Authors: Isabel Rojas, Bradley A. Barth, Jeremy W. Stewart
Categories: Mini Review, artificial intelligence, magnetically controlled capsule, patency capsule, pediatric gastroenterology, video capsule endoscopy, wireless motility capsule
Source: Frontiers in Pediatrics
Authors: Isabel Rojas, Bradley A. Barth, Jeremy W. Stewart
Video capsule endoscopy (VCE) has revolutionized the evaluation of small bowel pathology, offering a safe, non-invasive, radiation-free diagnostic modality with broad clinical utility. Patency capsule use has further improved safety by minimizing the risk of retention in patients with suspected strictures. Since its introduction, its applications have expanded from obscure gastrointestinal bleeding and Crohn's disease to celiac disease, polyposis syndromes, and small bowel tumors among other indications. Emerging artificial intelligence (AI) integration promises to enhance diagnostic accuracy, streamline image analysis, and reduce interobserver variability. Furthermore, advancements in capsule design, including magnetic-assisted navigation and extended battery life, enable precise control and complete small bowel evaluation, even in cases of delayed gastrointestinal motility. High-definition imaging further allows for the identification of subtle mucosal abnormalities, such as vascular lesions, inflammation, and erosions, that might otherwise go undetected. Beyond diagnosis, novel applications, such as motility capsule studies and wireless capsule drug delivery systems, are unlocking new possibilities for functional and therapeutic interventions. Future innovations combining diagnostic and interventional capabilities promise to reduce the need for invasive procedures, optimize outcomes, and significantly enhance the quality of life for pediatric patients.
Video capsule endoscopy (VCE) has transformed the diagnostic landscape of small bowel evaluation since its initial introduction in 2000. Its non-invasive nature, lack of ionizing radiation, and ability to directly visualize the entire small intestine make it particularly well suited for use in pediatric populations, where minimizing procedural risk and patient discomfort is paramount (1, 2). Initially developed to investigate obscure gastrointestinal bleeding in adults, VCE has since found widespread applications in children, including the evaluation of suspected Crohn's disease, celiac disease, polyposis syndromes, and small bowel tumors (3).
Pediatric gastrointestinal disorders often present with nonspecific symptoms and require comprehensive evaluation of the small bowel; an area historically considered a diagnostic blind spot. Traditional modalities such as radiographic imaging, enteroscopy, and surgery carry limitations in sensitivity, invasiveness, and feasibility in children. VCE bridges this gap by enabling detailed mucosal assessment with high diagnostic yield, particularly in cases where other tests have been inconclusive (1, 4, 5).
In recent years, technological advances, including high-definition imaging, extended battery life, and novel delivery systems, have further expanded the scope and effectiveness of VCE in pediatric care (6, 7). The integration of artificial intelligence (AI) into image analysis platforms is poised to enhance lesion detection, reduce reading times, and minimize interobserver variability (8, 9). Simultaneously, safety innovations such as the use of patency capsules help mitigate the risk of capsule retention in children with suspected strictures, broadening the patient populations for whom VCE can be safely performed (10, 11).
As VCE continues to evolve, its role is shifting from purely diagnostic to potential therapeutic and functional applications. This review provides an update on the current clinical use of VCE in children, highlights key technological and safety advancements, and explores future directions that may redefine its place in pediatric gastroenterology.
Video capsule endoscopy (VCE) has become an indispensable diagnostic tool for evaluating small bowel pathology in children (Figure 1). Its ability to provide high-resolution mucosal visualization without the need for sedation or ionizing radiation makes it especially valuable in children, where conventional modalities are limited. Although pediatric data remains more limited than adult studies, evidence consistently supports its safety, feasibility, and diagnostic yield across a range of clinical indications (1, 12–14).

Obscure gastrointestinal bleeding remains the most frequent indication for VCE in children, particularly those under eight years of age. This modality is highly effective in identifying sources of occult or overt bleeding that go undetected following negative upper and lower endoscopies. Reported diagnostic yields vary widely from approximately 19%–95% in small series, with a multicenter pediatric study demonstrating a yield of around 53% (1, 2). Common findings include vascular malformations, ulcers, erosions, and occasionally Meckel's diverticulum (12, 13). In many cases, VCE findings directly inform subsequent management, including therapeutic intervention via device-assisted enteroscopy. Although adult data suggest that earlier timing of capsule administration, within 24–72 h of a bleeding episode, increased diagnostic yield, pediatric evidence supporting this interval is limited; thus, its application should be considered extrapolated from adult studies or center-specific practice (1).
Video capsule endoscopy (VCE) plays a pivotal role in the evaluation and management of suspected or established Crohn's disease. It enables direct visualization of early or isolated small bowel lesions that may be missed by conventional endoscopy or imaging. The North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) recognizes VCE as a key modality for identifying small bowel involvement, assessing disease extent and recurrence, and guiding treatment decisions. Capsule endoscopy demonstrates good sensitivity compared with magnetic resonance enterography (MRE) and computed tomography enterography (CTE), with variable specificity (1). Prospective data support complementary roles for VCE and MRE; VCE offers higher specificity for mucosal disease and can influence management decisions, including monitoring for mucosal healing within treat-to-target paradigms (15, 16). Pediatric cohorts further document that VCE findings lead to therapeutic modifications in a substantial proportion of inflammatory bowel disease cases (14).
While histology remains the diagnostic standard, VCE can identify characteristic features such as villous atrophy, scalloping, and mosaic pattern, and is particularly useful when esophagogastroduodenoscopy (EGD) is not feasible or when assessing disease extent or complications. Clinical studies report high performance for detecting villous atrophy (17). A contemporary meta-analysis further supports substantial detection of these features, with higher diagnostic yield in refractory cases, reinforcing VCE's role as an adjunctive tool (18).
Capsule endoscopy is employed for small bowel surveillance in pediatric hamartomatous polyposis syndromes, particularly Peutz-Jegher's syndrome, allowing assessment of small-bowel polyp burden and guiding the timing of device-assisted enteroscopy. The European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) recommends small bowel surveillance using VCE and/or MRE beginning no later than 8 years, with VCE demonstrating greater sensitivity than imaging modalities (19).
Direct visualization of mucosal abnormalities, including masses, ulcerations, and polypoid lesions, is achievable with VCE, making it particularly useful for detecting lesions missed by radiography or cross-sectional imgaing and facilitating early diagnosis as well as guiding subsequent management, such as device-assisted enteroscopy or surgery (20). In pediatric cohorts, VCE has demonstrated high completion rates and diagnostic yield across a range of small bowel pathologies, including rare tumors, with abnormal findings reported in up to 59% of cases (12–14). NASPGHAN recommends VCE as a complementary tool when other modalities fail to identify a lesion (1).
Additional indications for VCE in children include protein-losing enteropathy, graft-vs.-host disease, and surveillance in post-operative or radiation-induced enteropathy (1, 21). With ongoing innovation and increasing familiarity, the range of indications is expected to expand further.
Capsule retention in pediatric VCE occurs in approximately 1%–2% of procedures overall, with higher rates reported among children with known or suspected Crohn's disease, small bowel strictures, or post-surgical anatomy. Retention is most often asymptomatic but may present with abdominal pain, nausea, or signs of bowel obstruction. The use of a patency capsule offers a safe and effective method to assess small bowel patency in at-risk pediatric populations, such as those with Crohn's disease, previous intestinal surgery, or suspected stricturing disease. Traditional screening approaches, based on clinical history or cross-sectional imaging, demonstrate limited sensitivity, whereas the patency capsule provides superior specificity and a lower false-negative rate.
Large pediatric series and meta-analyses have reported overall capsule retention rates between 1.4% and 2.3%, with higher frequencies observed in Crohn's disease (up to 2.2%–2.6%) and in patients with strictures. Retention most commonly results from inflammatory or post-surgical narrowing, and rates are determined primarily by indication rather than age (1, 4, 11, 22). Although often clinically silent, capsule retention can cause obstruction or, rarely, perforation; symptoms may include abdominal pain, nausea, or vomiting, and surgical intervention is occasionally required (1).
Conventional risk stratification methods, including review of clinical history, MR or CT enterography, and fluoroscopic studies, can help identify children at risk for retention, but these techniques have limited sensitivity for detecting functional patency. Cross-sectional imaging has a pooled sensitivity of 54% and a specificity of 88%, for predicting retention, while the patency capsule demonstrates higher specificity (94%) and lower false-negative rates of 2.7% (23).
The patency capsule (e.g., Agile Patency Capsule) is a dissolvable, radio-opaque capsule that mimics the size and shape of a VCE device. It contains an internal time plug that causes the capsule to disintegrate after a predetermined time, typically 30–40 h, breaking down into small fragments, allowing it to pass safely through the intestine, thereby minimizing the risk of obstruction. It contains a radiofrequency identification (RFID) tag or similar marker that permits noninvasive detection of the capsule's location using an external scanner or imaging (24).
It is recommended for children at increased risk of small bowel strictures, including those with Crohn's disease, prior bowel surgery, or post-inflammatory or radiation-induced narrowing. It is safe and feasible in older children and adolescents, with multicenter studies demonstrating high successful passage rates and minimal adverse events (4, 24–26). Overall, the patency capsule provides a more reliable assessment of small bowel patency compared with imaging alone and effectively predicts safe passage of the diagnostic capsule (23, 26).
Third-generation capsules, such as PillCam SB3 (Figure 2), incorporate high-resolution sensors and adaptive frame rate technology, substantially improving the visualization of subtle mucosal abnormalities, including aphthous ulcers, erosions, or vascular lesions. Enhanced image clarity facilitates earlier and more accurate detection of small bowel disease and increases interobserver agreement in pediatric interpretation. Integration of advanced optics and software-based image enhancement has further optimized contrast and mucosal detail, improving diagnostic confidence in children with suspected inflammatory or vascular pathology (1, 6, 7) (Table 1).

Extended battery life in newer capsule models (Pillcam SB2-ex, SB3) is associated with higher rates of complete small bowel transit and fewer incomplete studies, which is especially important in children with slower transit times or anatomical variations (7). Large pediatric cohorts report completion rates exceeding 95%, supporting the clinical impact of longer battery duration (12–14).
Magnetically controlled capsule endoscopy (MCE) represents a significant innovation, enabling active navigation and targeted visualization of the upper gastrointestinal tract, overcoming the limitations of passive capsule transit, and without the need for sedation. While its application in pediatrics remains limited, feasibility studies demonstrate safe use and excellent visualization, facilitating gastric emptying and transpyloric passage, including children as young as 6 years (27, 28). The ability to retrieve or reposition capsules in real-time enhances procedural safety, a crucial advantage in pediatrics (Figure 2).
Real-time viewing allows immediate assessment of capsule location and mucosal findings, improving procedural efficiency, and enabling prompt intervention if needed. This feature is highlighted in the NASPGHAN report and recent pediatric studies as beneficial for workflow and safety (1).
The development of smaller capsule sizes and improved delivery accessories has expanded the use of VCE in children as young as 8 months of age and as light as 7.9 kg (1, 29, 30). Dedicated pediatric delivery devices, such as endoscopic capsule deployment systems (e.g., AdvanCE), enable safe administration in patients unable to swallow the capsule, including those under six years of age (2, 30–32). These tailored designs have increased procedural success and safety, broadening the applicability of capsule accuracy across the pediatric age spectrum.
Artificial intelligence (AI) has rapidly emerged as one of the most transformative developments in VCE. With each study producing tens of thousands of images, manual review is time-sensitive and subject to human fatigue and variability. AI-driven algorithms, particularly those based on convolutional neural networks (CNNs), are being developed to assist in automatic detection, classification, and localization of small bowel lesions. These systems are designed to augment, rather than replace, physician interpretation, thereby improving efficiency, diagnostic consistency, and consistency across readers. Multiple multicenter studies and systematic reviews have confirmed that deep learning models, especially CNN-based architectures, can automatically detect and classify small bowel lesions and localize findings within the gastrointestinal tract, providing strong evidence for their use as reliable physician-assistive tools (8, 33).
AI integration in capsule endoscopy primarily focuses on automatic lesion recognition, including ulcers, erosions, angioectasias, and polyps. Algorithms trained on large annotated datasets can identify pathological features with sensitivities and specificities comparable to expert reviewers. Deep-learning systems have demonstrated excellent accuracy in detecting small bowel bleeding and vascular lesions, automatically flagging relevant frames for clinician review and prioritizing high-yield segments of the video. Some systems also provide lesion localization and quantification, facilitating longitudinal disease monitoring in conditions such as Crohn's disease. Meta-analyses and large validation studies consistently report that CNN-based algorithms achieve sensitivities and specificities exceeding 95% for the detection of ulcers, bleeding, and polyps, often matching or surpassing expert human performance. These AI models reliably identify a wide range of lesions relevant to both pediatric and adult populations, reinforcing their role as powerful assistive tools in capsule interpretation (8, 34).
AI-assisted reading has demonstrated substantial improvements in efficiency and diagnostic consistency. Multiple studies report dramatic reductions in interpretation time, from 60 to 96 min to as little as 5–15 min, without compromising, and in some cases improving, diagnostic accuracy (8, 35, 36). By automatically highlighting clinically relevant frames, AI minimizes the risk of missed lesions and enhances detection consistency across readers with varying experience levels. This standardization of review reduces interobserver variability and increases reproducibility, benefits that are particularly valuable in pediatric centers where capsule interpretation may be performed by clinicians with differing levels of expertise (9).
Commercial AI-enabled capsule platforms, including Omni Mode (Medtronic) and SmartScan (IntroMedic), are now integrated into clinical software, offering automated bleeding detection, lesion identification, and image triage capabilities (Figure 3) (36, 37). Most validation studies have been conducted in adult cohorts, where these systems have shown strong diagnostic performance and substantial reductions in review time. Although pediatric-specific data remain limited, extrapolation of results to children is reasonable given the similarity in lesion morphology and imaging characteristics. Preliminary pediatric experience suggests that AI-assisted analysis can enhance workflow efficiency and diagnostic confidence, but larger, dedicated pediatric studies are needed to validate performance across younger age groups and rare disease contexts (37).
![Figure 3: Findings detected by SmartScan-assisted Reading [Ref. (36), CC-BY license]. (A) Protruding lesion (tumor). (B) Flat lesion (angioectasia). (C) Mucosa (erythematous). (D) Excavated lesion (aphtha). (E) Content (parasite). (F) Content (blood).](fped-13-1738998-g003.jpg)
Wireless motility capsules (Figure 2), such as the SmartPill, measure intraluminal pH, pressure, and temperature to evaluate regional and whole gut transit times, offering valuable data for the assessment of gastrointestinal motility disorders without radiation exposure. Their use in pediatrics remains limited due to capsule size, regulatory restrictions, and the absence of validated pediatric normative data. Studies in adolescents and older children (ages 8–17) have demonstrated feasibility and safety, but practical limitations, including difficulty swallowing the capsule and variable gastric emptying, restrict use in younger patients. Ongoing efforts to miniaturize capsule devices and establish age-appropriate reference ages are underway, with the potential to integrate motility assessment into routine capsule-based diagnostics in the near future (38–41) (Table 2).
Beyond diagnostics, capsule technology is advancing toward targeted therapy. Wireless drug delivery capsules are being developed to release medication at specific gastrointestinal sites using pH, enzymatic, or magnetic triggers. Experimental models have demonstrated the feasibility of site-specific delivery for anti-inflammatory agents, antibiotics, and biologics, offering distinct advantages for diseases with localized pathology such as Crohn's disease. For pediatric patients, these systems could minimize systemic exposure and improve adherence by combining diagnosis and treatment within a single, noninvasive platform. Although still in preclinical or early transitional stages, continued progress in miniaturization and biocompatible materials to facilitate future pediatric applications (42–44).
Next-generation “active” capsule systems aim to integrate diagnostic imaging with interventional functions such as biopsy, electrocautery, and hemostatic delivery (Figure 2). Prototype devices have demonstrated proof of concept for tissue sampling, bleeding control, and localized therapy without the need for conventional endoscopy. These multifunctional capsules represent a potential paradigm shift in minimally invasive gastroenterology, particularly valuable in pediatrics, where reducing anesthesia exposure and procedural invasiveness is a major priority. While translational to clinical use remains forthcoming, rapid progress in micro-robotics, wireless actuation, and energy control continues to expand the feasibility of these platforms (42–44).
Tethered capsule endoscopes are under active investigation, offering real-time and controlled movement through the gastrointestinal tract via external magnetic or manual guidance (Figure 2). These devices typically include a thin tether for retrieval or manipulation and can be combined with magnetic steering for enhanced precision. Cable-transmission magnetically controlled capsule endoscopy (CT-MCCR) systems have shown high diagnostic yield and safety for upper GI evaluation, with performance comparable to conventional gastroscopy and improved patient tolerance (28, 45, 46).
Video capsule endoscopy has become an essential, minimally invasive tool for evaluating small bowel pathology in children, particularly for obscure gastrointestinal bleeding, inflammatory bowel disease, celiac disease, and hereditary polyposis syndromes. Advances in capsule design, including smaller size, extended battery life, and patency testing, have improved safety and broadened applicability. However, most technical validation originates from adult populations, and pediatric-specific data remain limited. Differences in anatomy, motility, and disease presentation highlight the need for dedicated pediatric studies to refine protocols and establish normative standards.
Artificial intelligence (AI)-assisted reading represents one of the most promising developments in capsule endoscopy. Early data show that AI can reduce reading time, enhance diagnostic accuracy, and minimize interobserver variability. Commercial systems with integrated AI triage and lesion detection are now available, but pediatric validation is scarce. The creation of large, annotated pediatric image databases through multicenter collaboration will be essential to ensure algorithm reliability across different ages and disease phenotypes.
Emerging technologies are transforming VCE from a passive diagnostic tool into an active, multifunctional platform. Motility capsules enable radiation-free assessment of gastrointestinal transit, while wireless drug delivery and hybrid diagnostic-interventional capsules may one day permit targeted therapy and tissue sampling without conventional endoscopy. Magnetically controlled and tethered capsule systems, especially when coupled with AI-based navigation, allow real-time steering, retrieval, and optimized visualization, reducing the need for sedation and improving procedural safety.
In summary, capsule endoscopy continues to evolve rapidly within pediatric gastroenterology. The convergence of miniaturization, smart navigation, and AI-driven interpretation promises to expand diagnostic reach and therapeutic potential while maintaining a noninvasive approach. Ongoing efforts to validate these technologies in children and ensure equitable access will be crucial to realizing the full clinical impact of next-generation capsule endoscopy.