The eye is a common site of granulomatosis with polyangiitis. A collaborative study.

BACKGROUND: Ocular manifestations of granulomatosis with polyangiitis (GPA) have been reported in a limited number of studies and with largely variable frequency. Here we report on the clinical, diagnostic, and therapeutic features of a cohort of 63 GPA patients, with particular regard to 22 of them with ophthalmic involvement (35%). METHODS: Clinical manifestations, results of immunological findings, histopathological pictures, imaging data, Birmingham Vasculitis Activity Score, therapeutic regimens, and outcomes were retrospectively analyzed. At diagnosis, in addition to a structured clinical assessment, all patients underwent a comprehensive ophthalmologic examination. RESULTS: The most frequently involved organs were kidneys, lungs, ear/nose/throat, and eyes. Ocular manifestations were bilateral in 32%. The three most commonly diagnosed ophthalmologic manifestations were scleritis (36%), retro-orbital pseudotumor or orbital mass (23%), and episcleritis (13%). Ocular and systemic involvement were simultaneously present at onset in 41% of the patients; systemic involvement was followed by ocular lesions in 36%; ocular inflammation was followed by systemic manifestations in 18%; and an orbital mass in the absence of systemic disease characterized 5%. Glucocorticoids plus cyclophosphamide and glucocorticoids plus rituximab were the combined therapies most frequently employed during remission induction and remission maintenance, respectively. Persistent ophthalmologic and extra-ocular remissions were achieved in 77 and 64% of the patients, respectively. One to three systemic relapses were diagnosed in 7 patients (31.8%). At the last follow-up, a visual outcome 20/40 or better in 31 (70%) of 44 eyes was determined. CONCLUSIONS: The eye was involved in over one third of our patients with GPA. Increased awareness, early diagnosis, and multi-specialty collaboration are critical in achieving a favorable outcome of GPA.

Impact of the systematic introduction of tomosynthesis on breast biopsies: 10 years of results.

Digital Breast Tomosynthesis (DBT) is a cutting-edge technology introduced in recent years as an in-depth analysis of breast cancer diagnostics. Compared with 2D Full-Field Digital Mammography, DBT has demonstrated greater sensitivity and specificity in detecting breast tumors. This work aims to quantitatively evaluate the impact of the systematic introduction of DBT in terms of Biopsy Rate and Positive Predictive Values for the number of biopsies performed (PPV-3). For this purpose, we collected 69,384 mammograms and 7894 biopsies, of which 6484 were Core Biopsies and 1410 were stereotactic Vacuum-assisted Breast Biopsies (VABBs), performed on female patients afferent to the Breast Unit of the Istituto Tumori "Giovanni Paolo II" of Bari from 2012 to 2021, thus, in the period before, during and after the systematic introduction of DBT. Linear regression analysis was then implemented to investigate how the Biopsy Rate had changed over the 10 year screening. The next step was to focus on VABBs, which were generally performed during in-depth examinations of mammogram detected lesions. Finally, three radiologists from the institute's Breast Unit underwent a comparative study to ascertain their performances in terms of breast cancer detection rates before and after the introduction of DBT. As a result, it was demonstrated that both the overall Biopsy Rate and the VABBs Biopsy Rate significantly decreased following the introduction of DBT, with the diagnosis of an equal number of tumors. Besides, no statistically significant differences were observed among the three operators evaluated. In conclusion, this work highlights how the systematic introduction of DBT has significantly impacted the breast cancer diagnostic procedure, by improving the diagnostic quality and thereby reducing needless biopsies, resulting in a consequent reduction in costs.

A Machine Learning Approach to Simulate Gene Expression and Infer Gene Regulatory Networks.

The ability to simulate gene expression and infer gene regulatory networks has vast potential applications in various fields, including medicine, agriculture, and environmental science. In recent years, machine learning approaches to simulate gene expression and infer gene regulatory networks have gained significant attention as a promising area of research. By simulating gene expression, we can gain insights into the complex mechanisms that control gene expression and how they are affected by various environmental factors. This knowledge can be used to develop new treatments for genetic diseases, improve crop yields, and better understand the evolution of species. In this article, we address this issue by focusing on a novel method capable of simulating the gene expression regulation of a group of genes and their mutual interactions. Our framework enables us to simulate the regulation of gene expression in response to alterations or perturbations that can affect the expression of a gene. We use both artificial and real benchmarks to empirically evaluate the effectiveness of our methodology. Furthermore, we compare our method with existing ones to understand its advantages and disadvantages. We also present future ideas for improvement to enhance the effectiveness of our method. Overall, our approach has the potential to greatly improve the field of gene expression simulation and gene regulatory network inference, possibly leading to significant advancements in genetics.

Ocular sarcoidosis: clinical experience and recent pathogenetic and therapeutic advancements.

PURPOSE: To describe the ocular manifestations in a cohort of patients with systemic sarcoidosis (SS). Recent advances in the pathophysiology, diagnosis, and therapy of SS are also discussed. METHODS: Data from 115 Italian patients diagnosed between 2005 and 2016 were retrospectively reviewed. All but the first 17 patients underwent a comprehensive ophthalmologic examination. The diagnosis was based on clinical features, the demonstration of non-caseating granulomas in biopsies from involved organs, and multiple imaging techniques. Data on broncho-alveolar lavage fluid analysis, calcemia, calciuria, serum angiotensin-converting enzyme levels and soluble interleukin-2 receptor levels were retrieved when available. RESULTS: Ocular involvement, detected in 33 patients (28.7%), was bilateral in 29 (87.9%) and the presenting feature in 13 (39.4%). Anterior uveitis was diagnosed in 12 patients (36.4%), Löfgren syndrome and uveoparotid fever in one patient each (3%), intermediate uveitis in 3 patients (9.1%), posterior uveitis in 7 (21.2%), and panuveitis in 9 (27.3%). First-line therapy consisted of corticosteroids, administered as eyedrops (10 patients), sub-Tenon's injections (1 patient), intravitreal implants (9 patients), or systemically (23 patients). Second-line therapy consisted of steroid-sparing immunosuppressants, including methotrexate (10 patients) and azathioprine (10 patients). Based on pathogenetic indications that tumor necrosis factor (TNF)-α is a central mediator of granuloma formation, adalimumab, targeting TNF-α, was employed in 6 patients as a third-line agent for severe/refractory chronic sarcoidosis. CONCLUSION: Uveitis of protean type, onset, duration, and course remains the most frequent ocular manifestation of SS. Diagnostic and therapeutic advancements have remarkably improved the overall visual prognosis. An ophthalmologist should be a constant component in the multidisciplinary approach to the treatment of this often challenging but intriguing disease.

Spatial distribution of mast cells and macrophages around tumor glands in human breast ductal carcinoma.

Macrophages and mast cells are usually present in the tumor microenvironment and play an important role as regulators of inflammation, immunological response and angiogenesis in the tumor microenvironment. In this study, we have evaluated macrophage, mast cell, and microvessel density in a selected group of different grade of invasive breast carcinoma tumor specimens. Furthermore, we have investigated the pattern of distribution of CD68-positive macrophages and tryptase-positive mast cells around tumor glands. Results have shown that: A) Macrophages are more numerous in G2 and G3 breast cancer stages respect to controls, the per cent of macrophages in G1 samples was comparable to the controls, and the spatial relationship between macrophages and glands (as indicated by the mean cell-to-gland distance) correlated with CD31-positive vessels. B) Mast cells in G2 and G3 tumor specimens show a significant increase in their number as compared to control samples, and their spatial distribution around the glands did not show any significant difference among groups. Overall, the results of this study confirm the important role of macrophages and mast cells in tumor progression and angiogenesis in human ductal breast cancer, and pointed out the spatial relationship between tumor macrophages and glands, and its correlation with microvascular density.

Alcoholic beverages and carbonated soft drinks: consumption and gastrointestinal cancer risks.

Alcoholic beverages (ABs) and carbonated soft drinks (CSDs) are widely consumed worldwide. Given the high consumption of these beverages, the scientific community has increased its focus on their health impact. There is epidemiological evidence of a causal association between AB intake and digestive cancer, but the role of alcohol in determining cancer is not fully defined. Experimental studies have so far identified multiple mechanisms involved in carcinogenesis; ethanol itself is not carcinogenic but available data suggest that acetaldehyde (AA) and reactive oxygen species-both products of ethanol metabolism-have a genotoxic effect promoting carcinogenesis. Other carcinogenetic mechanisms include nutritional deficits, changes in DNA methylation, and impaired immune surveillance. As CSDs are often suspected to cause certain gastrointestinal disorders, consequently, some researchers have hypothesized their involvement in gastrointestinal cancers. Of all the ingredients, carbon dioxide is prevalently involved in the alteration of gastrointestinal physiology by a direct mucosal effect and indirect effects mediated by the mechanical pressure determined by gas. The role of sugar or artificial sweeteners is also debated as factors involved in the carcinogenic processes. However, several surveys have failed to show any associations between CSDs and esophageal, gastric, or colon cancers. On the other hand, a slight correlation between risk of pancreatic cancer and CSD consumption has been found.

Native Aortic Valve Resection Using a Novel Blade-Based Device.

OBJECTIVE: The aim of this study was to validate the use of a new resection device in patient candidates for surgical aortic valve replacement. We evaluated the efficacy of this new circular blade to resect the aortic valve and the efficacy to collect the debris during the resection. METHODS: For this study, a single size instrument was used, with an external diameter of 22 mm, and patients were selected on the basis of the preoperative assessment of the aortic diameters. RESULTS: From October 2018 to June 2019, 10 patient candidates for surgical aortic valve replacement were selected to undergo native aortic valve resection using a new device, before surgical valve implantation. The mean age of the patients was 74 ± 7.6 years, and 8 of 10 were male. The mean aortic annulus diameter, measured before the procedure, was 25.7 ± 1.57 mm. The resection was complete in 9 (90%) patients. In 1 patient, due to an imprecise positioning of the device, the valve resection was partial. None of the patients showed signs or symptoms due to debris embolism. In all patients, the postoperative course was uneventful. CONCLUSIONS: These preliminary results show that resection of the aortic valve using a circular foldable blade is feasible. This prototype, used during conventional surgery even through a small incision, provided an efficient tool to easily resect the valve without debris release.

Value of multidetector computed tomography angiography in severe non-variceal upper gastrointestinal bleeding: a retrospective study in a referral bleeding unit.

BACKGROUND: Non-variceal upper gastrointestinal bleeding is a common gastroenterological emergency associated with significant morbidity and mortality. Upper gastrointestinal endoscopy is currently recommended as the gold standard modality for both diagnosis and treatment. As historically played a limited role in the diagnosis of acute non-variceal upper gastrointestinal bleeding, multidetector-row computed tomography angiography is emerging as a promising tool in the diagnosis of non-variceal upper gastrointestinal bleeding, especially for severe cases. However, to date, evidence concerning the role of multidetector-row computed tomography angiography in the non-variceal upper gastrointestinal bleeding diagnosis is still lacking. AIM: The purpose of this study was to retrospectively investigate the diagnostic performance of emergent multidetector-row computed tomography angiography performed prior to any diagnostic modality or following urgent upper endoscopy to identify the status, the site, and the underlying etiology of severe non-variceal upper gastrointestinal bleeding. METHODS: Institutional databases were reviewed in order to identify severe acute non-variceal upper gastrointestinal bleeding patients who were admitted to our bleeding unit and were referred for emergent multidetector-row computed tomography angiography prior to any hemostatic treatment (< 3 h) or following (< 3 h) endoscopy, between December 2019 and October 2022. The study aim was to evaluate the diagnostic performance of multidetector-row computed tomography angiography to detect the status, the site, and the etiology of severe non-variceal upper gastrointestinal bleeding with endoscopy, digital subtraction angiography, surgery, pathology, or a combination of them as reference standards. RESULTS: A total of 68 patients (38 men, median age 69 years [range 25-96]) were enrolled. The overall multidetector-row computed tomography angiography sensitivity, specificity, and accuracy to diagnose bleeding status were 77.8% (95% CI: 65.5-87.3), 40% (95% CI: 5.3-85.3), and 75% (95% CI: 63.0-84.7), respectively. Finally, the overall multidetector-row computed tomography angiography sensitivity to identify the bleeding site and the bleeding etiology were 92.4% (95% CI: 83.2-97.5) and 79% (95% CI: 66.8-88.3), respectively. CONCLUSION: Although esophagogastroduodenoscopy is the mainstay in the diagnosis and treatment of most non-variceal upper gastrointestinal bleeding cases, multidetector-row computed tomography angiography seems to be a feasible and effective modality in detecting the site, the status, and the etiology of severe acute non-variceal upper gastrointestinal bleeding. It may play a crucial role in the management of selected cases of non-variceal upper gastrointestinal bleeding, especially those clinically severe and/or secondary to rare and extraordinary rare sources, effectively guiding timing and type of treatment. However, further large prospective studies are needed to clarify the role of multidetector-row computed tomography angiography in the diagnostic process of acute non-variceal upper gastrointestinal bleeding.

Explainable prediction model for the human papillomavirus status in patients with oropharyngeal squamous cell carcinoma using CNN on CT images.

Several studies have emphasised how positive and negative human papillomavirus (HPV+ and HPV-, respectively) oropharyngeal squamous cell carcinoma (OPSCC) has distinct molecular profiles, tumor characteristics, and disease outcomes. Different radiomics-based prediction models have been proposed, by also using innovative techniques such as Convolutional Neural Networks (CNNs). Although some of these models reached encouraging predictive performances, there evidence explaining the role of radiomic features in achieving a specific outcome is scarce. In this paper, we propose some preliminary results related to an explainable CNN-based model to predict HPV status in OPSCC patients. We extracted the Gross Tumor Volume (GTV) of pre-treatment CT images related to 499 patients (356 HPV+ and 143 HPV-) included into the OPC-Radiomics public dataset to train an end-to-end Inception-V3 CNN architecture. We also collected a multicentric dataset consisting of 92 patients (43 HPV+ , 49 HPV-), which was employed as an independent test set. Finally, we applied Gradient-weighted Class Activation Mapping (Grad-CAM) technique to highlight the most informative areas with respect to the predicted outcome. The proposed model reached an AUC value of 73.50% on the independent test. As a result of the Grad-CAM algorithm, the most informative areas related to the correctly classified HPV+ patients were located into the intratumoral area. Conversely, the most important areas referred to the tumor edges. Finally, since the proposed model provided additional information with respect to the accuracy of the classification given by the visualization of the areas of greatest interest for predictive purposes for each case examined, it could contribute to increase confidence in using computer-based predictive models in the actual clinical practice.

Merkel Cell Polyomavirus targets SET/PP2A complex to promote cellular proliferation and migration.

Merkel Cell Carcinoma (MCC) is a rare neuroendocrine skin cancer. In our previous work, we decoded genes specifically deregulated by MCPyV early genes as opposed to other polyomaviruses and established functional importance of NDRG1 in inhibiting cellular proliferation and migration in MCC. In the present work, we found the SET protein, (I2PP2A, intrinsic inhibitor of PP2A) upstream of NDRG1 which was modulated by MCPyV early genes, both in hTERT-HK-MCPyV and MCPyV-positive (+) MCC cell lines. Additionally, MCC dermal tumour nodule tissues showed strong SET expression. Inhibition of the SET-PP2A interaction in hTERT-HK-MCPyV using the small molecule inhibitor, FTY720, increased NDRG1 expression and inhibited cell cycle regulators, cyclinD1 and CDK2. SET inhibition by shRNA and FTY720 also decreased cell proliferation and colony formation in MCPyV(+) MCC cells. Overall, these results pave a path for use of drugs targeting SET protein for the treatment of MCC.

Epstein-Barr Virus DNA Is Associated With Conjunctival Squamous Cell Carcinomas: A Case-Control Study From Zimbabwe.

Incidence of conjunctival squamous cell carcinoma (cSCC) in Zimbabwe is >30-fold higher than the global average. cSCC risk is notably higher among people with human immunodeficiency virus, implicating impaired immune response and a yet unknown infectious etiology. Formalin-fixed, paraffin-embedded blocks from Zimbabwe, comprising conjunctival precancer (n = 78), invasive cSCC cases (n = 148) and nonmalignant eye lesions (n = 119), were tested for multiple DNA viruses using Luminex bead-based technology. Epstein-Barr virus (EBV) type 1 positivity was strongly associated with cSCC diagnosis (adjusted odds ratio [aOR], 5.6 [95% confidence interval {CI}, 3.0-10.4) and marginally associated with precancer (aOR, 2.1 [95% CI, 1.0-4.5]). On analyzing EBV transcriptional activity with any of LMP1, EBNA1, and BZLF1, RNA transcripts were detected in 5 of 112 controls, 3 of 67 precancers, and 10 of 139 cases and none were associated with conjunctival case status. Our EBV DNA data suggest that EBV may play a role in cSCC. However, the low detection rate of EBV RNA supports further investigation to infer causality.

Changes in angiogenesis and hypoxia-inducible factor-1α protein expression in relapsed/refractory indolent non-Hodgkin lymphomas.

Angiogenesis is involved in the pathogenesis and progression of non-Hodgkin lymphomas (NHL), and hypoxia-inducible factor-1α (HIF-1α, also termed HIF1A) might contribute to this process. Currently, there is no direct evidence that the clinical progression of indolent NHL is associated with angiogenesis, and the expression of HIF-1α at recurrence is unknown. Matched lymph node biopsies at diagnosis and recurrence of relapsed/refractory indolent NHL patients were analysed by immunohistochemical and morphometric analysis. We observed an increased vascular network and HIF-1α protein expression in the second biopsy, providing direct evidence that angiogenesis is an essential process for disease progression.

Association of D-dimer levels with all-cause mortality in a healthy adult population: findings from the MOLI-SANI study.

Elevated D-dimer levels are reportedly associated with higher risk of total mortality in patients with different diseases. We investigated whether a similar association could be found in a large, apparently healthy population. A large sample of individuals (N=17,359, 47% men, age ≥ 35 years) free of clinically recognized cardiovascular and cancer disease, for whom baseline D-dimer level was available, were studied within the MOLI-SANI cohort, randomly recruited from the general adult population of Southern Italy. The cohort was followed for a median of 4.2 years (73,807 person-years). D-dimer was measured in fresh citrated plasma by an automated latex-enhanced immunoassay. Hazard ratios were calculated using three Cox-proportional hazard models. Two hundred and eighty deaths were recorded. When modeled as a continuous variable, D-dimer level at baseline showed a non-linear association with mortality, whose incidence increased only in the upper quartile of the distribution (D-dimer ≥ 221 ng/mL). Thus, the group of individuals with D-dimer <221 ng/mL (75% of the population) acted as the reference group, while the remaining individuals were subdivided in tertiles and compared with the former group. Multivariable hazard ratios for mortality were 1.06, 1.45 and 1.97, respectively (P for trend <0.0001) across the three categories of increasing D-dimer concentration. The association was slightly attenuated, but still highly significant (P for trend 0.0002), after further adjustment for white blood cell count and C-reactive protein. In conclusion, Elevated D-dimer levels were independently associated with increased risk of death from any cause in an apparently healthy adult population.

T-wave axis deviation and left ventricular hypertrophy interaction in diabetes and hypertension.

Electrocardiographic signs of left ventricular hypertrophy (ECG-LVH) and T-wave axis (TA) deviation are independent predictors of fatal and non fatal events. We assessed the prevalence of ECG-LVH, TA abnormalities and their combination according to the presence or absence of diabetes and/or hypertension in a large sample of the adult general Italian population. Data from 10,184 women (54 ± 11 years) and 8775 men (54 ± 11 years) were analyzed from the Moli-sani cohort, a database of randomly recruited adults (age >35) from the general population of Molise, a central region of Italy that includes collection of standard 12-lead resting ECG. Subjects with previous myocardial infarction, angina, cerebrovascular disease or left bundle brunch block or missing values for TA or ECG-LVH have been excluded. TA was measured from the standard 12-lead ECG and it was defined as the rotation of the T wave in the frontal plane as computed by a proprietary algorithm (CalECG/Bravo, AMPS-LLC, NY). ECG-LVH was defined as Sokolow Lyon voltage (SLv) >35 mm or Cornell voltage duration Product (CP) >= 2440 mm*ms. Among subjects with ECG-LVH, prevalence of hypertension was 59.0% and 49.7%, respectively for men and women, whereas that of diabetes was 10.7% and 5.7%. In hypertensives, TA was normal in 72.3% of subjects, borderline in 24.8% and abnormal in 2.9%. In diabetics, TA was normal in 70.4% of subjects, borderline in 26.5% and abnormal in 3.1%. In both hypertensive and diabetic subjects, the prevalence of ECG-LVH, was significantly greater in subjects with borderline or abnormal TA. Hypertension was an independent predictor of abnormal TA (odd ratio: 1.38, P = .025). These results suggest that hypertension might play a relevant role in the pathogenesis of TA deviation.

Discovering anomalies in big data: a review focused on the application of metaheuristics and machine learning techniques.

With the increase in available data from computer systems and their security threats, interest in anomaly detection has increased as well in recent years. The need to diagnose faults and cyberattacks has also focused scientific research on the automated classification of outliers in big data, as manual labeling is difficult in practice due to their huge volumes. The results obtained from data analysis can be used to generate alarms that anticipate anomalies and thus prevent system failures and attacks. Therefore, anomaly detection has the purpose of reducing maintenance costs as well as making decisions based on reports. During the last decade, the approaches proposed in the literature to classify unknown anomalies in log analysis, process analysis, and time series have been mainly based on machine learning and deep learning techniques. In this study, we provide an overview of current state-of-the-art methodologies, highlighting their advantages and disadvantages and the new challenges. In particular, we will see that there is no absolute best method, i.e., for any given dataset a different method may achieve the best result. Finally, we describe how the use of metaheuristics within machine learning algorithms makes it possible to have more robust and efficient tools.

Newly designed OTS Clip for preventing fully-covered self-expandable metal stent migration in the gastrointestinal tract.

Background and study aims Fully-covered self-expandable metal stents (FCSEMS) are frequently used for endoscopic management of gastrointestinal lesions. However, stent migration occurs in up to one-third of patients. Different tools are used to anchor stents to prevent migration. A specifically designed over-the-scope device (Stentfix OTS Clip system) was recently introduced to prevent fully covered SEMS migration in the gastrointestinal tract. The study aimed to evaluate technical success and stent migration rates with the Stentfix device. Patients and methods Data were collected from consecutive patients at four participating centers who were at high risk of FCSEMS migration and in whom the anchoring system was used to prevent migration. Results A total of 31 patients were enrolled. Technically successful clip placement was achieved in all cases. At follow-up, the distal part of the device dislocated from the duodenum into the antrum at 3 days in one patient, accounting for a 3.2 % (95 % CI = 0-9.4) rate of stent migration. The underlying lesion being treated healed in all patients, but 10 patients died before stent removal due to neoplastic progression. Conclusions A dedicated over-the-scope stent fixation device appears to be safe and effective in preventing fully-covered SEMS migration through the gastrointestinal tract.

Over-the-scope Technique for Partially Covered Self-expandable Metal Stent Placement to Treat Duodenal Perforation during EUS.

BACKGROUND AND AIMS: Although endoscopic ultrasound (EUS) is generally safe, duodenal perforation may occur during the procedure. When the iatrogenic break is wide, endoscopic positioning of a covered self- expandable metal stent is indicated to avoid a more invasive surgical approach. We evaluated the efficacy of the 'over-the-scope stenting' (OTSS) technique to treat iatrogenic duodenal perforations occurred during EUS. METHODS: Data of patients with large iatrogenic duodenal perforations treated with OTSS procedure were collected in 5 centers. Technical success was defined as a correct stent placement on the perforation site, and clinical success as complete healing of the duodenal leak at stent removal 3 weeks later. RESULTS: A total of 15 (7 males; median age: 78 years, range 47-91) patients were included in this series. A correct stent positioning was achieved in all cases (technical success: 100%), and the perforation was healed in all, apart from one patient at stent removal (clinical success: 93%). This patient was successfully treated with a novel stent placement. No immediate post-procedural adverse events occurred, and no need for emergency surgery was recorded. In one (6.7%) patient, stent migration occurred 10 days after positioning, and it was spontaneously expulsed with stool movement without complications two days later. CONCLUSIONS: Our data showed that the OTTS technique for partially covered self-expandable metal stent placement is feasible, safe and effective to tread large iatrogenic duodenal perforation occuring during EUS.

Tumor Cellularity Assessment of Breast Histopathological Slides via Instance Segmentation and Pathomic Features Explainability.

The segmentation and classification of cell nuclei are pivotal steps in the pipelines for the analysis of bioimages. Deep learning (DL) approaches are leading the digital pathology field in the context of nuclei detection and classification. Nevertheless, the features that are exploited by DL models to make their predictions are difficult to interpret, hindering the deployment of such methods in clinical practice. On the other hand, pathomic features can be linked to an easier description of the characteristics exploited by the classifiers for making the final predictions. Thus, in this work, we developed an explainable computer-aided diagnosis (CAD) system that can be used to support pathologists in the evaluation of tumor cellularity in breast histopathological slides. In particular, we compared an end-to-end DL approach that exploits the Mask R-CNN instance segmentation architecture with a two steps pipeline, where the features are extracted while considering the morphological and textural characteristics of the cell nuclei. Classifiers that are based on support vector machines and artificial neural networks are trained on top of these features in order to discriminate between tumor and non-tumor nuclei. Afterwards, the SHAP (Shapley additive explanations) explainable artificial intelligence technique was employed to perform a feature importance analysis, which led to an understanding of the features processed by the machine learning models for making their decisions. An expert pathologist validated the employed feature set, corroborating the clinical usability of the model. Even though the models resulting from the two-stage pipeline are slightly less accurate than those of the end-to-end approach, the interpretability of their features is clearer and may help build trust for pathologists to adopt artificial intelligence-based CAD systems in their clinical workflow. To further show the validity of the proposed approach, it has been tested on an external validation dataset, which was collected from IRCCS Istituto Tumori "Giovanni Paolo II" and made publicly available to ease research concerning the quantification of tumor cellularity.

Endoscopic Submucosal Dissection for Subepithelial Tumor Treatment in the Upper Digestive Tract: A Western, Multicenter Study.

Background/Aims: Endoscopic submucosal dissection (ESD) has been proposed for removal of gastrointestinal subepithelial tumors (GI-SETs), but data are still scanty. This study aimed to report a case series from a western country. Patients and Methods: Data of patients with upper GI-SETs suitable for ESD removal observed in 4 centers were retrospectively reviewed. Before endoscopic procedure, the lesion was characterized by endosonographic evaluation, histology, and CT scan. The en bloc resection and the R0 resection rates were calculated, as well as incidence of complications, and the 1-year follow-up was reported. Results: Data of 84 patients with esophageal (N = 13), gastric (N = 61), and duodenal (N = 10) GI-SETs were collected. The mean diameter of lesions was 26 mm (range: 12-110 mm). There were 17 gastrointestinal stromal tumors, 12 neuroendocrine tumors, 35 leiomyomas, 18 lipomas, and 2 hamartomas. En bloc and R0 resection were achieved in 83 (98.8%) and in 80 (95.2%) patients, respectively. Overall, a complication occurred in 11 (13.1%) patients, including bleeding (N = 7) and perforation (N = 4). Endoscopic approach was successful in all bleedings, but 1 patient who required radiological embolization, and in 2 perforations, while surgery was performed in the other patients. Overall, a surgical approach was eventually needed in 5 (5.9%), including 3 in whom R0 resection failed and 2 with perforation. Conclusions: Our study found that ESD may be an effective and safe alternative to surgical intervention for both benign and localized malignant GI-SETs.

Introdução/objetivos: A dissecção endoscópica da submucosa (ESD) tem sido proposta para a exérese de tumores subepiteliais gastrointestinais (GI-SETs), embora a literatura seja escassa. Este estudo teve como objetivo reportar uma série de casos de um país ocidental. Métodos: Coorte retrospectiva incluindo doentes com SETs do tubo digestivo superior submetidos a ESD em 4 centros (1 ano de follow-up). Antes do procedimento, a lesão foi caracterizada por ecoendoscopia, histologia e tomografia computadorizada. Foram avaliadas as taxas de ressecção em bloco e R0, bem como a incidência de complicações. Resultados: Incluídos 84 doentes com GI-SETs esofágicos (N = 13), gástricos (N = 61) e duodenais (N = 10). O diâmetro médio das lesões foi de 26 mm (intervalo 12­110 mm) ­ 17 tumores do estroma gastrointestinal, 12 tumores neuroendócrinos, 35 leiomiomas, 18 lipomas e 2 hamartomas. A resseção foi em bloco e R0 em 83 (98.8%) e em 80 (95.2%) doentes, respectivamente. Globalmente, ocorreram complicações em 11 (13.1%) doentes, incluindo hemorragia (N = 7) e perfuração (N = 4). A terapêutica endoscópica foi eficaz em todas as hemorragias exceto em 1 doente que necessitou de embolização radiológica e em 2 perfurações (submetidas a cirurgia). No geral, a abordagem cirúrgica foi necessária em 5 (5.9%) ­ 3 doentes com resseção R1 e 2 com perfuração. Conclusões: A ESD pode ser uma alternativa eficaz e segura à intervenção cirúrgica para GI-SETs benignos e malignos localizados.

The role of unpaired image-to-image translation for stain color normalization in colorectal cancer histology classification.

BACKGROUND: Histological assessment of colorectal cancer (CRC) tissue is a crucial and demanding task for pathologists. Unfortunately, manual annotation by trained specialists is a burdensome operation, which suffers from problems like intra- and inter-pathologist variability. Computational models are revolutionizing the Digital Pathology field, offering reliable and fast approaches for challenges like tissue segmentation and classification. With this respect, an important obstacle to overcome consists in stain color variations among different laboratories, which can decrease the performance of classifiers. In this work, we investigated the role of Unpaired Image-to-Image Translation (UI2IT) models for stain color normalization in CRC histology and compared to classical normalization techniques for Hematoxylin-Eosin (H&E) images. METHODS: Five Deep Learning normalization models based on Generative Adversarial Networks (GANs) belonging to the UI2IT paradigm have been thoroughly compared to realize a robust stain color normalization pipeline. To avoid the need for training a style transfer GAN between each pair of data domains, in this paper we introduce the concept of training by exploiting a meta-domain, which contains data coming from a wide variety of laboratories. The proposed framework enables a huge saving in terms of training time, by allowing to train a single image normalization model for a target laboratory. To prove the applicability of the proposed workflow in the clinical practice, we conceived a novel perceptive quality measure, which we defined as Pathologist Perceptive Quality (PPQ). The second stage involved the classification of tissue types in CRC histology, where deep features extracted from Convolutional Neural Networks have been exploited to realize a Computer-Aided Diagnosis system based on a Support Vector Machine (SVM). To prove the reliability of the system on new data, an external validation set composed of N = 15,857 tiles has been collected at IRCCS Istituto Tumori "Giovanni Paolo II". RESULTS: The exploitation of a meta-domain consented to train normalization models that allowed achieving better classification results than normalization models explicitly trained on the source domain. PPQ metric has been found correlated to quality of distributions (Fréchet Inception Distance - FID) and to similarity of the transformed image to the original one (Learned Perceptual Image Patch Similarity - LPIPS), thus showing that GAN quality measures introduced in natural image processing tasks can be linked to pathologist evaluation of H&E images. Furthermore, FID has been found correlated to accuracies of the downstream classifiers. The SVM trained on DenseNet201 features allowed to obtain the highest classification results in all configurations. The normalization method based on the fast variant of CUT (Contrastive Unpaired Translation), FastCUT, trained with the meta-domain paradigm, allowed to achieve the best classification result for the downstream task and, correspondingly, showed the highest FID on the classification dataset. CONCLUSIONS: Stain color normalization is a difficult but fundamental problem in the histopathological setting. Several measures should be considered for properly assessing normalization methods, so that they can be introduced in the clinical practice. UI2IT frameworks offer a powerful and effective way to perform the normalization process, providing realistic images with proper colorization, unlike traditional normalization methods that introduce color artifacts. By adopting the proposed meta-domain framework, the training time can be reduced, and the accuracy of downstream classifiers can be increased.