Detection of Intestinal Protozoa in Trichrome-Stained Stool Specimens by Use of a Deep Convolutional Neural Network.

Intestinal protozoa are responsible for relatively few infections in the developed world, but the testing volume is disproportionately high. Manual light microscopy of stool remains the gold standard but can be insensitive, time-consuming, and difficult to maintain competency. Artificial intelligence and digital slide scanning show promise for revolutionizing the clinical parasitology laboratory by augmenting the detection of parasites and slide interpretation using a convolutional neural network (CNN) model. The goal of this study was to develop a sensitive model that could screen out negative trichrome slides, while flagging potential parasites for manual confirmation. Conventional protozoa were trained as "classes" in a deep CNN. Between 1,394 and 23,566 exemplars per class were used for training, based on specimen availability, from a minimum of 10 unique slides per class. Scanning was performed using a 40× dry lens objective automated slide scanner. Data labeling was performed using a proprietary Web interface. Clinical validation of the model was performed using 10 unique positive slides per class and 125 negative slides. Accuracy was calculated as slide-level agreement (e.g., parasite present or absent) with microscopy. Positive agreement was 98.88% (95% confidence interval [CI], 93.76% to 99.98%), and negative agreement was 98.11% (95% CI, 93.35% to 99.77%). The model showed excellent reproducibility using slides containing multiple classes, a single class, or no parasites. The limit of detection of the model and scanner using serially diluted stool was 5-fold more sensitive than manual examinations by multiple parasitologists using 4 unique slide sets. Digital slide scanning and a CNN model are robust tools for augmenting the conventional detection of intestinal protozoa.

X-Ray of Excised Cancerous Breast Tissue Does Not Affect Clinical Biomarker Expression.

CONTEXT: College of American Pathologists (CAP) and the American Society of Clinical Oncology have emphasized the need to reduce preanalytic variables for evaluating predictive biomarker expression in breast cancer. Postoperative x-ray of excised breast tissue is commonplace, yet is a variable that has not been investigated previously. We asked whether such radiation affects expression of relevant biomarkers. DESIGN: A previous study found that human breast cancers grown in mice demonstrate the same immunohistochemical and molecular profiles as the original tumors. Thirteen patient-derived xenografts were harvested fresh and divided for specimen radiography and a matched nonirradiated control, while following CAP/ASCO guidelines for cold ischemia time and fixation. Samples were processed in a tissue microarray for immunohistochemistry. Estrogen receptor (ER), progesterone receptor (PR), p53, and Ki67 staining was evaluated using an optimized scoring algorithm performed on digitally scanned slides. Samples were also scored manually by a blinded pathologist using the H-score method, and HER2 by the CAP/ASCO 2013 protocol. Histologic scores were compared by analysis of variance. RESULTS: There was no significant difference in quantity or intensity of staining between irradiated and nonirradiated samples for estrogen receptor (P=0.28), p53 (P=0.96), and Ki67 (P=0.94). A small but statistically significant difference was observed for PR (P=0.0058). HER2 staining was similarly unchanged in the 1 tumor exhibiting 3+ staining. CONCLUSIONS: Our study demonstrates that x-ray of breast carcinomas does not significantly affect the expression of predictive biomarkers, with the exception of PR for unclear reasons. It also highlights the utility of the patient-derived xenograft model for biomarker studies.

MYC Drives Progression of Small Cell Lung Cancer to a Variant Neuroendocrine Subtype with Vulnerability to Aurora Kinase Inhibition.

Loss of the tumor suppressors RB1 and TP53 and MYC amplification are frequent oncogenic events in small cell lung cancer (SCLC). We show that Myc expression cooperates with Rb1 and Trp53 loss in the mouse lung to promote aggressive, highly metastatic tumors, that are initially sensitive to chemotherapy followed by relapse, similar to human SCLC. Importantly, MYC drives a neuroendocrine-low "variant" subset of SCLC with high NEUROD1 expression corresponding to transcriptional profiles of human SCLC. Targeted drug screening reveals that SCLC with high MYC expression is vulnerable to Aurora kinase inhibition, which, combined with chemotherapy, strongly suppresses tumor progression and increases survival. These data identify molecular features for patient stratification and uncover a potential targeted treatment approach for MYC-driven SCLC.

Calcium regulation of endothelin-1 synthesis in rat inner medullary collecting duct.

Collecting duct-derived endothelin-1 (ET-1) reduces blood pressure and inhibits Na and water reabsorption. Collecting duct ET-1 production is increased by volume expansion; however, the mechanism by which this occurs is unknown. We hypothesized that intracellular calcium, which is likely to be increased by volume expansion, regulates collecting duct ET-1 synthesis. Rat inner medullary collecting ducts (IMCD) were studied in primary culture. ET-1 release was decreased by 50-70% after chelation of intracellular calcium (BAPTA) or inhibition of CaM (W7) or CaMK (KN-93). These agents reduced ET-1 mRNA to a similar degree. CaM inhibition did not affect ET-1 mRNA stability. Transfection of IMCD with rat ET-1 promoter-luciferase constructs revealed maximal activity within 1.7 kb 5' to the transcription start site; 5, 20, 35, and 90% of this activity were in the 0.08-, 0.37-, 1.0-, and 3.0-kb promoter regions, respectively. W7 markedly inhibited activity of the 3.0-kb but not 0.37- or 1.0-kb promoter regions. In contrast, W7 did not affect ET-1 release by rat aortic endothelial cells. Furthermore, transfected endothelial cells had maximal activity in the 0.37-kb region (as compared with the 1.7- and 3.0-kb regions), whereas W-7 had no effect on the activity of any of these promoter regions. In summary, IMCD ET-1 synthesis is regulated by calcium/CaM/CaMK-dependent pathways. The calcium/CaM-sensitive pathway is active in IMCD, but not endothelial cells. This suggests that IMCD-specific enhancer elements exist within the ET-1 promoter that confer unique calcium responsiveness.