Machine Learning and 3D Reconstruction of Materials Surface for Nondestructive Inspection.

During the steel pipeline installation, special attention is paid to the butt weld control performed by fusion welding. The operation of the currently popular automated X-ray and ultrasonic testing complexes is associated with high resource and monetary costs. In this regard, this work is devoted to the development of alternative and cost-effective means of preliminary quality control of the work performed based on the visual testing method. To achieve this goal, a hardware platform based on a single board Raspberry Pi4 minicomputer and a set of available modules and expansion cards is proposed, and software whose main functionality is implemented based on the systemic application of computer vision algorithms and machine learning methods. The YOLOv5 object detection algorithm and the random forest machine learning model were used as a defect detection and classification system. The mean average precision (mAP) of the trained YOLOv5 algorithm based on extracted weld contours is 86.9%. A copy of YOLOv5 trained on the images of control objects showed a mAP result of 96.8%. Random forest identifying of the defect precursor based on the point clouds of the weld surface achieved a mAP of 87.5%.

Defining Ewing and Ewing-like small round cell tumors (SRCT): The need for molecular techniques in their categorization and differential diagnosis. A study of 200 cases.

BACKGROUND: Differentiation of Ewing sarcoma family of tumors (ESFT) and Ewing-like tumors remains problematic. Certain ESFT with morphological and immunohistochemical (IHC) profiles lack the EWSR1-ETS transcript. To improve diagnostic accuracy we investigated the presence of several specific transcripts in 200 small round cell tumors (SRCT) displaying ESFT morphology and immunophenotype in which EWSR1 FISH analysis was non-informative or negative. DESIGN: 200 tumors (formalin-fixed, paraffin-embedded) were analyzed by RT-PCR. All tumors were tested for EWSR1-ETS, EWSR1/WT1, PAX3/7-FOX01 or SYT/SSX transcripts, and the negative tumors were subsequently analyzed for CIC/DUX4, BCOR/CCNB3 and CIC/FOX04 transcripts. RESULTS: 133 (66.5%) ESFT displayed one of the above EWSR1-ETS translocations. Three cases (1.5%) revealed the SYT-SSX transcript for Synovial sarcoma, and one (0.5%) a EWSR1-WT1 transcript for Desmoplastic Small Round Cell tumor. The CIC-DUX4 translocation was found in six Ewing-like tumors (3%) with CD99 positivity. The BCOR-CCNB3 gene fusion was observed in 5 tumors (2.5%) displaying round or spindle cells with strong CCNB3 IHC expression in 3 tumors. Moreover, RT-PCR failed to detect any gene fusion transcripts in 19 tumors (9.5%) and were considered "undifferentiated small round cell sarcoma" (SRCS). Molecular biology results were non-informative in 33 SRCTs (16.5%) due to RNA degradation through inadequate fixation and/or decalcification. CONCLUSION: Our analysis of 200 SRCTs confirms the molecular heterogeneity of neoplasms with ESFT morphology and highlight that molecular studies with RT-PCR including new emerging gene fusion transcripts are mandatory for the diagnosis when EWSR1 FISH is negative or non-informative. The incidence of CIC-DUX4, BCOR-CCNB3 and CIC-FOX04 transcripts was relatively low. A small group of Ewing-like sarcomas or undifferentiated SRCS remains unclassified. Adopting appropriate tissue fixation and processing protocols is important to avoid degradation of fixed/embedded tissue when no frozen tumor is available.

Histological heterogeneity of Ewing's sarcoma/PNET: an immunohistochemical analysis of 415 genetically confirmed cases with clinical support.

Ewing's sarcoma (ES)/peripheral neuroectodermal tumor (PNET) are malignant neoplasms affecting children and young adults. We performed a study to typify the histological diversity and evaluate antibodies that may offer diagnostic/prognostic support. In total, 415 cases of genetically confirmed paraffin-embedded ES/PNET were analyzed on whole sections and in tissue microarrays. This study confirms the structural heterogeneity of ES/PNET, distinguishing three major subtypes: conventional ES (280 cases); PNET (53 cases); and atypical ES/PNET (80), including large cells, vascular-like patterns, spindle pattern, and adamantinoma-like configuration. All cases presented positivity for at least three of the four tested antibodies (CD99, FLI1, HNK1, and CAV1). CAV1 appeared as a diagnostic immunomarker of ES/PNET being positive in CD99-negative cases. Hence, the immunohistochemical analysis confirmed the diagnostic value of all four antibodies, which together cover more than 99% of the tumors, independently of the histological variety. The univariate analysis for survival revealed atypical ES as the only histological parameter apparently associated with less favorable clinical outcome, particularly in the subgroup of patients treated with surgery. In conclusion, the diagnosis of atypical ES is a challenge for the pathologist and needs support from molecular techniques to perform an optimal differential diagnosis with other small round cell tumors.