Dr. Riew concludes, “Stay tuned for more studies on this topic by our team.”
Machine Learning Transforms Meniscal Imaging
In the assessment of osteoarthritis (OA) progression, says new research from the UK, it would help to utilize 3D imaging to determine which meniscal pathologies undergo the most change. The study, “Where does meniscal damage progress most rapidly? An analysis using three-dimensional shape models on data from the Osteoarthritis Initiative,” is published in the January 2018 edition of Osteoarthritis and Cartilage.
Philip Conaghan M.B.B.S. Ph.D., professor of musculoskeletal medicine at the University of Leeds and deputy director of the National Institute for Health Research Leeds Biomedical Research Centre and co-author of the study, told OTW, “The meniscus is an integral part of the osteoarthritis process but much less studied than cartilage or even subchondral bone. In part this has been because it’s difficult to visualize even with MRI, which relies on a reader looking at a sequence of 2D images and which often fails to adequately show its 3-dimensional appearance.”
“This study included MRI images from people with definite OA and careful manual segmentation of the menisci in these images—but with the added value of supervised machine learning to enable correct placement of the menisci relevant to the tibia. And we were able to examine a number of different meniscal shapes as they degenerate over time.”
The authors wrote, “Knee images were selected from the progression cohort of the Osteoarthritis Initiative choosing participants with risk factors for medial OA progression. Medial and lateral menisci were manually segmented then analysed using a statistical shape model of the tibia as a reference surface.”
“Responsiveness was assessed at 1 year using standardized response means (SRMs) for four constructs: meniscal volume, extrusion volume, thickness and tibial coverage; anatomical sub-regions of these constructs were also explored. Paired images from 86 participants (median age 61.5, 49% female, 56% obese) were included.”
According to Dr. Conaghan, “Posterior medial meniscus was the location of most pathology. Despite relatively small numbers for a 12-month OA follow-up, there was responsiveness demonstrated for two of the meniscal measures.”
“There has been little investment in the OA field because demonstration of progression has been difficult in feasible time frames.”
“Modern image analysis has already provided the most responsive measures of OA progression in clinical trials, using either cartilage thickness or bone shape. This study provides a third responsive measure, meniscal shape. And it may add to the other measures.”
“This work underpins the benefits of modern 3D image analysis, which is not only being used for pre-joint replacement planning, but for understanding the importance of specific OA pathologies. In future we will be able to look at the range of meniscal pathologies and their relationship with symptoms.”
