ABSTRACT
Resumen Antecedentes: La resonancia magnética (RM) de próstata es uno de los métodos diagnósticos para la identificación del carcinoma de próstata. La escala PI-RADS (Prostate Imaging and Reporting Data System) es el sistema usado para la interpretación de estas imágenes. Es importante, para su reproducibilidad, la estandarización y la evaluación de dicha escala. Objetivo: Determinar la concordancia inter- e intraobservador de la versión 2.1 del PI-RADS. Material y métodos: Estudio observacional retrospectivo, evaluando 129 RM de pacientes con sospecha de cáncer de próstata por tres radiólogos con diferentes años de experiencia y en dos momentos del tiempo, usando el puntaje PI-RADS 2.1. Se evaluó la concordancia intra- e interobservador. Resultados: La concordancia interobservador fue sustancial (kappa > 0,6) en todos los observadores, siendo la categoría 5 la de mayor acuerdo interobservador. Se observó una alta reproducibilidad intraobservardor, con la mayor kappa siendo de 0,856. Cuando se realizó el análisis según años de experiencia de los radiólogos, la concordancia interobservador fue significativa en todos los casos. Conclusiones: El sistema de clasificación PI-RADS 2.1 es reproducible para las diferentes categorías y aumenta la concordancia cuando se trata de lesiones con mayor probabilidad de cáncer clínicamente significativo.
Abstract Background: Magnetic Resonance Imaging (MRI) of the prostate is a key diagnostic tool for identifying prostate carcinoma. The Prostate Imaging-Reporting and Data System (PI-RADS) scale is the standard system for interpreting these images. Standardizing and evaluating this scale is crucial for ensuring consistent and reproducible results. Objective: This study aims to assess both the interobserver and intraobserver agreement of the PI-RADS version 2.1. Material and methods: In this retrospective observational study, 129 prostate MRI scans from patients with suspected prostate cancer were evaluated. Three radiologists, each with different levels of experience, analyzed these scans at two separate times using the PI-RADS 2.1 scoring system. Both intraobserver and interobserver agreements were measured. Results: The study found substantial interobserver agreement (kappa > 0.6) across all categories, with category 5 showing the highest level of agreement. Intraobserver reproducibility was also high, with the highest kappa value reaching 0.856. Further analysis based on the radiologists years of experience revealed significant interobserver agreement in all instances. Conclusions: The PI-RADS 2.1 classification system demonstrates high reproducibility across different categories, particularly for lesions more likely to be clinically significant cancers. This underscores its reliability in varied diagnostic scenarios.
ABSTRACT
The liver is a frequent site of benign and malignant, primary and metastatic tumors. Hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) are the most common primary liver cancers, and colorectal liver metastasis (CRLM) is the most common secondary liver cancer. Although the imaging characteristic of these tumors is central to optimal clinical management, it relies on imaging features that are often non-specific, overlap, and are subject to inter-observer variability. Thus, in this study, we aimed to categorize liver tumors automatically from CT scans using a deep learning approach that objectively extracts discriminating features not visible to the naked eye. Specifically, we used a modified Inception v3 network-based classification model to classify HCC, ICC, CRLM, and benign tumors from pretreatment portal venous phase computed tomography (CT) scans. Using a multi-institutional dataset of 814 patients, this method achieved an overall accuracy rate of 96%, with sensitivity rates of 96%, 94%, 99%, and 86% for HCC, ICC, CRLM, and benign tumors, respectively, using an independent dataset. These results demonstrate the feasibility of the proposed computer-assisted system as a novel non-invasive diagnostic tool to classify the most common liver tumors objectively.
