Newly development of capsule endoscopy provides a comfortable and minimal invasive modality which is an alternative to conventional esophagogastroduodenoscopy(EGD). The use of capsule endoscopy beyond the small bowel is increasing and several capsule endoscopy systems have been introduced for the examination of the esophagus and colon. The current capsule endoscopy systems are less effective for the upper gastrointestinal tract examination. Short transit time in the esophagus and the passive movement of the capsule makes it more difficult to identify or visualize the lesion comparing with traditional EGD. The sensitivity rate of esophageal varices detection from capsule endoscopy was ranging from 65% to 80%. In order to control the capsule in the gastrointestinal tract for better visualization, many methods are invented. Magnetically assisted capsule endoscopy systems and string-mounted capsule endoscopy are applied in many studies. Magnetically assisted capsule endoscopy system and string-mounted capsule endoscopy are used to control the capsule endoscopy for elongating esophagus transit time to have a better visualization of the esophagus. The InsightEyes EGD System combines the string and magnetic assisted capsule endoscopy system to provide a real-time high-quality image during the examination. On the other hand, 3D image processing can be used for distinguishing the esophageal varices and normal folds well, theoretically. Thus, in this study, the investigators combine string, magnetically assisted capsule endoscopy systems, and 3D image processing together to form a new system for improving the detection of esophageal varices and other gastric lesions.

This is a prospective and observational study. Sample size calculation and patient selection The investigators plan to include a total of 140 participants in this study. Sixty-five cirrhotics with portal hypertensive symptoms (such as platelets count less than 100,000) will be enrolled to evaluate for the presence and stage of esophageal varices by using a magnetic-assisted capsule endoscope system. Thirty-five cirrhotics patients with history of endoscopy-confirmed esophageal varices will be included for the follow-up examination by using the magnetic-assisted capsule endoscope system. Another 40 healthy volunteers without known gastrointestinal disease will be enrolled as the control group. Those participants with less than 20-year-old, having implanted electronic or magnetic devices, pregnant, or unable to swallow the capsule, will be excluded from the study.

Intervention The patient will undergo both diagnostic tests in two days, with a magnetic-assisted capsule endoscope system first and then conventional EGD will be carried on at least 4 hours afterward.

Magnetic-assisted capsule endoscope system (MACE, InsightEyes EGD System) with or without 3D images("MedicalTek" Endoscopic Imaging system and accessories) will be used for detecting esophageal varices and other possible lesions. As standard per-endoscopy preparation, patients undergo a 6-8 hour fast. The participants take 400 mg of N-acetylcysteine solution in 200 mL of water 1 h before the MACE examination and an additional 300 mL of clear water after 30 min to remove gastric mucus. With the upright(sitting) position, the patient is instructed to swallow the capsule with 20mL of clear water. The hand-held magnet is held above the sternum in order to catch the capsule and examine the esophagus. Capsule endoscopy is manipulated by changes in the patient's position and hand-held magnet. The magnetic field navigator is placed in front of the patient's abdomen and helped the capsule moving and rotating by moving the magnetic field navigator.

The capsule endoscopy will end up in the second portion of the duodenum as the conventional EGD does.

The lesion is detected using 2D and 3D imaging at the same time during the examination.

Esophagogastroduodenoscopy (EGD) EGDs (Olympus GIF-H290) are conducted by experienced endoscopists. These practitioners are able to check the indication of the procedure and previous findings of EGD. The EGDs are conducted by standard techniques. The patient undergoes pharyngeal anesthesia and then EGD.

Preparation evaluation The quality of gastric preparation includes the degree of mucosal visibility and stomach distention. The degree of mucosal visibility is classified into four grades: excellent = no adherent mucus on the gastric mucosa; good = small amount of mucus on the gastric mucosa but without obscured vision; fair = small amount of mucus on the gastric mucosa with partially obscured vision; and poor = profuse mucus on the gastric mucosa with obscured vision.

Stomach distention is classified into three grades: good = satisfactory stomach distention with the remaining small amount of collapsed gastric folds; fair = considerable amount of collapsed gastric folds that obscure some parts of the stomach; and poor = insufficient stomach inflation, and examination of the area of interest is impossible. The investigators evaluate the proximal (cardia, fundus, and upper body) and distal (lower body and antrum) portions of the stomach separately.

Lesion detection Esophageal varices detection Esophageal varices will be evaluated and described according to the Japanese Research Society for Portal Hypertension.

Other lesion detection Another lesion is evaluated including portal hypertensive gastropathy, peptic ulcers, and polyps.

Questionnaire on comfort and acceptance After the examinations, the patient will receive a questionnaire, including the pain during the examination by using a visual analog scale (VAS), scoring from 1(no pain) to 10(badly painful, barely tolerable), other discomforts such as nausea and abdominal fullness. The questionnaire also includes the preferred choice of two modalities.

Study outcomes Primary outcome

1. The investigators will compare the detection rate of esophageal varices in the cirrhotic patients in the MACE system and 3D image processing with conventional EGD.
2. The investigators will compare the grading consistency in the patients who have the esophageal varices between the two modalities, using the conventional EGD as the gold standard.

Secondary outcome

1. The investigators will evaluate the image quality in the MACE system and 3D image processing in all participants.
2. The investigators will compare the detection rate of other gastric lesions such as portal hypertensive gastropathy, gastric polyp, and ulcer in the MACE system plus 3D image processing with conventional EGD in all participants.
3. The investigators will evaluate the discomfort using the questionnaire in all participants.

Inclusion Criteria:

• The investigators plan to include a total of 140 participants in this study, aged from 20 to 75-year-old.

Exclusion Criteria:

• Those participants with less than 20-year-old, having implanted electronic or magnetic devices, pregnant, or unable to swallow the capsule, will be excluded from the study.

The investigators plan to include a total of 140 participants in this study. Sixty-five cirrhotic patients with portal hypertensive symptoms (such as platelets count less than 100,000) will be enrolled to evaluate for the presence and stage of esophageal varices by using a magnetic-assisted capsule endoscope system with or without 3D image processing. Thirty-five cirrhotic patients with a history of endoscopy-confirmed esophageal varices will be included for the follow-up examination by using the magnetic-assisted capsule endoscope system with or without 3D images. Another 40 healthy volunteers without known gastrointestinal disease will be enrolled as the control group.