Computer-aided characterization of early cancer in Barrett's esophagus on i-scan magnification imaging: a multicenter international study.

BACKGROUND AND AIMS: We aimed to develop a computer-aided characterization system that could support the diagnosis of dysplasia in Barrett's esophagus (BE) on magnification endoscopy. METHODS: Videos were collected in high-definition magnification white-light and virtual chromoendoscopy with i-scan (Pentax Hoya, Japan) imaging in patients with dysplastic and nondysplastic BE (NDBE) from 4 centers. We trained a neural network with a Resnet101 architecture to classify frames as dysplastic or nondysplastic. The network was tested on 3 different scenarios: high-quality still images, all available video frames, and a selected sequence within each video. RESULTS: Fifty-seven patients, each with videos of magnification areas of BE (34 dysplasia, 23 NDBE), were included. Performance was evaluated by a leave-1-patient-out cross-validation method. In all, 60,174 (39,347 dysplasia, 20,827 NDBE) magnification video frames were used to train the network. The testing set included 49,726 i-scan-3/optical enhancement magnification frames. On 350 high-quality still images, the network achieved a sensitivity of 94%, specificity of 86%, and area under the receiver operator curve (AUROC) of 96%. On all 49,726 available video frames, the network achieved a sensitivity of 92%, specificity of 82%, and AUROC of 95%. On a selected sequence of frames per case (total of 11,471 frames), we used an exponentially weighted moving average of classifications on consecutive frames to characterize dysplasia. The network achieved a sensitivity of 92%, specificity of 84%, and AUROC of 96%. The mean assessment speed per frame was 0.0135 seconds (SD ± 0.006). CONCLUSION: Our network can characterize BE dysplasia with high accuracy and speed on high-quality magnification images and sequence of video frames, moving it toward real-time automated diagnosis.

A new artificial intelligence system successfully detects and localises early neoplasia in Barrett's esophagus by using convolutional neural networks.

BACKGROUND AND AIMS: Seattle protocol biopsies for Barrett's Esophagus (BE) surveillance are labour intensive with low compliance. Dysplasia detection rates vary, leading to missed lesions. This can potentially be offset with computer aided detection. We have developed convolutional neural networks (CNNs) to identify areas of dysplasia and where to target biopsy. METHODS: 119 Videos were collected in high-definition white light and optical chromoendoscopy with i-scan (Pentax Hoya, Japan) imaging in patients with dysplastic and non-dysplastic BE (NDBE). We trained an indirectly supervised CNN to classify images as dysplastic/non-dysplastic using whole video annotations to minimise selection bias and maximise accuracy. The CNN was trained using 148,936 video frames (31 dysplastic patients, 31 NDBE, two normal esophagus), validated on 25,161 images from 11 patient videos and tested on 264 iscan-1 images from 28 dysplastic and 16 NDBE patients which included expert delineations. To localise targeted biopsies/delineations, a second directly supervised CNN was generated based on expert delineations of 94 dysplastic images from 30 patients. This was tested on 86 i-scan one images from 28 dysplastic patients. FINDINGS: The indirectly supervised CNN achieved a per image sensitivity in the test set of 91%, specificity 79%, area under receiver operator curve of 93% to detect dysplasia. Per-lesion sensitivity was 100%. Mean assessment speed was 48 frames per second (fps). 97% of targeted biopsy predictions matched expert and histological assessment at 56 fps. The artificial intelligence system performed better than six endoscopists. INTERPRETATION: Our CNNs classify and localise dysplastic Barrett's Esophagus potentially supporting endoscopists during surveillance.

Efficacy and safety of a new low-volume PEG with citrate and simethicone bowel preparation for colonoscopy (Clensia): a multicenter randomized observer-blind clinical trial vs. a low-volume PEG with ascorbic acid (PEG-ASC).

BACKGROUND AND STUDY AIMS: Quality of inspection during colonoscopy is strictly related to the level of cleansing. High-volume (PEG-based) solutions are highly effective and safe, but their high volume affects tolerability and compliance. The aim of this study was to compare a new low-volume PEG with citrate and simethicone solution (PMF 104,Clensia) with a low-volume PEG with ascorbic acid solution (PEG-ASC; Moviprep). PATIENTS AND METHODS: This was a multicenter, randomized, observer-blind, parallel-group, phase 3 clinical trial, where patients were randomized between PMF 104 and PEG-ASC. In both groups, patients were instructed to take a full-dose regimen the evening before if colonoscopy was scheduled before 11 am to 12 pm, or to take a split regimen if colonoscopy was scheduled after 11 am to 12 pm. The primary end-point was an equivalence between PMF104 and PEG-ASC in the rate of adequate level of cleansing (Ottawa scale ≤ 6), with safety, mucosal visibility, tolerability, acceptance and compliance being also assessed. RESULTS: Of the 403 enrolled, 367 patients (Mean age [SD]: 55.6 (14.4) years; male:166 [45.2 %]) were included in the per protocol (PP) analysis: 184 being randomized in the PMF 104 group and 183 in the PEG-ASC group. Successful bowel cleansing was 78.3 % and 74.3 % in PMF104 and in PEG-ASC, respectively ( P  = 0.37). Both preparations were equally safe (mild adverse events were observed in 9.2 % and 9.3 % of patients in the PMF104 and in the PEG-ASC group, respectively) and acceptable (no or mild distress during the intake in 81.4 % and 80.8 % in the PMF104 in the PEG-ASC, respectively [ P  = 0.74]). CONCLUSION: The new low-volume product Clensia is equivalent to the reference low-volume PEG-ASC in terms of bowel cleansing, safety and acceptance.

A novel noninvasive genotyping method of Helicobacter pylori using stool specimens.

BACKGROUND & AIMS: The source(s) of the infection and the route(s) of transmission of Helicobacter pylori have not yet been clarified. This is to introduce a noninvasive protocol allowing molecular typing of H pylori using stool specimens. METHODS: The genotyping method is based on 2 H pylori-specific biprobe real-time polymerase chain reaction assays using fragments of the glmM and the recA genes as target sequences. Discrimination between strains results from differences in the melting temperature during melting curve analysis. In case of identical melting temperatures in both assays, sequence analysis of the glmM amplicon was performed to confirm strain identity. The method was validated using gastric biopsy specimens and stool specimens of 97 unrelated individuals suffering from abdominal pain and stool specimens of members of 10 families in Austria (infected index child and family members) and 8 African households. RESULTS: Of the 97 patients, 27 were infected as shown by culture, histology, and rapid urease test. The sensitivity of each of the assays was 100% in gastric biopsy specimens and 92.2% in stool specimens; the specificity was 100%. The discriminatory capacity of the method was 100%. Clonal identities were found in 9 of 10 (90%) European and 7 of 8 (87.5%) African households. In 2 African households, 2 different clonal lineages each were found. CONCLUSIONS: The genotyping protocol introduced allows for both accurate detection and discrimination of H pylori strains in stool samples. Large-scale studies using this protocol may contribute to the clarification of the transmission pathways of infection with H pylori.

Histopathology of columnar-lined esophagus in patients with gastroesophageal reflux disease.

BACKGROUND AND AIMS: The question of whether an endoscopically normal-appearing esophagogastric junction should be biopsied in patients with gastroesophageal reflux disease is controversial. We have addressed this issue using endoscopy and histopathology. METHODS: A total of 114 consecutive patients (58 males) with symptoms of gastroesophageal reflux disease prospectively underwent endoscopy, including biopsy sampling from the esophagogastric junction. Endoscopically visible columnar-lined esophagus was defined by the presence of gastric-type mucosa above the level of the rise of the gastric folds. Histopathology was conducted using the Paull-Chandrasoma classification. RESULTS: Of the 114 patients, 85 (74.6%) had endoscopically visible columnar-lined esophagus of length < or =0.5 cm (n = 82), 1 cm (n = 2) and 7 cm (n = 1); 29 patients (25.4%) had a normal endoscopic junction. All patients had histopathologic columnar-lined esophagus. Intestinal metaplasia and low-grade dysplasia was identified in 26 (22.8%) and 5 (4.4%) individuals, respectively, and was not statistically different in endoscopically normal vs. abnormal junction (P = 0.408 for intestinal metaplasia, P = 0.775 for low grade dysplasia). Intestinal metaplasia was independent from endoscopic esophagitis (P = 0.398) and hiatal hernia (P = 0.405). CONCLUSIONS: Columnar-lined esophagus cannot be excluded by endoscopy. In patients with gastroesophageal reflux disease, biopsy sampling of normal-appearing junction is recommended for histopathologic exclusion of intestinal metaplasia and low-grade dysplasia.

Prescintigraphic morphine application for abdominal adenocarcinoma imaging, with technetium-99m methoxy isobutyl isonitrile.

Imaging of tumors with cationic tracers, especially with technetium-99 methoxy isobutyl isonitrile ((99m)Tc-MIBI), revealed high specificity for the diagnosis and follow up of various malignancies. However, these radiopharmaceuticals are of limited value for the diagnosis of malignancies of the abdominal region due to the immediate biliary secretion of the tracer and the associated high background activity. In a prospective, single-blinded protocol, patients with endoscopically diagnosed gastrointestinal malignancies were assigned to undergo (99m)Tc-MIBI imaging of the abdomen. To overcome biliary secretion of cationic tracer we administered 0.04 mg/kg morphine hydrochloride intravenously before the administration of 600 MBq (99m)Tc-MIBI. Planar images were performed in the anterior and posterior views with a double-headed gamma camera and with 3 min acquisition time, followed by single photon emission tomography images (3 degrees, 20 sec/frame). Results were compared to findings of endoscopy, computed tomography scan and surgery. Twenty four patients 17 male and 7 female , mean age 69 years, range 52-83, years were enrolled. All patients suffered from adenocarcinoma, (19 from colorectum, 3 from gastric, 1 from pancreatic and one patient had both gastric and colorectal adenocarcinoma, for a total of 25 tumor lesions). The primary objective- inhibition of biliary secretion- was achieved in 23 of the 24 patients. (99m)Tc-MIBI-imaging was accumulated intra-abdominally in 19 patients. In 2 patients the tumor was endoscopically completely removed before the scan. In these two patients (99m)Tc-MIBI imaging showed no intra-abdominal tracer accumulation. When compared to the endoscopic findings, (99m)Tc-MIBI imaging showed time positive results in 13 of the 23 remaining individual tumor lesions, false positive in 6 and false negative in 4. This study showed a sensitivity of 57% and a specificity of 20% of the above technique for the identification of intra-abnominal adenocarcinomas. Correct diagnosis did not correlate with tumor size. In conclusion, prescintigraphic morphine administration inhibits background activity coming from biliary secretion, and enables better intra-abdominal (99m)Tc-MIBI imaging but with limited sensitivity and poor specificity.

Antibiotic treatment is not effective in patients infected with Helicobacter pylori suffering from extragastric MALT lymphoma.

PURPOSE: Apart from anecdotal reports implicating Helicobacter pylori (HP) in the development of extragastric mucosa associated lymphoid tissue (MALT) lymphoma, no large scale prospective studies have been performed on this topic. PATIENTS AND METHODS: A total of 77 patients with extragastric MALT lymphoma were prospectively studied. The presence or absence of HP was tested by histology, urease breath test, and serology. Patients were also tested for hepatitis A, B, and C and autoimmune conditions along with assessment of MALT lymphoma-specific genetic changes. RESULTS: Evidence for infection with HP was present in 35 of 77 patients (45%), and three of 75 patients tested (4%) were positive for hepatitis C and one for hepatitis B. All patients with HP-infection underwent eradication, 16 before initiation of further therapy. Apart from one patient with lymphoma involving parotid and colon, who achieved regression of the colonic lesions, none of these 16 patients showed regression of the lymphoma after a median follow-up of 14 months (range, 8 to 48+ months) before initiation of definitive treatment. No correlation between HP-status, localization, stage, autoimmune diseases, and genetic findings was seen. CONCLUSION: In our series, HP-eradication was ineffective for treatment of extragastric MALT lymphomas. This finding, along with an infection rate of 45%-as could also be expected in the general Austrian population-suggests that HP does not play a role in the development of these lymphomas. Antibiotic treatment targeting HP should, therefore, be discouraged in patients with extragastric MALT lymphomas.