Embracing Predictive Models in Non-Surgical Scoliosis Treatment during growth: A Game-Changer for Patient Care

Scoliosis is defined in 80–85% of cases as “idiopathic”, which is another way of saying that we currently don’t know the cause. Over 80% of these cases are diagnosed during adolescence, affecting about 2-3% of the population, with seven in ten cases being female.

This condition, characterized by its unpredictability, poses a significant challenge in determining the most appropriate treatment options. However, recent advances in predictive modelling may significantly improve how we approach non-surgical scoliosis rehabilitation (so-called conserva-tive treatment), providing a beacon of hope for affected families.

What is Scoliosis?

Scoliosis is a three-dimensional deformity of the spine. While it cannot be healed (i.e. going back to zero degrees), what matters is having a functional spine (or a spine that does not create health issues). Even if the spine is not perfectly straight, a slight curve is acceptable. Therapy begins once the deformity is discovered and consists of observation, specific exercises and bracing. Depending on the severity, treatment can also include surgery. In most cases, scoliosis is asymptomatic and discovered by chance by either parents or teachers since healthcare professionals stopped doing school screenings in the last century. Treatment for scoliosis ends at the end of bone growth and is followed by a monitoring phase during adulthood, with periodic checks depending on the person.

Understanding the natural history of scoliosis, and its progression when untreated, is crucial

In 2018 Isico published a meta-analysis in the American Journal of Physical Medicine and Rehabilitation, which focused on 13 studies in scientific literature examining the natural history of idiopathic scoliosis and how scoliosis evolves without any treatment. The meta-analysis revealed high progression rates for all idiopathic scoliosis forms with the data on infantile scoliosis showing the highest variability partly because of the pathology, partly because of bad quality of the studies. Suffice to say that three studies referring to infantile scoliosis showed progression rates ranging from 5 to 80%! Furthermore, scoliosis curve progression can vary widely between individuals, with factors such as age, curve pattern, initial Cobb angle, and skeletal maturity playing pivotal roles. This variability has historically made it difficult to predict how a specific case might evolve, especially during crucial growth periods like puberty.

As an example, Cindy was diagnosed with scoliosis when she was eight years old. Along with her parents, she began a challenging journey of extended brace therapy without being able to predict how the curve would evolve during the pubertal growth spurt in adolescence and what treatments would be required. Despite wearing the brace, Cindy continued to participate in various sports, and today, nearly ten years later, she works as a model. Learn more about Cindy’s story here.

Effective predictive models could have been a game-changer for Cindy and her parents

The potential benefits of an effective predictive model are numerous. By providing more information it could have reduced the psychological burden throughout Cindy’s years long treatment. Moreover, beyond providing optimal results, a predictive model could have done a better job at fine-tuning the proposed treatment according to the expected results.

By analysing data from natural history conditions and incorporating multiple clinical variables, these models could forecast the aggressiveness of scoliosis with remarkable precision. This predictive capability is not just a technical achievement; it’s a tool that empowers clinicians, patients and their families to make informed decisions about treatment options ranging from observa-tion and scoliosis-specific exercises to bracing, without resorting to invasive surgery.

“Awareness of what could happen to my child helps us make the right decision at the right time,” shares the mother of a young patient currently undergoing brace therapy. “Lavinia is now 12 years old; she has been in therapy for two years, and the road is still long. Understanding the potential progression of her scoliosis through predictive models would give us clarity and con-fidence in choosing a treatment path that’s both effective and least disruptive to her life.”

A personalised approach to scoliosis management

Predictive models should offer a personalised approach to scoliosis management. For instance, the Bracing in Adolescent Idiopathic Scoliosis Trial (BrAIST) highlighted the effectiveness of brac-ing in reducing progression to surgery. This insight, derived from predictive analysis, reinforces the importance of early and tailored intervention strategies.

Isico has developed models predicting Idiopathic Scoliosis (IS) curve progression, underscoring the approach’s potential. Using systematic data collected from untreated patients, we’ve aimed to create predictive tools to guide clinical decisions from the initial examination to future goals. Based on simple clinical (i.e., external appearance and measures) and radiographic predictors (such as curve severity, bone age, spine balance), these models seek to offer precision in forecasting future curve angles, significantly impacting treatment planning and outcomes.

The evolution of predictive models in scoliosis treatment is a testament to the power of data-driven healthcare. By providing a clearer picture of scoliosis progression, these models enhance the precision of non-surgical interventions and foster shared decision-making between clinicians, patients, and their families. As we refine these tools, using clinical decision support systems using artificial intelligence (AI), the potential to improve the clinical management of those affected by scoliosis is immense, offering a path to more predictable, personalised, and effective care.

In conclusion, integrating AI-based predictive models into managing idiopathic scoliosis is a significant leap forward. It’s an approach that aligns with the principles of modern healthcare — personalised, data-informed, and person-centred. As we look to the future, the continued development and application of these models promise to transform the treatment landscape for scoliosis, ensuring that patients can lead fuller, healthier lives without the burden of uncertainty that scoliosis once cast. This is the aim of PREPARE – creating data-driven tools to assist with prognosis and pave the way for personalised rehabilitation. Through these tools, PREPARE will better inform clinicians and patients, contributing to better shared decision making and preparation for the patient’s future.

Author: Isico