Publicación: A Deep Neural Network For Describing Breast Ultrasound Images in Natural Language
Cargando...
Fecha
2022-09-01
Autores
Editor/a
Director/a
Tutor/a
Coordinador/a
Prologuista
Revisor/a
Ilustrador/a
Derechos de acceso
Atribución-NoComercial-SinDerivadas 4.0 Internacional
info:eu-repo/semantics/openAccess
info:eu-repo/semantics/openAccess
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Nacional de Educación a Distancia (España). Escuela Técnica Superior de Ingeniería Informática. Departamento de Inteligencia Artificial
Resumen
El cáncer de mama es el tipo de cáncer más común y la principal causa de mortalidad en la población femenina. Sin embargo, su detección temprana puede incrementar la tasa de supervivencia relativa a cinco años del 29% al 99%. La ecografía es una de las técnicas más utilizadas para el diagnóstico de cáncer de mama, pero es necesario un experto para interpretar sus resultados de forma correcta. Esto no es común en algunos países que no cuentan con un programa de cribado apropiado, suponiendo una bajada de la tasa al 20%. Los diagnósticos asistidos por ordenador (CAD) tratan de ayudar a los médicos en este proceso, mejorando los resultados y ahorrando tiempo. Los expertos en cáncer de mama emplean la clasificación BI-RADS para describir tumores, estimar su malignidad y establecer el tratamiento a seguir. Mientras la mayoría de sistemas CAD se limitan a clasificar imágenes según su malignidad, presentamos un modelo basado en dos sistemas para la detección y descripción en lenguaje BI-RADS de tumores en tiempo real. El primer sistema es un algoritmo de detección basado en YOLO que obtiene una precisión de 0.965, una exhaustividad de 0.95 y un área bajo la curva de precisión-exhaustividad de 0.97. El segundo es un sistema de descripción que recibe el tumor detectado y devuelve, en lenguaje natural, su descripción en BI-RADS y una estimación de su malignidad. Para este sistema hemos realizado tres experimentos en colaboración con una radióloga experta en mama y hemos obtenido unos valores de concordancia con sus diagnósticos que se encuentran entre los valores de intercorrelación e intracorrelación entre expertos que hemos encontrado en la literatura. Además, observamos que entrenar los modelos con los descriptores BI-RADS mejora la clasificación según malignidad y los acerca al razonamiento experto.
Breast Cancer is the most common cancer and the first cause of mortality among female population. However, its early detection can increase the 5-year survival rate from 29% to 99%. Ultrasound is one of the most used techniques for breast cancer diagnosis, but an expert in the field is necessary to interpret the test correctly. This is not common in some countries that can not afford a proper screening program, resulting in a drop in the 5-year survival rate to 20%. Computer Aided Diagnosis (CAD) systems aim to help physicians during this process, improving results and saving time. Breast cancer experts use Breast Imaging-Reporting and Data System (BI-RADS) classification to describe tumors, estimate their malignancy and establish the treatment to follow. While most CAD systems are limmited to classifying ultrasound images as benign or malignant, giving an explanation via a Region of Interest or an attention mechanism, we have developed a two-system-based model for real-time tumor detection and description using BI-RADS language. The first system is a YOLO-based detection algorithm, which obtains a precision of 0.965, a recall of 0.95, and an area under the precision-recall curve of 0.97. The second is a description system, which uses detected tumor and outputs, in natural language, its description in BI-RADS, and an estimation of the malignancy. For this system, we have carried out three different experiments in collaboration with an expert radiologist in breast cancer and obtained agreement values with her diagnoses that lay between expert intercorrelation and intracorrelation. We also show how training the models with BI-RADS descriptors improves malignancy classification and brings the model closer to expert reasoning.
Breast Cancer is the most common cancer and the first cause of mortality among female population. However, its early detection can increase the 5-year survival rate from 29% to 99%. Ultrasound is one of the most used techniques for breast cancer diagnosis, but an expert in the field is necessary to interpret the test correctly. This is not common in some countries that can not afford a proper screening program, resulting in a drop in the 5-year survival rate to 20%. Computer Aided Diagnosis (CAD) systems aim to help physicians during this process, improving results and saving time. Breast cancer experts use Breast Imaging-Reporting and Data System (BI-RADS) classification to describe tumors, estimate their malignancy and establish the treatment to follow. While most CAD systems are limmited to classifying ultrasound images as benign or malignant, giving an explanation via a Region of Interest or an attention mechanism, we have developed a two-system-based model for real-time tumor detection and description using BI-RADS language. The first system is a YOLO-based detection algorithm, which obtains a precision of 0.965, a recall of 0.95, and an area under the precision-recall curve of 0.97. The second is a description system, which uses detected tumor and outputs, in natural language, its description in BI-RADS, and an estimation of the malignancy. For this system, we have carried out three different experiments in collaboration with an expert radiologist in breast cancer and obtained agreement values with her diagnoses that lay between expert intercorrelation and intracorrelation. We also show how training the models with BI-RADS descriptors improves malignancy classification and brings the model closer to expert reasoning.
Descripción
Categorías UNESCO
Palabras clave
cáncer de mama, BI-RADS, ecografía, diagnósticos asistidos por ordenador, mecanismo de atención, inteligencia artificial explicable, aprendizaje profundo, descripción de imágenes médicas, breast cancer, deep learning, medical image captioning, ultrasound image, computer aided diagnosis, attention mechanism, explainable artificial intelligence
Citación
Centro
E.T.S. de Ingeniería Informática
Departamento
Inteligencia Artificial