Invasive stratified mucin-producing carcinoma (ISMC) is a recently described entity of human papillomavirus (HPV)-associated endocervical adenocarcinoma with phenotypic plasticity and aggressive clinical behavior. To identify the cell of origin of ISMC, we investigated the immunohistochemical expression of cervical epithelial cell markers (CK7, PAX8, CK5/6, p63, and CK17), stemness markers (ALDH1 and Nanog), and epithelial-mesenchymal transition (EMT) markers (Snail, Twist, and E-cadherin) in 10 pure and mixed type ISMCs with at least 10% of ISMC component in the entire tumor, seven usual type endocervical adenocarcinomas (UEAs), and seven squamous cell carcinomas (SCCs). In addition, targeted sequencing was performed in 10 ISMCs. ISMC was significantly associated with larger tumor size (p = 0.011), more frequent lymphovascular invasion and lymph node metastasis (p  less then  0.001), higher FIGO stage (p = 0.022), and a tendency for worse clinical outcomes (p = 0.056) compared to other HPV-associated subtypes. ISMC showed negative or borderline positivity for PAX8, CK5/6, and p63, which were distinct from UEA and SCC (p  less then  0.01). Compared to UEA and SCC, ISMC showed higher expression for ALDH1 (p = 0.119 for UEA and p = 0.009 for SCC), Snail (p = 0.036), and Twist (p = 0.119), and tended to show decreased E-cadherin expression (p = 0.083). In next-generation sequencing analysis, ISMC exhibited frequent STK11, MET, FANCA, and PALB2 mutations compared to conventional cervical carcinomas, and genes related to EMT and stemness were frequently altered. EMT-prone and stemness characteristics and peripheral expression of reserve cell and EMT markers of ISMC suggest its cervical reserve cell origin. We recommend PAX8, CK5/6, and p63 as diagnostic triple biomarkers for ISMC. These findings highlight the distinct biological basis of ISMC.Renal oncocytoma is the most common benign epithelial renal neoplasm. Several adverse features that would typically increase the stage of renal cell carcinomas are not uncommon in renal oncocytoma, including perinephric, sinus fat, or renal vein invasion. Herein, we report the largest single institutional series of renal oncocytoma with adverse pathologic features. The cohort comprised 50 patients, 38 were men (76%) and 12 were women (24%), with a mean age of 68 years (range, 50-87 years). All cases were diagnosed on nephrectomy specimens. No laterality predilection was noted. The tumors ranged in size from 1.5-15.7 cm (mean, 5.3 cm). Adverse pathologic features included perinephric fat invasion (n = 25; 50%), renal sinus fat invasion (n = 9; 18%), and renal vein invasion (n = 5; 10%). More than one adverse feature was seen in 11 tumors (22%). All tumors showed diffuse reactions to KIT (n = 40; 100%) and cyclin D1 (n = 27; 100%). Keratin 7 highlighted rare ( less then 5%) scattered cells, as well as entrapped renal tubules (n = 21; 100%). Reaction to DOG1 was patchy in three tumors (n = 27; 11%) while reactions to vimentin (n = 31) and Hale colloidal iron special stain (n = 30) were negative. On follow-up, no tumor recurrence or metastasis was observed over a follow-up range of 1-144 months (mean, 54 months; median, 60 months). Our data suggest that adverse pathologic features in renal oncocytoma do not alter their benign course.Testicular Leydig cell tumor (LCT), the most common sex-cord stromal tumor in men, represents a small fraction of all testicular tumors (~1 to 3%). Although most testicular LCTs are indolent and cured by radical orchiectomy, 5-10% have aggressive biology and metastatic potential. In primary LCTs, large size, cytologic atypia, necrosis, increased mitotic activity, and vascular invasion have been associated with clinically aggressive tumors. From a molecular perspective, the characteristics of aggressive LCTs and the differences between aggressive and nonaggressive LCTs remain largely unexplored. This study compares the genomic landscape of aggressive and nonaggressive testicular LCTs. Twenty-six cases were analyzed using next-generation DNA sequencing (NGS) and immunohistochemistry. Cases were classified as aggressive LCT if they met published criteria for malignancy in primary (i.e., testicular) tumors or if they had pathology-proven metastatic disease; otherwise, cases were considered nonaggressive. This mul. Nuclear translocation of β-catenin and Wnt pathway activation appear to be early driver events that provide an environment conducive for progression to aggressive biology in a subset of LCTs.Detection of nodal micrometastasis (tumor size 0.2-2.0 mm) is challenging for pathologists due to the small size of metastatic foci. Since lymph nodes with micrometastasis are counted as positive nodes, detecting micrometastasis is crucial for accurate pathologic staging of colorectal cancer. Previously, deep learning algorithms developed with manually annotated images performed well in identifying micrometastasis of breast cancer in sentinel lymph nodes. However, the process of manual annotation is labor intensive and time consuming. Multiple instance learning was later used to identify metastatic breast cancer without manual annotation, but its performance appears worse in detecting micrometastasis. Here, we developed a deep learning model using whole-slide images of regional lymph nodes of colorectal cancer with only a slide-level label (either a positive or negative slide). Capmatinib inhibitor The training, validation, and testing sets included 1963, 219, and 1000 slides, respectively. A supercomputer TAIWANIA 2 was used to train a deep learning model to identify metastasis. At slide level, our algorithm performed well in identifying both macrometastasis (tumor size > 2.0 mm) and micrometastasis with an area under the receiver operating characteristics curve (AUC) of 0.9993 and 0.9956, respectively. Since most of our slides had more than one lymph node, we then tested the performance of our algorithm on 538 single-lymph node images randomly cropped from the testing set. At single-lymph node level, our algorithm maintained good performance in identifying macrometastasis and micrometastasis with an AUC of 0.9944 and 0.9476, respectively. Visualization using class activation mapping confirmed that our model identified nodal metastasis based on areas of tumor cells. Our results demonstrate for the first time that micrometastasis could be detected by deep learning on whole-slide images without manual annotation.