Sarah Galante, Anna Corti, Riccardo Stuani, Katia Chiappetta, Mattia Loppini, Valentina D. A. Corino
Abstract
The growing prevalence of total hip arthroplasty (THA) revisions, along with their generally poorer outcomes compared to primary procedures, emphasizes the urgent need for early detection of primary THA failure. This study proposes a model that integrates radiographic, clinical, and comorbidity data to automatically detect pathological signs within one year after primary THA.
Introduction
Total hip arthroplasty (THA) is widely recognized as the gold standard surgical intervention for the treatment of various hip conditions including end-stage osteoarthritis [1,2]. With the increase of life expectancy and an aging population, combined with improved access to surgical care, the demand for THA has significantly increased [3].
Materials and methods
This study involved two independent cohorts of patients who underwent primary THA. The first cohort, retrospectively enrolled, (cohort A) was used for model development and internal testing, while the second cohort, prospectively collected, (cohort B) served for external validation of the model.
Results
Considering results on the validation set, the RF achieved the best performances in terms of balanced accuracy, F1 score and AUC, both in the clinical and demographic models, thus it was selected for the next steps in the combined multi-domain approach.
Discussion
This study emphasizes the value of a multi-domain analysis that combines post-operative radiographs with clinical, demographic and comorbidity data to improve the early detection of pathological signs after primary THA. The main findings of this study are: i) the implementation of a multi-domain approach that integrates imaging with clinical and comorbidity information relevant to implant function.
Conclusion
This study introduces the first integrated DL and ML framework for detecting pathological signs within one year after primary THA, leveraging radiographic, clinical, demographic, and comorbidity data to enable a virtual follow-up. The combination of the three models through probability averaging enhanced the overall predictive performance, compared to single models.
Citation: Galante S, Corti A, Stuani R, Chiappetta K, Loppini M, Corino VDA (2026) Automatic early detection of pathological signs following primary total hip arthroplasty using radiographs, clinical scores, and comorbidities. PLoS One 21(6): e0348790. https://doi.org/10.1371/journal.pone.0348790
Editor: Pawan Acharya, University of Alabama at Birmingham, UNITED STATES OF AMERICA
Received: November 12, 2025; Accepted: April 21, 2026; Published: June 15, 2026
Copyright: © 2026 Galante et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data cannot be shared publicly because they contain potentially identifiable and sensitive patient information, and are therefore subject to ethical and privacy restrictions. Access to the data may be requested by qualified researchers from the Ethics Committee via cetlombardia5@humanitas.it, for researchers who meet the criteria for access to confidential data. The code used for the analyses is publicly available in the project GitHub repository and archived in Zenodo at https://doi.org/10.5281/zenodo.19064049.
Funding: This research received funding from the Ministry of University and Research (MUR), PRIN: progetti di ricerca di rilevante interesse nazionale – bando 2022 prot. 2022YME9N3. AC is funded by the National Plan for NRRP Complementary Investments (PNC, established with the decree-law 6 May 2021, n. 59, converted by law n. 101 of 2021) in the call for the funding of research initiatives for technologies and innovative trajectories in the health and care sectors (Directorial Decree n. 931 of 06-06-2022) - project n. PNC0000003 - AdvaNced Technologies for Human-centrEd Medicine (project acronym: ANTHEM).
Competing interests: The authors have declared that no competing interests exist.
