy (DVIU) sets having the next highest availability and Flexible UreteroRenoScope(URS) and laser lithotripters the least accessible. Urology in Ethiopia is in its infancy, where the lack of advanced medical equipment combined with a paucity of qualified urologists have created a huge challenge for the provision of these services.

Palate shape is associated with Unilateral Hypoglossal Nerve Stimulation Outcomes.

The aim was to determine the potential association between palate shape and unilateral hypoglossal nerve stimulation (HNS) outcomes. Preoperative drug-induced sleep endoscopy (DISE) videos were reviewed and scored by 3 blinded reviewers to determine airway narrowing at the hard-soft palate junction (HP), soft palate genu, and inferior velum, as described by Woodson (2014). Scoring was as follows: 1-open airway, 2-narrow, 3-severe narrowing. Overall palate shape (oblique, intermediate, or vertical) was determined based on prior criteria. Successful surgical treatment was defined by the HNS titration polysomnogram as a reduction of ≥50% in the apnea-hypopnea index (AHI) to <15 events/h. Of 332 adults, the majority was male (77%) with an average BMI of 29.2 ± 3.6 kg/m2 . Overall success rate was 73%. Success rate was lower in patients with vertical palate shape compared with the other shapes (56% vs. 75%, p = 0.029). HP score 3 compared with scores 2 and 1 was associated with lower success rates (60% vs. 76%, p = 0.028), but genu and velum scores were not associated with outcomes. Patients with both HP score 3 and complete oropharyngeal lateral wall-related obstruction had notably worse outcomes (22% vs. 74%, p = 0.026). HP score 3 (OR 0.45, 95%CI 0.22-0.92) and vertical palate shape (OR 0.33, 95%CI 0.15-0.78) were independently associated with lower odds of surgical response after adjustment for DISE findings, age, gender, and BMI. Vertical palate shape and narrowing at the hard-soft palate junction are independently associated with lower HNS surgical success rates. 3 Laryngoscope, 134:981-986, 2024.

Automatic summarization of endoscopic skull base surgical videos through object detection and hidden Markov modeling.

Endoscopic endonasal surgery is a medical procedure that utilizes an endoscopic video camera to view and manipulate a surgical site accessed through the nose. Despite these surgeries being video recorded, these videos are seldom reviewed or even saved in patient files due to the size and length of the video file. Editing to a manageable size may necessitate viewing 3 h or more of surgical video and manually splicing together the desired segments. We suggest a novel multi-stage video summarization procedure utilizing deep semantic features, tool detections, and video frame temporal correspondences to create a representative summarization. Summarization by our method resulted in a 98.2% reduction in overall video length while preserving 84% of key medical scenes. Furthermore, resulting summaries contained only 1% of scenes with irrelevant detail such as endoscope lens cleaning, blurry frames, or frames external to the patient. This outperformed leading commercial and open source summarization tools not designed for surgery, which only preserved 57% and 46% of key medical scenes in similar length summaries, and included 36% and 59% of scenes containing irrelevant detail. Experts agreed that on average (Likert Scale = 4) that the overall quality of the video was adequate to share with peers in its current state.

Vocal Vibration Opening Onset Position Abnormality in Patients With Adductor Spasmodic Dystonia by Using High-Speed Endoscopy

To analyze vocal fold vibration onset in patients with adductor laryngeal dystonia (ADLD) by analyzing vocal vibration opening onset position (VVOOP). Case-control study SETTING: A voice center. Eleven patients with ADLD diagnosed in our voice center were enrolled in the ADLD group. Eleven healthy subjects matched by exact age and gender to the ADLD patients were selected as the control group. All subjects underwent laryngeal high-speed video endoscopy. VVOOP and its change were assessed by two otolaryngologists. The multiline video kymography was used to analyze the openquotient (OQ) and standard deviation of OQ. VVOOP had more than one position in 54.6% (6/11) of the patients with ADLD, which was higher than the control group (P<0.05). VVOOP appeared in the front of the vocal fold in 54.6% (6/11) of patients with ADLD and in the back of the vocal fold in 81.8% (9/11) of patients with ADLD. VVOOP can be abnormal in 90.9% (10/11) of patients with ADLD, and the rate of VVOOP abnormality was higher than that of the control group (P<0.05). Of 11, 6 (54.6%) patients with ADLD had a variable VVOOP; thevariability rate of VVOOP was higher than that in the control group(P<0.05). OQ and OQ standard deviation in the ADLD group were significantly greater than in the control group (P<0.05). In patients with ADLD, vocal fold vibration was irregular, and VVOOP was abnormal and had a variable position and could reflect variability of the vocal vibration. Level 4.

Laparoscopic surgery augmentation through vibro-acoustic sensing of instrument-tissue interactions

Abstract This paper presents the use of vibro-acoustic sensing to augment laparoscopic surgery procedures by analyzing the signals produced during cutting and palpation tasks on various tissue samples. Vibro-acoustic signals were acquired during an experiment on a dedicated phantom covered in dense foam, where three trocars were inserted into the phantom to place the endoscopic camera and two laparoscopic instruments. The results of the signals analysis demonstrate the potential of this approach for making laparoscopic interactions audible, differentiating between tissue types, and detecting variations in tissue properties. Vibro-acoustic sensing could be a valuable tool for integrating sound into the current clinical workflow for enhancing endoscope video images.

Detection and Classification of Gastrointestinal Diseases by using Machine Learning: A Review

Currently, gastrointestinal diseases claim the lives of up to two million people worldwide. GI disease treatment can be challenging, time-consuming, and expensive. One of the most recent advancements in medical imaging is the use of video endoscopy to diagnose gastrointestinal illnesses such stomach ulcers, bleeding, and polyps. Doctors require a lot of time to review all the images produced by medical video endoscopy since there are so many of them. This makes manual diagnosis difficult and has encouraged research into computer-aided approaches to diagnose all of the generated images quickly and accurately. The innovative aspect of the suggested methodology is the creation of a system for the diagnosis of digestive disorders. Machine learning techniques have the potential to significantly lower the cost of examination procedures while increasing the accuracy and speed of diagnosis. This paper describes a method for classifying GI illnesses using machine learning techniques.

406. A SYSTEMATIC REVIEW AND META-ANALYSIS ON THE USE OF ARTIFICIAL INTELLIGENCE IN THE ENDOSCOPIC DIAGNOSIS OF ESOPHAGEAL CANCER

Abstract Introduction Endoscopy plays a key role in the surveillance of patients with Barrett’s esophagus and the diagnosis of esophageal cancer. The subtleties of endoscopic interpretation are subject to clinical expertise, diagnostic skill, and thus human error. Therefore, artificial intelligence (AI) is increasingly being incorporated into endoscopy in order to improve diagnostic accuracy and early detection. This systematic review and meta-analysis consolidates the evidence on the use of AI in the endoscopic diagnosis of esophageal cancer. Methods The systematic review was carried out using Pubmed, MEDLINE and Ovid EMBASE databases as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Studies which described the role of AI in supporting endoscopic video or image diagnosis of esophageal cancer were included. A meta-analysis was also performed. Results 45 studies formed the qualitative and quantitative review. 14 studies with 1590 patients in total assessed the use of AI in endoscopic diagnosis of esophageal squamous cell carcinoma (ESCC)- the pooled sensitivity and specificity were 91.2% (84.3–95.2%) and 80% (64.3–89.9%). A further, 9 studies consisting of 478 patients overall explored the capabilities of AI in supporting the endoscopic diagnosis of esophageal adenocarcinoma (EAC), with a pooled sensitivity and specificity of 93.1% (86.8–96.4%) and 86.9% (81.7–90.7%). AI was often superior to endoscopists when diagnosing ESCC and improved the overall accuracy of both novice and expert endoscopists. Conclusion AI technology, as an adjunct to endoscopy has proven to be beneficial in early, accurate detection of esophageal malignancy. Multiple studies have shown superior diagnostic capabilities when comparing AI technology to endoscopists alone. Despite promising results, the application in real-time endoscopy is limited, and further multicenter trials are required to accurately assess its use in routine practice.