Skip Navigation
Skip to contents

Clin Endosc : Clinical Endoscopy

Search
Advanced Search
OPEN ACCESS
mobile search button mobile menu button

Articles

Page Path
HOME > Clin Endosc > Volume 59(2); 2026 > Article
Review Recent technological advances in video capsule endoscopy: a comprehensive review
Minjee Kim1orcid, Hyun Joo Jang2orcid
Clinical Endoscopy 2026;59(2):182-193.
DOI: https://doi.org/10.5946/ce.2025.135
Published online: September 29, 2025

1Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea

2Department of Medicine, Hallym University Dongtan Sacred Heart Hospital, Hwaseong, Korea

Correspondence: Hyun Joo Jang Department of Internal Medicine, Hallym University Dongtan Sacred Heart Hospital, Hallym University College of Medicine, 7 Keunjaebong-gil, Hwaseong 18450, Korea E-mail: jhj1229@hallym.or.kr
• Received: April 29, 2025   • Revised: July 7, 2025   • Accepted: July 8, 2025

© 2026 Korean Society of Gastrointestinal Endoscopy

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

  • 12,994 Views
  • 413 Download
  • 2 Web of Science
  • 2 Crossref
prev next
Full Article

Download PDF Download PDF

  • Video capsule endoscopy (VCE) originally revolutionized gastrointestinal imaging by providing a noninvasive method for evaluating small bowel diseases. Recent technological innovations, including enhanced imaging systems, artificial intelligence (AI), and improved localization, have significantly improved VCE’s diagnostic accuracy, efficiency, and clinical utility. This review aims to summarize and evaluate recent technological advances in VCE, focusing on system comparisons, image enhancement, localization technologies, and AI-assisted lesion detection.
  • Keywords: Artificial intelligence; Capsule endoscopy; Deep learning; Diagnostic imaging; Image enhancement
Video capsule endoscopy (VCE) is a noninvasive method for visualizing the gastrointestinal tract. It offers several advantages, including the absence of ionizing radiation exposure and sedation-related risks associated with conventional endoscopic procedures. Furthermore, VCE can detect subtle mucosal pathologies that are often missed on radiological imaging, such as angiodysplasia, minor inflammatory changes, and flat polyps. The lesion detection rate of VCE was significantly higher than that of intestinal ultrasound in patients with suspicious small bowel bleeding (62% vs. 14%, p<0.05) and complicated celiac disease (55% vs. 35%, p<0.05).1 Accordingly, practical guidelines outline comprehensive diagnostic strategies and examination techniques, such as VCE and balloon enteroscopy, that are applicable to patients with suspected small bowel bleeding in routine clinical settings.2
Small-bowel capsule endoscopy (SBCE) was introduced in 2001 and is a standard method for evaluating the small bowel, a region that is challenging to evaluate using traditional endoscopy methods. The most common indications for SBCE are obscure gastrointestinal blood loss, iron-deficiency anemia, suspected or established Crohn’s disease (CD), small bowel tumors, inherited polyposis syndromes, abnormal radiological imaging of the small bowel, and refractory celiac disease.3,4 SBCE has a diagnostic yield of approximately 50% across various clinical indications.5,6 Scaramella et al.7 reported that antithrombotic therapies had no significant impact on the diagnostic yield or overt bleeding rates of VCE and, therefore, appeared to be safe with respect to the risk of small bowel bleeding.
Over the past two decades, VCE technology has evolved considerably, addressing the initial limitations of VCE, such as poor resolution, short battery life, long reading time, and difficulty in the precise localization of the detected lesions. In this study, we aimed to review the recent technological developments in VCE.
Since the first commercial use of SBCE in 2001, dual-camera capsules were introduced in 2006.8 The capsule endoscope systems currently in use include PillCam SB3 (Medtronic), MiroCam (IntroMedic), CapsoCam (CapsoVision), EndoCapsule (Olympus), and OMOM Capsule (Jinshan Science and Technology). Recent studies have assessed the diagnostic performance, image quality, reading time, battery life, and patient tolerability of various capsule endoscopy (CE) systems. This section summarizes the key findings of the head-to-head comparative studies on major capsule systems (Table 1).
PillCam SB3 vs. CapsoCam Plus
CapsoCam Plus uses a unique 360° panoramic imaging system with four lateral cameras, generating a maximum of 20 images per second, whereas PillCam SB3 has a field of view of 156° and generates two to six images per second. The second-generation CapsoCam SV-2 was introduced for internal image storage, eliminating the need for external receivers. In a randomized multicenter study involving patients with suspected small bowel bleeding, CapsoCam Plus demonstrated a slightly higher diagnostic yield (39.7%) compared to PillCam SB3 (34.6%), although the difference was not statistically significant.9 However, the evaluation time was longer for CapsoCam Plus (40 minutes) than for PillCam SB3 (27 minutes), potentially owing to the increased image data from the panoramic view. Physician satisfaction was high for both systems, and patient acceptance of CapsoCam Plus was unexpectedly high despite the requirement to retrieve the capsule after excretion.
Hirata et al.10 evaluated the usefulness and acceptability of CapsoCam Plus in Japanese patients and reported similar results.9 CapsoCam required a longer interpretation time; however, it was well accepted by patients and provided superior visualization of both the entire small bowel (97% vs. 73%, p=0.006) and the ampulla of Vater (82% vs. 15%, p<0.001). The diagnostic yields for P2 lesions were 18% and 21%, respectively. The reading time of CapsoCam was significantly longer than that of SB3 (30 vs. 25 minutes, p<0.001), and the time from capsule endoscope swallowing to read completion was longer for CapsoCam than for SB3 (37 vs. 12 hours, p<0.001). Its ability to detect bleeding sources was similar to that of the comparative device in cases of occult and delayed overt small bowel bleeding (>48 hours post-bleeding). These findings suggest that CapsoCam is a valuable diagnostic tool for evaluating patients with suspected small bowel bleeding.
PillCam SB3 vs. MiroCam MC2000
Among the various CE systems, the PillCam SB3 and MiroCam are predominantly utilized in Korea. Specifically, MiroCam, a dual-camera CE system, has expanded the field of view from 170° to 340°, thereby reducing blind spots and potentially enhancing diagnostic accuracy. The comparison between these two systems is as follows: although comparative studies between dual- and single-camera CEs are limited, some studies have compared these CEs. The quality of the small bowel images did not significantly differ between the two groups, and both were deemed to provide sufficient quality for interpretation.11,12 The detection rates for the MC2000 and SB3 were as high as 75% and 41.7%, respectively, and the MC2000 had increased diagnostic yields for neoplastic lesions. Kim et al.13 demonstrated the superiority of using the MC2000 with a dual camera over the SB3 with a single camera for detecting small bowel structures, such as the ampulla of Vater and diverticula, as well as potential lesions causing small bowel bleeding. Because bowel movements may alter the capsule’s orientation, resulting in random forward or backward image acquisition, a single-camera capsule can miss individual lesions.
OMOM vs. MiroCam MC1600
A matched case-control study compared the OMOM HD and MiroCam MC1600 CE systems.14 Both devices showed similar diagnostic yields (55.0%) and comparable global bowel preparation quality. However, the OMOM HD capsule had a shorter small bowel transit time (265±118 minutes) compared to MiroCam MC1600 (307±87 minutes, p=0.020). This study suggests that differences in capsule characteristics, such as resolution and illumination, may affect the perception of preparation quality but do not affect diagnostic yield (Table 2).
OMOM Smart Capsule 2 vs. PillCamSB3
In a randomized head-to-head study involving patients with small bowel bleeding, both the OMOM Smart Capsule 2 and PillCam SB3 demonstrated similar diagnostic yields.15 PillCam SB3 had a longer battery life (816.5 minutes) than the OMOM Smart Capsule 2 (700.5 minutes, p<0.001), whereas OMOM had a shorter download time (33 vs. 132 minutes, p<0.001).16 The agreement between the two devices for detecting P2 lesions was moderate (κ=0.628), indicating that both are safe and effective for clinical use (Table 3).
The introduction of a panenteric capsule allows for comprehensive evaluation of both the entire small bowel and colon with a single procedure. Inherent limitations, such as the inability to insufflate the lumen, affect visualization across all anatomical regions. In the esophagus, rapid capsule transit, particularly in the upright position, can lead to a reduced number of captured mucosal images and an incomplete Z-line assessment.17 Similarly, the stomach’s nontubular anatomy and reliance on passive peristalsis, even with dual-headed capsules, often result in incomplete examination, particularly in the proximal regions.18
In the current guidelines, SBCE is recommended as the third test after performing upper endoscopy and colonoscopy with negative results in patients with gastrointestinal bleeding. Although there are limited results, some recent studies have suggested that panenteric CE, as the second test for patients with gastrointestinal bleeding and negative upper endoscopic findings, provides a high diagnostic yield and shortens the time to diagnosis.
Technological advancements have enabled the creation of capsule systems specifically designed to assess the upper gastrointestinal tract and colon. A novel panenteric CE system, known as the PillCam Crohn’s Capsule (Medtronic), was recently developed. Equipped with dual cameras, it enables the comprehensive visualization of the entire gastrointestinal tract within a single CE session by providing an expanded field of view. CE is an important tool for the diagnostic investigation of patients with suspected CD and for evaluating the extent of small bowel involvement in patients with established CD.19 Initial studies have demonstrated that deep learning algorithms can accurately detect intestinal strictures and categorize ulcer severity in patients with CD. In a pilot investigation led by Ferreira et al.,20 a convolutional neural network (CNN) was developed to analyze images from the PillCam Crohn’s capsule, achieving 83% sensitivity and 98% specificity for ulcer detection and 91% sensitivity and 93% specificity for identifying erosions in the small and large bowel, with a frame analysis rate of 68 frames per second. Zhang et al.21 introduced a magnetically guided robotic capsule system designed for comprehensive gastric evaluation. This innovation marks the beginning of a new phase in which gastric assessment can be followed by small bowel examination using a single device, complemented by artificial intelligence (AI)-assisted interpretation.22 These advancements have transformed CE into a fully integrated AI-based diagnostic modality.23
Recent technological advancements in VCE have significantly improved image enhancement, leading to improved diagnostic accuracy and efficiency. One of the key developments in VCE is the integration of hyperspectral imaging in VCE. The spectrum aided visual enhancer technique converts white-light images into hyperspectral images, enhances mucosal visualization, and improves the detection of subtle lesions. This method has shown close similarity to the narrow-band imaging used in conventional endoscopy, with a structural similarity index metric of 91%.24 In addition, post-processing algorithms have been developed to simulate narrow-band imaging from white-light images. These algorithms enhance the contrast and features of the mucosa, facilitating the early detection of esophageal cancer and other lesions.25
Spectral and dynamic range enhancement (flexible spectral imaging color enhancement and high dynamic range)
VCE traditionally produces standard white-light images; however, newer image-enhanced endoscopy techniques have been adapted from conventional endoscopy to improve visualization. One example is flexible spectral imaging color enhancement (FICE), a digital post-processing method that emphasizes specific light wavelengths to highlight subtle mucosal features. The FICE was originally developed for standard endoscopy in 2005 and has since been incorporated into CE reading systems.26 By reconstructing images with selected red, green, and blue wavelength ranges, the FICE can enhance surface patterns and improve the detection of lesions in the small intestine. Clinical studies have suggested that applying FICE during capsule image review is feasible and may increase the visibility of minute lesions, such as angioectasias or erosions.26-29
Another advancement in VCE imaging is the use of high dynamic range (HDR) techniques to address the capsule’s lighting limitations. HDR imaging combines multiple exposure levels to create a single image with a wide range of brightness and contrast, preventing overexposure of bright areas and elucidating details in dark or shadowed regions.30 In endoscopy, HDR processing yields a more uniformly exposed view of the gastrointestinal mucosa, enhancing the subtle details across varying lighting conditions. This approach has been implemented in newer capsule systems; for instance, one capsule platform uses HDR in conjunction with multiple LEDs to produce clearer images, thereby improving physician confidence in identifying small bowel pathology.
Together, spectral enhancement (such as FICE) and HDR imaging broaden the visual spectrum and dynamic range of CE, leading to improved contrast, color differentiation, and overall image clarity for diagnosis.
Noise reduction techniques
CE images are often hampered by noise owing to low lighting, sensor limitations, and compression. Image noise can obscure mucosal details and mimic pathological textures, thus reducing the diagnostic accuracy. Although basic smoothing filters (such as Gaussian blur or simple bilateral filters) can suppress noise, they tend to blur important details and produce artifacts.31 To address this issue, advanced noise-reduction algorithms have been explored. Techniques such as non-local means filtering, block-matching three-dimensional (3D) filtering, K-nearest neighbor filtering, and adaptive median (AM) filtering have been applied to VCE frames. In a comparative analysis, the AM filter demonstrated superior performance in removing impulsive noise while preserving fine image details, outperforming other methods in maintaining sharpness. Furthermore, researchers have developed noise-reduction methods based on wavelet transforms. For example, Gopi et al.32 introduced a double-density dual-tree complex wavelet transform algorithm for capsule image denoising.
By cleaning up graininess and sensor noise while retaining subtle textures (such as the villous pattern of the small bowel mucosa), these noise-reduction tools enhance image clarity and ensure that diagnostically important features are not masked or distorted. This improved clarity can support clinicians by making abnormalities (e.g., blood spots or tiny ulcers) more conspicuous against the background, thereby aiding their interpretation.
Stereo camera-based capsule endoscopy
Another significant development is the use of stereo camera-based CE for the 3D reconstruction of small bowel lesions. The 3D reconstruction of CE images has long been attempted to obtain more information on small bowel structures. However, because of the limited hardware resources of capsule size and battery capacity, software approaches have been studied, but have mainly exhibited inherent limitations. Recently, stereo camera-based CE, which can perform hardware-enabled 3D reconstructions, has been developed. This technology provides detailed spatial information and enhances the characterization of subepithelial tumors and other abnormalities.33
High-resolution cameras and wide-angle lenses
Advances in imaging technology have led to the development of high-resolution cameras and wide-angle lenses. These innovations improve the image quality and field of view, facilitate comprehensive examinations, and better diagnostic outcomes.34,35
Clinical impact and future directions
Improvements in image enhancement directly translate into better diagnostic support for clinicians using CE. By reducing noise and blur and augmenting image details, these techniques yield clearer visualization of the gastrointestinal mucosa, which is critical for detecting subtle pathologies. Studies have documented that enhanced images lead to quantifiable gains in quality metrics (e.g., higher clarity and similarity to the ground truth), and qualitatively, endoscopists can more readily identify lesions when images are optimized.36 Generally, the clinical utility of these enhancements is evidenced by their contribution to higher diagnostic yields. Park et al.36 noted that computational image processing can correct imaging errors and improve the overall image quality, ultimately boosting the diagnostic performance of CE without any hardware changes.
Enhanced clarity indicates that inflammatory lesions, bleeding sources, or even tiny polyps are less likely to be missed because of poor image quality. Moreover, clearer images can shorten review times, both by making manual reviews more efficient and providing better inputs for computer-aided detection algorithms. In addition to the current state-of-the-art methods, AI is poised to further advance image enhancement in VCE. Deep learning techniques, which have already shown remarkable success in automated lesion detection in capsule images, are beginning to be applied in image-enhancement tasks as well.37 In theory, CNN can be trained to perform denoising, deblurring, or super-resolution by learning from large capsule image datasets. Early explorations of AI-based enhancements are promising, as these data-driven models could potentially learn the complex patterns of noise or blur in CE and correct them in real time. The introduction of such deep learning methods in CE analysis has been highlighted as a major future direction and is expected to bring outstanding improvements in both image quality and interpretative accuracy. Ultimately, the combination of advanced image enhancement algorithms and AI-assisted analysis will augment the clinicians’ ability to interpret CE examinations. By providing sharper and more diagnostic images, these technologies strive to maximize the clinical yield of VCE, enabling more reliable detection of gastrointestinal diseases, improving patient outcomes, and reducing the need for repeated diagnostic procedures. As multidisciplinary collaboration between gastroenterologists and imaging scientists continues, we expect an ongoing refinement of these enhancement tools and their integration into routine CE practice for superior visualization and diagnostic confidence.
In summary, recent advancements in AI, hyperspectral imaging, and super-resolution techniques have significantly improved image enhancement in VCE, leading to better diagnostic accuracy and efficiency in gastrointestinal evaluations.
The European Society of Gastrointestinal Endoscopy (ESGE) advises the use of a modified diet along with purgative agents, most commonly 2 L of polyethylene glycol (PEG), before performing SBCE to enhance mucosal visualization.38 Evidence from systematic reviews and randomized controlled trials supports the effectiveness of purgatives in improving small bowel visibility.39-43 For large bowel preparation, the administration of 1 L of PEG combined with ascorbate (PEG-ASC) has been shown to significantly improve patient adherence and tolerability.44,45 The low-volume 1 L PEG-ASC regimen has shown efficacy as a substitute for traditional high-volume preparations prior to colonoscopy and may thus serve as a potential new standard for bowel preparation in patients undergoing SBCE.46 Recently, a study compared the efficacy of the standard 2 L PEG solution with a 1 L PEG-ASC preparation and proved that the 1 L PEG-ASC bowel preparation is comparable to 2 L PEG in terms of visibility in SBCE.47
The accurate localization of lesions remains critical for clinical utility. Recent technological enhancements have integrated precise localization techniques and efficient data handling mechanisms. Numerous studies have shown that panenteric CE demonstrates high performance for assessing CD mucosal activity and extent compared to more invasive methods such as magnetic resonance enterography and/or ileocolonoscopy, without the need for multiple tests in non-stricturing Crohn’s.48 Recently, magnetically controlled CE (MCE), radio frequency (RF), video, and hybrid-based localization have shown promise, particularly when integrated with on-chip technologies and AI-based models.
The MCE represents a leap forward in active navigation and targeted diagnostics. Assisted by robotic arms, MCE allows precise control and positioning within the gastrointestinal tract, improving visualization and enabling potential therapeutic interventions.49,50 It utilizes static or dynamic magnetic fields to track the capsule’s position and orientation.51 It exhibited high precision (errors <5 mm and <5°), was robust against tissue heterogeneity, and was suitable for active locomotion. However, they can be sensitive to environmental magnetic noise; the integration of magnets and sensors is space-constrained, and external magnetic fields may interfere. Recent innovations include sensor fusion and miniaturized chips, which enhance the feasibility of real-world applications.51
Radiofrequency-based CE enables localization without direct visual access, and some techniques integrate wireless power transfer. RF-based systems provide noninvasive positioning but face challenges with tissue interference and time synchronization. AI-enhanced models and RF identification-based designs have shown promise for improving accuracy. RF-based localization exploits the interaction between electromagnetic waves and human tissues.52 These interactions, which include attenuation, reflection, and phase changes, generate measurable signal characteristics that can be used for positioning. The growing trend in this field is toward integrating AI, machine learning, and advanced optimization algorithms, which help reduce errors despite the complexities of in-body signal propagation.
Video-based localization uses visual data captured by a capsule to estimate position and detect anomalies. Although video-based approaches can visually pinpoint the pathology and orientation, they suffer from limited image clarity and require high computational resources. AI-driven enhancements have significant diagnostic potential. Deep learning-type (CNNs) can efficiently analyze visual data, and principal component analysis reduces the dimensionality of endoscopic images to extract the essential features, which results in enhanced accuracy of localization.51,53,54
Hybrid systems synergize multiple modalities for enhanced accuracy, but raise practical concerns regarding device size, power use, and system complexity. For example, there is a fusion of video+RF for simultaneous localization and mapping and magnetic and inertial measurement unit-based systems with sensor fusion for better pose estimation.
There are also ESGE guidelines with a statement on CE completeness. “For acceptance of AI in evaluating the completeness of an SBCE investigation, the AI-assisted definition of completeness should be comparable to the identification of the cecum, colon, or stoma by experienced endoscopists.” Accurate computer-aided identification of anatomical landmarks such as the cecum, colon, or stoma—and thereby precise localization of the capsule within the gastrointestinal tract—remains a significant challenge. Only a few studies have explored the use of AI in automated capsule localization.55,56 These investigations primarily relied on visual odometry, a technique that estimates the position and orientation of a device by analyzing sequential camera images.
One notable advancement is the development of computer-aided analysis (CADx) methods that utilize deep learning models, such as swine transformers. These models classify endoscopic images into different segments of the gastrointestinal tract (stomach, small intestine, and large intestine) with high precision, followed by algorithms such as K-means to correct classification outliers, and localization algorithms to determine the capsule's position. This approach has demonstrated high precision rates (93.46% for the stomach, 97.28% for the small intestine, and 98.68% for the large intestine) and minimal transition time errors, significantly reducing the need for real-time visual inspection by medical staff.57
Another significant development is the use of AI models based on CNNs and deep neural networks (DNNs) to detect and characterize lesions. For instance, AI models such as Inception-Resnet-V2 have been trained to identify clinically significant lesions in CE images, improve lesion detection rate, and reduce reading times for both expert and trainee reviewers.58 In addition, explainable AI models have been developed to autonomously detect, localize, and characterize colorectal polyps, achieve high sensitivity and specificity, and integrate them seamlessly into clinical workflows.59
In conclusion, recent technological advances in VCE, including AI integration, 3D reconstruction, magnetic control, high-resolution imaging, and improved compression techniques, have significantly enhanced the diagnostic and therapeutic capabilities of this noninvasive modality. These innovations promise to improve patient outcomes and expand the clinical applications of wireless capsule endoscopy.
CE image analysis involves prolonged and often monotonous review periods. Moreover, clinically relevant lesions are present in only a minor proportion of total images.60 Capsule reading is a labor-intensive and frequently monotonous task, with reported average interpretation times ranging from 30 to 120 minutes.38 A sustained, undistracted focus is required to ensure that pathological findings are not overlooked.61 Consequently, an extended duration of image review may contribute to elevated miss rates, even among experienced and well-trained interpreters.62 Selecting the most relevant images for review can significantly reduce interpretation time and enhance diagnostic accuracy. Nevertheless, the automated detection of pathological findings in CE has historically posed considerable challenges.54,63-66 Since the advent of deep learning techniques in computer vision, there has been a substantial improvement in image recognition performance across large-scale datasets. Previous studies have used AI-assisted detection.
Current evidence indicates that several algorithms have demonstrated both effectiveness and reliability in detecting small bowel mucosal abnormalities, achieving a diagnostic accuracy comparable to that of expert readers. However, most studies in this area have assessed the performance of computer-assisted systems with a focus on detecting a single lesion type at a time (e.g., bleeding, vascular lesions, ulcers, or protruding masses).63,64,67 However, in routine clinical settings, the specific nature of potential small intestinal lesions is often unpredictable prior to capsule ingestion. More recently, multiclass detection algorithms have shown the ability to not only identify a variety of abnormalities but also prioritize those with clinical significance.68-71 Aoki et al.68 aimed to develop a deep-learning-based AI system to automatically detect erosions and ulcerations in wireless CE images. Using a deep CNN trained on 5,360 annotated images and tested on an independent set of 10,440 images, the system achieved high diagnostic performance. The CNN processed the test set in 233 seconds. Xie et al.72 recently reported the development and validation of AI software for reviewing SBCE videos based on a multicenter study in China. The system was trained using CNNs to identify 17 distinct findings defined by CE structured terminology, including venous structures, nodules, masses or tumors, polyps, angioectasia, red and white plaques, red spots, abnormal villi, lymphangiectasia, erythematous and edematous erosions, ulcers, aphthae, bleeding, and parasites.73 The AI algorithm demonstrated a remarkably low miss rate of 1.0% when compared to conventional reading. Conversely, the conventional reading group exhibited a miss rate of 9.6% relative to the AI-assisted interpretation. Recently, the U.S. Food and Drug Administration (FDA) approved the NaviCam small bowel system integrated with the ProScan DNN (developed by ANKON and AnX Robotica), marking the first commercially available AI-assisted reading platform. Subsequently, a prospective, multicenter, non-inferiority study evaluating the system’s ability to detect small bowel bleeding lesions was published. The AI model was trained to differentiate the normal small bowel mucosa from various abnormalities, including hemorrhagic sites, vascular lesions, protrusions, inflammatory changes, ulcers, and lymphangiectasia. Notably, the diagnostic yield for bleeding lesions using the AI-assisted system was not only noninferior but superior to conventional readings, achieving significantly faster interpretation times (3.8 minutes compared to 33.7 minutes).74
Numerous studies have consistently demonstrated that CNN-based systems are significantly faster than human readers in processing and analyzing CE images.63,64,67-71 Recently, a systematic review and meta-analysis analyzed six studies involving proprietary AI platforms (e.g., the NaviCam ProScan and OMOM SmartScan) using CNNs.75 This review compared the diagnostic performance of AI-assisted SBCE with conventional human-only reading.75 The AI-assisted SBCE showed higher sensitivity than human readers (up to 99.9% vs. 75%–89%), comparable specificity, significantly reduced reading time (average 4.7 vs. 56.7 minutes), and improved diagnostic odds ratio (10.3 AI vs. 7.4 conventional). AI was particularly effective in reducing the image burden and improving the detection of subtle or rare findings, with consistent performance across studies. NaviCam ProScan is an AI-assisted reading tool for small bowel VCE. Anx Robotica has received approval from the FDA for its NaviCam ProScan. Another meta-analysis analyzed 14 studies comparing AI-assisted CE and conventional CE for the detection of small bowel lesions.76 The aggregated diagnostic accuracy for conventional CE was 0.966 (95% confidence interval [CI], 0.925–0.988), slightly higher than that of AI-assisted CE, which was 0.919 (95% CI, 0.914–0.923). In terms of sensitivity, AI-assisted CE demonstrated superior performance, with a pooled value of 0.924 (95% CI, 0.865–0.987) compared to 0.860 (95% CI, 0.786–0.934) for conventional CE. The positive predictive value was higher in conventional CE at 0.982 (95% CI, 0.976–0.987), while AI-assisted CE yielded a positive predictive value of 0.893 (95% CI, 0.755–0.999). Conversely, pooled specificity was markedly higher for conventional CE at 0.998 (95% CI, 0.996–0.999) compared to 0.537 (95% CI, 0.524–0.549) for AI-assisted CE. Notably, AI-assisted CE outperformed in terms of negative predictive value, which was 0.943 (95% CI, 0.939–0.946), surpassing the negative predictive value of conventional CE at 0.760 (95% CI, 0.577–0.943).
The CNN model effectively identified clinically significant lesions that were overlooked in initial human interpretation of SBCE, with diagnostic alterations in nearly 10% of cases.77 AI detects missed findings without overlooking previous recognitions by human readers. Meaningful lesions were identified in 61.2% of cases previously deemed negative. The rebleeding rate was 23.6% in patients with meaningful findings and 16.1% in those without meaningful findings. These findings suggest that deep learning can mitigate human errors and enhance diagnostic accuracy in routine practice. However, its definitive effect on patient outcomes remains to be clarified in future prospective trials. This study advocates AI as a supportive tool for easing endoscopy workloads and enhancing novice training.
Several studies have developed AI-based algorithms for irritable bowel disease diagnosis using small-bowel and colonic CE videos with varying numbers of training images and compared the results of endoscopists, experts, and fellows. Among them, one study evaluated the role of AI in video-CE for ulcerative colitis (UC), and three were prospective studies. In a prospective study including UC lesions, the use of the DL ResNet50 framework, with a computational performance of 25 fps, achieved diagnostic accuracy rates of 99.2% and 98.3% for the training and validation datasets, respectively.78 This DL model has been proven to be a useful tool for reducing the burden of image interpretation on endoscopists. The other two prospective studies, which included CD lesions, used the DL ResNet50 and AXARO (Augmented Endoscopy) frameworks, applying the ResNet50 framework with a patient-dependent split of images for training, validation, and testing. The diagnostic sensitivity, specificity, and accuracy for CD-related ulcers were 95.7% (95% CI, 93.4%–97.4%), 99.8% (95% CI, 99.2%–100.0%), and 98.4% (95% CI, 97.6%–99.0%), respectively, with two expert readers as comparators.79 In this study, the diagnostic accuracy was equally high for both the small bowel and colon. The AXARO framework, applied in a prospective multicenter study of patients with suspected CD, achieved a 97.1% reduction in analyzable images and up to a 94% reduction in the reading time (pooled median review time=3.2 minutes per patient) compared to fully read capsules. It also demonstrated a sensitivity and specificity of 92%–96% and 90%–93%, respectively, and an area under the curve of 0.91–0.94, highlighting its potential as a rapid tool for ruling out irritable bowel disease in patients undergoing panenteric video CE. The reported diagnostic sensitivity, specificity, and accuracy of CD-related lesions from the other retrospective studies assessing different CNN and DL models on video CE images varied from 88.2% to 98%, 89% to 99.9%, and 90.5% to 99%, respectively.80
However, the ESGE guidelines state that an increased sensitivity may lead to a greater number of flagged areas requiring verification by the reader, potentially prolonging the reading process. Additionally, there is an ongoing debate regarding the optimal integration of these AI systems into routine SBCE workflows: whether they should be employed prior to the clinician’s review (by preselecting suspected areas), concurrently during the review process (similar to AI applications in colonoscopy), or as a secondary review following the initial reading. Nevertheless, with the continued advancement of AI technology, it is anticipated that the guidelines related to SBCE reading times and protocols will be increasingly refined and standardized.
Recent technological advancements in VCE, including AI-assisted analysis, enhanced imaging modalities, and improved localization techniques, have significantly expanded its diagnostic capabilities. These innovations have reduced interpretation time, increased diagnostic accuracy, and enabled panenteric evaluation in a single procedure. As these technologies continue to evolve, VCE has become an efficient, reliable, and integral tool for gastrointestinal diagnostics and disease monitoring.
Table 1.
Comparison of capsule endoscopy
Feature PillCam SB3 MiroCam MC2000 CapsoCam Plus OMOM HD EndoCapsule
Company Medtronic IntroMedic CapsoVision Jinshan Science and Technology Olympus
Size (mm) 11.4×26.2 10.8×30 11×31 11×25 11×26
Weight (g) 3 3.25–4.70 3.8 4.5 3.3
No. of cameras 1 2 (dual-headed) 4 (side of capsule) 1 1
Field of view (º) 156 170 (×2) 360 (panoramic imaging) 172 160
Battery life (hr) 8–15 12 15 10 12
Transmission type RF telemetry Human body communication N/A RF telemetry RF telemetry
Real-time viewing Yes Yes No Yes Yes
Storage type External receiver External receiver Onboard storage (capsule retrieval required) External receiver External receiver
Frame velocity (frames/s) 2–6 3 fps per camera, total 6 fps 12–20 2–6 2

RF, radio frequency; N/A, not available; fps, frames per second.

Table 2.
Comparison of OMOM HD and MiroCam MC1600
Parameter OMOM HD Mirocam MC1600 p-value
Complete examination (%) 97.2 98.1 0.762
Capsule retention (%) 0.9 0.9 0.999
Transit times (min, mean±SD)
 Gastric 53.0±43.0 56.0±47.0 0.867
 Small bowel 265.0±118.0 307.0±87.0 0.032
Bowel preparation (mean±SD)a)
 Global small bowel 8.2±1.2 7.9±1.0 0.113
 First tertile 8.5±1.1 8.0±1.1 0.017
 Second tertile 8.2±1.3 7.5±1.1 0.004
 Third tertile 7.9±1.6 7.1±1.3 0.003
Findings (%)
 Diagnostic yield 54.2 56.1 0.613
 Ulcers/erosions 30.8 32.7 0.561
 Angioectasia 20.5 15.9 0.103
 Subepithelial lesion 2.8 0.9 0.099
 Adenocarcinoma 0.9 0.0 0.394
 Extra-small bowel findings 6.1 5.6 0.112

SD, standard deviation.

a)Classified according to the adapted Brotz scale.

Table 3.
Comparison of PillCam SB3 and OMOM Smart Capsule 2
Feature PillCam SB3 OMOM Smart Capsule 2 p-value
Complete small bowel visualization (%) 95.5 95.5 1
Small bowel transit time (min) 310.5 295.5 0.653
Download time (min) 132.0 33.0 <0.001
Detection rate for all P2 lesions (%) 72.7 65.9 0.784
Diagnostic yield 88.6 77.3 0.256
Extraintestinal lesions 9.1 6.8 0.694
  • 1. Elli L, Centorrino E, Costantino A, et al. Capsule enteroscopy versus small-bowel ultrasonography for the detection and differential diagnosis of intestinal diseases. Clin Endosc 2022;55:532–539.ArticlePubMedPMCPDF
  • 2. Kim SE, Kim HJ, Koh M, et al. A practical approach for small bowel bleeding. Clin Endosc 2023;56:283–289.ArticlePubMedPMCPDF
  • 3. Chetcuti Zammit S, Sidhu R. Capsule endoscopy: recent developments and future directions. Expert Rev Gastroenterol Hepatol 2021;15:127–137.ArticlePubMed
  • 4. Park J, Kim JS, Song JH, et al. Discordance rate and risk factor of other diagnostic modalities for small bowel tumors detected by device-assisted enteroscopy: a Korean Association for the Study of Intestinal Disease (KASID) multicenter study. Gut Liver 2024;18:686–694.ArticlePubMedPMC
  • 5. Hong SM, Jung SH, Baek DH. Diagnostic yields and clinical impacts of capsule endoscopy. Diagnostics (Basel) 2021;11:1842.ArticlePubMedPMC
  • 6. Romeo S, Neri B, Mossa M, et al. Diagnostic yield of small bowel capsule endoscopy in obscure gastrointestinal bleeding: a real-world prospective study. Intern Emerg Med 2022;17:349–358.ArticlePubMedPMCPDF
  • 7. Scaramella L, Zammit SC, Sidhu R, et al. Effect of antithrombotic therapies on small bowel bleeding: an European multicenter retrospective study. Clin Endosc 2025;58:102–111.ArticlePubMedPMCPDF
  • 8. Eliakim R, Fireman Z, Gralnek IM, et al. Evaluation of the PillCam Colon capsule in the detection of colonic pathology: results of the first multicenter, prospective, comparative study. Endoscopy 2006;38:963–970.ArticlePubMed
  • 9. Zwinger LL, Siegmund B, Stroux A, et al. CapsoCam SV-1 versus PillCam SB 3 in the detection of obscure gastrointestinal bleeding: results of a prospective randomized comparative multicenter study. J Clin Gastroenterol 2019;53:e101–e106.ArticlePubMed
  • 10. Hirata I, Tsuboi A, Matsubara Y, et al. Clinical usefulness and acceptability of small-bowel capsule endoscopy with panoramic imaging compared with axial imaging in Japanese patients. DEN Open 2024;5:e389.ArticlePubMedPMC
  • 11. Yung DE, Robertson AR, Davie M, et al. Double-headed small-bowel capsule endoscopy: real-world experience from a multi-centre British study. Dig Liver Dis 2021;53:461–466.ArticlePubMed
  • 12. Kim SH, Choi HS, Chun HJ, et al. Diagnostic benefit of simultaneous capsule endoscopy using two different systems. Gastroenterol Res Pract 2018;2018:9798546.ArticlePubMedPMCPDF
  • 13. Kim JE, Kim ER, Park JJ, et al. Comparative study of MiroCam MC2000 and PillCam SB3 in detecting small bowel bleeding: a multicenter prospective randomized crossover study. Gut Liver 2025;19:569–578.ArticlePubMed
  • 14. Estevinho MM, Pinho R, Rodrigues A, et al. Capsule enteroscopy using the Mirocam® versus OMOM® systems: a matched case-control study. Life (Basel) 2023;13:1809.ArticlePubMedPMC
  • 15. Liao Z, Gao R, Li F, et al. Fields of applications, diagnostic yields and findings of OMOM capsule endoscopy in 2400 Chinese patients. World J Gastroenterol 2010;16:2669–2676.ArticlePubMedPMC
  • 16. Blanco-Velasco G, Zamarripa-Mottú RA, Solórzano-Pineda OM, et al. Comparison in the diagnostic yield between "Pillcam SB3" capsule endoscopy and "OMOM Smart Capsule 2" in small bowel bleeding: a randomized head-to-head study. Dig Dis 2021;39:211–216.ArticlePubMedPDF
  • 17. Park J, Cho YK, Kim JH. Current and future use of esophageal capsule endoscopy. Clin Endosc 2018;51:317–322.ArticlePubMedPMCPDF
  • 18. Kim JH, Nam SJ. Capsule endoscopy for gastric evaluation. Diagnostics (Basel) 2021;11:1792.ArticlePubMedPMC
  • 19. Maaser C, Sturm A, Vavricka SR, et al. ECCO-ESGAR guideline for diagnostic assessment in IBD part 1: initial diagnosis, monitoring of known IBD, detection of complications. J Crohns Colitis 2019;13:144–164.ArticlePubMed
  • 20. Ferreira JP, de Mascarenhas Saraiva MJ, Afonso JP, et al. Identification of Ulcers and erosions by the Novel Pillcam™ Crohn's capsule using a convolutional neural network: a multicentre pilot study. J Crohns Colitis 2022;16:169–172.ArticlePubMedPDF
  • 21. Zhang Y, Zhang Y, Huang X. Development and application of magnetically controlled capsule endoscopy in detecting gastric lesions. Gastroenterol Res Pract 2021;2021:2716559.ArticlePubMedPMCPDF
  • 22. Rey JF. Gastric examination by guided capsule endoscopy: a new era. Lancet Gastroenterol Hepatol 2021;6:879–880.ArticlePubMed
  • 23. Rey JF. As how artificial intelligence is revolutionizing endoscopy. Clin Endosc 2024;57:302–308.ArticlePubMedPMCPDF
  • 24. Wang YP, Karmakar R, Mukundan A, et al. Spectrum aided vision enhancer enhances mucosal visualization by hyperspectral imaging in capsule endoscopy. Sci Rep 2024;14:22243.ArticlePubMedPMCPDF
  • 25. Yang KY, Fang YJ, Karmakar R, et al. Assessment of narrow band imaging algorithm for video capsule endoscopy based on decorrelated color space for esophageal cancer. Cancers (Basel) 2023;15:4715.ArticlePubMedPMC
  • 26. Ogata N, Ohtsuka K, Ogawa M, et al. Image-enhanced capsule endoscopy improves the identification of small intestinal lesions. Diagnostics (Basel) 2021;11:2122.ArticlePubMedPMC
  • 27. Pohl J, Aschmoneit I, Schuhmann S, et al. Computed image modification for enhancement of small-bowel surface structures at video capsule endoscopy. Endoscopy 2010;42:490–492.ArticlePubMed
  • 28. ASGE Technology Committee, Manfredi MA, Abu Dayyeh BK, et al. Electronic chromoendoscopy. Gastrointest Endosc 2015;81:249–261.ArticlePubMed
  • 29. Mishkin DS, Chuttani R, Croffie J, et al. ASGE technology status evaluation report: wireless capsule endoscopy. Gastrointest Endosc 2006;63:539–545.ArticlePubMed
  • 30. Kumar GA, G S R, S P P, et al. Improving endoscopic image quality through the use of high dynamic range imaging-like method with real-time performance. Annu Int Conf IEEE Eng Med Biol Soc 2023;2023:1–4.ArticlePubMed
  • 31. Obukhova N, Motyko A, Pozdeev A, et al. Review of noise reduction methods and estimation of their effectiveness for medical endoscopic images processing. In: Proceedings of the 2018 22nd Conference of Open Innovations Association (FRUCT); 2018 May 15-18; Jyvaskyla, Finland. p. 204–210.Article
  • 32. Gopi VP, Palanisamy P, Niwas SI. Capsule endoscopic colour image denoising using complex wavelet transform. In: Proceedings of the Wireless Networks and Computational Intelligence (ICIP 2012); 2012 August 10-12; Berlin, Heidelberg. p. 220–229.ArticlePubMed
  • 33. Nam SJ, Lim YJ, Nam JH, et al. 3D reconstruction of small bowel lesions using stereo camera-based capsule endoscopy. Sci Rep 2020;10:6025.ArticlePubMedPMCPDF
  • 34. Tontini GE, Wiedbrauck F, Cavallaro F, et al. Small-bowel capsule endoscopy with panoramic view: results of the first multicenter, observational study (with videos). Gastrointest Endosc 2017;85:401–408.ArticlePubMed
  • 35. Moghaddasi M, Hu B, Abrinaei F. Highly robust performance of a compact, wide-angle, nano-optic capsule endoscopy camera. J Opt Soc Am A Opt Image Sci Vis 2023;40:2298–2306.ArticlePubMed
  • 36. Park J, Hwang Y, Yoon JH, et al. Recent development of computer vision technology to improve capsule endoscopy. Clin Endosc 2019;52:328–333.ArticlePubMedPMCPDF
  • 37. Hwang Y, Park J, Lim YJ, et al. Application of artificial intelligence in capsule endoscopy: where are we now? Clin Endosc 2018;51:547–551.ArticlePubMedPMCPDF
  • 38. Rondonotti E, Spada C, Adler S, et al. Small-bowel capsule endoscopy and device-assisted enteroscopy for diagnosis and treatment of small-bowel disorders: European Society of Gastrointestinal Endoscopy (ESGE) technical review. Endoscopy 2018;50:423–446.ArticlePubMed
  • 39. Kotwal VS, Attar BM, Gupta S, et al. Should bowel preparation, antifoaming agents, or prokinetics be used before video capsule endoscopy?: a systematic review and meta-analysis. Eur J Gastroenterol Hepatol 2014;26:137–145.ArticlePubMed
  • 40. Niv Y. Efficiency of bowel preparation for capsule endoscopy examination: a meta-analysis. World J Gastroenterol 2008;14:1313–1317.ArticlePubMedPMC
  • 41. Rokkas T, Papaxoinis K, Triantafyllou K, et al. Does purgative preparation influence the diagnostic yield of small bowel video capsule endoscopy?: a meta-analysis. Am J Gastroenterol 2009;104:219–227.ArticlePubMedPDF
  • 42. Spada C, McNamara D, Despott EJ, et al. Performance measures for small-bowel endoscopy: a European Society of Gastrointestinal Endoscopy (ESGE) quality improvement initiative. Endoscopy 2019;51:574–598.ArticlePubMed
  • 43. Wu S, Gao YJ, Ge ZZ. Optimal use of polyethylene glycol for preparation of small bowel video capsule endoscopy: a network meta-analysis. Curr Med Res Opin 2017;33:1149–1154.ArticlePubMed
  • 44. Cash BD, Moncrief MB, Epstein MS, et al. Patient experience with NER1006 as a bowel preparation for colonoscopy: a prospective, multicenter US survey. BMC Gastroenterol 2021;21:70.ArticlePubMedPMCPDF
  • 45. Maida M, Macaluso FS, Sferrazza S, et al. Effectiveness and safety of NER1006 versus standard bowel preparations: a meta-analysis of randomized phase-3 clinical trials. Dig Liver Dis 2020;52:833–839.ArticlePubMed
  • 46. Pontone S, Angelini R, Standoli M, et al. Low-volume plus ascorbic acid vs high-volume plus simethicone bowel preparation before colonoscopy. World J Gastroenterol 2011;17:4689–4695.ArticlePubMedPMC
  • 47. Caccia R, Rimondi A, Elli L, et al. Bowel preparation for small bowel capsule endoscopy: standard regimen with 2 L polyethylene glycol versus 1 L polyethylene glycol plus ascorbate. Clin Endosc 2025;58:285–290.ArticlePubMedPMCPDF
  • 48. Bruining DH, Oliva S, Fleisher MR, et al. Panenteric capsule endoscopy versus ileocolonoscopy plus magnetic resonance enterography in Crohn's disease: a multicentre, prospective study. BMJ Open Gastroenterol 2020;7:e000365.ArticlePubMedPMC
  • 49. Mascarenhas M, Martins M, Afonso J, et al. The future of minimally invasive capsule panendoscopy: robotic precision, wireless imaging and AI-driven insights. Cancers (Basel) 2023;15:5861.ArticlePubMedPMC
  • 50. Rosa B, Cotter J. Capsule endoscopy and panendoscopy: a journey to the future of gastrointestinal endoscopy. World J Gastroenterol 2024;30:1270–1279.ArticlePubMedPMC
  • 51. Li K, Xu Y, Zhao Z, et al. Leveraging three orthogonal magnetic field sources for localization of a magnetic capsule endoscope with unique solution. In: Proceedings of the 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO); 2022 Dec 5–9; Jinghong, China. p. 1929–1933.Article
  • 52. Umay I, Fidan B, Barshan B. Localization and tracking of implantable biomedical sensors. Sensors (Basel) 2017;17:583.ArticlePubMedPMC
  • 53. Muruganantham P, Balakrishnan S. Attention aware deep learning model for wireless capsule endoscopy lesion classification and localization. J Med Biol Eng 2022;42:157–168.ArticlePDF
  • 54. Trasolini R, Byrne MF. Artificial intelligence and deep learning for small bowel capsule endoscopy. Dig Endosc 2021;33:290–297.ArticlePubMedPDF
  • 55. Dimas G, Spyrou E, Iakovidis DK, et al. Intelligent visual localization of wireless capsule endoscopes enhanced by color information. Comput Biol Med 2017;89:429–440.ArticlePubMed
  • 56. Lowe DG. Distinctive image features from scale-invariant keypoints. Int J Comput Vis 2004;60:91–110.ArticlePDF
  • 57. Zhang R, Peng B, Liu Y, et al. Localization of capsule endoscope in alimentary tract by computer-aided analysis of endoscopic images. Sensors (Basel) 2025;25:746.ArticlePubMedPMC
  • 58. Park J, Hwang Y, Nam JH, et al. Artificial intelligence that determines the clinical significance of capsule endoscopy images can increase the efficiency of reading. PLoS One 2020;15:e0241474.ArticlePubMedPMC
  • 59. Nadimi ES, Braun JM, Schelde-Olesen B, et al. Towards full integration of explainable artificial intelligence in colon capsule endoscopy's pathway. Sci Rep 2025;15:5960.ArticlePubMedPMCPDF
  • 60. Faigel DO, Baron TH, Adler DG, et al. ASGE guideline: guidelines for credentialing and granting privileges for capsule endoscopy. Gastrointest Endosc 2005;61:503–505.ArticlePubMed
  • 61. Rondonotti E, Pennazio M, Toth E, et al. How to read small bowel capsule endoscopy: a practical guide for everyday use. Endosc Int Open 2020;8:E1220–E1224.ArticlePubMedPMC
  • 62. Zheng Y, Hawkins L, Wolff J, et al. Detection of lesions during capsule endoscopy: physician performance is disappointing. Am J Gastroenterol 2012;107:554–560.ArticlePubMedPDF
  • 63. Dray X, Iakovidis D, Houdeville C, et al. Artificial intelligence in small bowel capsule endoscopy: current status, challenges and future promise. J Gastroenterol Hepatol 2021;36:12–19.ArticlePubMedPDF
  • 64. Soffer S, Klang E, Shimon O, et al. Deep learning for wireless capsule endoscopy: a systematic review and meta-analysis. Gastrointest Endosc 2020;92:831–839.ArticlePubMed
  • 65. Tsuboi A, Oka S, Aoyama K, et al. Artificial intelligence using a convolutional neural network for automatic detection of small-bowel angioectasia in capsule endoscopy images. Dig Endosc 2020;32:382–390.ArticlePubMedPDF
  • 66. Vieira PM, Freitas NR, Lima VB, et al. Multi-pathology detection and lesion localization in WCE videos by using the instance segmentation approach. Artif Intell Med 2021;119:102141.ArticlePubMed
  • 67. Qin K, Li J, Fang Y, et al. Convolution neural network for the diagnosis of wireless capsule endoscopy: a systematic review and meta-analysis. Surg Endosc 2022;36:16–31.ArticlePubMedPMCPDF
  • 68. Aoki T, Yamada A, Kato Y, et al. Automatic detection of various abnormalities in capsule endoscopy videos by a deep learning-based system: a multicenter study. Gastrointest Endosc 2021;93:165–173.ArticlePubMed
  • 69. Ding Z, Shi H, Zhang H, et al. Gastroenterologist-level identification of small-bowel diseases and normal variants by capsule endoscopy using a deep-learning model. Gastroenterology 2019;157:1044–1054.ArticlePubMed
  • 70. Mascarenhas Saraiva MJ, Afonso J, Ribeiro T, et al. Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network. BMJ Open Gastroenterol 2021;8:e000753.ArticlePubMedPMC
  • 71. Piccirelli S, Milluzzo SM, Bizzotto A, et al. Small bowel capsule endoscopy and artificial intelligence: first or second reader? Best Pract Res Clin Gastroenterol 2021;52-53:101742.ArticlePubMed
  • 72. Xie X, Xiao YF, Zhao XY, et al. Development and validation of an artificial intelligence model for small bowel capsule endoscopy video review. JAMA Netw Open 2022;5:e2221992.ArticlePubMedPMC
  • 73. Korman LY, Delvaux M, Gay G, et al. Capsule endoscopy structured terminology (CEST): proposal of a standardized and structured terminology for reporting capsule endoscopy procedures. Endoscopy 2005;37:951–959.ArticlePubMed
  • 74. Spada C, Piccirelli S, Hassan C, et al. AI-assisted capsule endoscopy reading in suspected small bowel bleeding: a multicentre prospective study. Lancet Digit Health 2024;6:e345–e353.ArticlePubMed
  • 75. Cortegoso Valdivia P, Fantasia S, Kayali S, et al. Conventional small-bowel capsule endoscopy reading vs proprietary artificial intelligence auxiliary systems: systematic review and meta-analysis. Endosc Int Open 2025;13:a25442863.ArticlePubMedPMC
  • 76. Dhali A, Kipkorir V, Maity R, et al. Artificial intelligence-assisted capsule endoscopy versus conventional capsule endoscopy for detection of small bowel lesions: a systematic review and meta-analysis. J Gastroenterol Hepatol 2025;40:1105–1118.ArticlePubMedPMC
  • 77. Choi KS, Park D, Kim JS, et al. Deep learning in negative small-bowel capsule endoscopy improves small-bowel lesion detection and diagnostic yield. Dig Endosc 2024;36:437–445.ArticlePubMed
  • 78. Higuchi N, Hiraga H, Sasaki Y, et al. Automated evaluation of colon capsule endoscopic severity of ulcerative colitis using ResNet50. PLoS One 2022;17:e0269728.ArticlePubMedPMC
  • 79. Majtner T, Brodersen JB, Herp J, et al. A deep learning framework for autonomous detection and classification of Crohn's disease lesions in the small bowel and colon with capsule endoscopy. Endosc Int Open 2021;9:E1361–E1370.ArticlePubMedPMC
  • 80. Brodersen JB, Jensen MD, Leenhardt R, et al. Artificial intelligence-assisted analysis of pan-enteric capsule endoscopy in patients with suspected Crohn's disease: a study on diagnostic performance. J Crohns Colitis 2024;18:75–81.ArticlePubMedPDF

Figure & Data

REFERENCES

    Citations

    Citations to this article as recorded by  
    • Pillcam Genius Pilot Experience – An Eruditive Innovative -Time to ditch the belt?
      N Nandi, FW D Tai, X Dray, M Keuchel, P Baltes, L Elli, L Scaramella, A Cosenza, R Sidhu
      Endoscopy.2026; 58(S 03): S364.     CrossRef
    • Isolated cavernous hemangioma of the hepatic flexure mimicking a colonic neoplasm: A case report
      Tya Youssef, Ahmad Karim Mourad, Philippe Attieh, Karam Karam, Jessy Fadel, Mazen Farhat, Elias Fiani, Mona Hallak
      Medical Reports.2026; 19: 100491.     CrossRef

    • Cite
      CITE
      export Copy Download
      Close
      Download Citation
      Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.
      Format:
      • RIS — For EndNote, ProCite, RefWorks, and most other reference management software
      • BibTeX — For JabRef, BibDesk, and other BibTeX-specific software
      Include:
      • Citation for the content below
      Recent technological advances in video capsule endoscopy: a comprehensive review
      Clin Endosc. 2026;59(2):182-193.   Published online September 29, 2025
      DOI: https://doi.org/10.5946/ce.2025.135
      Close
    • XML DownloadXML Download
    Related articles
    Recent technological advances in video capsule endoscopy: a comprehensive review
    Recent technological advances in video capsule endoscopy: a comprehensive review
    Feature PillCam SB3 MiroCam MC2000 CapsoCam Plus OMOM HD EndoCapsule
    Company Medtronic IntroMedic CapsoVision Jinshan Science and Technology Olympus
    Size (mm) 11.4×26.2 10.8×30 11×31 11×25 11×26
    Weight (g) 3 3.25–4.70 3.8 4.5 3.3
    No. of cameras 1 2 (dual-headed) 4 (side of capsule) 1 1
    Field of view (º) 156 170 (×2) 360 (panoramic imaging) 172 160
    Battery life (hr) 8–15 12 15 10 12
    Transmission type RF telemetry Human body communication N/A RF telemetry RF telemetry
    Real-time viewing Yes Yes No Yes Yes
    Storage type External receiver External receiver Onboard storage (capsule retrieval required) External receiver External receiver
    Frame velocity (frames/s) 2–6 3 fps per camera, total 6 fps 12–20 2–6 2
    Parameter OMOM HD Mirocam MC1600 p-value
    Complete examination (%) 97.2 98.1 0.762
    Capsule retention (%) 0.9 0.9 0.999
    Transit times (min, mean±SD)
     Gastric 53.0±43.0 56.0±47.0 0.867
     Small bowel 265.0±118.0 307.0±87.0 0.032
    Bowel preparation (mean±SD)a)
     Global small bowel 8.2±1.2 7.9±1.0 0.113
     First tertile 8.5±1.1 8.0±1.1 0.017
     Second tertile 8.2±1.3 7.5±1.1 0.004
     Third tertile 7.9±1.6 7.1±1.3 0.003
    Findings (%)
     Diagnostic yield 54.2 56.1 0.613
     Ulcers/erosions 30.8 32.7 0.561
     Angioectasia 20.5 15.9 0.103
     Subepithelial lesion 2.8 0.9 0.099
     Adenocarcinoma 0.9 0.0 0.394
     Extra-small bowel findings 6.1 5.6 0.112
    Feature PillCam SB3 OMOM Smart Capsule 2 p-value
    Complete small bowel visualization (%) 95.5 95.5 1
    Small bowel transit time (min) 310.5 295.5 0.653
    Download time (min) 132.0 33.0 <0.001
    Detection rate for all P2 lesions (%) 72.7 65.9 0.784
    Diagnostic yield 88.6 77.3 0.256
    Extraintestinal lesions 9.1 6.8 0.694
    Table 1. Comparison of capsule endoscopy

    RF, radio frequency; N/A, not available; fps, frames per second.

    Table 2. Comparison of OMOM HD and MiroCam MC1600

    SD, standard deviation.

    a)Classified according to the adapted Brotz scale.

    Table 3. Comparison of PillCam SB3 and OMOM Smart Capsule 2

    Table 1.

    Table 2.

    Table 3.

    Clin Endosc : Clinical Endoscopy Twitter Facebook
    © Korean Society of Gastrointestinal Endoscopy.
    Close layer
    TOP