Magnetically guided gastric capsule endoscopy: a review and new developments

Jean-Francois Rey

doi:10.5946/ce.2025.062

Articles

Page Path: HOME > Clin Endosc > Volume 58(6); 2025 > Article

Review Magnetically guided gastric capsule endoscopy: a review and new developments: Jean-Francois Rey; Clinical Endoscopy 2025;58(6):797-807.
DOI: https://doi.org/10.5946/ce.2025.062
Published online: July 29, 2025

Arnault Tzanck Institut, Saint Laurent du var, France

Correspondence: Jean-Francois Rey Arnault Tzanck Institut, 706 chemin du Suy Blanc 06610 Lagaude, Saint Laurent du var, France E-mail: jean-francois.rey@orange.fr

• Received: February 26, 2025 • Revised: April 11, 2025 • Accepted: April 13, 2025

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.

20,938 Views
326 Download
3 Web of Science
5 Crossref
4 Scopus

prev next

Full Article

Download PDF

Abstract

Since 2001, capsule endoscopy has been the primary test used to diagnose small-intestinal diseases. However, video capsule endoscopy of the stomach was considered impractical because visualizing the entire stomach was deemed impossible and would require a steerable capsule. Magnetically controlled gastric capsule endoscopy has been increasingly used for the diagnosis of gastric diseases, with significant developments in China. This noninvasive, hygienic, and comfortable method has gained popularity as an alternative to traditional electronic gastroscopy owing to its disposable nature and recent hardware upgrades (resolution, brightness, and field of view). Important steps forward with artificial intelligence and robots allow for the automated detection and characterization of gastric lesions. As it was restricted in China, questions have been raised about its cost-effectiveness worldwide, particularly in countries where early gastric cancer is not a priority. In this paper, I review the initial trials with this innovative capsule and the important technical updates in the last 5 years: robots for capsule guidance and artificial intelligence for the detection and characterization of gastric lesions.
Keywords: Artificial intelligence; Capsule endoscopy; Robot capsule; Stomach

INTRODUCTION

Since early 2001, capsule endoscopy (CE) has been the first-line test for diagnosing small-bowel diseases.¹ However, gastric examination using video CE was initially considered impractical because of the difficulty in visualizing the entire stomach, which requires a steerable capsule. Nevertheless, magnetically guided gastric capsules (MGCE) were successfully tested in models, animal trials,²^,³ and one human volunteer⁴ before our first human clinical trial in 2010.⁵ Over the past few years, magnetically controlled gastric CE has been increasingly used to diagnose gastric diseases.⁶^,⁷ This method is noninvasive and comfortable, making it a popular alternative to traditional electronic gastroscopy. It is primarily utilized for health examinations, gastric cancer screening, and for individuals who cannot tolerate gastroscopy. Studies have shown that the sensitivity and specificity of MGCE are comparable to those of electronic gastroscopy. The examination, during which the physician controls the device and records images, requires approximately 6,000 to 10,000 frames per patient, with an average reading time of 20 to 25 minutes for the specialist. This results in a significant workload. The application of artificial intelligence (AI)-powered computer-aided automatic analysis technology has significantly enhanced the detection of gastrointestinal (GI) abnormalities during CE. AI can assist in the diagnosis of CE, reduce the risk of human error, alleviate the burden on physicians, and improve diagnostic efficiency. It boasts high levels of sensitivity and specificity, typically reaching approximately 80% to 90% accuracy.⁸^,⁹ Finally, the ultimate approach includes direct gastric examinations conducted by a robot and readings using AI software.¹⁰ This opens up a new era¹¹ in which the same software allows for the simultaneous assessment of the stomach and small intestine.¹²

TECHNICAL GUIDANCE: FROM MAGNETIC RESONANCE WITH JOYSTICKS TO ROBOTS WITH AI

The initial trial developed with the Olympus-Siemens system was based on a specific low-level magnetic resonance (the magnetic field produced scores of times smaller than that used for magnetic resonance imaging). The magnet of the guidance system (Fig. 1) has a footprint of 1 m×2 m and generates dynamic magnetic fields and field gradients in a three-dimensional space over the entire stomach.⁵

The capsule is 31 mm long and 11 mm in diameter. It uses two image sensors with charge-coupled devices. The images are visualized and transmitted at 4 frames/sec (fps). The patient is equipped with multiple antennas to capture real-time gastric images. The dual-monitor panel simultaneously displays images from both sensors, with each sensor’s output displayed on a separate screen. The panel also indicates the possible maneuvers and settings for moving the capsule (forward, backward, diving, tilting, or jumping). The physician controls and directs the capsule, conducting its movement using two joysticks. Notably, due to commercial drawbacks, this system was abandoned.

Ankon, a Chinese company, created a more streamlined guidance device with a magnetic pole, similar to an X-ray table.⁶ They also developed a smaller and more office-friendly control system (Table 1). Multiple studies have been published since 2016, mainly by the Shanghai Physician Group led by Zhao-Shen Li.⁶

The Intromedic-Mirocam capsule from Korea resembles the original Paul Swain prototype, which is guided by a magnetic pad that captures and directs the capsule over the patient, albeit with less precise control.¹³

Jinshan designed a breakthrough robot guidance system with AI in 2019.¹⁰ The third-generation system, the fully automated magnetically controlled capsule endoscopy (FAMCE) system, can autonomously guide a magnetic capsule with an adaptive frame to perform gastroscopy without requiring a human operator to manually control the capsule. The robot, with the help of AI monitoring, can complete a comprehensive gastric examination in approximately 15 minutes. In addition to guidance, AI has been used to detect and characterize gastric lesions. Following completion of the gastric examination, the capsule is navigated to the duodenum, where a routine small bowel examination is conducted.

The small magnetic capsule endoscopy system, developed by JIFU Medical Technologies in China, consists of a capsule endoscope, a magnetic guidance robot, and an image-processing computer. The magnetic robot is a standalone system without arms and is equipped with wireless receivers.¹⁴

CAPSULE MANEUVERABILITY

Based on manual guidance (Olympus-Siemens, Ankon, and OMOM systems), the capsule can be moved with five independent mechanical degrees of freedom: two rotational and three translational (in three dimensions). It can be tilted (equivalent to large steering-wheel movements of the endoscope tip) or rotated (equivalent to small steering-wheel movements). The tilting command enabled the capsule to be oriented at a fixed point. The guided capsule can be navigated to the water surface in the stomach, or made to dive to the bottom of the stomach. When the capsule lies on the stomach wall, it can be directed to crawl, and if the capsule is blocked between gastric folds, it can be dislodged by being made to “jump.”⁵

Jinshan has made significant advancements in magnetic control technology. A robot can now perform a comprehensive gastric examination in just 15 minutes by leveraging AI software for evaluation and interpretation.¹⁰^,¹² These upgrades significantly reduce the reading time.¹⁵^,¹⁶ In addition, AI is utilized for guidance, precise identification, and accurate diagnosis. Although JIFU’s features are similar, its development and implementation are currently limited to China.¹⁴

IMAGE QUALITY

Compared to conventional endoscopy, video capsule images are closer to those provided by endoscopes. New sensors have enhanced the resolution, field of view, and depth of field, resulting in high-definition images. The brightness has improved, and close-up views provide remarkable details of the gastric mucosal pattern.¹⁰ In some examinations, the capsule view may have had a foggy appearance; however, this quickly resolved, revealing a clear view of the gastric cavity. In some cases, the oral mucus adheres to the front of the capsule sensor lens. It could be removed using a “capsule-shaking” command; however, this remains an important limitation.

GASTRIC VISUALIZATION

The initial hurdle was recognizing the anatomy of the stomach, as the capsule captures images from different positions, including those without inflation. The capsule’s movement is unique compared to that of a conventional endoscope because it can rotate in four directions, producing two simultaneous views from the sensors at each end of the capsule, owing to the Olympus-Siemens prototype. Traditional gastroscopy is simpler, with the examination performed in the forward direction in a distended stomach or in retroversion for observation of the cardia, fornix, and fundus. The gastric capsule provides an excellent new panoramic view of the lesser curve. This is one of the main advantages when the gastric capsule is directed to dive near the greater curve in front of the angulus, giving an overview of the lesser curve anatomy for diagnosis and orientation. The fundus and antrum are easy to assess in larger or close-up views. Similar to traditional endoscopy, this key landmark can be easily identified when the gastroscope passes through the cardiac orifice in an inverted position. Well-established differences in mucosal patterns are also major aids in navigating the capsule. Although the pylorus is easily identifiable, one must be aware that a cardia that is closed or only slightly open most of the time presents an unfamiliar aspect of capsule visualization (Fig. 2).

Gastric motility also presented challenges. The power of the capsule was inadequate to overcome gastric movements, particularly in the antrum, where powerful contractions prevented progression to the pylorus. However, in certain instances, the opposite was observed: a fast journey through the stomach into the duodenum.

The FAMCE robot overcomes most of these difficulties by utilizing AI to monitor and perform full gastric examinations, detect gastric abnormalities, and operate without human manipulations (Fig. 3).¹⁰

CLINICAL PROCEDURE

To conduct a thorough examination of the stomach using the Olympus-Siemens, Ankon, and Intromedic-Mirocam systems, it was essential to shift the patient from one position to another, allowing water to fill the various gastric regions and facilitating capsule movement. In the initial left lateral position, the cardia, fornix, fundus, and part of the antrum were typically visible; this part of the examination typically took approximately 10 minutes. Subsequently, the patient was moved to the supine position for a more complete examination of the fundus and antrum. The right lateral position was useful only for obtaining general and close-up appearance of the antrum and pylorus. However, in many cases, it was also possible to navigate the capsule back to the cardia and repeat the full procedure. The magnetic steering function does not depend on the patient’s position; however, changing of the position is needed to fill each part of the stomach with water. The patient while fasting, drank approximately 500 mL of clear water approximately 1 hour before the procedure, followed by 400 mL of water just before swallowing the capsule, to create an air/water interface in the stomach immediately before capsule ingestion. The patient was then asked to lie down on the table of the guidance equipment. The position of the table in relation to the magnetic pole was predetermined to allow for optimal gastric imaging and maximal magnetic force for capsule navigation. When it was difficult to move or navigate the capsule in a specific position, the patient was turned around to another position and sometimes made to lie in the prone position. If necessary, additional water was ingested to create optimal conditions for examination because the capsule requires some water volume for navigation.⁶^,¹⁶^,¹⁷

With the Jinshan robot (FAMCE), the patients are requested to lie on their backs. This position is more efficient for guidance by robots and AI. This is beneficial for the elderly and obese patients.¹⁰^,¹²

OPERATORS’ LEARNING CURVE

To begin, the operators honed their skills in manipulating the technology through hands-on practice using artificial “stomachs” with marked sections and real pig stomachs. This initial phase allowed them to grasp the fundamentals of the procedure, explore the potential of the video capsule with magnetic field guidance, and gain insights into the limitations of the technology and patient safety. This stage was completed using simulation software because the practitioners were unfamiliar with joystick handling. This stage was useful, but limited because only the basic functions were tested. Known lesions were identified by using the capsule. Initially, because of the unfamiliar appearance of the uninflated stomach, it was difficult to identify some structures, such as the closed cardia. The handling of the MGCE procedure could also be assumed by the nurse.¹⁸

The Jinshan robot equipment streamlined the procedure by eliminating the need for the patient to move beyond the initial position. The robot’s guidance was the only factor to consider, and joysticks were only used to adjust the guidance for clinical reasons.

EVALUATION OF STUDY FINDINGS

The primary outcome measures were the percentage of patients in whom the gastric surface was fully visualized in the antrum, body, and fundus and the identification of the cardia and pylorus. Secondary outcomes of the studies were examination time and the percentage of abnormal findings seen on gastroscopy that were reproducible by guided capsules and vice versa (Table 2).⁵^,⁶^,¹⁶ Most studies have identified favorable outcomes, including specific lesions, such as erosion, polyps, ulcers, and even submucosal tumors, as well as widespread conditions, such as atrophic gastritis. Normal gastric mucosa or mild inflammation was defined as a negative finding. The results of both examinations were considered consistent when the location, characteristics, and size of the lesions were consistent. If more than one lesion was detected in a patient, the most serious lesion, such as an ulcer, polyp, or submucosal tumor, was selected. The final assessment was based on these findings (Table 3). The capsule and gastroscopy results were blindly assessed by different operators.¹⁹ The final outcome was outstanding, with a remarkable success rate of 90% in visualizing the stomach (Fig. 2).⁵^,¹⁶

However, with the help of AI and improved technologies, Pan et al.⁹ overcame these challenges. They trained the SDSS-AI system using 34,062 MGCE images of 856 patients treated at Shanghai Hospital between January 2016 and October 2019. In addition, 50 patients referred to MGCE at Shanghai Hospital from December 2019 to January 2020 were enrolled to evaluate the diagnostic accuracy of SDSS-AI, using expert readings as the gold standard. The overall sensitivity of SDSS-AI for detecting gastric lesions was 98.9 % (95 % confidence interval [CI], 93.3 %–99.9 %), with sensitivities of 98.7 % (95 % CI, 91.9 %–99.9 %) and 100 % (95 % CI, 77.1 %–100 %) for detecting gastric erosion/bleeding/ulcer and polyp/submucosal tumor, respectively. The overall accuracy of SDSS-AI for identifying gastric anatomical landmarks was 94.2 % (95 % CI, 92.9 %–95.2 %), with accuracies of 97.8 % (95 % CI, 95.7 %–98.9 %), 96.5 % (95 % CI, 94.2 %–98.0 %), 73.8 % (95 % CI, 69.2 %–77.8 %), 96.0 % (95 % CI, 93.6 %–97.6 %), 98.0 % (95 % CI, 96.0 %–99.1 %), 96.0 % (95 % CI, 93.6 %–97.6 %), 96.8 % (95 % CI, 94.5 %–98.2 %), and 98.8 % (95 % CI, 97.0 %–99.6 %) for identifying the cardia, fundus, body, greater curvature, lesser curvature, angulus, antrum and pylorus, respectively.

A similar procedure was used for most studies that included various capsules.⁸^,⁹^,¹⁴

PATIENT ACCEPTABILITY

Regardless of its significance in the acceptance of GI examinations or its perceived medical importance, the high level of patient acceptance in studies is due to the absence of sedation and its potential drawbacks (such as fear of anesthesia or inability to drive or work on the day of the procedure).¹⁹^-²² Although gastroscopy examinations are now less uncomfortable because of various sedation methods or ultrathin gastroscopes for nasal insertion, we still need to improve patient acceptance and compliance in cases of repeated examinations.⁶^,⁷ In our initial study, all gastroscopies were performed with propofol sedation.⁵^,¹⁶^,¹⁹ Although the time required for capsule examination is currently longer than that for gastroscopy, the acceptance rate for guided capsule examination is very high. This noninvasive method of gastric examination is potentially important in countries where the use of a gastric screening tool is essential because of the high incidence of gastric cancer.

Similar to CE, capsule gastroscopy may pose a risk of impaction or retention. Reports in patients, mostly those with obscure bleeding, have described them as low risk.¹⁷^,²¹ Capsule impaction is rare in thousands of capsule procedures performed since 2001.²³^,²⁴

CLINICAL EVALUATION GUIDED CAPSULE VERSUS GASTROSCOPY

At present, high-definition endoscopy continues to be the reference method for gastric examinations as small bowel capsules fail to offer a comprehensive view. Only 58.3% of gastric lesions (63/108) were detected using both techniques.⁵^,¹⁶ Additionally, some of the findings from this initial trial were surprising. Because high-definition endoscopy serves as the gold standard for stomach examinations and the guided capsule was an early prototype, the primary focus of this study was the number of missed diagnoses associated with its use. This was observed in 14 patients (Table 2). However, even more striking was the discovery of 31 findings that were detected exclusively in the capsule. These findings do not have significant clinical importance in the detection of benign gastric polyps or angiomas. The superior ability of the capsule to identify tiny lesions, such as tiny hyperplastic polyps and angiomas, may be attributed to the longer examination time. This was because of our testing of the capsule’s guidance capabilities. The main discrepancy between these two methods is in the detection of inflammation or erosion. At this point, we can only offer two possible explanations: the capsule technology and length of the examinations, which were mentioned earlier, and the potential bias related to the order of the blinded and randomized examinations in the protocol. However, for the main clinical diagnostic outcomes in this trial, gastroscopy and guided capsule administration yielded similar results.⁸

Nearly 10 years later, trials with robots and AI with similar outcomes were described using FAMCE.¹⁰ The main distinctions were related to automated guidance, resulting in a significantly shorter examination time of only 15 minutes.

OVERALL CLINICAL STUDIES

In the first prospective, randomized, and strictly blinded study, MGCE was shown to be feasible in clinical practice and was clearly preferred by patients.¹⁷ A total of 189 symptomatic patients (105 male; mean age, 53 years) from two subsequently blinded groups underwent capsule and conventional gastroscopy by nine and six examiners, respectively. As a result, 23 major lesions were found in 21 patients. Capsule accuracy was 90.5% (95% CI, 85.4%–94.3%), with a specificity of 94.1% (95% CI, 89.3%–97.1%) and a sensitivity of 61.9% (95% CI, 38%–82%). The accuracy did not correlate with lesion location, gastric luminal visibility, examiner case volume, or examination time. Of the remaining 168 patients, 94% of patients had minor and mostly multiple lesions; the capsule made a correct diagnosis in 88.1% of patients (95% CI, 82.2%–92.6%), with gastric visibility and lesion location in the proximal stomach having significant influence.

Since then, multiple studies have been conducted using various types of equipment with similar clinical outcomes and improved technologies.⁶^,¹² Clinical benefits have been assessed in various conditions, such as first-line examination for upper GI bleeding,¹³^,²³^,²⁴ direct visualization of drug behaviors in the upper GI tract,²⁵ noncontact endoscopy for infection-free gastric examination during the coronavirus disease 2019 pandemic,²⁶ screening for gastric cancer,²⁷^,²⁸ recurrent and refractory iron deficiency,²⁹ gastric focal lesions,¹⁴ aspirin gastric injury in the asymptomatic patient,³⁰ investigating the digestion process,³¹ or global gastric transit time.³² The detection value of MGCE in patients with suspected small-intestinal disease was also assessed.³³ Few studies have been conducted in children without difficulty.³⁴^,³⁵ In high-risk patients, follow-up using traditional gastroscopy did not reveal any missing lesions 1 year after the initial gastric examination with a guided capsule.³⁶ Using a modified Pillcam prototype, Ohmiya et al.³⁷ performed esophageal, stomach, and colon screening for tumors.

NEW DEVELOPMENTS IN CE: AI AND ROBOT

During the last 5 years, important improvements for all capsule endoscopies have been carried out from technical points of view: high-resolution sensors, a wider angle of view, a deeper field of view, and long-lasting batteries.³⁸ The second-generation Ankon Navicam has a higher image resolution of 720×720 pixels and an improved adaptive frame rate of 8 fps. Ankon MGCE significantly improved image quality (p<0.001) and maneuverability (p<0.01). The total procedural time was also decreased from a mean of 7.78 mins to 5.27 mins (p<0.001).

However, the most important aspect was the introduction of AI in all capsule software for detection, and multiple studies have focused only on one or two features.³⁹ Ding’s study was a major step in the detection and characterization of small bowel abnormalities, reading numerous images with detailed diagnosed lesions, and faster reading time.⁴⁰

With FAMCE, Xiao et al.¹⁰ performed a prospective comparative study to investigate the safety and efficacy of OMOM robot capsules (Jinshan). A total of 114 patients with a mean age of 44.0 years were recruited for this study. The OMOM robot capsule and traditional gastroscopy had a high concordance rate of 99.61% for the detection of the five types of gastric lesions (gastritis, polyps, submucosal protuberances, mucosal erosion, and xanthoma). Five gastric lesions could not be identified using the capsule, whereas 16 were missed using traditional gastroscopy. For the detection of gastritis, the capsule identified 12 cases that were not detected by traditional gastroscopy, whereas traditional gastroscopy detected 4 cases that were missed by the capsule. Subsequently, Xie et al.¹² used AI for MGCE monitoring and reading software for both the stomach and small intestine. They achieved full magnetic steering using a simple device, which was described as an important improvement in gastric CE.¹⁰^,⁴¹ This full digestive tract examination is now a new possibility, especially for non-hematemesis acute GI bleeding²³ and obscure GI bleeding,²⁴ even though complete esophageal evaluation remains a challenge.

On the Navicam platform, Xia et al.⁸ created a cutting-edge, fully automatic system for detecting gastric lesions that can identify five common types of lesions (erosion, polyps, ulcers, submucosal tumors, and xanthogranuloma) with a sensitivity of 96.2% and specificity of 76.2%. In terms of the image processing speed, the processing time of the system per image was 44 ms per image, compared to 0.38±0.29 seconds per image for clinicians; thus, the system also greatly reduces the diagnostic time and improves diagnostic accuracy.

MAJOR DEVELOPMENTS

A comprehensive review revealed that MGCE technology has made significant progress; however, research on its detection and treatment remains in its early stages.⁴²^-⁴⁴

Over the past 5 years, multiple reviews have been published. They highlighted the improvement of MGCE with AI; however, they found it difficult to assess the global impact of this technology, as the equipment and software are similar but not identical.⁴⁵

Although many research groups have achieved remarkable results using AI in the field of CE, it has yet to be applied to real-world patient management beyond clinical studies.⁴⁶^-⁴⁹ Several obstacles must be overcome before the clinical implementation of AI. Most published studies to date have been performed retrospectively and have used data from a single center or a small number of centers, which leads to inherent selection and spectrum biases.⁵⁰ High-quality pictures and video data without debris bubbles or low-resolution recording are essential for model development. In addition, because the mechanism of the AI system is difficult to explain (black box, lack of interpretability), validation of the AI system is an important step in evaluating AI performance.⁵¹ For meticulous evaluation and verification of the clinical relevance of the CNN system, further multicenter prospective studies and external validation with irrelevant data for model development are required. We must also consider the technical variations among platforms that can make the software less reliable.⁵² There are several challenges in integrating AI solutions into clinical practice, such as ensuring that algorithms are transparent, understandable, and explainable. Developing appropriate governance and payment models for AI in CE is challenging.

As MGCE is a digestive endoscopy technique, the same positive and negative aspects of AI apply to CE⁵³: (1) Ensuring that AI algorithms are transparent, interpretable, and explainable is key to their clinical adoption; (2) Developing appropriate governance structures and procuring high-quality data are critical to the successful implementation of AI; (3) Establishing reliable cost-effectiveness studies and payment models to guide purchasing decisions and reimbursements for AI systems; (4) Relevant clinical outcomes and performance metrics for AI in gastroenterology are not well defined, necessitating steps to improve the research quality and interpretability; (5) Active collaboration among the medical technology industry, computer scientists, gastroenterologists, and researchers is essential for meaningful advancements; and (6) Do not forget that in clinical practice, the final decision and responsibility rely on the physician.

CHINA: AN INCREDIBLE DEVELOPMENT

Most MGCE techniques were developed in China with guidelines for clinical use.⁵⁴ Clinical trials in this area of research are limited in Western countries.⁵⁵^,⁵⁶ The clinical use of videogastroscopy versus MGCE presents an important challenge due to its high cost. It is usually fully covered by the healthcare system but varies depending on the country. In addition, the disparities in upper GI diseases must be considered. In Asia, early gastric cancer is a significant concern, whereas in the West, reflux disease and Barrett’s esophagus are the primary indications.⁵⁷

LIMITATIONS AND DRAWBACKS

The most significant difficulty is obtaining a clean stomach. During gastroscopy, gastric mucus, air bubbles, and other residues not only affect the imaging of the gastric mucosa, but also affect the accuracy of its diagnosis. Antifoams prevent the formation of air bubbles in the stomach by reducing the surface tension of the air bubbles and destroying them. For example, pronase can break down protein–peptide bonds with a strong proteolytic effect. Therefore, during routine MGCE, these substances can reduce the amount of mucus and air bubbles and improve the diagnosis of gastric lesions. Zhu et al.⁵⁸ proposed using Sprite Zero to clear the stomach, as we did in our initial trial with various liquids or 37°C water. However, none of these solutions led to a significant improvement in visibility.⁵

Strong gastric motility could also induce rapid gastric passage without a full detailed examination; unfortunately, MGCE cannot be moved from the duodenum to the stomach.

It is essential to conduct larger, international, multicenter studies to confirm diagnostic accuracy and provide a robust analysis of the cost-effectiveness.⁵⁷ Esophageal examination is also a major challenge in the West.⁵⁹ In terms of patient tolerance, we are not convinced of the Jiang et al., Rey and Monfort, and Marmo et al. string system⁵⁴^-⁵⁶^,⁶⁰ because we were unable to replicate the same level of acceptance as our patients.⁵⁵

FUTURE OUTCOMES

The emergence of capsule technology is particularly suitable for addressing the anticipated constraints on gastroenterological resources in many nations. Capturing and examining images using software that accelerates reading could allow nurses or assistants to conduct future capsule examinations with the practitioner only needing to review the recordings.⁵⁵ This is a crucial consideration, especially if this technology is to be integrated into gastric cancer screening programs in Asia.

Financial considerations

As this new technology becomes widely adopted, the cost of the capsule is expected to decrease, similar to the trend observed with digital cameras. However, an initial investment in the magnetic guidance system is required. Endoscopists also entail a new business model and a new way of organizing the endoscopy suite. In addition, it is important to highlight the benefits of CE for green endoscopy.⁶¹^-⁶³

The key points for magnetically guided capsule are the following (Table 3)¹⁶: (1) improvement in the detection and characterization of gastric lesions; (2) full monitoring of gastric examinations with AI and robots: guidance, detection, characterization, and initialized reporting; (3) gastric examination followed by traditional small bowel examination with the same capsule; (4) technical drawbacks: lack of complete esophageal examination and impaired gastric viewing due to mucus; (5) costs; and (6) collaboration among stakeholders: active collaboration among the medical industry, computer scientists, gastroenterologists, and scientific societies is mandatory for meaningful advancement.

CONCLUSIONS

We comprehensively examined the development of MGCE, starting with our initial clinical trials. Specifically, we extensively investigated its ability to detect GI lesions as well as the factors influencing image quality and the application of AI in GI endoscopy diagnosis through numerous clinical trials conducted in China. However, it is important to note that the results concerning the performance of the MGCE in detecting common GI lesions are based on limited evidence, and they are not widely accepted in the West. As a painless, noninvasive, safe, and hygienic examination method, it has gained widespread public acceptance, particularly in Asia, where it can be used as an effective tool for early gastric cancer screening. The cost and worldwide acceptance remain long journeys.

Conflicts of Interest

The author has no potential conflicts of interest.

Funding

None.

Fig. 1.

Olympus-Siemens prototype.

Fig. 2.

Visualisation of the stomach.

Fig. 3.

A fully automated magnetically controlled capsule endoscopy equipment in my unit.

Table 1.

Characteristics of various magnetically guided capsule endoscopy techniques

	Manufacturer
	Olympus-Siemens	Given Imaging	Ankon	Jinshan	Intromedic Mirocam	JIFU SMCE	Jinshan FAMCE	Ankon MGCE-2
Year	2010	2010	2012	2014	2015	2019	2019	2020
Contol method	MRI with joysticks	Handheld	Magnetic with joysticks	Magnetic with joysticks	Magnetic with pad	Standing-type guidance robot	Robot	Robot control
Magnetic field (mT)	100	272	200		341	200+500		200
Size (mm)	11×31	11×31	11.8×27	25.4×11	11×24	27×12		11.6×26.8
Weight (g)	4.2	7	5	6	4.2	2.7		5
Camera	2	1	1	1	1	1	1	1
Resolution	512×512	256×256	480×480	320×240	320×320	480×480	512×512	720×720
Frames (fps)	4	4	2	2	3	4	2–8 adaptative	0.5–6 up to 12
Field of view (°)	145	156	140	140	175	136	17	150
Depth of field (mm)	>20	0–30	0–30	0–35	0–30	0–50	0–50	0–30
Battery life	>60 mn	10 hr	>8 hr	8 hr	8 hr	30–40 mn	9 hr	>16

FAMCE, fully automated magnetically controlled capsule endoscopy; MGCE, magnetically guided gastric capsule; MRI, magnetic resonance imaging; fps, frames per second.

Table 2.

Comparative results between MGCE and conventional gastroscopy on 108 lesions diagnosed by MGCE and gastroscopy¹⁶

Finding	MGCE only	EGD only	Both EGD and MGCE
Polyps	11	2	8
Inflammation/erosion	10	1	44
Angioma	1	2	0
Ulceration	5	2	3
Atrophy	0	2	4
Important bile reflux	0	1	0
Hypertrophic folds	0	2	0
External compression	0	0	1
Fundic varices	0	0	1
Metaplasia	1	2	2
Bleeding	2	0	0
Hiatal hernia	1	0	0

MGCE, magnetically guided gastric capsule; EGD, esophagogastroduodenoscopy.

Table 3.

Key points for magnetically guided gastric capsule

No.	Key points
1	Improvement in the detection and characterization of gastric lesions
2	Full monitoring of gastric examination with artificial intelligence and robot: guidance, detection, characterization, and initialized report
3	Gastric examination then traditional small bowel examination with same capsule
4	Technical drawbacks: lack of esophageal complete examination, impaired gastric viewing by mucus
5	Costs
6	Collaborations among stakeholders: active collaboration between medical industry, computer scientist, gastroenterologists and scientific societies is mandatory for meaningful advancement

REFERENCES

1. Iddan G, Meron G, Glukhovsky A, et al. Wireless capsule endoscopy. Nature 2000;405:417.Article PDF
2. Ciuti G, Donlin R, Valdastri P, et al. Robotic versus manual control in magnetic steering of an endoscopic capsule. Endoscopy 2010;42:148–152.Article PubMed
3. Valdastri P, Quaglia C, Susilo E, et al. Wireless therapeutic endoscopic capsule: in vivo experiment. Endoscopy 2008;40:979–982.Article PubMed
4. Swain P, Toor A, Volke F, et al. Remote magnetic manipulation of a wireless capsule endoscope in the esophagus and stomach of humans (with videos). Gastrointest Endosc 2010;71:1290–1293.Article PubMed
5. Rey JF, Ogata H, Hosoe N, et al. Feasibility of stomach exploration with a guided capsule endoscope. Endoscopy 2010;42:541–545.Article PubMed
6. Liao Z, Hou X, Lin-Hu EQ, et al. Accuracy of magnetically controlled capsule endoscopy, compared with conventional gastroscopy, in detection of gastric diseases. Clin Gastroenterol Hepatol 2016;14:1266–1273.Article PubMed
7. Geropoulos G, Aquilina J, Kakos C, et al. Magnetically controlled capsule endoscopy versus conventional gastroscopy: a systematic review and meta-analysis. J Clin Gastroenterol 2021;55:577–585.Article PubMed
8. Xia J, Xia T, Pan J, et al. Use of artificial intelligence for detection of gastric lesions by magnetically controlled capsule endoscopy. Gastrointest Endosc 2021;93:133–139.Article PubMed
9. Pan J, Xia J, Jiang B, et al. Real-time identification of gastric lesions and anatomical landmarks by artificial intelligence during magnetically controlled capsule endoscopy. Endoscopy 2022;54:E622–E623.Article PubMed
10. Xiao YF, Wu ZX, He S, et al. Fully automated magnetically controlled capsule endoscopy for examination of the stomach and small bowel: a prospective, feasibility, two-centre study. Lancet Gastroenterol Hepatol 2021;6:914–921.Article PubMed
11. Rey JF. Gastric examination by guided capsule endoscopy: a new era. Lancet Gastroenterol Hepatol 2021;6:879–880.Article PubMed
12. Xie X, Xiao YF, Yang H, et al. A new artificial intelligence system for both stomach and small-bowel capsule endoscopy. Gastrointest Endosc 2024;100:878.Article
13. Ching HL, Hale MF, Sidhu R, et al. Magnetically assisted capsule endoscopy in suspected acute upper GI bleeding versus esophagogastroduodenoscopy in detecting focal lesions. Gastrointest Endosc 2019;90:430–439.Article PubMed
14. Lai HS, Wang XK, Cai JQ, et al. Standing-type magnetically guided capsule endoscopy versus gastroscopy for gastric examination: multicenter blinded comparative trial. Dig Endosc 2020;32:557–564.Article PubMed PDF
15. Beg S, Wronska E, Araujo I, et al. Use of rapid reading software to reduce capsule endoscopy reading times while maintaining accuracy. Gastrointest Endosc 2020;91:1322–1327.Article PubMed
16. Rey JF, Ogata H, Hosoe N, et al. Blinded nonrandomized comparative study of gastric examination with a magnetically guided capsule endoscope and standard videoendoscope. Gastrointest Endosc 2012;75:373–381.Article PubMed
17. Denzer UW, Rösch T, Hoytat B, et al. Magnetically guided capsule versus conventional gastroscopy for upper abdominal complaints: a prospective blinded study. J Clin Gastroenterol 2015;49:101–107.Article PubMed
18. Ching HL, Healy A, Thurston V, et al. Upper gastrointestinal tract capsule endoscopy using a nurse-led protocol: first reported experience. World J Gastroenterol 2018;24:2893–2901.Article PubMed PMC
19. Jiang X, Pan J, Li ZS, et al. Standardized examination procedure of magnetically controlled capsule endoscopy. VideoGIE 2019;4:239–243.Article PubMed PMC
20. Tai FW, Ching HL, Sloan M, et al. Comparison of patient tolerance and acceptability of magnet-controlled capsule endoscopy and flexible endoscopy in the investigation of dyspepsia. Endosc Int Open 2022;10:E735–E744.Article PubMed PMC
21. Meltzer AC, Kumar A, Kallus SJ, et al. Magnetically controlled capsule for assessment of the gastric mucosa in symptomatic patients: a prospective, single-arm, single-center, comparative study. iGIE 2023;2:139–147.Article
22. Nam SJ, Lee HS, Lim YJ. Evaluation of gastric disease with capsule endoscopy. Clin Endosc 2018;51:323–328.Article PubMed PMC PDF
23. Sung JJ, Tang RS, Ching JY, et al. Use of capsule endoscopy in the emergency department as a triage of patients with GI bleeding. Gastrointest Endosc 2016;84:907–913.Article PubMed
24. Yu Y, Liao Z, Jiang X, et al. The use of magnet-controlled capsule endoscopy as the initial diagnostic tool in patients with acute upper gastrointestinal bleeding. J Gastroenterol Hepatol 2023;38:2027–2034.Article PubMed
25. Wang YC, Pan J, Jiang B, et al. Direct visualization of drug behaviors in the upper GI tract via magnetically controlled capsule endoscopy. VideoGIE 2021;6:333–338.Article PubMed PMC
26. Pan J, Li Z, Liao Z. Noncontact endoscopy for infection-free gastric examination during the COVID-19 pandemic. VideoGIE 2020;5:402–403.Article PubMed PMC
27. Zhao AJ, Qian YY, Sun H, et al. Screening for gastric cancer with magnetically controlled capsule gastroscopy in asymptomatic individuals. Gastrointest Endosc 2018;88:466–474.Article PubMed
28. Qian YY, Zhu SG, Hou X, et al. Preliminary study of magnetically controlled capsule gastroscopy for diagnosing superficial gastric neoplasia. Dig Liver Dis 2018;50:1041–1046.Article PubMed
29. Ching HL, Hale MF, Kurien M, et al. Diagnostic yield of magnetically assisted capsule endoscopy versus gastroscopy in recurrent and refractory iron deficiency anemia. Endoscopy 2019;51:409–418.Article PubMed
30. Chen X, Gao F, Zhang J. Screening for gastric and small intestinal mucosal injury with magnetically controlled capsule endoscopy in asymptomatic patients taking enteric-coated aspirin. Gastroenterol Res Pract 2018;2018:2524698.Article PubMed PMC PDF
31. Fujimori S. Capsule endoscopy for investigating the digestion process. VideoGIE 2019;5:14–15.Article PubMed PMC
32. Jiang X, Qian YY, Liu X, et al. Impact of magnetic steering on gastric transit time of a capsule endoscopy (with video). Gastrointest Endosc 2018;88:746–754.Article PubMed
33. Chen XY, Da W, Liang R, et al. The detective value of magnetically controlled robotic capsule endoscopy in patients with suspected small intestinal disease. Front Med (Lausanne) 2021;8:610563.Article PubMed PMC
34. Gu Z, Wang Y, Lin K, et al. Magnetically controlled capsule endoscopy in children: a single-center, retrospective cohort study. J Pediatr Gastroenterol Nutr 2019;69:13–17.Article PubMed
35. Di Nardo G, Micheli F, Cozzi DA, et al. Magnetic-assisted capsule endoscopy in children with crohn disease: feasibility and impact on gastric transit time. J Pediatr Gastroenterol Nutr 2023;76:646–651.Article PubMed
36. Wu TT, Zhang MY, Tan ND, et al. Patients at risk for further examination with conventional gastroscopy after undergoing magnetically controlled capsule endoscopy. J Dig Dis 2023;24:522–529.Article PubMed
37. Ohmiya N, Omori T, Horiguchi N, et al. Diagnostic yields of magnetically-aided gastrointestinal capsule endoscopy for esophageal, gastric, and colorectal tumors: a randomized controlled trial. Gastrointest Endosc 2023;97(6 Supplement):AB1248–AB1249.Article
38. Jiang B, Qian YY, Pan J, et al. Second-generation magnetically controlled capsule gastroscopy with improved image resolution and frame rate: a randomized controlled clinical trial (with video). Gastrointest Endosc 2020;91:1379–1387.Article PubMed
39. 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.Article PubMed
40. 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.Article PubMed
41. Xu T, Li YY, Huang F, et al. A multi-task neural network for image recognition in magnetically controlled capsule endoscopy. Dig Dis Sci 2024;69:4231–4239.Article PubMed PDF
42. Yang YJ. The future of capsule endoscopy: the role of artificial intelligence and other technical advancements. Clin Endosc 2020;53:387–394.Article PubMed PMC PDF
43. Wang X, Hu X, Xu Y, et al. A systematic review on diagnosis and treatment of gastrointestinal diseases by magnetically controlled capsule endoscopy and artificial intelligence. Therap Adv Gastroenterol 2023;16:17562848231206991.Article PubMed PMC PDF
44. 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.Article PDF
45. O'Hara FJ, Mc Namara D. Capsule endoscopy with artificial intelligence-assisted technology: real-world usage of a validated AI model for capsule image review. Endosc Int Open 2023;11:E970–E975.Article PubMed PMC
46. 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.Article PubMed PMC
47. Cao Q, Deng R, Pan Y, et al. Robotic wireless capsule endoscopy: recent advances and upcoming technologies. Nat Commun 2024;15:4597.Article PubMed PMC PDF
48. Mota J, Almeida MJ, Mendes F, et al. From data to insights: how is AI revolutionizing small-bowel endoscopy? Diagnostics (Basel) 2024;14:291.Article PubMed PMC
49. Mascarenhas M, Martins M, Afonso J, et al. Deep learning and capsule endoscopy: Automatic multi-brand and multi-device panendoscopic detection of vascular lesions. Endosc Int Open 2024;12:E570–E578.Article PubMed PMC
50. Li Q, Xie W, Wang Y, et al. A deep learning application of capsule endoscopic gastric structure recognition based on a transformer model. J Clin Gastroenterol 2024;58:937–943.Article PubMed
51. Mascarenhas M, Mendes F, Ribeiro T, et al. Deep learning and minimally invasive endoscopy: automatic classification of pleomorphic gastric lesions in capsule endoscopy. Clin Transl Gastroenterol 2023;14:e00609.Article PubMed PMC
52. Byrne MF, Donnellan F. Artificial intelligence and capsule endoscopy: is the truly “smart” capsule nearly here? Gastrointest Endosc 2019;89:195–197.Article PubMed
53. ASGE AI Task Force, Parasa S, Berzin T, et al. Consensus statements on the current landscape of artificial intelligence applications in endoscopy, addressing roadblocks, and advancing artificial intelligence in gastroenterology. Gastrointest Endosc 2025;101:2–9.Article PubMed
54. Jiang B, Pan J, Qian YY, et al. Clinical guideline on magnetically controlled capsule gastroscopy (2021 edition). J Dig Dis 2023;24:70–84.Article PubMed
55. Rey JF, Monfort C. Magnetically guided capsules with robot and artificial intelligence for non-hematemesis gastrointestinal bleeding. Endoscopy 2023;55(S02):S339.Article
56. Marmo C, Riccioni ME, Pecere S, et al. Mini-invasive upper gastrointestinal examination with the new magnetically assisted capsule endoscopy: first experience in a tertiary care center. Endoscopy 2024;56(S02):S424.Article
57. Keuchel M. Will the steerable capsule replace upper endoscopy in the future? Endoscopy 2019;51:401–402.Article PubMed
58. Zhu JH, Qian YY, Liu X, et al. Sprite Zero® for better gastric preparation for capsule endoscopy examination: a randomized controlled pilot study. Gastrointest Endosc 2021;93(6 Supplement):AB37.Article
59. Beg S, Card T, Warburton S, et al. Diagnosis of Barrett’s esophagus and esophageal varices using a magnetically assisted capsule endoscopy system. Gastrointest Endosc 2020;91:773–781.Article PubMed
60. Liao Z, Gao R, Xu C, et al. Sleeve string capsule endoscopy for real-time viewing of the esophagus: a pilot study (with video). Gastrointest Endosc 2009;70:201–209.Article PubMed
61. Chen YZ, Pan J, Luo YY, et al. Detachable string magnetically controlled capsule endoscopy for complete viewing of the esophagus and stomach. Endoscopy 2019;51:360–364.Article PubMed
62. Mascarenhas M, Ribeiro T, Afonso J, et al. Smart endoscopy is greener endoscopy: leveraging artificial intelligence and blockchain technologies to drive sustainability in digestive health care. Diagnostics (Basel) 2023;13:3625.Article PubMed PMC
63. Pioche M, Cunha Neves JA, Pohl H, et al. The environmental impact of small-bowel capsule endoscopy. Endoscopy 2024;56:737–746.Article PubMed

Figure & Data

REFERENCES

Citations

Citations to this article as recorded by

Robotic Capsule Endoscopy: Simultaneous Gastric and Enteric Evaluation in Real-World Practice
Hélder Cardoso, Miguel Mascarenhas, Joana Mota, Miguel Martins, Maria João Almeida, Joana Frias, Catarina Cardoso Araújo, Francisco Mendes, Margarida Marques, Patrícia Andrade, Guilherme Macedo
Diagnostics.2026; 16(2): 334. CrossRef
Multimodal artificial intelligence in capsule endoscopy: Integrating video and sensor data for advanced gastrointestinal diagnostics
Rishi Chowdhary, Param Darpan Sheth, Insiya Mohammed Rampurawala, Chitresh Kapadia, Chirag Vohra, Rahul Chowdhary, Kirti Arora, Varna Taranikanti, Ashita Rukmini Vuthaluru, Omesh Goyal, Manjeet Kumar Goyal
Artificial Intelligence in Gastrointestinal Endoscopy.2026;[Epub] CrossRef
Selective Magnetic Field Generation Method for Effective Manipulation of Two-Dimensional Magnetic Microrobots Using a Triad of Electromagnetic Coils
Dongjun Lee, Yonghun Lee, Seungmun Jeon
Micromachines.2026; 17(3): 337. CrossRef
Magnetically Controlled Capsule Endoscopy: Innovations, Standardized Practice, and Emerging Clinical Roles
Xi Jiang, Chen He, Xiaoou Qiu, Yizhi Chen, Wei Zhou, Zhaoshen Li, Zhuan Liao
Clinical and Translational Gastroenterology.2026;[Epub] CrossRef
Artificial intelligence technologies in digestive system endoscopy: the state of the problem and prospects (literature review)
E. V. Shlyakhto, E. G. Solonitsyn, D. G. Baranov, B. V. Sigua, I. N. Danilov
Russian surgical journal.2025; 1(2): 8. CrossRef

Cite

CITE: export Copy Download Format; 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
Citation and abstract for the content below

Magnetically guided gastric capsule endoscopy: a review and new developments

Clin Endosc. 2025;58(6):797-807. Published online July 30, 2025

DOI: https://doi.org/10.5946/ce.2025.062

XML Download

Figure

Magnetically guided gastric capsule endoscopy: a review and new developments

Fig. 1. Olympus-Siemens prototype.

Fig. 2. Visualisation of the stomach.

Fig. 3. A fully automated magnetically controlled capsule endoscopy equipment in my unit.

Fig. 1.

Fig. 2.

Fig. 3.

Magnetically guided gastric capsule endoscopy: a review and new developments

	Manufacturer
	Olympus-Siemens	Given Imaging	Ankon	Jinshan	Intromedic Mirocam	JIFU SMCE	Jinshan FAMCE	Ankon MGCE-2
Year	2010	2010	2012	2014	2015	2019	2019	2020
Contol method	MRI with joysticks	Handheld	Magnetic with joysticks	Magnetic with joysticks	Magnetic with pad	Standing-type guidance robot	Robot	Robot control
Magnetic field (mT)	100	272	200		341	200+500		200
Size (mm)	11×31	11×31	11.8×27	25.4×11	11×24	27×12		11.6×26.8
Weight (g)	4.2	7	5	6	4.2	2.7		5
Camera	2	1	1	1	1	1	1	1
Resolution	512×512	256×256	480×480	320×240	320×320	480×480	512×512	720×720
Frames (fps)	4	4	2	2	3	4	2–8 adaptative	0.5–6 up to 12
Field of view (°)	145	156	140	140	175	136	17	150
Depth of field (mm)	>20	0–30	0–30	0–35	0–30	0–50	0–50	0–30
Battery life	>60 mn	10 hr	>8 hr	8 hr	8 hr	30–40 mn	9 hr	>16

Finding	MGCE only	EGD only	Both EGD and MGCE
Polyps	11	2	8
Inflammation/erosion	10	1	44
Angioma	1	2	0
Ulceration	5	2	3
Atrophy	0	2	4
Important bile reflux	0	1	0
Hypertrophic folds	0	2	0
External compression	0	0	1
Fundic varices	0	0	1
Metaplasia	1	2	2
Bleeding	2	0	0
Hiatal hernia	1	0	0

No.	Key points
1	Improvement in the detection and characterization of gastric lesions
2	Full monitoring of gastric examination with artificial intelligence and robot: guidance, detection, characterization, and initialized report
3	Gastric examination then traditional small bowel examination with same capsule
4	Technical drawbacks: lack of esophageal complete examination, impaired gastric viewing by mucus
5	Costs
6	Collaborations among stakeholders: active collaboration between medical industry, computer scientist, gastroenterologists and scientific societies is mandatory for meaningful advancement

Table 1. Characteristics of various magnetically guided capsule endoscopy techniques

FAMCE, fully automated magnetically controlled capsule endoscopy; MGCE, magnetically guided gastric capsule; MRI, magnetic resonance imaging; fps, frames per second.

Table 2. Comparative results between MGCE and conventional gastroscopy on 108 lesions diagnosed by MGCE and gastroscopy16