Three-dimensional measurement of the uterus on magnetic resonance images: development and performance analysis of an automated deep learning tool

Abstract Background The aim of our study was to develop, validate, and test a deep learning (DL) tool for fully automated measurement of the three-dimensional size of the uterus on magnetic resonance imaging (MRI) and to compare it to manual reference measurement. Materials and Methods In this single-centre retrospective study, 845 cases were included for training and validation. The ground truth was a manual measurement of the uterus on magnetic resonance (MR) images. A deep learning tool using a convolutional neural network (CNN) with VGG-16/VGG-11 architecture was developed. The performance of the model was evaluated using the objective keypoint similarity (OKS), mean difference in millimetres, and coefficient of determination R² on a new set of 100 patients. Results The OKS of our artificial intelligence (AI) model was 0.92 (validation) and 0.96 (test). These performances show a strong correspondence of the positioning of the measurement points between the algorithm and radiologists. The average deviation and R² coefficient between the AI measurements and the manual ones were respectively 3.9 mm and 0.93 for two-points length, 3.7 mm and 0.94 for three-points length, 2.6 mm and 0.93 for width, 4.2 mm and 0.75 for thickness. Inter-radiologist variability was 1.4 mm. A three-dimensional automated measurement was obtained in 1.6 s. Conclusion Our deep learning model can locate the uterus on MR images and place measurement points on it to obtain its three-dimensional measurement with a very good correlation with manual measurements.