Malignant melanoma is a severe and aggressive type of skin cancer, with a rapid decrease in survival rate if not diagnosed and treated at an early stage. Histopathological examination of hematoxylin and eosin stained tissue biopsies under a light microscope is currently the gold standard for diagnosis. However, this manual examination is a difficult and timeconsuming task, and diagnosis is often subject to intra-and inter-observer variability. With more pathology departments starting to convert conventional glass slides into digital resources, a Computer Aided Diagnostic (CAD) system that can automate part of the diagnostic process will help address these challenges. It is expected to reduce examination time, increase diagnostic accuracy, and reduce diagnostic variations. An important initial step in developing such a system is an automated epidermis segmentation algorithm, since several important diagnostic factors are within or seen relatively to the epidermis' location. In this paper, we propose a new epidermis segmentation technique built on Convolutional Neural Networks. We trained an U-net based architecture end-to-end, with ∼ 380k overlapping high resolution image patches at 512 × 512 pixels, extracted and augmented from 36 digitized histopathological images from two different clinical sites, to discriminate pixels as either epidermal or non-epidermal. The proposed technique was evaluated on 33 test images, where we achieved a mean Positive Predictive Value at 0.89 ± 0.16 , Sensitivity at 0.92 ± 0.1 , Dice Similarity Coefficient at 0.89 ± 0.13 and a Matthews Correlation Coefficient at 0.89 ± 0.11 , showing a superior performance when compared to existing techniques. Our algorithm also proves to be robust to variations in staining, tissue thickness and laboratory pre-processing.

Phosphohistone H3 assessed proliferation has strong prognostic value. Lymph vessel invasion by D2-40 is also prognostic, but D2-40 þ myoepithelial expression in small ducts completely filled by solid-pattern ductal carcinoma in situ can mimic lymphovascular invasion. As myoepithelial cells are also p63 positive, we have investigated whether lymph vessel invasion identified by combined D2-40/p63 is stronger prognostically than by D2-40 alone, and whether it has independent prognostic value to phosphohistone H3. In 240 operable T 1-2 N 0 M 0 node negative invasive breast cancer patients o71 years, phosphohistone H3 was determined by quantitative immunohistochemistry and lymph vessel invasion by D2-40/p63 double immunostaining. Correlation analysis between the clinico-pathologic factors and lymph vessel invasion, and univariate and multivariate prognostic survival analysis were performed. With median 117 (range: 12-192) months follow-up, 36 patients (15%) developed and 28 (12%) died of distant metastases. Ten of the 61 patients (16%) with cancer cells surrounded by D2-40 were p63 positive and none of these 'false lymph vessel invasion' recurred. D2-40 þ / p63À lymph vessel invasion occurred in 51/239 (21%) cases and correlated with grade, mitotic activity index, phosphohistone H3, ER, cytokeratin14, and HER2. D2-40 þ /p63À lymph vessel invasion was strongly prognostic, but far more in women Z55 than those o55 years (Po0.0001 and 0.04). With multivariate analysis, phosphohistone H3 proliferation was the strongest single prognosticator. Lymph vessel invasion had additional prognostic value to phosphohistone H3 only in women Z55. This group of patients, without/with lymph vessel invasion, had 10-year survival rates of 83 and 50%, respectively (hazard ratio-lymph vessel invasion ¼ 3.0, P ¼ 0.04; hazard ratio-phosphohistone H3 ¼ 6.9, P ¼ 0.002). Where age was o55 years, only phosphohistone H3 had independent prognostic value. Combinations of other features had no additional value. In conclusion, T 1-2 N 0 M 0 invasive breast cancer patients Z55 years with phosphohistone H3Z13, D2-40 þ /p63À defined lymph vessel invasion identifies a subgroup with a high risk of distant metastases.

PurposePrimary mesenchymal tumors of the pancreas are rare, with leiomyosarcomas the most encountered entities among the pancreatic sarcomas. With few exceptions, single case reports published over the last six decades constitute the entire scientific literature on this topic. Thus, evidence regarding clinical decision-making is scant.MethodsBased on a case report and an extensive literature search in PubMed, we discuss the clinical aspects and current management of this rare malignancy.ResultsWe identified only two papers with more than a single case presentation; these institutional patient series were limited to five and nine patients. Additionally, a few papers sought to summarize the individual case reports published in the English and/or Chinese language. The clinical presentation is rather non-specific. Moreover, modern imaging modalities are insufficiently accurate to diagnose leiomyosarcoma of the pancreas. Treatment goals include a complete resection with free margins. Proper morphologic examination using immunohistochemistry and the application of a grading system are clinically important for prognostication. The efficacy of adjuvant treatments has not been established.ConclusionPrimary pancreatic leiomyosarcoma is extremely rare, and the scientific literature is primarily based on single case reports. Conclusions on management and prognosis should be drawn with caution. A multidisciplinary team consultation is warranted to discuss a thorough individual treatment plan based on the available scientific literature, despite its low evidence level.

No abstract

No abstract