Computer-aided effective fracture risk stratification of patients with vertebral metastases for personalised treatment through robust computational models validated in clinical settings

Project description

Detailed Description of METASTRA

Cancer patients (2.7 million in Europe) with a positive prognosis face a high incidence of secondary tumours (approximately 1 million). Bone metastases extend to the spine in 30-70% of cases, diminishing the load-bearing capacity of the vertebrae and triggering fractures in 30% of instances. Clinicians have only two choices: either to undertake surgery to stabilize the spine or to leave the patient exposed to a substantial fracture risk. This decision is considerably subjective and can result in either unnecessary surgical procedures or fractures that significantly impact the patient's quality of life and cancer treatment.

The current standard-of-care for categorizing patients with vertebral metastasis relies on scoring systems rooted in radiographic images, with limited consideration of local biomechanics. Existing scoring systems fail to establish a clear indication for surgery in approximately 60% of cases. Consequently, there exists an unmet requirement to precisely and promptly quantify fracture risk to enhance patient stratification and determine the most suitable personalised treatment.

This interdisciplinary project will formulate biomechanical computational models, grounded in Artificial Intelligence (AI) and Physiology (VPH), to categorize patients with spine metastasis who are at a heightened risk of fractures and to identify the most appropriate personalised surgical interventions. The models will be amalgamated into a decision support system (DSS), empowering clinicians to proficiently categorize metastatic patients. Both the models and the DSS will be designed to meet regulatory standards and for potential future utilization.

METASTRA will recommend fresh guidelines for the stratification and management of metastatic patients. It is anticipated that the METASTRA approach will reduce the proportion of uncertain diagnoses from the current 60% to 20% of cases. This will not only alleviate patient suffering but also lead to potential cost savings of 2.4 billion euros per year.

Detailed description of the clinical gap, a scientific gap, and a technological gap

Clinical gap: SINS alone lacks the capability to accurately stratify the risk of fracture in patients with vertebral metastasis. Indeed, SINS exhibits low reliability for several reasons: (i) the score relies on a limited set of clinical parameters, with several factors that could enhance fracture risk assessment being overlooked; (ii) it relies on the expert opinions of clinicians, and (iii) its testing was conducted on a restricted number of cases (approximately 40). Currently, there is no robust decision support system (DSS) available to assist clinicians in the stratification and treatment of patients. Consequently, the assessment of fracture risk in vertebral metastases and their subsequent treatment largely hinges on the expertise of individual clinicians or working teams.

  • Scientific gap: Clinical data regarding fracture occurrences in patients with vertebral metastases are sparse and currently insufficient for constructing predictive models. While lytic tissue is acknowledged for its extreme fragility, predictions regarding the strength of blastic or mixed lesions based solely on bone density are unreliable. In fact, our understanding of the mechanical properties of metastatic tissue remains limited, and the influence of various metastasis factors (type, size, position) on the remaining strength of the affected vertebra is poorly understood.
  • Technological gap: Presently, the biomechanical assessment of metastatic features is imprecise and relies heavily on the subjective judgment of clinicians. There exists a pressing need for a dependable tool to forecast the fracture risk of metastatic vertebrae. This necessitates the development of an automated, validated, and robust toolkit, along with evidence-based, trustworthy solutions for identifying metastatic tissue from biomedical images, quantifying its pertinent biomechanical attributes, and assessing their impact on vertebral strength and fracture risk.
Project grant agreement number: 101080135