Percepción de la inteligencia artificial en la comunidad radiológica argentina / Perception of artificial intelligence in the Argentine radiological community

Resumen Introducción: La aplicación de la inteligencia artificial (IA) en radiología se encuentra en fases iniciales y estaría influenciada por la percepción de sus usuarios. Objetivo: Valorar el grado de conocimiento, interés y expectativas acerca de la IA en la comunidad radiológica argentina. Método: Entre septiembre y noviembre de 2022 se distribuyó una encuesta online acerca del conocimiento, la actitud y las expectativas con respecto a la IA. Resultados: De un total de 161 encuestados, la mayoría eran mujeres (55,3%), tenían entre 31 y 40 años (43,4%), eran médicos de planta (42%), trabajaban en el ámbito privado (55,6%) y estaban localizados en la CABA/Buenos Aires (59%). El 52,5% no habían utilizado software de IA y el 87,7% no habían realizado cursos de IA. El 88,2% manifestaron interés en la formación en este campo. Casi la totalidad concordaba en que la IA modificaría su labor diaria en los próximos 10 años (87,6%), aunque no los reemplazaría (56,8%), ni cambiarían de carrera (69,8%). Conclusiones: Aunque el grado de conocimiento acerca de la IA en la comunidad radiológica argentina es moderado, encontramos un alto interés y expectativas, y un bajo nivel de miedo o rechazo. Las sociedades radiológicas argentinas deberían desarrollar cursos de formación en IA.

Abstract Introduction: The application of artificial intelligence (AI) in radiology is in early stages and could be influenced by the perception of its users. Objectives: To assess the degree of knowledge, interest, and expectations about AI in the Argentine radiological community. Method: Between September and November 2022, an online survey about knowledge, attitude and expectations of AI was distributed. Results: Of a total of 161 respondents, the majority were women (55.3%), had between 31 and 40 years old (43.4%), were staff physicians (42%), worked in the private sector (55.6%) and were located in CABA/Buenos Aires (59%). 52.5% had not used AI software and 87.7% had not taken any previous AI courses. 88.2% were interested in training in this field. Almost all agreed that AI would modify their daily work in the next 10 years (87.6%), although it would not replace them (56.8%), nor would they change their careers (69.8%). Conclusions: Although the degree of knowledge about AI in the Argentine radiological community is moderate, we found high interest and expectations, and low level of fear or rejection. Argentine radiological societies should develop AI training courses.

Transbronchial biopsies' histopathological findings leading to successful late steroid therapy in Covid-19 acute respiratory failure.

We present results from clinical, radiologic, gas exchange, lung mechanics, and fibre-optic bronchoscopy-guided transbronchial biopsies in a case of acute respiratory failure due to SARS-CoV-2 (Covid-19). This report highlights the pulmonary, immunological, and inflammatory changes found during acute diffuse alveolar damage and the later organizing phase. An early diffuse alveolar damage pattern with predominant epithelial involvement with active recruitment of T cells and monocytes was observed followed by a late organizing pattern with pneumocyte hyperplasia, inflammatory infiltration, prominent endotheliitis, and secondary germinal centers. The patient's deterioration paralleling the late immuno-pathological findings based the decision to administer intravenous corticosteroids, resulting in clinical, gasometric, and radiologic improvement. We believe that real-time clinicopathological correlation, along with the description of the immunological processes at play, will contribute to the full clinical picture of Covid-19 and might lead to a more rational approach in the precise timing of anti-inflammatory, anti-cytokine, or steroid therapies.