Improved Bladder Tumor RNA Isolation from Archived Tissues Using Methylene Blue for Normalization, Multiplex RNA Hybridization, Sequencing and Subtyping.

Methylene blue (MB) is a dye used for histology with clinical importance and intercalates into nucleic acids. After MB staining of formalin fixed paraffin embedded (FFPE) muscle invasive bladder cancer (MIBC) and normal urothelium, specific regions could be microdissected. It is not known if MB influences RNA used for gene expression studies. Therefore, we analyzed MIBC using five different RNA isolation methods comparing patient matched FFPE and fresh frozen (FF) tissues pre-stained with or without MB. We demonstrate a positive impact of MB on RNA integrity with FF tissues using real time PCR with no interference of its chemical properties. FFPE tissues showed no improvement of RNA integrity, which we propose is due to formalin induced nucleotide crosslinks. Using direct multiplex RNA hybridization the best genes for normalization of MIBC and control tissues were identified from 34 reference genes. In addition, 5SrRNA and 5.8SrRNA were distinctive reference genes detecting <200 bp fragments important for mRNA analyses. Using these normalized RNAs from MB stained MIBC and applying multiplex RNA hybridization and mRNA sequencing, a minimal gene expression panel precisely identified luminal and basal MIBC tumor subtypes, important for diagnosis, prognosis and chemotherapy response.

Spatial expression of claudin 18.2 in matched primaries and metastases of tubo-ovarian carcinoma of all subtypes.

Physiologically, claudin 18 splice variant 2 (CLDN18.2) expression is restricted to the gastric epithelium, but its expression has been detected in solid cancers. Zolbetuximab, a chimeric IgG1 antibody targeting CLDN18.2, has demonstrated promising effects in patients suffering from CLDN18.2-positive, HER2-negative locally advanced gastric cancer and is currently being studied further. To date, little is known about CLDN18.2 expression in other histological subtypes of tubo-ovarian carcinoma (TOC) and their matching metastases.Using a cohort of all histological TOC subtypes, we investigated the immunohistochemical (IHC) CLDN18.2 expression in both TOCs (n = 536), their matching metastatic tissue (n = 385) and in 93 metastases without primary. Tissue microarrays comprised both the tumor center and periphery. IHC positivity was defined as biomarker expression of ≥ 75% in tumor cells with moderate-to-strong membranous staining.Overall CLDN18.2 positivity was 4.1% (21/515) in the TOC centers and 3.6% (18/498) in their peripheries. In primaries of mucinous tubo-ovarian carcinoma (MTOC), CLDN18.2 positivity rates were 45% (18/40) and 36.6% (15/41), respectively. Positivity rates for the corresponding metastases were 33% (4/12, center) and 27% (3/11, periphery). The expression was relatively homogenous throughout all tumor sites. With no expression in 99.5% of nonmucinous tumors, CLDN18.2 positivity was almost exclusively seen in the mucinous subtype.In tubo-ovarian carcinoma, CLDN18.2 expression was, with rare exceptions, restricted to the mucinous subtype. Among them, 33% of metastasized MTOCs presented with CLDN18.2 positivity. Hence, CLDN18.2 might display a promising target for personalized therapy in patients with advanced MTOC.

Tumor-Stroma Ratio in Colorectal Cancer-Comparison between Human Estimation and Automated Assessment.

The tumor-stroma ratio (TSR) has been repeatedly shown to be a prognostic factor for survival prediction of different cancer types. However, an objective and reliable determination of the tumor-stroma ratio remains challenging. We present an easily adaptable deep learning model for accurately segmenting tumor regions in hematoxylin and eosin (H&E)-stained whole slide images (WSIs) of colon cancer patients into five distinct classes (tumor, stroma, necrosis, mucus, and background). The tumor-stroma ratio can be determined in the presence of necrotic or mucinous areas. We employ a few-shot model, eventually aiming for the easy adaptability of our approach to related segmentation tasks or other primaries, and compare the results to a well-established state-of-the art approach (U-Net). Both models achieve similar results with an overall accuracy of 86.5% and 86.7%, respectively, indicating that the adaptability does not lead to a significant decrease in accuracy. Moreover, we comprehensively compare with TSR estimates of human observers and examine in detail discrepancies and inter-rater reliability. Adding a second survey for segmentation quality on top of a first survey for TSR estimation, we found that TSR estimations of human observers are not as reliable a ground truth as previously thought.