The pathology of biliary carcinomas is diverse with different gross and histological features in tumors arising in the different segments of the biliary system. Various epidemiological risk factors, varied genetic makeup, and tissue microenvironment are contributory factors. As biliary tumors have been shown to be a part of the Lynch syndrome tumor spectrum, it is plausible to speculate that DNA mismatch repair (MMR) deficiency plays a role in biliary tumors. Literature data suggest that DNA MMR deficiency indeed occurs in these tumors, albeit infrequently with the reported frequencies (weighted for sample size) of high level microsatellite instability (MSI) being 5% each for gallbladder carcinoma and carcinoma of extra-hepatic bile ducts, and 10% each for intrahepatic cholangiocarcinoma and ampullary carcinoma. Importantly, the presence of MMR deficiency in these tumors has been shown to have different implications with regard to its association with Lynch syndrome, tumor histological features, and other clinical characteristics, when compared with non-biliary tumors or among the biliary tumors from the different segments of the biliary system. Ongoing and future efforts that utilize large scale sequencing techniques and aim at detecting actionable molecular targets should emphasize a multidisciplinary approach that integrates genomic discoveries with not only functional studies but also studies of tumor pathology and the tumor's clinical and biological behavior.
Background
Mesonephric adenocarcinomas are rare neoplasms which most commonly arise in the lateral cervix and vagina. Tumors with similar morphologic, immunophenotypic, and molecular characteristics recently have been described in the uterine corpus and ovary. Herein, the authors sought to characterize the cytomorphologic features of adenocarcinomas exhibiting mesonephric‐like differentiation arising in the upper gynecologic tract.
Methods
Institutional databases were queried retrospectively for tumors of the upper gynecologic tract described as a “tumor of Wolffian origin” or “with mesonephric features” between 2007 and 2017. All available cytologic material was reviewed. Cytomorphologic characteristics were evaluated by 3 pathologists.
Results
The current study cohort consisted of 8 cases taken from 7 patients. Primary sites included the ovary (3 cases); endometrium (4 cases); and pelvis, not otherwise specified (1 case). All cases demonstrated tight 3‐dimensional clusters of overlapping cells. Additional architectural features included tubular (5 of 8 cases; 63%) and papillary (3 of 8 cases; 38%) formations. Cells were small with scant (7 of 8 cases; 88%) to moderate (1 of 8 cases; 12%) cytoplasm. Three of the 8 cases (38%) demonstrated extracellular hyaline globules. Nuclei were uniform in size (6 of 8 cases; 75%) or showed mild anisonucleosis (2 of 8 cases; 25%). Nuclear grooves and indentations were observed in all cases. Mitoses (5 of 8 cases; 63%) and apoptotic bodies (4 of 8 cases; 50%), when present, were rare. No necrosis was noted.
Conclusions
Adenocarcinomas exhibiting mesonephric‐like differentiation show a monotonous population of small cells with scant to moderate cytoplasm and abundant nuclear grooves arranged in tight, overlapping, 3‐dimensional clusters. Occasionally, papillary or tubular architecture, as well as extracellular hyaline globules, may be seen. These features should prompt further testing (eg, immunohistochemistry) to confirm the diagnosis and to exclude potential mimics.
In the field of computational pathology, the use of decision support systems powered by state-of-the-art deep learning solutions has been hampered by the lack of large labeled datasets. Until recently, studies relied on datasets in the order of few hundreds of slides which are not enough to train a model that can work at scale in the clinic. Here, we have gathered a dataset consisting of 12,160 slides, two orders of magnitude larger than previous datasets in pathology and equivalent to 25 times the pixel count of the entire ImageNet dataset. Given the size of our dataset it is possible for us to train a deep learning model under the Multiple Instance Learning (MIL) assumption where only the overall slide diagnosis is necessary for training, avoiding all the expensive pixelwise annotations that are usually part of supervised learning approaches. We test our framework on a complex task, that of prostate cancer diagnosis on needle biopsies. We performed a thorough evaluation of the performance of our MIL pipeline under several conditions achieving an AUC of 0.98 on a held-out test set of 1,824 slides. These results open the way for training accurate diagnosis prediction models at scale, laying the foundation for decision support system deployment in the clinic.
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.