BACK-to-MOVE: Machine learning and computer vision model automating clinical classification of non-specific low back pain for personalised management.

Publisher: Public Library of Science Country of Publication: United States NLM ID: 101285081 Publication Model: eCollection Cited Medium: Internet ISSN: 1932-6203 (Electronic) Linking ISSN: 19326203 NLM ISO Abbreviation: PLoS One Subsets: MEDLINE

Original Publication: San Francisco, CA : Public Library of Science

Low Back Pain*/therapy ; Low Back Pain*/diagnosis ; Low Back Pain*/classification ; Low Back Pain*/physiopathology ; Machine Learning*; Humans ; Male ; Female ; Adult ; Middle Aged ; Neural Networks, Computer ; Movement ; Precision Medicine/methods ; Patient Reported Outcome Measures

Background: Low back pain (LBP) is a major global disability contributor with profound health and socio-economic implications. The predominant form is non-specific LBP (NSLBP), lacking treatable pathology. Active physical interventions tailored to individual needs and capabilities are crucial for its management. However, the intricate nature of NSLBP and complexity of clinical classification systems necessitating extensive clinical training, hinder customised treatment access. Recent advancements in machine learning and computer vision demonstrate promise in characterising NSLBP altered movement patters through wearable sensors and optical motion capture. This study aimed to develop and evaluate a machine learning model (i.e., 'BACK-to-MOVE') for NSLBP classification trained with expert clinical classification, spinal motion data from a standard video alongside patient-reported outcome measures (PROMs).
Methods: Synchronised video and three-dimensional (3D) motion data was collected during forward spinal flexion from 83 NSLBP patients. Two physiotherapists independently classified them as motor control impairment (MCI) or movement impairment (MI), with conflicts resolved by a third expert. The Convolutional Neural Networks (CNNs) architecture, HigherHRNet, was chosen for effective pose estimation from video data. The model was validated against 3D motion data (subset of 62) and trained on the freely available MS-COCO dataset for feature extraction. The Back-to-Move classifier underwent fine-tuning through feed-forward neural networks using labelled examples from the training dataset. Evaluation utilised 5-fold cross-validation to assess accuracy, specificity, sensitivity, and F1 measure.
Results: Pose estimation's Mean Square Error of 0.35 degrees against 3D motion data demonstrated strong criterion validity. Back-to-Move proficiently differentiated MI and MCI classes, yielding 93.98% accuracy, 96.49% sensitivity (MI detection), 88.46% specificity (MCI detection), and an F1 measure of .957. Incorporating PROMs curtailed classifier performance (accuracy: 68.67%, sensitivity: 91.23%, specificity: 18.52%, F1: .800).
Conclusion: This study is the first to demonstrate automated clinical classification of NSLBP using computer vision and machine learning with standard video data, achieving accuracy comparable to expert consensus. Automated classification of NSLBP based on altered movement patters video-recorded during routine clinical examination could expedite personalised NSLBP rehabilitation management, circumventing existing healthcare constraints. This advancement holds significant promise for patients and healthcare services alike.
Competing Interests: The authors have declared that no competing interests exist.
(Copyright: © 2024 Hartley 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.)

Spine (Phila Pa 1976). 2012 Apr 15;37(8):E486-95. (PMID: 22024899)
Man Ther. 2015 Jun;20(3):456-62. (PMID: 25511448)
Lancet. 2017 Feb 18;389(10070):736-747. (PMID: 27745712)
Phys Ther. 2018 May 1;98(5):408-423. (PMID: 29669082)
Man Ther. 2009 Oct;14(5):555-61. (PMID: 18838331)
Br J Sports Med. 2010 Nov;44(14):1054-62. (PMID: 19996331)
Arthritis Rheum. 2012 Jun;64(6):2028-37. (PMID: 22231424)
Arthritis Rheum. 2008 May 15;59(5):632-41. (PMID: 18438893)
BMC Med Imaging. 2015 Aug 12;15:29. (PMID: 26263899)
Br J Sports Med. 2020 Jul;54(13):782-789. (PMID: 31630089)
Eur J Pain. 2019 Sep;23(8):1416-1424. (PMID: 30974479)
Man Ther. 2005 Nov;10(4):242-55. (PMID: 16154380)
Int J Environ Res Public Health. 2021 Oct 17;18(20):. (PMID: 34682647)
Clin Biomech (Bristol, Avon). 2019 Dec;70:237-244. (PMID: 31669957)
Spine (Phila Pa 1976). 2010 Jun 15;35(14):1387-95. (PMID: 20195206)
Sensors (Basel). 2016 Nov 25;16(12):. (PMID: 27898003)
J Neurosci Methods. 2015 Jul 30;250:126-36. (PMID: 25596422)
Lancet Reg Health Eur. 2022 Dec 06;24:100550. (PMID: 36643660)
Eur Spine J. 2011 Jun;20(6):826-45. (PMID: 21221663)
Front Bioeng Biotechnol. 2021 Nov 15;9:767974. (PMID: 34869281)
Spine (Phila Pa 1976). 2009 Jul 1;34(15):1610-8. (PMID: 19564772)
PLoS One. 2018 Aug 6;13(8):e0201904. (PMID: 30080866)
J Orthop Sports Phys Ther. 2019 Jun;49(6):380-388. (PMID: 29895232)
Spine (Phila Pa 1976). 2000 Nov 15;25(22):2940-52; discussion 2952. (PMID: 11074683)
Lancet. 2018 Nov 10;392(10159):1789-1858. (PMID: 30496104)
Eur Spine J. 2018 Jan;27(1):163-170. (PMID: 28733722)
Bioinformatics. 2006 Sep 1;22(17):2059-65. (PMID: 16820428)
SN Comput Sci. 2021;2(6):461. (PMID: 34549197)
Man Ther. 2013 Oct;18(5):410-7. (PMID: 23518039)
Eur J Pain. 2007 Feb;11(2):153-63. (PMID: 16446108)
Eur J Neurosci. 2021 Dec;54(11):7989-8005. (PMID: 34719827)
Eur J Pain. 2004 Oct;8(5):495-502. (PMID: 15324781)
Phys Ther. 2015 Nov;95(11):1478-88. (PMID: 25929536)
BMJ. 2021 Feb 12;372:n291. (PMID: 33579691)
Man Ther. 2011 Feb;16(1):9-14. (PMID: 21094624)
Pain. 2011 Oct;152(10):2399-2404. (PMID: 21856077)
Lancet. 2018 Jun 9;391(10137):2384-2388. (PMID: 29573871)
Man Ther. 2006 Feb;11(1):28-39. (PMID: 15936976)
Best Pract Res Clin Rheumatol. 2013 Oct;27(5):649-61. (PMID: 24315146)
BMC Musculoskelet Disord. 2018 Aug 28;19(1):309. (PMID: 30153815)
BMC Musculoskelet Disord. 2014 Jul 10;15:229. (PMID: 25012528)
Best Pract Res Clin Rheumatol. 2007 Feb;21(1):77-91. (PMID: 17350545)

Date Created: 20240510 Date Completed: 20240510 Latest Revision: 20240512

20240512

PMC11086851

10.1371/journal.pone.0302899

38728282