EP4.7 Machine Learning Clustering and Classification of Pre-operative Patient Symptoms: Demographics and Orthopedic Radiographs Only Tell Half the storY (DOROTHY)

Clicks: 10
ID: 331139
2026
Free PDF
Article Quality & Performance Metrics
Overall Quality
Not rated
Combines reader engagement with the AI quality analysis. This article has not been analysed, so there is no overall score — reader engagement is measured and shown alongside.
AI Quality Assessment
Not analyzed
Readership in this journal
Steady

Ranked #279 of 444 articles by views in Journal of hip preservation surgery

Most read Least read

Bar heights use a square-root scale. Only the 120 most-read articles are drawn; the journal has 444 in total.

Mint this article as an NFT
Not yet minted

Create a permanent, verifiable on-chain record of this article on the Scimatic Network. The NFT is held in your Journament account, and you can withdraw it to your own wallet at any time.

5 SUSD one-off · no wallet required
Sign in to mint
Abstract
Abstract Purpose Pre-operative symptom severity of patients with femoroacetabular impingement syndrome (FAIS) and/or hip instability are complex and influenced by both bony/soft tissue morphology and patient-specific factors. Patient-reported outcome measures (PROMs), such as the iHOT-12 (a hip-specific PROM), PROMIS Pain Interference (PI), and PROMIS Physical Function (PF), help inform diagnoses, treatment decisions, and outcomes. This study applied machine learning (ML) clustering and classification models to identify patient clusters based on PROMs and the radiographic and demographic features that most influence preoperative symptom severity. Methods A retrospective chart review identified patients with FAIS and hip instability at a single center. A ML K-means clustering model was developed to determine the number of patient clusters using a combination of PROMs. Next, demographics and radiographic measurements from standing AP view x-rays and 3D CT scans were used as inputs to an ML classification model to classify the PROMs-based clusters. These classification models (n=14 total) included linear (logistic regression, support vector) models, tree-based (random forest, extra trees) models, boosting models, bagging models, and a neural network. All classification models followed standard ML procedures with hyperparameter tuning and were evaluated using repeated cross-validation, balanced accuracy, F1-score, and the area under the receiver operating characteristic curve (AUC). Feature importance via Shapley Additive exPlanations analysis was used to identify the most important predictors of symptom cluster assignment. All ML model development and feature importance were performed using ElucidML software. Results There was a total of 616 hips in 555 patients (Age 33±10 years, BMI 25.4±4.5, 72% Female). The ML clustering model identified 3 patient clusters, interpreted as Mild, Moderate, and Severe symptoms, based on a combination of PROMs (iHOT-12, PROMIS PI and PF, silhouette score = 0.307). However, the ML classification models had poor accuracy, approaching random guessing, with the best model, Extra Trees (ET), achieving a balanced accuracy of 0.4, an F1-score of 0.4, and an AUC of 0.57. Given the poor accuracy, feature importance is challenging to interpret and should be perceived cautiously. For ET, the top features were patient sex at birth, BMI, pubic symphysis to sacrococcygeal joint distance, acetabular version, and age. Conclusion Patients with hip-related pain clustered into 3 symptom severity groups based on a combination of PROMs. However, demographics and radiographic measurements alone were not sufficient to classify these preoperative symptom clusters. These findings suggest that preoperative symptom severity is not well explained by bony morphology alone, and future work should incorporate more patient-specific, soft tissue, and clinical factors for PROM prediction in hip preservation.
Reference Key
openalex_W7220403956 Use this key to autocite in the manuscript while using SciMatic Manuscript Manager or Thesis Manager
Authors Brianna Rowe, Lauren Heylmun, Michael B. Ellman, James Genuario, Michael David
Are you one of these authors? Claim your profile
Journal Journal of hip preservation surgery
Year 2026
DOI
10.1093/jhps/hnag043.341
URL
https://doi.org/10.1093/jhps/hnag043.341
Keywords Keywords not found

Citations

No citations found. To add a citation, contact the admin at info@scimatic.org

Comments

No comments yet. Be the first to comment on this article.