ORIGINAL RESEARCH article
Front. Surg.
Sec. Neurosurgery
This article is part of the Research TopicEndoscopy, Navigation, Robotics, Current Trends and Newer Technologies in the Management of Spinal Disorders. Towards a Paradigm Change in the Clinical Practice.View all 12 articles
Validity and accuracy of a machine learning predictive model in the exploitation of Patients Related Outcomes in Spine Surgery
Provisionally accepted
- 1Hôpitaux Universitaires Pitié Salpêtrière, Paris, France
- 2Ramsay Sante SA, Paris, France
- 3Cortexx Medical Intelligence SAS, Paris, France
- 4Polyclinique Jean Villar de Bruges, Bruges, France
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Background : Lumbar spine disorders are amongst the most prevalent and disabling conditions worldwide. Patient selection remains very complex and the benefits of surgical interventions uncertain and may depend on patient baseline health characteristics. Patient Related Outcome Measurements (PROMs) is a standard way to measure success in lumbar surgery. The aim of this study is to prospectively validate accuracy of a deep-learning algorithm to predict the clinical outcomes of patients undergoing lumbar surgery (MCID /no-MCID). Material and methods: This study is multicentric, longitudinal and prospective over a 16 months period (September 2021 to December 2022). Patients with a surgical indication for lumbar decompression were included preoperatively. All patients were enrolled in the SuMO© dedicated mobile application to fill in the preoperative and postoperative data. Patients were classified in two categories according to their postoperative outcomes. Minimal clinically important difference (MCID) is defined by the ODI combined with the intake of opioids and motor loss of patients. These results were then compared to the prediction of the algorithm based on the preoperative data in order to determine the accuracy of the algorithm. Results: 119 patients enrolled preoperatively, follow-up was obtained in 103 patients postoperatively. Mean preoperative Oswestry Disability Index (ODI) was 0.43 (SD 0.17). Postoperative mean ODI was 0.28 (SD 0.18) at 1-month, and 0.14 (SD 0.16) at 3-months.The mean postoperative ODI in the MCID group was 0.12 while it was 0.26 in tne no-MCID group at 8 months. The algorithm predicted the outcome with an accuracy of 81,6% (ROC score). Conclusion: This study confirmed the validity and accuracy of the algorithm in predicting postoperative outcomes prospectively, as well as the sensitivity of MCID definition, especially when coupled with a remote patient centered follow-up. Artificial Intelligence algorithms may help physicians in their future daily practice to address personalized care.
Keywords: Spine surgeries, artificial intelligence, machine learning, PROM (Patient reported outcome measurement), MCID (minimal clinically important differences)
Received: 22 Sep 2025; Accepted: 29 Nov 2025.
Copyright: © 2025 Andre, peyrou, Vignaux, Boissière and Obeid. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Arthur Andre
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.