Search

Summary: Non-small cell lung cancer (NSCLC) is one of the most common malignant tumors. In this study, we develop a clinically useful computational pathology platform for NSCLC that can be a foundation for multiple downstream applications and provide immediate value for patient care optimization and individualization. We train the primary multi-class tissue segmentation algorithm on a substantial, high-quality, manually annotated dataset of whole-slide images with lung adenocarcinoma and squamous cell carcinomas. We investigate two downstream applications. NSCLC subtyping algorithm is trained and validated using a large, multi-institutional (n = 6), multi-scanner (n = 5), international cohort of NSCLC cases (slides/patients 4,097/1,527). Moreover, we develop four AI-derived, fully explainable, quantitative, prognostic parameters (based on tertiary lymphoid structure and necrosis assessment) and validate them for different clinical endpoints. The computational platform enables the high-precision, quantitative analysis of H&E-stained slides. The developed prognostic parameters facilitate robust and independent risk stratification of patients with NSCLC.

Abstract Objective Although targeted approaches have become available in second‐ and third‐line settings, platinum‐based chemotherapy remains the standard first‐line treatment for advanced muscle‐invasive bladder cancer (MIBC). Therefore, the prediction of platinum resistance is of utmost clinical importance. Methods In this study, we established a routine compatible method for the molecular classification of MIBC samples according to various classification systems and applied this method to evaluate the impact of subtypes on survival after adjuvant chemotherapy. This retrospective study included 191 patients with advanced MIBC (pT≥3 or pN+) who underwent radical cystectomy, with or without adjuvant chemotherapy. A 48‐gene panel and classifier rule set were established to determine molecular subtypes according to TCGA, MDA, LundTax, and Consensus classifications. Additionally, 12 single platinum‐predictive candidate genes were assessed. The results were correlated with patients' clinicopathological and follow‐up data and were validated using independent data sets. Results Our final evaluation of 159 patients demonstrated better survival in the luminal groups for those who received chemotherapy compared with those who did not. In contrast, no such differences were observed in basal subtypes. The use of chemotherapy was associated with better survival in patients with high APOBEC3G expression (p

5q-related Spinal muscular atrophy (SMA) is a hereditary multi-systemic disorder leading to progressive muscle atrophy and weakness caused by the degeneration of spinal motor neurons (MNs) in the ventral horn of the spinal cord. Three SMN-enhancing drugs for SMA treatment are available. However, even if these drugs are highly effective when administrated early, several patients do not benefit sufficiently or remain non-responders, e.g., adults suffering from late-onset SMA and starting their therapy at advanced disease stages characterized by long-standing irreversible loss of MNs. Therefore, it is important to identify additional molecular targets to expand therapeutic strategies for SMA treatment and establish prognostic biomarkers related to the treatment response. Using high-throughput nCounter NanoString technology, we analyzed serum samples of late-onset SMA type 2 and type 3 patients before and six months under nusinersen treatment. Four genes (AMIGO1, CA2, CCL5, TLR2) were significantly altered in their transcript counts in the serum of patients, where differential expression patterns were dependent on SMA subtype and treatment response, assessed with outcome scales. No changes in gene expression were observed six months after nusinersen treatment, compared to healthy controls. These alterations in the transcription of four genes in SMA patients qualified those genes as potential SMN-independent therapeutic targets to complement current SMN-enhancing therapies.

Malignant pleural effusion (MPE) is a common complication of lung adenocarcinoma (LADC) which is associated with a dismal prognosis. We investigated the prognostic role of PD-L1 and other immunomodulators expression in the immune compartment of MPE immune composition. MPE cytologic cell blocks of 83 LADC patients were analysed for the mRNA expression of 770 cancer-immune genes by the NanoString nCounter platform. The expression of relevant immune cell lineage markers was validated by immunohistochemistry (IHC) using quantitative pathology. The mRNA immune profiling identified four MPE patient clusters (C). C1/2 (adaptive+, hot) showed better overall survival (OS) than C3/4 (adaptive-, cold). Additionally, cold immunity profiles (adaptive-), C4 (innate+) were associated with worse OS than C3 (innate-). High PD-L1 expression was linked to the regulation of T cell activation and interferon signalling pathways. Genes of pattern recognition receptor and type I interferon signalling pathways were specifically upregulated in the long-survival (≥90 days) patient group. Moreover, immunomodulators were co-activated and highly expressed in hot adaptive immunity patient clusters, whereas CD274 (PD-L1), TNFRSF9 (4-1BB), VEGFA (VEGF-A) and CD276 (B7-H3) were upregulated in the groups referred as cold. The patient cluster, age and PD-L1 expression were independent prognosticators for LADC MPE patients (p-value < 0.05). Our study sheds light on the variances of immune contexture regarding different PD-L1 expression and survival conditions. It revealed four distinct prognostic patient clusters with specific immune cell components and immunomodulator expression profiles, which, collectively, is supportive for future therapeutic and prognosis for cancer management.