Elevated Expression of the Immune Checkpoint Ligand CD276 (B7-H3) in Urothelial Carcinoma Cell Lines Correlates Negatively with the Cell Proliferation.

The cell surface molecule CD276 (B7-H3) is an immune checkpoint antigen. The elevated expression of CD276 on tumors contributes to the suppression of anti-tumor T-cell responses and correlates with poor prognosis. METHODS: The expression of CD276 was explored in vitro on eight urothelial carcinoma cell lines (UM-UC) in comparison to eight normal urothelial cells (NUCs) by RT-qPCR, Western blotting, and flow cytometry. Cell proliferation was enumerated over consecutive passages. The expression of cancer stem cell markers CD24 and CD44, cytokeratins, and vimentin was investigated by immunofluorescence. The expression of CD276 in bladder tumor samples and metastases was explored by immunohistochemistry. RESULTS: Expression of CD276 on cell surfaces was elevated on UM-UCs when compared to NUCs. In UM-UCs, CD276 transcripts correlated moderately positive with CD276 protein expression (ρ = 0.660) and strongly positive with CD276 surface-expression (ρ = 0.810). CD276 mRNA expression (ρ = -0.475) and CD276 protein expression (ρ = -0.417) had a significant negative correlation with proliferation, while a significant correlation between proliferation and cell surface expression was not observed in UM-UCs. CONCLUSION: The expression of CD276 on UM-UC bladder tumor cell surfaces is elevated. Slow proliferating UM-UC cells express more CD276 mRNA and protein than fast proliferating cells. In patients, slow proliferating CD276high tumor (stem) cells may evade immune surveillance. However, cancer therapy targeting CD276 may be effective in the treatment of slow proliferating tumor cells.

Oncological validation of bone turnover markers c-terminal telopeptide of type I collagen (1CTP) and peptides n-terminal propeptide of type I procollagen (P1NP) in patients with prostate cancer and bone metastases.

BACKGROUND: Bone formation markers c-terminal telopeptide of type I collagen (1CTP) and peptides n-terminal propeptide of type I procollagen (P1NP) were reported to be increased in patients with prostate cancer (PC) and bone metastases. The objective of the presented study was to investigate the utility of serum 1CTP and P1NP values in the diagnosis of bone metastases and in predicting oncological outcome in patients with PC. METHODS: In total, serum samples of 186 patients were included retrospectively including 53 (28.50%) benign prostatic hyperplasia (BPH) patients and 133 (71.50%) PC-patients. The group of patients with PC consisted of 58 patients with non-metastatic PC (cM0) (43.61%) and 70 (52.63%) patients with bone metastases (cM1b). Serum 1CTP and P1NP were measured by radioimmunoassay (RIA). Results were compared to clinical variables including oncologic follow-up data by univariate and multivariate analyses. RESULTS: Median 1CTP concentrations were significantly higher in patients with PC compared to the BPH group [5.08 (range, 1.73-158.00) vs. 4.00 (range, 2.18-34.19) µg/L, P=0.019]. However, no significant difference of P1NP levels could be shown for these groups. With median values of 6.04 (1.73-158.00) and 3.91 µg/L (2.04-34.51) for 1CTP and 48.60 (9.12-1,074.37) and 33.90 (8.72-149.30) for P1NP both markers were altered in cM1b patients compared to cM0 patients (P=0.001 each). Furthermore, cancer-specific survival (CSS) and overall survival (OS) were significantly shorter in cM1b patients with higher 1CTP concentrations (P=0.037 and P=0.019, respectively), whereas no associations of P1NP and outcomes were observed. CONCLUSIONS: The present study confirms that increased levels of 1CTP and P1NP concentrations are associated with presence of metastatic disease in the bone. Moreover, these markers are able to predict clinical course in PC patients with bone metastases. The potential use of these markers for treatment selection in advanced PC remains to be determined.

Deep Neural Network Driven Speech Classification for Relevance Detection in Automatic Medical Documentation.

The automation of medical documentation is a highly desirable process, especially as it could avert significant temporal and monetary expenses in healthcare. With the help of complex modelling and high computational capability, Automatic Speech Recognition (ASR) and deep learning have made several promising attempts to this end. However, a factor that significantly determines the efficiency of these systems is the volume of speech that is processed in each medical examination. In the course of this study, we found that over half of the speech, recorded during follow-up examinations of patients treated with Intra-Vitreal Injections, was not relevant for medical documentation. In this paper, we evaluate the application of Convolutional and Long Short-Term Memory (LSTM) neural networks for the development of a speech classification module aimed at identifying speech relevant for medical report generation. In this regard, various topology parameters are tested and the effect of the model performance on different speaker attributes is analyzed. The results indicate that Convolutional Neural Networks (CNNs) are more successful than LSTM networks, and achieve a validation accuracy of 92.41%. Furthermore, on evaluation of the robustness of the model to gender, accent and unknown speakers, the neural network generalized satisfactorily.