Min KJ, et al. J Gynecol Oncol. 2015 Jul;26(3):232-239. https://doi.org/10.3802/jgo.2015.26.3.232

Background: Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS) has been demonstrated to be useful for molecular profiling of common solid tumors. Using recently developed MALDI matrices for lipid profiling, we evaluated whether direct tissue MALDI MS analysis on proteins and lipids may classify human breast cancer samples according to the intrinsic subtype. Methods: Thirty-four pairs of frozen, resected breast cancer and adjacent normal tissue samples were analyzed using histology-directed, MALDI MS analysis. Sinapinic acid and 2,5-dihydroxybenzoic acid/α-cyano-4-hydroxycinnamic acid were manually deposited on areas of each tissue section enriched in epithelial cells to identify lipid profiles, and mass spectra were acquired using a MALDI-time of flight instrument. Results: Protein and lipid profiles distinguish cancer from adjacent normal tissue samples with the median prediction accuracy of 94.1%. Luminal, HER2+, and triple-negative tumors demonstrated different protein and lipid profiles, as evidenced by permutation P values less than 0.01 for 0.632+ bootstrap cross-validated misclassification rates with all classifiers tested. Discriminatory proteins and lipids were useful for classifying tumors according to the intrinsic subtype with median prediction accuracies of 80.0-81.3% in random test sets. Conclusions: Protein and lipid profiles accurately distinguish tumor from adjacent normal tissue and classify breast cancers according to the intrinsic subtype.

Recent findings in molecular pathology suggest that genetic translocation and/oroverexpression of oncoproteins is important in salivary gland tumorigenesis and diagnosis. Weinvestigated PLAG1, SOX10, and Myb protein expression in various salivary gland neoplasm tissues. A total of 113 cases of surgically resected salivary gland neoplasms at the NationalCancer Center from January 2007 to March 2017 were identified. Immunohistochemical stainingof PLAG1, SOX10, and Myb in tissue samples was performed using tissue microarrays. Among the 113 cases, 82 (72.6%) were benign and 31 (27.4%) were malignant. PLAG1 showednuclear staining and normal parotid gland was not stained. Among 48 cases of pleomorphicadenoma, 29 (60.4%) were positive for PLAG1. All other benign and malignant salivary glandneoplasms were PLAG1-negative. SOX10 showed nuclear staining. In normal salivary gland tissuesSOX10 was expressed in cells of acinus and intercalated ducts. In benign tumors, SOX10 expressionwas observed in all pleomorphic adenoma (48/48), and basal cell adenoma (3/3), but not inother benign tumors. SOX10 positivity was observed in nine of 31 (29.0%) malignant tumors.Myb showed nuclear staining but was not detected in normal parotid glands. Four of 31 (12.9%)malignant tumors showed Myb positivity: three adenoid cystic carcinomas (AdCC) and onemyoepithelial carcinoma with focal AdCC-like histology. PLAG1 expression is specificto pleomorphic adenoma. SOX10 expression is helpful to rule out excretory duct origin tumor,but its diagnostic value is relatively low. Myb is useful for diagnosing AdCC when histology isunclear in the surgical specimen.

Berek JS, et al. J Gynecol Oncol. 2018 Aug;30:e40. https://doi.org/10.3802/jgo.2019.30.e40

Background: Artificial intelligence (AI) is increasingly being applied in pathology and cytology, showing promising results. We collected a large dataset of whole slide images (WSIs) of thyroid fine-needle aspiration cytology (FNA), incorporating z-stacking, from institutions across the nation to develop an AI model. Methods: We conducted a multicenter retrospective diagnostic accuracy study using thyroid FNA dataset from the Open AI Dataset Project that consists of digitalized images samples collected from 3 university hospitals and 215 Korean institutions through extensive quality check during the case selection, scanning, labeling, and reviewing process. Multiple z-layer images were captured using three different scanners and image patches were extracted from WSIs and resized after focus fusion and color normalization. We pretested six AI models, determining Inception ResNet v2 as the best model using a subset of dataset, and subsequently tested the final model with total datasets. Additionally, we compared the performance of AI and cytopathologists using randomly selected 1031 image patches and reevaluated the cytopathologists’ performance after reference to AI results. Results: A total of 10,332 image patches from 306 thyroid FNAs, comprising 78 malignant (papillary thyroid carcinoma) and 228 benign from 86 institutions were used for the AI training. Inception ResNet v2 achieved highest accuracy of 99.7%, 97.7%, and 94.9% for training, validation, and test dataset, respectively (sensitivity 99.9%, 99.6%, and 100% and specificity 99.6%, 96.4%, and 90.4% for training, validation, and test dataset, respectively). In the comparison between AI and human, AI model showed higher accuracy and specificity than the average expert cytopathologists beyond the two-standard deviation (accuracy 99.71% [95% confidence interval (CI), 99.38–100.00%] vs. 88.91% [95% CI, 86.99–90.83%], sensitivity 99.81% [95% CI, 99.54–100.00%] vs. 87.26% [95% CI, 85.22–89.30%], and specificity 99.61% [95% CI, 99.23–99.99%] vs. 90.58% [95% CI, 88.80–92.36%]). Moreover, after referring to the AI results, the performance of all the experts (accuracy 96%, 95%, and 96%, respectively) and the diagnostic agreement (from 0.64 to 0.84) increased. Conclusions: These results suggest that the application of AI technology to thyroid FNA cytology may improve the diagnostic accuracy as well as intra- and inter-observer variability among pathologists. Further confirmatory research is needed.

MicroRNA (miRNA) expression is known to be deregulated in cervical carcinomas. However, no data is available about the miRNA expression pattern for the minimal deviation adenocarcinoma (MDA) of uterine cervix. We sought to detect deregulated miRNAs in MDA in an attempt to find the most dependable miRNA or their combinations to understand their tumorigenesis pathway and to identify diagnostic or prognostic biomarkers. We also investigated the association between those miRNAs and their target genes, especially Notch1 and Notch2. We evaluated miRNA expression profiles via miRNA microarray and validated them using.real-time PCR assays with 24 formalin-fixed, paraffin-embedded tissue blocks of MDA and 11 normal proliferative endocervical tissues as control. Expression for Notch1 and 2 was assessed by immunohistochemistry. MiRNA-135a-3p, 192-5p, 194-5p, and 494 were up-regulated, whereas miR-34b-5p, 204-5p, 299-5p, 424-5p, and 136-3p were down-regulated in MDA compared with normal proliferative endocervical tissues (all P <0.05). Considering the second-order Akaike Information Criterion consisting of likelihood ratio and number of parameters, miR-34b-5p showed the best discrimination power among the nine candidate miRNAs. A combined panel of miR-34b-5p and 194-5p was the best fit model to discriminate between MDA and control, revealing 100% sensitivity and specificity. Notch1 and Notch2, respective target genes of miR-34b-5p and miR-204-5p, were more frequently expressed in MDA than in control (63% vs. 18%; 52% vs. 18%, respectively, P <0.05). MiR-34b-5p expression level was higher in Notch1-negative samples compared with Notch1-positive ones (P <0.05). Down-regulated miR-494 was associated with poor patient survival (P = 0.036). MDA showed distinctive expression profiles of miRNAs, Notch1, and Notch2 from normal proliferative endocervical tissues. In particular, miR-34b-5p and 194-5p might be used as diagnostic biomarkers and miR-494 as a prognostic predictor for MDA. The miR-34b-5p/Notch1 pathway as well as Notch2 might be important oncogenic contributors to MDA.