ABSTRACT
Deep learning has recently been gaining attention as a promising technology to improve the identification of lesions, and deep-learning algorithms for lesion detection have been actively developed in small-bowel capsule endoscopy (SBCE). We developed a detection algorithm for abnormal findings by deep learning (convolutional neural network) the SBCE imaging data of 30 cases with abnormal findings. To enable the detection of a wide variety of abnormal findings, the training data were balanced to include all major findings identified in SBCE (bleeding, angiodysplasia, ulceration, and neoplastic lesions). To reduce the false-positive rate, "findings that may be responsible for hemorrhage" and "findings that may require therapeutic intervention" were extracted from the images of abnormal findings and added to the training dataset. For the performance evaluation, the sensitivity and the specificity were calculated using 271 detectable findings in 35 cases. The sensitivity was calculated using 68,494 images of non-abnormal findings. The sensitivity and specificity were 93.4% and 97.8%, respectively. The average number of images detected by the algorithm as having abnormal findings was 7514. We developed an image-reading support system using deep learning for SBCE and obtained a good detection performance.
ABSTRACT
BACKGROUND: Capsule endoscopy and balloon-assisted enteroscopy (BAE) enable visualization of rare small bowel conditions such as small intestinal malignant tumors. However, details of the endoscopic characteristics of small intestinal malignant tumors are still unknown. AIM: To elucidate the endoscopic characteristics of small intestinal malignant tumors. METHODS: From March 2005 to February 2017, 1328 BAE procedures were performed at Keio University Hospital. Of these procedures, malignant tumors were classified into three groups, Group 1: epithelial tumors including primary small intestinal cancer, metastatic small intestinal cancer, and direct small intestinal invasion by an adjacent organ cancer; Group 2: small intestinal malignant lymphoma; and Group 3, small intestinal gastrointestinal stromal tumors. We systematically collected clinical and endoscopic data from patients' medical records to determine the endoscopic characteristics for each group. RESULTS: The number of patients in each group was 16 (Group 1), 22 (Group 2), and 6 (Group 3), and the percentage of solitary tumors was 100%, 45.5%, and 100%, respectively (P < 0.001). Patients' clinical background parameters including age, symptoms, and laboratory data were not significantly different between the groups. Seventy-five percent of epithelial tumors (Group 1) were located in the upper small intestine (duodenum and ileum), and approximately 70% of gastrointestinal stromal tumors (Group 3) were located in the jejunum. Solitary protruding or mass-type tumors were not seen in malignant lymphoma (Group 2) (P < 0.001). Stenosis was seen more often in Group 1, (68.8%, 27.3%, and 0%; Group 1, 2, and 3, respectively; P = 0.004). Enlarged white villi inside and/or surrounding the tumor were seen in 12.5%, 54.5%, and 0% in Group 1, 2, and 3, respectively (P = 0.001). CONCLUSION: The differential diagnosis of small intestinal malignant tumors could be tentatively made based on BAE findings.
ABSTRACT
BACKGROUND: Dermatomyositis (DM) is an autoimmune disease characterized by cutaneous Gottron papules, heliotrope rash, and proximal myopathy. It may also present as a paraneoplastic syndrome that can complicate a variety of different cancers, such as lung, cervical, and breast cancer. However, the association with hepatocellular carcinoma (HCC) is extremely rare. Moreover, to our knowledge, there are no previous reports of colonic perforation following steroid pulse treatment for a DM patient. CASE SUMMARY: A 61-year-old male complained of a skin rash that began in his neck and spread to his face and abdomen. On physical examination, the patient was also found to have symmetrical proximal muscle weakness, abdominal pain, heliotrope rash in the periorbital skin, and poikiloderma on his face and abdomen. Serum level of muscle enzymes was remarkably increased. Muscle examination revealed symmetrical proximal weakness. The diagnosis of DM was made, and steroid treatment was started for symptomatic relief. A search for causative malignancy revealed HCC. Despite steroid therapy for DM, his symptoms did not improve. Additionally, C-reactive protein elevation was seen along with severe abdominal pain on day 14 of admission. Shortly after this, the patient died of septic shock due to suppurative peritonitis after perforation of the ascending colon. CONCLUSION: Here, we present a rare case of DM caused by non-hepatitis-associated advanced HCC with colonic perforation. The cause of colonic perforation is still unclear. This case demonstrates the need to carefully monitor abdominal pain in DM patients as symptoms can be masked by steroid therapy.
