Artificial Intelligence in the Management of Low Back Pain

International Journal of Biomedicine 15(3) (2025) 452-456 http://dx.doi.org/10.21103/Article15(3)_RA3 REVIEW ARTICLE INTERNATIONAL JOURNAL OF BIOMEDICINE Artificial Intelligence in the Management of Low Back Pain Dafina Milaj Cacaj1, Xhorxhina Peshku Alushaj1* 1 Department of Physiotherapy, Alma Mater Europea Campus College “Rezonanca,” Pristina, Kosovo Abstract Background: The last decade has witnessed a technological revolution driven mainly by the development of artificial intelligence (AI), a technology designed to replicate human thinking and behavior. AI has significantly penetrated almost all professional fields, including the medical sciences. The study aimed to review the literature data on the application of AI in the management of low back pain (LBP). Methods and Results: This study summarizes relevant data from PubMed, Google Scholar, and Scopus, published between 2000 and 2023. Only studies published in English were considered. Artificial intelligence showed great promise in improving the accuracy of LBP diagnosis, optimizing treatment approaches, and predicting clinical outcomes. Artificial intelligence has facilitated the development of personalized self-management programs and real-time symptom monitoring. AI models have outperformed traditional statistical methods in predicting long-term pain and functional recovery. Conclusion: Although current data suggest a promising role of artificial intelligence in managing LBP, ongoing research will be crucial to determine its clinical utility and broader integration into everyday clinical practice.(International Journal of Biomedicine. 2025;15(3):452-456.) Keywords: artificial intelligence • machine learning • low back pain • therapy • clinical outcomes For citation: Cacaj DM, Alushaj XP. Artificial Intelligence in the Management of Low Back Pain. International Journal of Biomedicine. 2025;15(3):452-456. doi:10.21103/Article15(3)_RA3 Introduction Low back pain (LBP) is a broad clinical term encompassing a spectrum of conditions characterized by pain and discomfort localized between the costal margin and the inferior gluteal folds.1,2 Low back pain affects approximately 50% of adults at some point in their lifetime, with peak prevalence occurring between the ages of 40 and 50,1 and a substantial burden is also observed among older adults.3 Low back pain refers to axial, non-radiating discomfort confined to the lumbar region, occurring in the absence of red-flag indicators suggestive of serious pathology—such as neoplastic processes, infectious etiologies, or cauda equina syndrome—as well as without evidence of specific spinal disorders, including spinal canal stenosis, radiculopathy, osteoporotic vertebral fractures, or seronegative spondyloarthropathies such as ankylosing spondylitis.4,5 The diagnosis of LBP is primarily established through a process of exclusion, ruling out identifiable etiologies such as intervertebral disc herniation, spinal infection, malignancy, and other red-flag conditions indicative of serious underlying pathology.6 Patients presenting with acute LBP are initially assessed for the presence of red flag indicators, which may suggest an underlying serious pathology necessitating prompt and comprehensive diagnostic evaluation.7 In the absence of alarm signals, doctors usually inform patients about the nonspecific nature of low back pain and the high probability of a favorable prognosis. Patients are instructed to avoid prolonged bed rest and to maintain physical activity within acceptable limits. It encourages an early return to work and daily activities to promote functional recovery.8 First-line pharmacologic management of low back pain typically includes acetaminophen, non-steroidal antiinflammatory drugs (NSAIDs), and muscle relaxants, with subsequent incorporation of physical therapy and rehabilitative interventions as indicated.9 Recent technological advances have made AI a key tool in modern healthcare, enabling secure management of patient data, improving medical image analysis, supporting D. M. Cacaj & X. P. Alushaj / International Journal of Biomedicine 15(3) (2025) 452-456 diagnostic decision making, and acting as virtual assistants for both physicians and patients.10 The concept of AI was first introduced by Professor John McCarthy at the Dartmouth Conference in 1956, where it was defined as the creation of intelligent machines capable of perceiving, understanding, reasoning, learning, and making decisions in a manner analogous to human cognition.11 Artificial intelligence, including machine learning algorithms, has quickly become an integral part of modern healthcare, and the field of rehabilitation is poised to benefit significantly from its analytical and predictive capabilities.12 The integration of AI into physical therapy and rehabilitation has been linked to improved patient compliance and faster recovery times, primarily through the implementation of personalized, data-driven intervention strategies.13 Beyond interpretation, AI has demonstrated utility in enhancing and reconstructing spinal imaging. AI algorithms can be trained to distinguish between high-quality and degraded MRI or CT images, enabling the reconstruction of clearer, diagnostically valuable images from suboptimal input data. This capability not only improves image quality but also has the potential to reduce the need for repeat imaging, thereby minimizing patient exposure to radiation and streamlining diagnostic workflows.14 Artificial intelligence can also be effectively utilized to detect pain through neurophysiological approaches.15 Electroencephalography (EEG), which records the brain’s electrical activity, has been employed in conjunction with AI algorithms to not only identify the presence of pain but also quantify its intensity. These advancements suggest a promising role for AI in developing objective, real-time pain assessment tools, particularly in clinical scenarios where patient selfreporting is limited or unreliable.15-17 This literature review aimed to explore emerging applications of AI in the management of LBP, with a particular focus on recent biomedical innovations and their clinical relevance. Material and Methods An electronic literature search was conducted using the biomedical databases PubMed/MEDLINE, Scopus, and the National Library of Medicine, covering publications from 2000 to 2023. Only studies published in English were considered. The keywords used in the search included “artificial intelligence,” “low back pain (LBP),” and “ LBP diagnosis.” Article selection was based on a review of titles and abstracts containing the phrase “artificial intelligence in low back pain management,” with a focus on clinical applications. The inclusion criteria for this review encompassed case reports, case series, original research articles, review papers, in vitro and in vivo studies, animal studies, and controlled clinical trials involving the use of AI in physiotherapy-related cont (...truncated)