Use of 2D U-Net Convolutional Neural Networks for Automated Cartilage and Meniscus Segmentation of Knee MR Imaging Data to Determine Relaxometry and Morphometry

Abstract

Purpose To analyze how automatic segmentation translates in accuracy and precision to morphology and relaxometry compared with manual segmentation and increases the speed and accuracy of the work flow that uses quantitative magnetic resonance (MR) imaging to study knee degenerative diseases such as osteoarthritis (OA). Materials and Methods This retrospective study involved the analysis of 638 MR imaging volumes from two data cohorts acquired at 3.0 T: (a) spoiled gradient-recalled acquisition in the steady state T1ρ-weighted images and (b) three-dimensional (3D) double-echo steady-state (DESS) images. A deep learning model based on the U-Net convolutional network architecture was developed to perform automatic segmentation. Cartilage and meniscus compartments were manually segmented by skilled technicians and radiologists for comparison. Performance of the automatic segmentation was evaluated on Dice coefficient overlap with the manual segmentation, as well as by the automatic segmentations’ ability to quantify, in a longitudinally repeatable way, re-laxometry and morphology. Results The models produced strong Dice coefficients, particularly for 3D-DESS images, ranging between 0.770 and 0.878 in the cartilage compartments to 0.809 and 0.753 for the lateral meniscus and medial meniscus, respectively. The models averaged 5 seconds to generate the automatic segmentations. Average correlations between manual and automatic quantification of T1ρ and T2 values were 0.8233 and 0.8603, respectively, and 0.9349 and 0.9384 for volume and thickness, respectively. Longitudinal precision of the automatic method was comparable with that of the manual one. Conclusion U-Net demonstrates efficacy and precision in quickly generating accurate segmentations that can be used to extract relaxation times and morphologic characterization and values that can be used in the monitoring and diagnosis of OA.

Fully automated patellofemoral MRI segmentation using holistically nested networks: Implications for evaluating patellofemoral osteoarthritis, pain, injury, pathology, and adolescent development

“…The current study did not have such limitations. Furthermore, previous studies evaluating the semiautomatic or automatic segmentation of PF cartilage reported accuracies (DSCs ranged from 63% to 84%) that were inferior to the current results for PF bone segmentation. The stability of the HNN architecture, its ability to capture texture variation across the full image context, and the large training set likely promoted the superior performance of the HNN model.…”

“…To the best of our knowledge, only 2 studies presented accuracies for segmenting the full patellar bone, given that the focus has been primarily on segmenting the tibiofemoral bone and cartilage surfaces . One study on patellar bone segmentation reported more‐accurate segmentation (Table ) than the current study.…”

A 3D deep convolutional neural network approach for the automated measurement of cerebellum tracer uptake in FDG PET‐CT scans

“…Besides the change of dimension, the 3D U‐Net also included batch normalization before the rectified linear unit (ReLU) activations for efficiency and reduced the number of tiers from five to four to reduce GPU memory requirements. U‐Nets have been utilized for 2D and 3D (volumetric) medical image segmentation . Inspired by the U‐Net, other similarly structured FCN networks were developed.…”

Clinical Utility of Advanced Imaging of the Knee

“…This method allows for a z ‐score of articular cartilage to determine expected status of cartilage relative to a healthy knee. Machine‐learning algorithms have also been applied with success to automatically defining cartilage for post‐processing . These techniques have the potential to allow for the more routine use of quantitative cartilage imaging in clinical practice by eliminating the need for laborious post‐processing.…”