Use of magnetically controlled capsule endoscopy for the diagnosis of gastric diseases in adults: a systematic review and meta-analysis

Introduction

Capsule endoscopy has become a very important tool in diagnosing small bowel diseases since it was firstly introduced in 2000 (1,2). Since then many attempts have been made to spread the role of capsule endoscopy to the upper gastrointestinal tract (3,4). To observe the gastric mucosal surface clearly and accurately, capsule endoscopy must overcome several difficulties such as the capacious and complicated anatomy of the stomach, peristaltic waves and a folded stomach, which are different from conventional gastroscopy (5,6). In order to solve these problems, magnetic manipulation of capsule endoscope was firstly manufactured in 2006 (7). With the development of technology, the magnetically controlled capsule endoscopy (MCCE) now becomes more feasible and safe (5,8). What is more, MCCE was greatly welcomed by patients for it can represent a more patient-friendly substitute, and can be utilized without notable discomfort (9). To date, several researches have shown promising information to support the potentiality of MCCE (10). Swain et al. (11) suggested that it was easy to control the capsule back from the pylorus to the cardia esophageal junction and hold the capsule at any position in the stomach and did not cause discomfort. Rey et al. (12) also believed that magnetically navigated video capsule endoscopy was feasible and sufficiently accurate for gastric examination. However, Keller et al. (13), on the other hand, showed that MCCE cannot hold the capsule endoscopy against peristalsis when it approached the gastroesophageal junction because of its weak magnetic forces. Although there are several studies that have directly compared MCCE with CG, most of the available researches are limited by their small sample sizes, and no meta-analysis has been established about it. And we still don’t know whether MCCE will exceed conventional gastroduodenoscopy (CG), which remains the gold standard diagnostic method in the upper gastrointestinal tract. Therefore, we here formulated the first systematic review and meta-analysis to clarify the sensitivity and specificity of MCCE and CG. We present the following article in accordance with the PRISMA reporting checklist (available at http://dx.doi.org/10.21037/dmr-20-141).

Methods

Search and strategy

PubMed, Embase, Cochrane Library and Web of Science were systematic searched independently by two authors (Jianlong Jiang and Hao Zhang) from established to December 2019. Language was not limited during the search. The terms “magnetic-controlled capsule endoscopy” “magnetically controlled capsule endoscopy” “magnetically guided capsule” “magnetic maneuverable capsule” “magnetic assisted capsule endoscopy” “magnetic maneuverable capsule” “conventional gastroscopy” “gastroscopy” “Standard video endoscope” “Traditional passive VCE” “Videoendoscope” “VCE” “sensitivity and specificity” “diagnosis” were used. Regardless of language, year of publication, type of articles, or publication status, all related documents were taken into account. We manually searched the references lists of studies of interest for further potentially relevant studies. Wherever there were conflicts, they were discussed by two authors. The whole process was made according to the PRISMA guidelines.

Study election and data extraction

The inclusion criteria were: (I) participants should be healthy volunteers or have upper abdominal complaint; (II) participants must be given both MCCE and CG; (III) true positive (TP), false positive (FP), true negative (TN), false negative (FN) could be extracted; (IV) the golden standard was CG. The exclusion criteria were as follows: (I) case reports; (II) animal experiments; (III) meta-analyses and reviews; (IV) studies with insufficient information; (V) repeated studies were eliminated expect the most recent one; (VI) patients who suspected dysphagia or symptoms of gastric outlet obstruction, had a history of upper gastrointestinal surgery or abdominal surgery altering gastrointestinal anatomy, congestive heart failure, renal failure, etc. The references were screened and the data were extracted independently by two authors (Jianlong Jiang and Hao Zhang), when conflicts were encountered, the third author was consulted for the final decisions.

We collected the baseline characteristics of included studies, including: the first author, year of publication, countries, participants, age and sex distribution, the type of magnetic manipulation, the time cost by MCCE and CG operation and sample size. The primary outcomes were TP, FP, TN, FN; the secondary outcomes were the diagnostic consistencies, visualization of major stomach landmarks and numbers of side effects or complications.

Definition of outcomes

CG was regarded as the golden diagnostic standard. Both major lesions that would require endoscopic therapy or biopsy (ulceration, polyps, important bile reflux, fundic varicose, bleeding, hiatal hernia) and minor lesions (do not require subsequent gastroscopy) could be diagnosed as positive. The primary outcome measurements in our study were the diagnostic accuracy, sensitivity, and specificity of MCCE in identifying gastric diseases in healthy volunteer or patients with upper abdominal pain.

Index test

The index text was the applying of MCCE with studies describing evidence of gastric diseases in our analysis. In the worldwide, there are three types of MCCE, which mainly differs in the way of magnetic manipulation and the amount of imaging sensor, as follows: robotic-arm magnetic manipulation (ANKON) with one imaging sensor, hand-held magnetic manipulation (Intromedic) with imaging sensor and robot-assisted magnetic manipulation (Olympus and Siemens) with two imaging sensors.

Quality assessment of included studies

Two authors independently evaluated each included literature according to QUADAS-2, and consensus was reached on all items. QUADAS-2 is composed of fourteen items and have already been structured so that four major issues are ranked for risk of bias and concerns regarding applicability to the research question was used to measure the included studies. Each key issue has a set of signaling questions to assess bias and applicability (14,15). We utilized tabular and graphical displays in Review Manager 5 (RevMan 5.3) to summarize the QUADAS-2 assessments.

Statistical analysis

To collect TP, FP, TN, FN, we formed a fourfold table for primary outcomes before we investigated each study. Then we performed the synthesis of the primary data within the bivariate mixed-effects binary regression modeling framework basing on the Cochrane DTA Working Group methodology. After obtaining all the data, we used Stata statistical software version 13.0 (Stata, College Station, TX) to pool the data and MetaDiSc statistical software (Meta-DiSc, version 1.4, Madrid, Spain) to estimate the threshold effect and meta-regression. A P value of less than 0.05 was considered statistically significant. We also calculated the positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR) and the summary receiver operating characteristics curve (SROC). Then we depicted a Fagan plot to evaluate meaningfulness or clinical efficacy. The Fagan nomogram is a very useful graphical instrument for assessing how much the effect of a diagnostic test adjusts the probability that a patient indeed has a disease (16). At last, we rated heterogeneity using forest plots and I² statistics. Larger values of I² prove increasing heterogeneity, with values of 25%, 50%, and 75% revealing low, moderate, and high degrees of heterogeneity, respectively. If there was significant heterogeneity, meta-regression was implemented to explain it. If not, then subgroup analysis was utilized.

Results

Characteristics of included studies

Four studies were finally selected (17-20). A total of 612 patients were assessed. A PRISMA flow chart of the search is shown in (Figure 1). At first, we collected a total of 832 documents through searching databases. Then we drop out duplications, case reports, animal experiments, reviews and articles that do not meet the included criteria. Finally, 4 studies were included. Furthermore, we collected baseline characteristics of included studies including the type of magnetic manipulation, time of MCCE and CG operation, distribution of gender and age, country, publication year and so on in Table 1. Three of the studies included only patients with upper abdominal discomfort and one study included only healthy volunteers. There were two kinds of MCCE as illustrated in (Table 1). Hand-held magnets are simple to use and portable. However, robotic control magnets are better at movement precision than hand-held magnets in porcine models (21). Both male and female patients were included. As mentioned in a study (22), the time of MCCE operation was greatly longer than that of CG, which leads to finding more minor lesions. Besides, the sequence of examinations could also cause bias when CG was given before MCCE, since CG can injury gastric mucosa. All participants in this meta-analysis were firstly given MCCE. In two study (16,18), the visualization of six gastric landmarks was recorded compared MCCE with CG.

Figure 1 Search flow for meta-analysis.

Quality assessment of included studies

The quality of the eligible studies was evaluated by QUADAS-2 criteria and is illustrated (Figure 2). In all included studies, a high risk of bias concerning the selection of patients was discovered, because both healthy volunteers and patients with upper abdominal discomforts should be recruited in clinical practice. There was a low risk of bias issues of the index text, reference standard, flow and timing.

Figure 2 Quality assessments of included studies.

Pooled sensitivity and specificity

The pooled measurements of sensitivity and specificity of MCCE in the diagnosis of gastric diseases were computed as (Figure 3). Four total studies assessing gastric diseases demonstrated pooled of sensitivity and specificity of 91% (95% CI, 0.87–0.93) and 90% (95% CI, 0.75–0.96), respectively. The diagnostic accuracy of MCCE was 91% (95% CI, 0.88–0.94), Area under the curve (AUC) for MCCE was 0.91 (0.88–0.94) by ROC analysis (Figure 4), the PLR and NLR of MCCE were 9.14 (95% CI, 3.44–24.29) and 0.10 (95% CI, 0.07–.15), respectively (Figure 5). The DOR was 87.2 (95% CI, 27.38–277.74) (Figure 6). There was no threshold effect (Spearman correlation coefficient =0.4, P=0.60), but high heterogeneity was found when pooling of specificity, we did not find the reason through meta-regression. Moreover, we did not observe significant publication bias among the studies as assessed according to the DEEK’s test (P=0.456) and Deeks plot (Figure 7).

Figure 3 Pooled sensitivity (A) and specificity (B) of total studies included for meta-analysis.

Figure 4 Diagnostic accuracy of MCCE for the diagnosis of gastric diseases.

Figure 5 Pooled positive (A) and negative (B) likelihood ratio of total studies included for meta-analysis.

Figure 6 Pooled diagnostic odds ratio of total studies included for meta-analysis.

Figure 7 Deek’s plot for publication bias.

The associated Fagan Nomogram is illustrated in (Figure 8). With a low pre-test probability (20%) of gastric diseases, if MCCE demonstrated the presence of gastric diseases, the post-test probability that the patient truly has gastric diseases would be approximately 70%. Alternatively, if the patient tests negative (i.e., no gastric diseases are seen on MCCE), the post-test probability that the patient indeed has gastric diseases would be lower than 3%.

Figure 8 Fagan nomogram for diagnosis of gastric diseases.

Adverse events

Two studies reported minor discomforts due to the process of drinking too much water and swallowing the capsule (3 patients had abdominal distension and nausea, 1 patient had headache and vomiting, 1 patient had foreign body sensations and 1 patient had chronic diarrhea).

Discussion

This review was intended to confirm the sensitivity and specificity of MCCE in the diagnosis of gastric diseases in adults. We demonstrate that MCCE has a high sensitivity (91%) and specificity (90%) in diagnosis of gastric disease. In addition, we also found that AUC and DOR of MCCE was 0.91 and 87.2, which indicate MCCE has good capability at distinguishing patients from healthy. Collectively, our data proved that MCCE is a suitable alternative for gastric diseases that do not require biopsy.

Several studies (16,17) in related fields demonstrated that MCCE was able to detect gastric lesions with comparable accuracy with conventional gastroscopy and can be an alternative for screening gastric diseases without sedation. Likewise, we quantitatively pooled all the related studies, and the result confirmed that MCCE serves as a useful tool in diagnosis gastric disease. The Fagan nomogram for MCCE demonstrated that even in participants with a low pre-test probability, MCCE is an effective tool for expelling the presence of gastric diseases. However, two studies (18,19) reported that the visualization of gastric fundus and cardia in the up 1/3 stomach was lower than bottom 2/3, which caused by insufficient expansion of the stomach. In our study, the distribution of lesions almost was in the bottom 2/3 of stomach in healthy volunteers and patients with upper abdominal discomforts. Studies about MCCE focus on detecting lesion in the proximal stomach were greatly needed.

There was considerable heterogeneity in the analysis of specificity between the studies. In two studies (17,19), the specificity of MCCE in detecting gastric diseases was only 78.0% and 71.0% respectively, while in two studies the specificity was 100% and 95.0% respectively (16,18). Meta-regression was undertaken to discern the cause of the heterogeneity, however we didn’t found the culprit through statistical methods in the three expected variables that is the operation time of MCCE (P=0.3244), the kind of participants (P=0.9584), the type of MCCE (P=0.3463). We speculated some hidden variables that may arouse heterogeneity. The most possible reason for the presence of heterogeneity may be connected with the experience of the endoscopists and the number of endoscopists who read the capsule images. Despite the fact that the MCCE investigators were all blinded to CG findings, their experience was not set out in detail.

However, the limitations of our study have to be acknowledged. Firstly, because MCCE is a new device which has not been commonly used for detecting gastric diseases, the number of related studies is insufficient. Secondly, the numbers of involved patients of included studies were significantly dissimilar, ranging from 26 to 350 patients. Thirdly, only 26 (26/612, 4.24%) healthy volunteers were recruited, which is not consistent with clinical practice. Fourthly, as reported in a study (17), luminal visibility and location of the lesion can dissatisfy the diagnostic accuracy of minor lesions and have no impact on major lesions. MCCE perform better than CG at diagnosing minor lesions, however the proportion of minor lesion is not sufficient, which may confuse researchers. Further studies concerning about minor lesions were demanded. At last, lesions distributing in the upper 1/3 of the stomach were not sufficient, which often were missed by MCCE, this may lead to highly diagnostic accuracy compared to CG.

In summary, MCCE have acceptable sensitivity and specificity in diagnosing gastric diseases in adults, however there are still many problems waiting to be resolved (23). A lens-cleaning system which is available with CG, a stronger guidance system which seems to be too weak at present and requires faster speed of movement with stronger force, a biopsy system (24) and a better capability of keeping the capsule more steadily have to be settled urgently in the future. All in all, further large scale studies are required to test and verify its value in the upper gastrointestinal tract. We believe that one day in the near future capsule endoscopy will change the way we practice gastroenterology.

Acknowledgments

Funding: This work was supported by the Science and Technology Planning Project of Guangdong Province, China (2017A020215014) and the Medical Scientific Research Foundation of Guangdong Province, China (A2017274).

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at http://dx.doi.org/10.21037/dmr-20-141

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at http://dx.doi.org/10.21037/dmr-20-141). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Iddan G, Meron G, Glukhovsky A, et al. Wireless capsule endoscopy. Nature 2000;405:417. [Crossref] [PubMed]
2. Rex DK, Adler SN, Aisenberg J, et al. Accuracy of capsule colonoscopy in detecting colorectal polyps in a screening population. Gastroenterology 2015;148:948-957.e2. [Crossref] [PubMed]
3. Hale M, McAlindon M. Capsule endoscopy as a panenteric diagnostic tool. Br J Surg 2014;101:148-9. [Crossref] [PubMed]
4. Hosoe N, Naganuma M, Ogata H. Current status of capsule endoscopy through a whole digestive tract. Dig Endosc 2015;27:205-15. [Crossref] [PubMed]
5. Rahman I, Afzal N, Patel P. The role of magnetic assisted capsule endoscopy (MACE) to aid visualisation in the upper GI tract. Comput Biol Med 2015;65:359-63. [Crossref] [PubMed]
6. Liao Z, Duan XD, Xin L, et al. Feasibility and safety of magnetic-controlled capsule endoscopy system in examination of human stomach: a pilot study in healthy volunteers. J Interv Gastroenterol 2012;2:155-60. [Crossref] [PubMed]
7. Carpi F, Galbiati S, Carpi A. Magnetic shells for gastrointestinal endoscopic capsules as a means to control their motion. Biomed Pharmacother 2006;60:370-4. [Crossref] [PubMed]
8. Ching HL, Hale MF, McAlindon ME. Current and future role of magnetically assisted gastric capsule endoscopy in the upper gastrointestinal tract. Therap Adv Gastroenterol 2016;9:313-21. [Crossref] [PubMed]
9. Keller J, Fibbe C, Volke F, et al. Inspection of the human stomach using remote-controlled capsule endoscopy: a feasibility study in healthy volunteers (with videos). Gastrointest Endosc 2011;73:22-8. [Crossref] [PubMed]
10. Shamsudhin N, Zverev VI, Keller H, et al. Magnetically guided capsule endoscopy. Med Phys 2017;44:e91-e111. [Crossref] [PubMed]
11. 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;71:1290-3. [Crossref] [PubMed]
12. Rey JF, Ogata H, Hosoe N, et al. Feasibility of stomach exploration with a guided capsule endoscope. Endoscopy 2010;42:541-5. [Crossref] [PubMed]
13. Keller J, Fibbe C, Volke F, et al. Remote magnetic control of a wireless capsule endoscope in the esophagus is safe and feasible: results of a randomized, clinical trial in healthy volunteers Gastrointest Endosc 2010;72:941-6. [Crossref] [PubMed]
14. Thompson SG, Higgins JP. How should meta-regression analyses be undertaken and interpreted? Stat Med 2002;21:1559-73. [Crossref] [PubMed]
15. Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25. [Crossref] [PubMed]
16. Fagan TJ. Letter: Nomogram for Bayes theorem. N Engl J Med 1975;293:257. [Crossref] [PubMed]
17. 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.e1. [Crossref] [PubMed]
18. 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;49:101-7. [Crossref] [PubMed]
19. Rahman I, Pioche M, Shim CS, et al. Magnetic-assisted capsule endoscopy in the upper GI tract by using a novel navigation system (with video). Gastrointest Endosc 2016;83:889-895.e1. [Crossref] [PubMed]
20. Zou WB, Hou XH, Xin L, et al. Magnetic-controlled capsule endoscopy vs. gastroscopy for gastric diseases: a two-center self-controlled comparative trial. Endoscopy 2015;47:525-8. [Crossref] [PubMed]
21. Ciuti G, Donlin R, Valdastri P, et al. Robotic versus manual control in magnetic steering of an endoscopic capsule. Endoscopy 2010;42:148-52. [Crossref] [PubMed]
22. 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;75:373-81. [Crossref] [PubMed]
23. Hale MF, Rahman I, Drew K, et al. Magnetically steerable gastric capsule endoscopy is equivalent to flexible endoscopy in the detection of markers in an excised porcine stomach model: results of a randomized trial. Endoscopy 2015;47:650-3. [Crossref] [PubMed]
24. Chen X, Ran ZH, Tong JL. A meta-analysis of the yield of capsule endoscopy compared to double-balloon enteroscopy in patients with small bowel diseases. World J Gastroenterol 2007;13:4372-8. [Crossref] [PubMed]