Search

Abstract In this paper we present a deep learning segmentation approach to classify and quantify the two most prevalent primary liver cancers – hepatocellular carcinoma and intrahepatic cholangiocarcinoma – from hematoxylin and eosin (H&E) stained whole slide images. While semantic segmentation of medical images typically requires costly pixel-level annotations by domain experts, there often exists additional information which is routinely obtained in clinical diagnostics but rarely utilized for model training. We propose to leverage such weak information from patient diagnoses by deriving complementary labels that indicate to which class a sample cannot belong to. To integrate these labels, we formulate a complementary loss for segmentation. Motivated by the medical application, we demonstrate for general segmentation tasks that including additional patches with solely weak complementary labels during model training can significantly improve the predictive performance and robustness of a model. On the task of diagnostic differentiation between hepatocellular carcinoma and intrahepatic cholangiocarcinoma, we achieve a balanced accuracy of 0.91 (CI 95%: 0.86-0.95) at case level for 165 hold-out patients. Furthermore, we also show that leveraging complementary labels improves the robustness of segmentation and increases performance at case level.

Abstract Purpose To assess the eligibility of patients with advanced or recurrent solid malignancies presented to a molecular tumor board (MTB) at a large precision oncology center for inclusion in trials with the endpoints objective response rate (ORR) or duration of response (DOR) based on Response Evaluation Criteria in Solid Tumors (RECIST version 1.1). Methods Prospective patients with available imaging at the time of presentation in the MTB were included. Imaging data was reviewed for objectifiable measurable disease (MD) according to RECIST v1.1. Additionally, we evaluated the patients with MD for representativeness of the identified measurable lesion(s) in relation to the overall tumor burden. Results 262 patients with different solid malignancies were included. 177 patients (68%) had MD and 85 (32%) had non-measurable disease (NMD) at the time point of MTB presentation in accordance with RECIST v1.1. MD was not representative of the overall tumor burden in eleven patients (6%). The main reasons for NMD were lesions with longest diameter shorter than 10 mm (22%) and non-measurable peritoneal carcinomatosis (18%). Colorectal cancer and malignant melanoma displayed the highest rates of MD (> 75%). In contrast, gastric cancer, head and neck malignancies, and ovarian carcinoma had the lowest rates of MD (

Abstract Head and Neck Squamous Cell Carcinoma (HNSCC) is a heterogeneous malignancy that remains a significant challenge in clinical management due to frequent treatment failures and pronounced therapy resistance. While metabolic dysregulation appears to be a critical factor in this scenario, comprehensive analyses of the metabolic HNSCC landscape and its impact on clinical outcomes are lacking. This study utilized transcriptomic data from four independent clinical cohorts to investigate metabolic heterogeneity in HNSCC and define metabolic pathway-based subtypes (MPS). In HPV-negative HNSCCs, MPS1 and MPS2 were identified, while MPS3 was enriched in HPV-positive cases. MPS classification was associated with clinical outcome post adjuvant radio(chemo)therapy, with MPS1 consistently exhibiting the highest risk of therapeutic failure. MPS1 was uniquely characterized by upregulation of glycan (particularly chondroitin/dermatan sulfate) metabolism genes. Immunohistochemistry and pilot mass spectrometry imaging analyses confirmed this at metabolite level. The histological context and single-cell RNA sequencing data identified the malignant cells as key contributors. Globally, MPS1 was distinguished by a unique transcriptomic landscape associated with increased disease aggressiveness, featuring motifs related to epithelial-mesenchymal transition, immune signaling, cancer stemness, tumor microenvironment assembly, and oncogenic signaling. This translated into a distinct histological appearance marked by extensive extracellular matrix remodeling, abundant spindle-shaped cancer-associated fibroblasts, and intimately intertwined populations of malignant and stromal cells. Proof-of-concept data from orthotopic xenotransplants replicated the MPS phenotypes on the histological and transcriptome levels. In summary, this study introduces a metabolic pathway-based classification of HNSCC, pinpointing glycan metabolism-enriched MPS1 as the most challenging subgroup that necessitates alternative therapeutic strategies.

Abstract Background Peroxisome proliferator-activated receptor gamma (PPARG) is a member of the nuclear receptor family. It is involved in the regulation of adipogenesis, lipid metabolism, insulin sensitivity, vascular homeostasis and inflammation. In addition, PPARG agonists, known as thiazolidinediones, are well established in the treatment of type 2 diabetes mellitus. PPARGs role in cancer is a matter of debate, as pro- and anti-tumour properties have been described in various tumour entities. Currently, the specific role of PPARG in patients with colorectal cancer (CRC) is not fully understood. Material and methods The prognostic impact of PPARG expression was investigated by immunohistochemistry in a case-control study using a matched pair selection of CRC tumours (n = 246) with either distant metastases to the liver (n = 82), lung (n = 82) or without distant metastases (n = 82). Its effect on proliferation as well as the sensitivity to the chemotherapeutic drug 5-fluorouracil (5-FU) was examined after activation, inhibition, and transient gene knockdown of PPARG in the CRC cell lines SW403 and HT29. Results High PPARG expression was significantly associated with pulmonary metastasis (p = 0.019). Patients without distant metastases had a significantly longer overall survival with low PPARG expression in their tumours compared to patients with high PPARG expression (p = 0.045). In the pulmonary metastasis cohort instead, a trend towards longer survival was observed for patients with high PPARG expression in their tumour (p = 0.059). Activation of PPARG by pioglitazone and rosiglitazone resulted in a significant dose-dependent increase in proliferation of CRC cell lines. Inhibition of PPARG by its specific inhibitor GW9662 and siRNA-mediated knockdown of PPARG significantly decreased proliferation. Activating PPARG significantly increased the CRC cell lines sensitivity to 5-FU while its inhibition decreased it. Conclusion The prognostic effect of PPARG expression depends on the metastasis localization in advanced CRC patients. Activation of PPARG increased malignancy associated traits such as proliferation in CRC cell lines but also increases sensitivity towards the chemotherapeutic agent 5-FU. Based on this finding, a combination therapy of PPARG agonists and 5-FU-based chemotherapy constitutes a promising strategy which should be further investigated.

Lung cancer is a paradigm for a genetically driven tumor. A variety of drugs were developed targeting specific biomarkers requiring testing for tumor genetic alterations in relevant biomarkers. Different next-generation sequencing technologies are available for library generation: 1) anchored multiplex-, 2) amplicon based- and 3) hybrid capture-based-PCR. Anchored multiplex PCR-based sequencing was investigated for routine molecular testing within the national Network Genomic Medicine Lung Cancer (nNGM). Four centers applied the anchored multiplex ArcherDX-Variantplex nNGMv2 panel to re-analyze samples pre-tested during routine diagnostics. Data analyses were performed by each center and compiled centrally according to study design. Pre-defined standards were utilized, and panel sensitivity was determined by dilution experiments. nNGMv2 panel sequencing was successful in 98.9% of the samples (N = 90). With default filter settings, all but two potential MET exon 14 skipping variants were identified at similar allele frequencies. Both MET variants were found with an adapted calling filter. Three additional variants (KEAP1, STK11, TP53) were called that were not identified in pre-testing analyses. Only total DNA amount but not a qPCR-based DNA quality score correlated with average coverage. Analysis was successful with a DNA input as low as 6.25 ng. Anchored multiplex PCR-based sequencing (nNGMv2) and a sophisticated user-friendly Archer-Analysis pipeline is a robust and specific technology to detect tumor genetic mutations for precision medicine of lung cancer patients.

Abstract Tumor stem cells play a pivotal role in carcinogenesis and metastatic spread in colorectal cancer (CRC). Olfactomedin 4 (OLFM4) is co‐expressed with the established stem cell marker leucine‐rich repeat‐containing G protein‐coupled receptor 5 at the bottom of intestinal crypts and has been suggested as a surrogate for cancer stemness and a biomarker in gastrointestinal tumors associated with prognosis. Therefore, it was the aim of the present study to clarify whether OLFM4 is involved in carcinogenesis and metastatic spread in CRC. We used a combined approach of functional assays using forced OLFM4 overexpression in human CRC cell lines, xenograft mice, and an immunohistochemical approach using patient tissues to investigate the impact of OLFM4 on stemness, canonical Wnt signaling, properties of metastasis and differentiation as well as prognosis. OLFM4 expression correlated weakly with tumor grade in one patient cohort (metastasis collection: p = 0.05; pooled analysis of metastasis collection and survival collection: p = 0.19) and paralleled the expression of differentiation markers (FABP2, MUC2, and CK20) (p = 0.002) but did not correlate with stemness‐associated markers. Further analyses in CRC cells lines as well as xenograft mice including forced overexpression of OLFM4 revealed that OLFM4 neither altered the expression of markers of stemness nor epithelial–mesenchymal transition, nor did OLFM4 itself drive proliferation, migration, or colony formation, which are all prerequisites of carcinogenesis and tumor progression. In line with this, we found no significant correlation between OLFM4 expression, metastasis, and patient survival. In summary, expression of OLFM4 in human CRC seems to be characteristic of differentiation marker expression in CRC but is not a driver of carcinogenesis nor metastatic spread.

Sinonasal tumour diagnosis can be complicated by the heterogeneity of disease and classification systems. Here, the authors use machine learning to classify sinonasal undifferentiated carcinomas into 4 molecular classe with differences in differentiation state and clinical outcome.

Abstract Histological sections of the lymphatic system are usually the basis of static (2D) morphological investigations. Here, we performed a dynamic (4D) analysis of human reactive lymphoid tissue using confocal fluorescent laser microscopy in combination with machine learning. Based on tracks for T-cells (CD3), B-cells (CD20), follicular T-helper cells (PD1) and optical flow of follicular dendritic cells (CD35), we put forward the first quantitative analysis of movement-related and morphological parameters within human lymphoid tissue. We identified correlations of follicular dendritic cell movement and the behavior of lymphocytes in the microenvironment. In addition, we investigated the value of movement and/or morphological parameters for a precise definition of cell types (CD clusters). CD-clusters could be determined based on movement and/or morphology. Differentiating between CD3- and CD20 positive cells is most challenging and long term-movement characteristics are indispensable. We propose morphological and movement-related prototypes of cell entities applying machine learning models. Finally, we define beyond CD clusters new subgroups within lymphocyte entities based on long term movement characteristics. In conclusion, we showed that the combination of 4D imaging and machine learning is able to define characteristics of lymphocytes not visible in 2D histology.

Abstract Chimeric fusion transcription factors are oncogenic hallmarks of several devastating cancer entities including pediatric sarcomas, such as Ewing sarcoma (EwS) and alveolar rhabdomyosarcoma (ARMS). Despite their exquisite specificity, these driver oncogenes have been considered largely undruggable due to their lack of enzymatic activity. Here, we show in the EwS model that – capitalizing on neomorphic DNA-binding preferences – the addiction to the respective fusion transcription factor EWSR1-FLI1 can be leveraged to express therapeutic genes. We genetically engineered a de novo enhancer-based, synthetic and highly potent expression cassette that can elicit EWSR1-FLI1-dependent expression of a therapeutic payload as evidenced by episomal and CRISPR-edited genomic reporter assays. Combining in silico screens and immunohistochemistry, we identified GPR64 as a highly specific cell surface antigen for targeted transduction strategies in EwS. Functional experiments demonstrated that anti-GPR64-pseudotyped lentivirus harboring our expression cassette can specifically transduce EwS cells to promote the expression of viral thymidine kinase sensitizing EwS for treatment to otherwise relatively non-toxic (Val)ganciclovir and leading to strong anti-tumorigenic, but no adverse effects in vivo. Further, we prove that similar vector designs can be applied in PAX3-FOXO1-driven ARMS, and to express immunomodulatory cytokines, such as IL-15 and XCL1, in tumor entities typically considered to be immunologically ‘cold’. Collectively, these results generated in pediatric sarcomas indicate that exploiting, rather than suppressing, the neomorphic functions of chimeric transcription factors may open inroads to innovative and personalized therapies, and that our highly versatile approach may be translatable to other cancers addicted to oncogenic transcription factors with unique DNA-binding properties.

Abstract Understanding the pathological properties of dysregulated protein networks in individual patients’ tumors is the basis for precision therapy. Functional experiments are commonly used, but cover only parts of the oncogenic signaling networks, whereas methods that reconstruct networks from omics data usually only predict average network features across tumors. Here, we show that the explainable AI method layer-wise relevance propagation (LRP) can infer protein interaction networks for individual patients from proteomic profiling data. LRP reconstructs average and individual interaction networks with an AUC of 0.99 and 0.93, respectively, and outperforms state-of-the-art network prediction methods for individual tumors. Using data from The Cancer Proteome Atlas, we identify known and potentially novel oncogenic network features, among which some are cancer-type specific and show only minor variation among patients, while others are present across certain tumor types but differ among individual patients. Our approach may therefore support predictive diagnostics in precision oncology by inferring “patient-level” oncogenic mechanisms.

Purpose: In prostate cancer, the indication to irradiate the pelvic lymphatic pathways in clinical node-negative patients is solely based on clinical nomograms. To define biological risk patterns of lymphatic spread, we studied DNA-methylation and genomic copy number in primary tumors and corresponding lymph nodes metastases. Methods/Patients: DNA-methylation and genomic copy number profiles of primary tumors (PT) and paired synchronous lymph node metastases (LN) from Gleason Score (GS)-6/7a (n = 20 LN-positive, n = 20 LN-negative) and GS-9/10 patients (LN-positive n = 20) after prostatectomy and lymphonodectomy were analyzed. Results: GS-6/7a pN0 PTs and GS-6/7a pN1 PTs differed in histone H3K27me3/H3K9me3 mediated methylation. PTs compared to LNs, in both, GS-6/7a pN1 and GS-9/10 pN1 patients showed large differences in DNA-methylation mediated by histones H3K4me1/2, in addition to copy number changes of chromosomal regions 11q13.1, 14q11.2 and 15q26.1. Between GS-6/7a pN1 and GS-9/10 pN1 patients, methylation levels differed more when comparing LNs than PTs. 16q21-22.1 was specifically lost in GS-9/10 pN0 PTs. Immune system-related pathways characterized the differences between PTs and LNs in both GS-6/7a pN1 and GS-9/10 pN1 patients. Comparing PTs and LKs between GS-6/7a pN1 and GS-9/10 pN1 patients revealed altered transmembrane and G-protein-coupled receptor signaling. Conclusions: Our data suggest that progression of prostate cancer, including lymphatic spread, is associated with histone-mediated DNA-methylation and we hypothesize a methylation signature predicting lymphatic spread in GS-6/7a patients from primary tumors. Lymphatic spread in GS-6/7a patients, flanked by DNA-methylation and CNA alterations, appears to be more complex than in GS-9/10 patients, in whom the primary tumors already appear to bear lymph node metastasis-enabling alterations.

In renal cell carcinoma (RCC), accurate imaging methods are required for treatment planning and response assessment to therapy. In addition, there is an urgent need for new therapeutic options, especially in metastatic RCC. One way to combine diagnostics and therapy in a so-called theranostic approach is the use of radioligands directed against surface antigens. For instance, radioligands against prostate-specific membrane antigen (PSMA) have already been successfully used for diagnosis and radionuclide therapy of metastatic prostate cancer. Recent studies have demonstrated that PSMA is expressed not only in prostate cancer but also in the neovasculature of several solid tumors, which has raised hopes to use PSMA-guided theranostic approaches in other tumor entities, too. However, data on PSMA expression in different histopathological subtypes of RCC are sparse. Because a better understanding of PSMA expression in RCC is critical to assess which patients would benefit most from theranostic approaches using PSMA-targeted ligands, we investigated the expression pattern of PSMA in different subtypes of RCC on protein level. Immunohistochemical staining for PSMA was performed on formalin-fixed, paraffin-embedded archival material of major different histological subtypes of RCC (clear cell RCC (ccRCC)), papillary RCC (pRCC) and chromophobe RCC (cpRCC). The extent and intensity of PSMA staining were scored semi-quantitatively and correlated with the histological RCC subtypes. Group comparisons were calculated with the Kruskal–Wallis test. In all cases, immunoreactivity was detected only in the tumor-associated vessels and not in tumor cells. Staining intensity was the strongest in ccRCC, followed by cpRCC and pRCC. ccRCC showed the most diffuse staining pattern, followed by cpRCC and pRCC. Our results provide a rationale for PSMA-targeted theranostic approaches in ccRCC and cpRCC.

Formalin-fixed, paraffin-embedded (FFPE) specimens are an attractive resource for retrospective clinical proteomics studies. Here, the authors benchmark and optimize proteomics workflows for FFPE sample analysis and provide guidelines on which methods to use for different samples and applications.

ObjectivesThe use of liquid biopsies (LB) in patients with solid malignancies enables comprehensive genomic profiling (CGP) of circulating tumor DNA (ctDNA) and has the potential to guide therapy stratification and support disease monitoring. To examine clinical uptake of LB in a real-world setting, LB implementation was analyzed at two German cancer centers (LMU Munich and Charité - Universitätsmedizin Berlin) between 2017 and 2021, with focus on colorectal cancer (CRC) patients.MethodsIn this retrospective analysis, all patients who received a LB between January 2017 and December 2021 as part of routine clinical management were included. To provide adequate context, we collected disease characteristics and technical specifications of the LB methods applied. Additionally, we examined the concordance of RAS status in tumor tissue and LB. Finally, we discuss the potential of LB as a diagnostic tool to drive personalized treatment in CRC patients and how to implement LB in clinical routine.ResultsIn total, our cohort included 86 CRC patients and 161 LB conducted in these patients between 2017 and 2021. In 59 patients, comparison between tissue-based and liquid-based molecular diagnostics, revealed a divergence in 23 (39%) of the evaluable samples.ConclusionOur real-world data analysis indicates that the possibilities of LB are not yet exploited in everyday clinical practice. Currently, the variety of methods and lack of standardization, as well as restricted reimbursement for liquid based CGP hinder the use of LB in clinical routine. To overcome these issues, prospective clinical trials are needed to provide evidence driving the implementation of LB into the management of CRC patients and to support their implementation into clinical guidelines.

BackgroundDespite the success of standard magnetic resonance enterography (MRE) in detecting Crohn’s disease (CD), characterization of strictures and, thus, therapy guidance is still limited. The aim of the study was to determine diagnostic accuracy of MRE in detecting or ruling out active inflammation and identifying fibrotic lesions in patients with terminal ileal CD with histopathology as reference.MethodsSixty-seven consecutive patients (median age 32 years, range 19–79 years) with terminal ileal CD were retrospectively enrolled between January 2015 and October 2020. The median interval between MRE and surgery was 9 days (range 0–86 days). Sensitivity, specificity, positive and negative predictive value (PPV and NPV, respectively), and area under the curve (AUC) with 95% confidence intervals (CIs) were calculated for the MRE-based AIS (acute inflammation score) using the histopathology of surgical specimens as the reference standard.ResultsSensitivity, specificity, PPV, and NPV for detecting or ruling out active inflammation were 100% (CI, 0.94–1.00; 0.44–1.00; 0.93–1.00; 0.31–1.00) using an AIS cut-off of >4.1. AUC was 1.00 (CI, 1.00–1.00; p

The receptor tyrosine kinase c-KIT (CD117) has a key role in hematopoiesis and is a marker for endothelial and cardiac progenitor cells. In vivo, deficiency of c-KIT is lethal and therefore using CRISPR/Cas9 editing we generated heterozygous and homozygous c-KIT knockout human embryonic stem cell (ES cell) lines. The c-KIT knockout left ES cell pluripotency unaffected as shown by immunofluorescence and trilineage differentiation potential. Heterozygous and homozygous c-KIT knockouts showed complete loss of exon 17, resulting in ablation of c-KIT protein from the cell surface. c-KIT knockout ES cells provide a valuable tool for further investigating c-KIT biology.

Targeting the tumor-associated carbonic anhydrase XII (CA XII) is considered a promising strategy to improve cancer treatment. As such progress is highly demanded for ovarian carcinomas, the present study aimed to provide deeper information about their CA XII expression profile. A large collection of tissue specimens was stained immunohistochemically with a specific anti-CA XII antibody to evaluate the expression in neoplastic and non-neoplastic epithelial ovarian cells. In addition, flow cytometry was used to measure CA XII expression on tumor cells from malignant ascites fluid. Binding of the antibody revealed a significant CA XII expression in most ovarian carcinoma tissue samples and ascites-derived ovarian carcinoma cells. Moreover, CA XII was expressed at higher levels in ovarian carcinomas as compared to borderline ovarian tumors and non-neoplastic ovarian epithelia. Within the carcinoma tissues, high expression of CA XII was associated with higher tumor grading and a trend towards shorter overall survival. Our results indicate that CA XII plays a crucial role for the malignancy of ovarian carcinoma cells and emphasize the potential of CA XII as a diagnostic marker and therapeutic target in the management of ovarian carcinomas.

Abstract Background Clinical application of cancer immunotherapy requires a better understanding of tumor immunogenicity and the tumor microenvironment. HLA class I molecules present antigens to CD8+ cytotoxic cells. Their loss or downregulation is frequently found in tumors resulting in reduced T cell responses and worse prognosis. Methods We evaluated HLA class I heavy chain expression by immunohistochemistry in 863 biopsies (GeparTrio trial). Patients received neoadjuvant chemotherapy and adjuvant endocrine treatment if tumors were hormone receptor-positive (HR+). In parallel, the expression of HLA-A was analyzed using a microarray cohort of 320 breast cancer patients from the MD Anderson Cancer Center. We evaluated its association with clinical outcome, tumor-infiltrating lymphocytes (TILs), and immune cell metagenes. Results In HR+/HER2− breast cancer, HLA class I heavy chain expression was associated with increased TILs and better response to chemotherapy (7% vs. 14% pCR rate, P = 0.029), but worse disease-free survival (hazard ratio (HR) 1.6 (1.1–2.4); P = 0.024). The effect was significant in a multivariate model adjusted for clinical and pathological variables (HR 1.7 (1.1–2.6); P = 0.016) and was confirmed by analysis of HLA-A in a microarray cohort. HLA-A was correlated to most immune cell metagenes. There was no association with response or survival in triple-negative or HER2+ disease. Conclusions The study confirms the negative prognostic role of lymphocytes in HR+ breast cancer and points at a complex interaction between chemotherapy, endocrine treatment, and tumor immunogenicity. The results point at a subtype-specific and potentially treatment-specific role of tumor-immunological processes in breast cancer with different implications in triple-negative and hormone receptor-positive disease.

Background: Transplant glomerulopathy (TG) is one of the main causes of post-transplant proteinuria (PU). The features and possible risk factors for proteinuria in TG patients are uncertain.Methods: We investigated all patients who had biopsy-proven TG from 2000 to 2018 in our center. The clinical and histological data were compared between two groups with or without PU (cut-off = 0.3 g/day). Spearman correlation analysis was used to evaluate the relationship between PU and pathological changes. The risk factors for PU in TG patients were determined by multivariable logistic regression analysis.Results: One hundred and twenty-five (75.76%) of all enrolled 165 TG patients had proteinuria ≥0.3 g/day at the time of biopsy. TG patients' PU level was significantly correlated with Banff lesion score cg (ρ = 0.247, P = 0.003), and mm (ρ = 0.257, P = 0.012). Systolic blood pressure ≥140 mmHg (OR 2.72, 95% CI 1.04–7.10, P = 0.041), diastolic blood pressure ≥90 mmHg (OR 4.84, 95% CI 1.39–16.82, P = 0.013), peak PRA ≥5% (OR 6.47, 95% CI 1.67–25.01, P = 0.007), positive C4d staining (OR 4.55, 95% CI 1.29–16.11, 0.019), tacrolimus-based regimen (OR 3.5, 95% CI 1.28–9.54, P = 0.014), and calcium channel blocker usage (OR 4.38, 95% CI 1.59–12.09, P = 0.004) were independent risk factors for PU.Conclusions: Proteinuria is common in TG patients. systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, peak PRA ≥5%, positive C4d staining, tacrolimus-based regimen, and calcium channel blocker usage are associated with proteinuria in TG patients.

Background: Onychomycosis numbers among the most common fungal infections in humans affecting finger- or toenails. Histology remains a frequently applied screening technique to diagnose onychomycosis. Screening slides for fungal elements can be time-consuming for pathologists, and sensitivity in cases with low amounts of fungi remains a concern. Convolutional neural networks (CNNs) have revolutionized image classification in recent years. The goal of our project was to evaluate if a U-NET-based segmentation approach as a subcategory of CNNs can be applied to detect fungal elements on digitized histologic sections of human nail specimens and to compare it with the performance of 11 board-certified dermatopathologists. Methods: In total, 664 corresponding H&E- and PAS-stained histologic whole-slide images (WSIs) of human nail plates from four different laboratories were digitized. Histologic structures were manually annotated. A U-NET image segmentation model was trained for binary segmentation on the dataset generated by annotated slides. Results: The U-NET algorithm detected 90.5% of WSIs with fungi, demonstrating a comparable sensitivity with that of the 11 board-certified dermatopathologists (sensitivity of 89.2%). Conclusions: Our results demonstrate that machine-learning-based algorithms applied to real-world clinical cases can produce comparable sensitivities to human pathologists. Our established U-NET may be used as a supportive diagnostic tool to preselect possible slides with fungal elements. Slides where fungal elements are indicated by our U-NET should be reevaluated by the pathologist to confirm or refute the diagnosis of onychomycosis.

Abstract Background Comprehensive mutational profiling data now available on all major cancers have led to proposals of novel molecular tumor classifications that modify or replace the established organ- and tissue-based tumor typing. The rationale behind such molecular reclassifications is that genetic alterations underlying cancer pathology predict response to therapy and may therefore offer a more precise view on cancer than histology. The use of individual actionable mutations to select cancers for treatment across histotypes is already being tested in the so-called basket trials with variable success rates. Here, we present a computational approach that facilitates the systematic analysis of the histological context dependency of mutational effects by integrating genomic and proteomic tumor profiles across cancers. Methods To determine effects of oncogenic mutations on protein profiles, we used the energy distance, which compares the Euclidean distances of protein profiles in tumors with an oncogenic mutation (inner distance) to that in tumors without the mutation (outer distance) and performed Monte Carlo simulations for the significance analysis. Finally, the proteins were ranked by their contribution to profile differences to identify proteins characteristic of oncogenic mutation effects across cancers. Results We apply our approach to four current proposals of molecular tumor classifications and major therapeutically relevant actionable genes. All 12 actionable genes evaluated show effects on the protein level in the corresponding tumor type and showed additional mutation-related protein profiles in 21 tumor types. Moreover, our analysis identifies consistent cross-cancer effects for 4 genes (FGFR1, ERRB2, IDH1, KRAS/NRAS) in 14 tumor types. We further use cell line drug response data to validate our findings. Conclusions This computational approach can be used to identify mutational signatures that have protein-level effects and can therefore contribute to preclinical in silico tests of the efficacy of molecular classifications as well as the druggability of individual mutations. It thus supports the identification of novel targeted therapies effective across cancers and guides efficient basket trial designs.

Despite the overall poor prognosis of pancreatic cancer there is heterogeneity in clinical courses of tumors not assessed by conventional risk stratification. This yields the need of additional markers for proper assessment of prognosis and multimodal clinical management. We provide a proof of concept study evaluating the feasibility of Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI) to identify specific peptide signatures linked to prognostic parameters of pancreatic cancer. On 18 patients with exocrine pancreatic cancer after tumor resection, MALDI imaging analysis was performed additional to histopathological assessment. Principal component analysis (PCA) was used to explore discrimination of peptide signatures of prognostic histopathological features and receiver operator characteristic (ROC) to identify which specific m/z values are the most discriminative between the prognostic subgroups of patients. Out of 557 aligned m/z values discriminate peptide signatures for the prognostic histopathological features lymphatic vessel invasion (pL, 16 m/z values, eight proteins), nodal metastasis (pN, two m/z values, one protein) and angioinvasion (pV, 4 m/z values, two proteins) were identified. These results yield proof of concept that MALDI-MSI of pancreatic cancer tissue is feasible to identify peptide signatures of prognostic relevance and can augment risk assessment.

Assessment of patients with synchronous primary cancers and metastases is challenging, as it can be difficult to assign the metastases to the correct primary due to low differentiation, high similarity on histology or inaccessibility of tumour tissue. Systemic treatment for metastatic disease, however, needs to be directed at the leading histology or cover multiple tumour types with the same regimen. Considering the additional obstacles in cancer management, including tumour heterogeneity and clonal evolution, blood-based genomic profiling (‘liquid biopsy’) is suggested to be a useful tool to provide accessible tumour-derived biomarkers. We herein report a case of a patient with independent primary tumours of the colon and pancreas, as well as liver metastases. All lesions were resected and genotyped revealing KRAS mutations G12C and G12D in the primary tumours, respectively. The G12D mutation detected in the pancreatic tumour was retrieved in the metastasis, thus confirming the pancreatic cancer to be the origin of the liver lesions. The prevalence of the pancreatic tumour was additionally verified by the detection of the G12D variant in circulating cell-free DNA (cfDNA). This case demonstrates the utility of liquid biopsy to identify the predominant tumour burden in patients with multiple primary cancers, based on the detection of the tumour-associated gene mutation in the plasma. Serial monitoring through liquid biopsies might allow disease surveillance to guide cancer management. The review of the literature highlights the importance of liquid biopsies in personalised oncology, even though only one case report refers to the benefit of cfDNA analysis in a patient affected by synchronous primary tumours.

Liver fibrosis is a critical complication of obesity-induced fatty liver disease. Wnt1 inducible signaling pathway protein 1 (WISP1/CCN4), a novel adipokine associated with visceral obesity and insulin resistance, also contributes to lung and kidney fibrosis. The aim of the present study was to investigate the role of CCN4 in liver fibrosis in severe obesity. For this, human liver biopsies were collected from 35 severely obese humans (BMI 42.5 ± 0.7 kg/m2, age 46.7 ± 1.8 y, 25.7% males) during bariatric surgery and examined for the expression of CCN4, fibrosis, and inflammation markers. Hepatic stellate LX-2 cells were treated with human recombinant CCN4 alone or in combination with LPS or transforming growth factor beta (TGF-β) and examined for fibrosis and inflammation markers. CCN4 mRNA expression in the liver positively correlated with BMI and expression of fibrosis markers COL1A1, COL3A1, COL6A1, αSMA, TGFB1, extracellular matrix turnover enzymes TIMP1 and MMP9, and the inflammatory marker ITGAX/CD11c. In LX-2 cells, the exposure to recombinant CCN4 caused dose-dependent induction of MMP9 and MCP1. CCN4 potentiated the TGF-β-mediated induction of COL3A1, TIMP1, and MCP1 but showed no interaction with LPS treatment. Our results suggest a potential contribution of CCN4 to the early pathogenesis of obesity-associated liver fibrosis.

The contribution of intra-tumour heterogeneity is increasingly associated with resistance to therapy. Here, the authors use genomic analyses to study heterogeneity in colorectal cancer and perform in-depth reconstruction of heterogeneity in one sample.

Knowledge on immune and stromal cells in medulloblastoma microenvironment is still limited as previous work was frequently restricted by low sample size and the lack of molecular subgroup information. We characterized 10 microenvironment cell populations as well as PD-L1 from gene expression in 1422 brain tumors and 763 medulloblastomas. All in all, medulloblastomas showed low expression of immune markers. Still, there were substantial differences with a clustering of medulloblastoma subgroups according to their microenvironment profile. Specifically, SHH medulloblastomas displayed strong signatures of fibroblasts, T cells and macrophages, while markers of cytotoxic lymphocytes were enriched in Group 4 tumors. PD-L1 gene expression appeared to be relatively high in single SHH and WNT cases but was undetectable by immunohistochemistry. In addition, two diverse immuno-stromal patterns were identified, indicating distinct types of local tumor immunosuppression, which were primarily controlled by either macrophage and regulatory T cell-mediated mechanisms or immunosuppressive cytokines and checkpoints, respectively. None of the immune cell signatures had an independent prognostic value in the present dataset after multiple testing correction. These results suggest a mild, but subgroup-specific infiltration of immune cells in medulloblastoma.

In this contribution we report on the possibilities of dry and lubricated friction modification introduced by different laser surface texturing methods. We compare the potential of Laser-Induced Periodic Surface Structures and Laser Beam Interference Ablation on 100Cr6 steel in a linear reciprocating ball-on-disc configuration using 100Cr6 steel and tungsten carbide balls with load forces between 50 mN and 1000 mN. For dry friction, we find a possibility to reduce the coefficient of friction and we observe a pronounced direction dependency for surfaces fabricated by Laser Beam Interference Ablation. Furthermore, Laser-Induced Periodic Surface Structures result in a load-dependent friction reduction for lubricated linear reciprocating movements. This work helps to identify the modification behaviour of laser generated micro structures with feature sizes of approximately 1 µm and reveals new possibilities for surface engineering.

Soft-tissue sarcomas (STS) are rare malignancies that account for 1% of adult cancers and comprise more than 50 entities. Current therapeutic options for advanced-stage STS are limited. Immune checkpoint inhibitors targeting the PD-1/PD-L1 signaling axis are being explored as new treatment modality in STS; however, the determinants of response to these agents are largely unknown. Using the sarcoma data set of The Cancer Genome Altas (TCGA) and an independent cohort of untreated high-grade STS, we analyzed DNA copy number status and mRNA expression of PD-L1 in a total of 335 STS cases. Copy number gains (CNG) were detected in 54 TCGA cases (21.1%), of which 21 (8.2%) harbored focal PD-L1 CNG and that were most prevalent in myxofibrosarcoma (35%) and undifferentiated pleomorphic sarcoma (34%). In the untreated high-grade STS cohort, we detected CNG in six cases (7.6%). Analysis of co-amplified genes identified a 5.6-Mb core region comprising 27 genes, including JAK2. Patients with PD-L1 CNG had higher PD-L1 expression compared with STS without CNG (fold change, 1.8; p = 0.02), an effect that was most pronounced in the setting of focal PD-L1 CNG (fold change, 3.0; p = 0.0027). STS with PD-L1 CNG showed a significantly higher mutational load compared with tumors with a diploid PD-L1 locus (median number of mutated genes; 58 vs. 40; p = 3.6E-06), and PD-L1 CNG were associated with inferior survival (HR = 1.82; p = 0.025). In contrast, T-cell infiltrates quantified by mRNA expression of CD3Z were associated with improved survival (HR = 0.88; p = 0.024) and consequently influenced the prognostic power of PD-L1 CNG, with low CD3Z levels conferring poor survival in cases with PD-L1 CNG (HR = 1.8; p = 0.049). These data demonstrate that PD-L1 GNG and elevated expression of PD-L1 occur in a substantial proportion of STS, have prognostic impact that is modulated by T-cell infiltrates, and thus warrant investigation as response predictors for immune checkpoint inhibition.

Understanding and interpreting classification decisions of automated image classification systems is of high value in many applications, as it allows to verify the reasoning of the system and provides additional information to the human expert. Although machine learning methods are solving very successfully a plethora of tasks, they have in most cases the disadvantage of acting as a black box, not providing any information about what made them arrive at a particular decision. This work proposes a general solution to the problem of understanding classification decisions by pixel-wise decomposition of nonlinear classifiers. We introduce a methodology that allows to visualize the contributions of single pixels to predictions for kernel-based classifiers over Bag of Words features and for multilayered neural networks. These pixel contributions can be visualized as heatmaps and are provided to a human expert who can intuitively not only verify the validity of the classification decision, but also focus further analysis on regions of potential interest. We evaluate our method for classifiers trained on PASCAL VOC 2009 images, synthetic image data containing geometric shapes, the MNIST handwritten digits data set and for the pre-trained ImageNet model available as part of the Caffe open source package.

BackgroundAutopsy rates in Western countries consistently decline to an average of Methods and findingsOf all adult autopsy cases of the Charité Institute of Pathology from the years 1988, 1993, 1998, 2003 and 2008, the pre- and postmortal diagnoses and all demographic data were analyzed retrospectively. Based on power analysis, 1,800 cases were randomly selected to perform discrepancy classification (class I-VI) according to modified Goldman criteria. The rate of discrepancies in major diagnoses (class I) was 10.7% (95% CI: 7.7%-14.7%) in 2008 representing a reduction by 15.1%. Subgroup analysis revealed several influencing factors to significantly correlate with the discrepancy rate. Cardiovascular diseases had the highest frequency among class-I-discrepancies. Comparing the 1988-data of East- and West-Berlin, no significant differences were found in diagnostic discrepancies despite an autopsy rate differing by nearly 50%. A risk profile analysis visualized by intuitive heatmaps revealed a significantly high discrepancy rate in patients treated in low or intermediate care units at community hospitals. In this collective, patients with genitourinary/renal or infectious diseases were at particularly high risk.ConclusionsThis is the current largest and most comprehensive study on diagnostic discrepancies worldwide. Our well-powered analysis revealed a significant rate of class-I-discrepancies indicating that autopsies are still of value. The identified risk profiles may aid both pathologists and clinicians to identify patients at increased risk for a discrepant diagnosis and possibly suboptimal treatment intra vitam.

Gene or protein expression data are usually represented by metric or at least ordinal variables. In order to translate a continuous variable into a clinical decision, it is necessary to determine a cutoff point and to stratify patients into two groups each requiring a different kind of treatment. Currently, there is no standard method or standard software for biomarker cutoff determination. Therefore, we developed Cutoff Finder, a bundle of optimization and visualization methods for cutoff determination that is accessible online. While one of the methods for cutoff optimization is based solely on the distribution of the marker under investigation, other methods optimize the correlation of the dichotomization with respect to an outcome or survival variable. We illustrate the functionality of Cutoff Finder by the analysis of the gene expression of estrogen receptor (ER) and progesterone receptor (PgR) in breast cancer tissues. This distribution of these important markers is analyzed and correlated with immunohistologically determined ER status and distant metastasis free survival. Cutoff Finder is expected to fill a relevant gap in the available biometric software repertoire and will enable faster optimization of new diagnostic biomarkers. The tool can be accessed at http://molpath.charite.de/cutoff.

The type III secretion system (TTSS) is a key mechanism for host cell interaction used by a variety of bacterial pathogens and symbionts of plants and animals including humans. The TTSS represents a molecular syringe with which the bacteria deliver effector proteins directly into the host cell cytosol. Despite the importance of the TTSS for bacterial pathogenesis, recognition and targeting of type III secreted proteins has up until now been poorly understood. Several hypotheses are discussed, including an mRNA-based signal, a chaperon-mediated process, or an N-terminal signal peptide. In this study, we systematically analyzed the amino acid composition and secondary structure of N-termini of 100 experimentally verified effector proteins. Based on this, we developed a machine-learning approach for the prediction of TTSS effector proteins, taking into account N-terminal sequence features such as frequencies of amino acids, short peptides, or residues with certain physico-chemical properties. The resulting computational model revealed a strong type III secretion signal in the N-terminus that can be used to detect effectors with sensitivity of approximately 71% and selectivity of approximately 85%. This signal seems to be taxonomically universal and conserved among animal pathogens and plant symbionts, since we could successfully detect effector proteins if the respective group was excluded from training. The application of our prediction approach to 739 complete bacterial and archaeal genome sequences resulted in the identification of between 0% and 12% putative TTSS effector proteins. Comparison of effector proteins with orthologs that are not secreted by the TTSS showed no clear pattern of signal acquisition by fusion, suggesting convergent evolutionary processes shaping the type III secretion signal. The newly developed program EffectiveT3 (http://www.chlamydiaedb.org) is the first universal in silico prediction program for the identification of novel TTSS effectors. Our findings will facilitate further studies on and improve our understanding of type III secretion and its role in pathogen-host interactions.

Background: Cervical spinal cord injury (SCI) models in rats have become increasingly useful because of their translational potential. The goal of this study was to design, develop and validate a quick and reliable forelimb locomotor rating scale for adult rats with unilateral cervical SCI injury., New Method: Adult female rats were subjected to a C5 unilateral mild contusion (n=10), moderate contusion (n=10) or hemisection injury (n=9). Forelimb locomotion was evaluated before injury, four times during the first week (Days 2, 3, 4 and 7) and weekly for up to 8 weeks post-injury. Scoring categories were identified and animals were ranked based on their performance in these categories. The scale was validated for its usefulness by comparing animals with different injury models (dorsolateral funiculotomy C3/4), levels of injury (moderate contusion C4) and sex (male - moderate contusion C3/4) and also by correlating FLS scores with other established behavioral tests (grid walking and kinetic tests)., Results and Comparison With Existing Methods: Forelimb performance on both the grid-walking and kinetic tests was positively correlated with the forelimb locomotor rating scale (FLS). Histological analysis established a positive correlation between the spared tissue and the observed FLS score. Our results show that the new rating scale can reliably detect forelimb deficits and recovery predicted by other behavioral tests. Furthermore, the new method provides reproducible data between trained and naïve examiners., Conclusion: In summary, the proposed rating scale is a useful tool for assessment of injury and treatments designed to enhance recovery after unilateral cervical SCI., (Copyright © 2014 Elsevier B.V. All rights reserved.)

Individually, motor training, pharmacological interventions, and housing animals in an enriched environment (EE) following spinal cord injury (SCI) result in limited functional improvement but, when combined, may enhance motor function. Here, we tested amphetamine (AMPH)-enhanced skilled motor training following a unilateral C3-C4 contusion injury on the qualitative components of reaching and on skilled forelimb function, as assessed using single-pellet and staircase reaching tasks. Kinematic analysis evaluated the quality of the reach, and unskilled locomotor function was also tested. Animals receiving AMPH and skilled forelimb training performed better than operated control animals on qualitative reaching, but not on skilled reaching. Those that received the combination treatment and were housed in EE cages showed significantly less improvement in qualitative reaching and grasping. Kinematic analysis revealed a decrease in digit abduction during skilled reaching among all groups, with no differences among groups. Kinematics provided no evidence that improved function was related to improved quality of reach. There was no evidence of neuroprotection in the cervical spinal cord. The absence of evidence for kinematic improvement or neuroprotection suggested that AMPH-enhanced motor training is due primarily to supraspinal effects, an enhancement of attention during skilled motor training, or plasticity in supraspinal circuitry involved with motor control.

The hypothesis was tested that carotid baroreflex gain is increased after 20% hemorrhage. The baroreceptor reflex responses to changes in carotid sinus pressure (CSP) were measured in control, 20% hemorrhage, and reinfusion conditions in three experimental groups: conscious intact (n = 7), anesthetized intact (n = 8), and anesthetized vagotomized (n = 8) dogs. Mean arterial pressure (MAP), heart rate (HR), cardiac output (CO), stroke volume (SV), and calculated total peripheral resistance (TPR) responses to changes in CSP were measured. At any given CSP, MAP, CO, and SV all decreased significantly with the 20% hemorrhage, as reflected by a downward shift in the reflex characteristic curve with no change in overall reflex range or gain. In contrast, TPR and HR responses to CSP were not significantly altered by 20% hemorrhage; reflex curves and gains were comparable to control conditions. In the conscious intact dogs, the maximal reflex gain, Gmax, for the MAP response was -1.365 +/- 0.25, -1.298 +/- 0.33, and -1.324 +/- 0.25 in control, 20% hemorrhage, and reinfusion conditions, respectively, and was not significantly altered by hemorrhage. In the same group, the Gmax for the HR response was -1.792 +/- 0.65, -1.709 +/- 0.33, and -1.986 +/- 0.67 in control, 20% hemorrhage, and reinfusion conditions, respectively; baroreflex gain on HR was not increased with hemorrhage. Plasma arginine vasopressin (AVP), an increase in which has been proposed to augment baroreflex gain, increased from a control level of 0.98 +/- 0.27 to 9.66 +/- 2.67 pg/ml during 20% hemorrhage in the conscious intact dogs; despite the increase in plasma AVP during hemorrhage, augmentation of baroreflex gain was not observed.(ABSTRACT TRUNCATED AT 250 WORDS)

The activity of five adrenocortical steroidogenic enzymes, 3 beta-hydroxysteroid dehydrogenase/isomerase (3 beta-HSD), 17-hydroxylase (17-OHase) 17,20 desmolase (17,20D), 21-hydroxylase (21-OHase) and 11-hydroxylase (11-OHase), were measured in vitro in purified mitochondria or microsomes from rhesus monkey (Macaca mulata) and human adrenal tissue in the presence and absence of graded concentrations of ketoconazole. Rhesus 3 beta-HSD activity was unaffected by ketoconazole at concentrations up to 5000 microM. However, human adrenal 3 beta-HSD was inhibited by approximately 40% (p less than .01) at concentrations of 500 microM and by 80% at 100 microM. 17-OHase and 17,20D were significantly inhibited in the human at 5 microM (p less than .001) and in the rhesus monkey at 50 microM (p less than .001). A similar inhibitory effect was found on microsomal 21-OHase, with significant inhibition at 5 microM ketoconazole in the human and rhesus monkey (p less than 0.001). Mitochondrial 11-OHase was also significantly inhibited by ketoconazole in both the human (p less than .005) and rhesus (p less than .001) at 2.0 microM. These results represent documentation of the specific adrenal steroidogenic steps affected by ketoconazole and confirm the observations that this imidazole derivative is a powerful inhibitor of enzymes in the glucocorticoid pathway.

The in vitro inhibition of Leydig cell microsomal steroidogenesis by ketoconazole, a potent P-450 dependent enzyme blocker, was evaluated in the human, stallion and pig. Purified Leydig cells were isolated by mechanical dispersion of teased, decapsulated whole testes and sieving through a 0.25 mm stainless steel mesh. The activity of 3 beta-hydroxysteroid dehydrogenase/isomerase (3 beta-HSD), 17-hydroxylase (17-OHase), 17,20-desmolase (17,20D), 17-ketosteroid reductase (17-KSR) and aromatase were measured using a constant amount (50 microM) of 14C-labelled substrates in the presence of varying concentrations of pure ketoconazole. Products were isolated by thin layer chromatography and verified by derivative formation. 17-OHase and 17,20D activities were significantly inhibited (p less than .001) by ketoconazole at concentrations as low as 5 microM. 3 beta-HSD, 17-KSR and aromatase activities were only significantly inhibited by ketoconazole at concentrations of 500 and 5000 microM. These data describe the specific loci of inhibition of ketoconazole on testicular steroidogenesis and confirm the observations that ketoconazole is an effective inhibitor of androgen biosynthesis in several species.