BackgroundFor virtually every patient with colorectal cancer (CRC), hematoxylin–eosin (HE)–stained tissue slides are available. These images contain quantitative information, which is not routinely used to objectively extract prognostic biomarkers. In the present study, we investigated whether deep convolutional neural networks (CNNs) can extract prognosticators directly from these widely available images.Methods and findingsWe hand-delineated single-tissue regions in 86 CRC tissue slides, yielding more than 100,000 HE image patches, and used these to train a CNN by transfer learning, reaching a nine-class accuracy of >94% in an independent data set of 7,180 images from 25 CRC patients. With this tool, we performed automated tissue decomposition of representative multitissue HE images from 862 HE slides in 500 stage I–IV CRC patients in the The Cancer Genome Atlas (TCGA) cohort, a large international multicenter collection of CRC tissue. Based on the output neuron activations in the CNN, we calculated a “deep stroma score,” which was an independent prognostic factor for overall survival (OS) in a multivariable Cox proportional hazard model (hazard ratio [HR] with 95% confidence interval [CI]: 1.99 [1.27–3.12], p = 0.0028), while in the same cohort, manual quantification of stromal areas and a gene expression signature of cancer-associated fibroblasts (CAFs) were only prognostic in specific tumor stages. We validated these findings in an independent cohort of 409 stage I–IV CRC patients from the “Darmkrebs: Chancen der Verhütung durch Screening” (DACHS) study who were recruited between 2003 and 2007 in multiple institutions in Germany. Again, the score was an independent prognostic factor for OS (HR 1.63 [1.14–2.33], p = 0.008), CRC-specific OS (HR 2.29 [1.5–3.48], p = 0.0004), and relapse-free survival (RFS; HR 1.92 [1.34–2.76], p = 0.0004). A prospective validation is required before this biomarker can be implemented in clinical workflows.ConclusionsIn our retrospective study, we show that a CNN can assess the human tumor microenvironment and predict prognosis directly from histopathological images.

Automatic recognition of different tissue types in histological images is an essential part in the digital pathology toolbox. Texture analysis is commonly used to address this problem; mainly in the context of estimating the tumour/stroma ratio on histological samples. However, although histological images typically contain more than two tissue types, only few studies have addressed the multi-class problem. For colorectal cancer, one of the most prevalent tumour types, there are in fact no published results on multiclass texture separation. In this paper we present a new dataset of 5,000 histological images of human colorectal cancer including eight different types of tissue. We used this set to assess the classification performance of a wide range of texture descriptors and classifiers. As a result, we found an optimal classification strategy that markedly outperformed traditional methods, improving the state of the art for tumour-stroma separation from 96.9% to 98.6% accuracy and setting a new standard for multiclass tissue separation (87.4% accuracy for eight classes). We make our dataset of histological images publicly available under a Creative Commons license and encourage other researchers to use it as a benchmark for their studies.

Introduction:The rarity of thymomas and lack of multi-institutional studies have hampered therapeutic progress for decades. To overcome this, the members of the International Thymic Malignancy Interest Group created a worldwide retrospective database. This database was analyzed regarding the demographic and geographic distribution of thymomas and the impact of different variables on survival and recurrence.Methods:This study analyzed 4221 thymomas diagnosed between 1983 and 2012 with World Health Organization histotype information from the International Thymic Malignancy Interest Group database. Associations to survival and recurrence were studied by univariate and multivariate analyses.Results:Type B2 thymoma is the most common (28%) and type A the least common (12%) histotypes. They are significantly more frequent in Europe and the United States than Asia. Type A and AB occur at significantly higher age than other thymomas (64 and 57 years, respectively). There are no differences in gender distribution. Stage is lower in type A (90% in stages I–II) and AB than B1 to B3 thymomas (38% of type B3 in stage III). In univariate analysis, recurrence is significantly less frequent among stage I/II tumors, in type A and AB (recurrence rates, 1–2%) than B1 to B3 thymomas (2–7%). Multivariate analysis reveals an impact of age, stage, and resection status on survival and recurrence, whereas for histology there is only a significant impact on recurrence.Conclusion:New findings are (1) geographic differences such as a lower incidence of type A and B2 thymoma in Asia; and (2) impact of stage and histology, the latter partially limited to early stage disease, on recurrence.

Introduction:The WHO classification of pulmonary neuroendocrine tumors (PNETs) is also used to classify thymic NETs (TNETs) into typical and atypical carcinoid (TC and AC), large cell neuroendocrine carcinoma (LCNEC), and small cell carcinoma (SCC), but little is known about the usability of alternative classification systems.Methods: One hundred seven TNET (22 TC, 51 AC, 28 LCNEC, and 6 SCC) from 103 patients were classified according to the WHO, the European Neuroendocrine Tumor Society, and a grading-related PNET classification. Low coverage whole-genome sequencing and immunohistochemical studies were performed in 63 cases. A copy number instability (CNI) score was applied to compare tumors. Eleven LCNEC were further analyzed using targeted next-generation sequencing. Morphologic classifications were tested against molecular features.Results: Whole-genome sequencing data fell into three clusters: CNI low , CNI int , and CNI high . CNI low and CNI int comprised not only TC and AC, but also six LCNECs. CNI high contained all SCC and nine LCNEC, but also three AC. No morphologic classification was able to predict the CNI cluster. Cases where primary tumors and metastases were available showed progression from low-grade to higher-grade histologies. Analysis of LCNEC revealed a subgroup of intermediate NET G3 tumors that differed from LCNEC by carcinoid morphology, expression of chromogranin, and negativity for enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2).Conclusions: TNETs fall into three molecular subgroups that are not reflected by the current WHO classification. Given the large overlap between TC and AC on the one hand, and AC and LCNEC on the other, we propose a

Current outcome prediction markers for localized prostate cancer (PCa) are insufficient. The impact of the lipid-modifying Sphingomyelin Phosphodiesterase Acid Like 3B (SMPDL3B) in PCa is unknown. Two cohorts of patients with PCa who underwent radical prostatectomy (n = 40, n = 56) and benign prostate hyperplasia (BPH) controls (n = 8, n = 11) were profiled for SMPDL3B expression with qRT-PCR. Publicly available PCa cohorts (Memorial Sloane Kettering Cancer Centre (MSKCC; n = 131, n = 29 controls) and The Cancer Genome Atlas (TCGA; n = 497, n = 53 controls)) served for validation. SMPDL3B’s impact on proliferation and migration was analyzed in PC3 cells by siRNA knockdown. In both cohorts, a Gleason score and T stage independent significant overexpression of SMPDL3B was seen in PCa compared to BPH (p < 0.001 each). A lower expression of SMPDL3B was associated with a shorter overall survival (OS) (p = 0.005) in long term follow-up. A SMPDL3B overexpression in PCa tissue was confirmed in the validation cohorts (p < 0.001 each). In the TCGA patients with low SMPDL3B expression, biochemical recurrence-free survival (p = 0.011) and progression-free interval (p < 0.001) were shorter. Knockdown of SMPDL3B impaired PC3 cell migration but not proliferation (p = 0.0081). In summary, SMPLD3B is highly overexpressed in PCa tissue, is inversely associated with localized PCa prognosis, and impairs PCa cell migration.

In VUAS tissue, severe alterations on mRNA and miRNA level are found. These consistent changes give insights into the pathogenesis of VUAS after radical prostatectomy and point to future options for transcriptomics-based risk stratification and targeted therapies.

No abstract

No abstract