Virtualising patients’ spines for better decision-making on back treatments

At the moment, patients rely on the judgment of clinicians about what treatment is appropriate; however, decisions made in the clinic cannot take into account the many factors – often not apparent at the time – that vary from patient to patient, and which can affect the outcomes of deciding to operate or not. Among these factors, biomechanical effects are particularly difficult to assess, since no force measurements can be performed directly on the tissues, inside a patient.

Now MySpine, an EU-funded project coordinated by Barcelona’s Institute for Bioengineering of Catalonia (IBEC) and involving seven other partners across Europe, has made a huge step forward towards addressing this problem. The project, which received excellent final reviews at its closure in November, involved creating a platform to help doctors decide the best treatment for specific back patients by anticipating the likely long-term responses – depending on patient weight, lumbar spine geometry, motion, and disc condition, among other variables – of various types of treatments, from the least to the most invasive.

“The MySpine computing platform uses patient-specific models for its simulations, which come from combining two types of medical images – CT and MRI scans – taken from the patient, “ explains Jérôme Noailly, Principal Investigator at IBEC. “Until now, these two complementary methods were used in parallel, and not together. MySpine also creates a patient-specific 3D geometrical model which integrates all the tissues of the lumbar spine in order to anticipate the local effects of a range of simulated treatments, and to determine which of these treatments promises the most long-term success.”

Treatments for low back pain – which is often a sign of intervertebral disc degeneration and/or disc herniation – can range from non-invasive methods such as physiotherapy to more serious operations such as discectomy (removal of herniated disc material) or spinal fusion (bridging two vertebra together with an implant). Often, further treatment may become necessary if the original treatment fails, leaving the patient faced with multiple surgeries and uncertain outcomes. Not only that, but every intervention changes the biomechanics of the spine, redistributing the mechanical loads among the different spine tissues and potentially having negative effects.

“MySpine’s calculations aim to quantify tissue load distributions in the patient models. The results obtained offer a unique way to anticipate all the possible biomechanically-driven outcomes in an educated way, and to understand their origin,” explains Jérôme. “In this way, it provides the clinician with enough information and rationale to estimate how individual patients may respond to treatments, and to decide when or which surgery is most appropriate, and at what time.” The most advanced models prepared for the MySpine platform were also designed to simulate the effects of intervertebral disc aging or vertebral bone remodelling over time, he adds.

MySpine is being tested in a clinical setting using real data from about 200 patients. For the moment, the platform has been validated by comparing the simulation results of 65 patients to the clinical decisions in cases where the diagnosis and decision-making about treatment was already very clear. In most cases, the simulations backed up the clinician’s instinct about what treatment was right, with testers saying that the platform is far less complex than other clinical software. The researchers now want to test MySpine on more complicated scenarios, where the rationale behind the decision-making might not be so clear.

Ultimately, they’d like to find an industry partner willing to help bring the product to market and make it widely available for use. “Clinics won’t need a high-level computing facility to run MySpine, as the idea is to make it accessible from any PC or even from a tablet,” says Jérôme.