ECP 2022 Abstract Book

Virchows Archiv (2022) 481 (Suppl 1):S1–S364 13 Cavalcante Alves, J. Vieira e Freitas Lourenço, L. Bedê Bruno, M. Macêdo Militão Mendonça, G. Benevides Lima *University of Fortaleza, Brazil Background & objectives: Maternal infections by some microor- ganisms include several unfavourable outcomes. Therefore, the study aimed to analyse anatomopathological alterations in ovular remains of spontaneous abortions of pregnant women with positive IgM serology for dengue, chikungunya, zika or positive treponemal test for syphilis. Methods: From June 2020/January 2021, 259 cases of pregnant women in spontaneous abortion were admitted, in which blood samples were collected for IgM ELISA serology for zika, dengue, chikungunya and rapid test for syphilis. Ovular remains were stored and fixed in 10% buffered formalin and sent for anatomopathologi- cal evaluation. Results: Among the cases admitted, regarding IgM, were reagent 17 (7.8%) in dengue, 32 (14.1%) in chikungunya and 9 (3.8%) in zika. In the rapid test for syphilis, 15 (5.9%) tested positive. In Dengue, fibrinoid deposits (p=0.000), syncytial nodes (p=0.001) and calcifications (p=0.034) showed statistical significance with the outcome. In Chikungunya, fibrin thrombus showed borderline statistical significance (p=0.053). In Zika, the prevalent change was poor vascularity, however no findings showed statistical significance. In syphilis, calcifications (p=0.020) and haemorrhagic spots (p=0.057) showed statistical significance, the latter being borderline. In all ovular remnants, it was found a rich inflammatory environment. Conclusion: The results found in this investigation suggest a direct effect of the aforementioned viral and bacterial infections during the gestational period, particularly in the first 12 weeks of gestation, leading to unfortunate outcomes. Therefore, further studies on the subject are needed to understand the pathogenesis of the identified alterations and their repercussions, since, as observed in other studies, they may be related to early termination of pregnancy or to some complication related to the foetus. Funding: Research funded by the Government of the State of Ceará, through the Secretary of Science, Technology and Higher Education and the Ceará Foundation for the Support of Scientific and Technologi- cal Development (FUNCAP) and the Directorate of Research, Develop- ment and Innovation (DPDI) of the University of Fortaleza. E-PS-14 | E-Posters IT in Pathology / Computational Pathology Symposium E-PS-14-001 Short training significantly improves ganglion cell detection using an algorithm-assisted approach A. Greenberg, B. Samueli, I. Fahoum, S. Farkash, O. Greenberg, V. Zemser-Werner, E. Sabo, R. Hagege, D. Hershkovitz* *Pathology Department, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel Background & objectives: Acquiring sufficient teaching material for rare diseases, such as Hirschsprung’s disease may be difficult, espe- cially in smaller institutes, limiting training. The Objective of this study is to assess the effect of a short training session on algorithm-assisted HSCR diagnosis. Methods: Five pathologists reviewed a dataset of 568 image sets (1704 images in total) selected from 50 cases by the DSA, and were tasked with scoring the images for the presence or absence of ganglion cells. The task was repeated a total of three times. Each pathologist had to complete a short educational presentation between the second and third iterations. Results: The training resulted in a significantly increased rate of correct diagnoses (true positive/negative) and a decreased need for referrals for expert consultation. No statistically significant changes in the rate of false positives/negatives were detected. Conclusion: A very short (<10min) training session can greatly improve the pathologist’s performance in the algorithm-assisted diagnosis of HSCR. The same approach may be feasible in training for the diagnosis of other rare diseases. E-PS-14-002 Artificial intelligence in dermatopathology: can an algorithm replace the pathologist? G. Cazzato*, A. Massaro, A. Colagrande, A. Vacca, E. Maiorano, G. Ingravallo, L. Resta *University of Bari "Aldo Moro", Italy Background & objectives: Artificial intelligence is a very current topic in this period. We tried to train an AI algorithm using basic histopathological criteria that indicate and differentiate with a good probability a malignant melanoma from a severe dysplastic nevus, starting fromWSI. Methods: The artificial intelligence image processing algorithm used to classify and to enhance anomalies contained in the micro- scope image is the Fast Random Forest (FRF). The learning process of the algorithm is based on a preliminary classification of cluster of pixels of the same image. The FRF testing provides as output the processed image with coloured enhanced Melanoma pixel clusters. Results: For five pixel clusters of the same dimensions, occurs a number of about 300 instances (computational cycles) to achieve the maximum precision (equals to 1), with a computational cost of about 2 minutes using a processor Intel®Core™ i5-7200U CPU, 2.71 GHz. The minimum recall performance parameter (near to 0) is achieved about 392 instances. The ROC curve (representing in the plane the true positive rate versus the false positive rate) is matching with the ideal curve of a perfect classifier. The performance indicators confirm the correct setting of the FRF hyperparam- eters. The FRF images have been processed by following a specific image diagnostic protocol, oriented on reading and algorithm error minimization. Conclusion: An important tool for melanoma diagnosis is the probability image estimated by the processed FRF output image. The probability image is useful to better discriminate information about ambiguous lesions. A single probability image is referred to a particular class of “defect”, and enhances, by the white colour, the defect distribution in the whole analysed image. By knowing the dimension of the acquired microscope image, it is also possible to estimate the defect distribution percentage. E-PS-14-003 Use of Anki flashcards tool for reviewing pathology and radiol- ogy of medicine contents J. Carneiro Melo*, L. Mosca de Carvalho Pinheiro, V. Tavares Morais, L. de Albuquerque Rocha, V. de Sousa Araújo Farias, V. Sena Apolinário, G. Cruz Monteiro, P. Lucena de Aquino, B. Vitória Mota barbosa, D. Nunes Oliveira, N. Selma Santos Costa *University of Fortaleza, Brazil Background & objectives: Pathology and Radiology are essential for medical education. Therefore, new study methods are in development, such as electronic flashcards, in order to make the process more dynamic. In this regard, Anki platform has been gaining popularity because it is accessible. Methods: In this study Anki platform was used, a cross-platform software to develop an electronic pathology/radiology flashcard S298