Signet ring cell carcinoma (SRCC) of the stomach is a rare type of cancer with a slowly rising incidence. It tends to be more difficult to detect by pathologists, mainly due to its cellular morphology and diffuse invasion manner, and it has poor prognosis when detected at an advanced stage. Computational pathology tools that can assist pathologists in detecting SRCC would be of a massive benefit. In this paper, we trained deep learning models using transfer learning, fully-supervised learning, and weakly-supervised learning to predict SRCC in Whole Slide Images (WSIs) using a training set of 1,765 WSIs. We evaluated the models on two different test sets (n = 999, n = 455). The best model achieved a ROC-AUC of at least 0.99 on all two test sets, setting a top baseline performance for SRCC WSI classification.
Esophagogastric junction (EGJ) adenocarcinoma has been on the increase in Western countries. However, in Asian countries, data on the incidence of EGJ adenocarcinoma are evidently lacking. In the present review, we focus on the current clinical situation of EGJ adenocarcinoma in three Asian countries: Japan, Hong Kong, and Malaysia. The incidence of EGJ adenocarcinoma has been reported to be gradually increasing in Malaysia and Japan, whereas it has stabilized in Hong Kong. However, the number of cases in these countries is comparatively low compared with Western countries. A reason for the reported difference in the incidence and time trend of EGJ adenocarcinoma among the three countries may be explained by two distinct etiologies: one arising from chronic gastritis similar to distal gastric cancer, and the other related to gastroesophageal reflux disease similar to esophageal adenocarcinoma including Barrett's adenocarcinoma. This review also shows that there are several concerns in clinical practice for EGJ adenocarcinoma. In Hong Kong and Malaysia, many EGJ adenocarcinomas have been detected at a stage not amenable to endoscopic resection. In Japan, histological curability criteria for endoscopic resection cases have not been established. We suggest that an international collaborative study using the same definition of EGJ adenocarcinoma may be helpful not only for clarifying the characteristics of these cancers but also for improving the clinical outcome of these patients.
Liquid-based cytology (LBC) for cervical cancer screening is now more common than the conventional smears, which when digitised from glass slides into whole-slide images (WSIs), opens up the possibility of artificial intelligence (AI)-based automated image analysis. Since conventional screening processes by cytoscreeners and cytopathologists using microscopes is limited in terms of human resources, it is important to develop new computational techniques that can automatically and rapidly diagnose a large amount of specimens without delay, which would be of great benefit for clinical laboratories and hospitals. The goal of this study was to investigate the use of a deep learning model for the classification of WSIs of LBC specimens into neoplastic and non-neoplastic. To do so, we used a dataset of 1605 cervical WSIs. We evaluated the model on three test sets with a combined total of 1468 WSIs, achieving ROC AUCs for WSI diagnosis in the range of 0.89–0.96, demonstrating the promising potential use of such models for aiding screening processes.
Various techniques including cold snare polypectomy and endoscopic mucosal resection are used for the removal of small colorectal polyps. Specimens of resected polyps are prepared in pathology laboratories and analyzed to make a pathological diagnosis. However, reports on how different resection methods influence the pathological diagnosis are limited. This article discusses the problems associated with the failure of polyp retrieval and fragmentation of small specimens during collection and the effects of certain parameters on the pathological diagnosis, particularly with regard to surgical margins. In the future, although pathologists are expected to encounter problems as a result of minor findings that are not clinically problematic, relatively rare cases such as submucosal invasion by a small carcinoma should not be overlooked.
Rationale:Gastric adenocarcinoma of fundic gland mucosa type (GA-FGM) is a rare tumor composed of atypical cells with differentiation toward the fundic gland as well as the foveolar epithelium. Including our case, only 9 cases of GA-FGMs were reported from 2010 to 2016.Concerns of the patient:An 87-year-old man was referred to our institution for endoscopic resection of a gastric lesion. The tumor was classified as type 0-I + IIa according to the Paris classification. Magnifying endoscopy with narrow band imaging (ME-NBI) revealed different structures of crypts and vessels among the components, illustrating the collision of 2 types of gastric cancer.Interventions:We performed endoscopic submucosal dissection and successfully removed the tumor en bloc.Outcomes:The histological findings differed markedly between the 0-I lesion and the 0-IIa lesion. The superficial part of the 0-I lesion consisted of a papillary structure, and the deeper part consisted of a tubular structure that showed inverted downward growth to the submucosal layer with the lamina muscularis mucosae. Immunohistochemically, the superficial part of the 0-I lesion was positive for MUC5AC, which had differentiated to foveolar epithelium. The deeper part was positive for pepsinogen-I and MUC6, which had differentiated to fundic gland. The 0-I lesion was diagnosed as gastric phenotype of adenocarcinoma differentiated to fundic gland mucosa with upward growth in the superficial part and downward growth in the deeper part. The 0-IIa lesion was composed of a tubular structure positive for MUC2, and it was diagnosed as an intestinal phenotype of well differentiated adenocarcinoma. The boundary was clear, and no transitional tissue was observed between the 0-I and 0-IIa lesions, suggesting that the 0-I + IIa lesion was a gastric collision tumor of GA-FGM and well differentiated adenocarcinoma.Lessons:We herein report the first case of inverted GA-FGM colliding with well differentiated adenocarcinoma. ME-NBI can be used to diagnose GA-FGM even if the lesion collides with other types of adenocarcinoma.
The pathological differential diagnosis between breast ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC) is of pivotal importance for determining optimum cancer treatment(s) and clinical outcomes. Since conventional diagnosis by pathologists using microscopes is limited in terms of human resources, it is necessary to develop new techniques that can rapidly and accurately diagnose large numbers of histopathological specimens. Computational pathology tools which can assist pathologists in detecting and classifying DCIS and IDC from whole slide images (WSIs) would be of great benefit for routine pathological diagnosis. In this paper, we trained deep learning models capable of classifying biopsy and surgical histopathological WSIs into DCIS, IDC, and benign. We evaluated the models on two independent test sets (n=1,382, n=548), achieving ROC areas under the curves (AUCs) up to 0.960 and 0.977 for DCIS and IDC, respectively.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.