Patients with hereditary diffuse gastric cancer often undergo prophylactic gastrectomy to minimize cancer risk. Because intramucosal poorly cohesive carcinomas in this setting are typically not grossly visible, many pathologists assess the entire gastrectomy specimen microscopically. With 150 or more slides per case, this is a major time burden for pathologists. This study utilizes deep learning methods to analyze digitized slides and detect regions of carcinoma. Prophylactic gastrectomy specimens from seven patients with germline CDH1 mutations were analyzed (five for training/validation and two for testing, with a total of 133 tumor foci). All hematoxylin and eosin slides containing cancer foci were digitally scanned, and patches of size 256×256 pixels were randomly extracted from regions of cancer as well as from regions of normal background tissue, resulting in 15,851 images for training/validation and 970 images for testing. A model with DenseNet-169 architecture was trained for 150 epochs, then evaluated on images from the test set. External validation was conducted on 814 images scanned at an outside institution. On individual patches, the trained model achieved a receiver operating characteristic (ROC) area under the curve (AUC) of 0.9986. This enabled it to maintain a sensitivity of 90% with a false-positive rate of less than 0.1%. On the external validation dataset, the model achieved a similar ROC AUC of 0.9984. On whole slide images, the network detected 100% of tumor foci and correctly eliminated an average of 99.9% of the non-cancer slide area from consideration. Overall, our model shows encouraging progress towards computer-assisted diagnosis of hereditary diffuse gastric cancer.

Intestinal type villous adenomas are uncommon in the genitourinary tract. Most reported cases have been located in the urinary bladder or urachus. Villous adenoma arising in the renal pelvis or ureteris very rare. We present a case of an 81-year-old female who presented with difficulty voiding and mucosuria. A computed tomography scan identified right-sided hydronephrosis, renal parenchymal atrophy, nonobstructing calculi and a lower pole renal mass. She underwent open right nephrectomy. Histopathologic examination of the kidney revealed an intestinal type villous adenoma of the renal pelvis with high-grade dysplasia and focal areas suspicious for invasive adenocarcinoma. We review the four previously reported cases of intestinal type villous adenoma in the renal pelvis and discuss diagnosis and management of this unusual neoplasm.

Pancreatic ductal adenocarcinoma (PDAC) remains highly lethal, with a five-year survival rate below 15 %. Claudin 18.2 (CLDN18.2) has emerged as a novel potential therapeutic target in PDAC. Zolbetuximab, a monoclonal antibody targeting CLDN18.2, has demonstrated therapeutic benefit in gastric cancers and is now in phase 2 clinical trials for PDAC. Trial eligibility for zolbetuximab requires tumor membranous immunohistochemical staining with the pan-claudin 18 companion diagnostic antibody clone 43-14A. However, the expression of CLDN18 detected using this clone has only been evaluated in 62 patients from a single retrospective study. Herein, we report immunohistochemical staining using 43-14A on surgically resected PDAC samples (n = 120). Samples were stained following the protocol used in clinical trials, using the 43-14A VENTANA antibody in a prediluted kit, and according to the manufacturer's recommended protocol. Positive cases were defined as ≥ 75 % of tumor cells exhibiting membranous staining with an intensity of ≥ 2+. Out of 120 PDAC cases, 39 (32.5 %) stained positive for CLDN18 with 43-14A. A significant association was observed between lower tumor grade and higher 43-14A staining (p < 0.05). CLDN18-positive cases demonstrated significantly improved survival at the cohort's median overall survival (23 months, p < 0.05), suggesting that claudin expression could serve as a both a diagnostic and prognostic marker. Our findings indicate that 32.5 % of PDAC tumors in this cohort are positive for CLDN18, suggesting that a significant proportion of patients with PDAC could benefit from zolbetuximab and other CLDN18.2 targeted immunotherapies if pancreatic cancer therapeutic trials prove successful.

Ciliated foregut cysts typically arise in the liver and pancreas, with rare cases reported to arise in the gallbladder wall. Herein, we describe an unusual case of a patient with a ciliated foregut cyst arising from the wall of the common hepatic duct, which has not been reported previously. As a result of its unique anatomical location, the cyst was mistaken for a type II choledochal cyst on preoperative imaging. The cyst caused intermittent biliary obstruction that resolved following surgical resection.

Context.—: Pathologists routinely assess Ki67 immunohistochemistry to grade gastrointestinal and pancreatic neuroendocrine tumors. Unfortunately, manual counts of the Ki67 index are very time consuming and eyeball estimation has been criticized as unreliable. Manual Ki67 counts performed by cytotechnologists could potentially save pathologist time and improve accuracy. Objective.—: To assess the concordance between manual Ki67 index counts performed by cytotechnologists versus eyeball estimates and manual Ki67 counts by pathologists. Design.—: One Ki67 immunohistochemical stain was retrieved from each of 18 archived gastrointestinal or pancreatic neuroendocrine tumor resections. We compared pathologists' Ki67 eyeball estimates on glass slides and printed color images with manual counts performed by 3 cytotechnologists and gold standard manual Ki67 index counts by 3 pathologists. Results.—: Tumor grade agreement between pathologist image eyeball estimate and gold standard pathologist manual count was fair (κ = 0.31; 95% CI, 0.030–0.60). In 9 of 20 cases (45%), the mean pathologist eyeball estimate was 1 grade higher than the mean pathologist manual count. There was almost perfect agreement in classifying tumor grade between the mean cytotechnologist manual count and the mean pathologist manual count (κ = 0.910; 95% CI, 0.697–1.00). In 20 cases, there was only 1 grade disagreement between the 2 methods. Eyeball estimation by pathologists required less than 1 minute, whereas manual counts by pathologists required a mean of 17 minutes per case. Conclusions.—: Eyeball estimation of the Ki67 index has a high rate of tumor grade misclassification compared with manual counting. Cytotechnologist manual counts are accurate and save pathologist time.