Arjun Parmar , Anthony A. Gatti , Ryan Fajardo , Matthew S. Harkey
{"title":"Ultrasound-based statistical shape modeling for quantifying femoral trochlear bone shape post-ACLR","authors":"Arjun Parmar , Anthony A. Gatti , Ryan Fajardo , Matthew S. Harkey","doi":"10.1016/j.ostima.2024.100255","DOIUrl":null,"url":null,"abstract":"<div><h3>Objective</h3><div>Traditional assessments of femoral bone shape are not always available and do not adequately describe the full complexity of concave bone shape. We aimed to develop and validate an ultrasound-based statistical shape model (SSM) and a derived bone shape score (B-score) to quantify the femoral trochlear morphology associated with anterior cruciate ligament reconstruction (ACLR).</div></div><div><h3>Design</h3><div>This was a cross-sectional investigation involving 20 individuals with and 28 individuals without a history of ACLR. Bilateral ultrasound images of the femoral trochlear groove were acquired and analyzed. Both the SSM and B-score were validated using 5-fold cross-validation, assessing reconstruction and classification accuracy, respectively.</div></div><div><h3>Results</h3><div>In held-out test data, the SSM captured over 99% of the bone shape variance with minimal reconstruction error (RMSE = 0.027 ± 0.004 mm). On test data, the B-score accurately quantified bone shape associated with ACLR, demonstrating high accuracy (92.42%), sensitivity (97.37%), specificity (85.71%), and AUROC (0.95). A B-score threshold of 1.41 standard deviations from the mean healthy bone shape was identified for classifying ACLR history.</div></div><div><h3>Conclusions</h3><div>The ultrasound-based SSM and derived B-score provide a valid and accessible method for quantifying femoral trochlear bone shape changes post-ACLR. This approach offers potential for early detection of bone shape changes associated with disease and injury, improving long-term outcomes for ACLR patients. Future research should focus on enhancing model generalizability and assessment of bone shape changes longitudinally.</div></div>","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 1","pages":"Article 100255"},"PeriodicalIF":0.0000,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Osteoarthritis imaging","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2772654124000898","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Objective
Traditional assessments of femoral bone shape are not always available and do not adequately describe the full complexity of concave bone shape. We aimed to develop and validate an ultrasound-based statistical shape model (SSM) and a derived bone shape score (B-score) to quantify the femoral trochlear morphology associated with anterior cruciate ligament reconstruction (ACLR).
Design
This was a cross-sectional investigation involving 20 individuals with and 28 individuals without a history of ACLR. Bilateral ultrasound images of the femoral trochlear groove were acquired and analyzed. Both the SSM and B-score were validated using 5-fold cross-validation, assessing reconstruction and classification accuracy, respectively.
Results
In held-out test data, the SSM captured over 99% of the bone shape variance with minimal reconstruction error (RMSE = 0.027 ± 0.004 mm). On test data, the B-score accurately quantified bone shape associated with ACLR, demonstrating high accuracy (92.42%), sensitivity (97.37%), specificity (85.71%), and AUROC (0.95). A B-score threshold of 1.41 standard deviations from the mean healthy bone shape was identified for classifying ACLR history.
Conclusions
The ultrasound-based SSM and derived B-score provide a valid and accessible method for quantifying femoral trochlear bone shape changes post-ACLR. This approach offers potential for early detection of bone shape changes associated with disease and injury, improving long-term outcomes for ACLR patients. Future research should focus on enhancing model generalizability and assessment of bone shape changes longitudinally.