Computational Bone Health Phenotyping in Multiple Myeloma

Multiple myeloma is a cancer of the bone marrow that affects thousands of people worldwide and accounts for about one in ten blood cancers. A defining feature of this disease is damage to the bones, which can lead to pain, fractures, reduced mobility, and a major loss in quality of life. However, multiple myeloma is not the same in every patient. It varies widely at the genetic, biological, and structural levels, making it difficult to predict how the disease will progress or how well a patient will respond to a specific treatment. Today, doctors routinely use whole-body CT scans to assess bone damage in multiple myeloma. While these scans are excellent at showing bone lesions, their evaluation is still largely visual and manual. This limits consistency, reproducibility, and the ability to capture subtle patterns across the entire skeleton. Important information about bone health, disease spread, and treatment effects may therefore remain unused. This project aims to change that. By combining state-of-the-art medical imaging, artificial intelligence, and clinical and genetic data, the project will develop an automated, comprehensive bone health profile for patients with multiple myeloma. Advanced computer algorithms will analyze whole-body CT scans to automatically identify bones and lesions, measure bone density, and characterize the size, shape, and spatial distribution of bone damage. Importantly, these features will be tracked over time, allowing researchers to observe how the disease evolves and how bones respond to treatment. The imaging-based bone health profile will then be linked with clinical information (such as laboratory values and treatment details) and genetic markers of the disease. By integrating these different data sources, the project aims to discover imaging biomarkers that reflect tumor biology, disease aggressiveness, and treatment effectiveness. This could help identify patients at higher risk earlier, predict outcomes more accurately, and support more personalized treatment decisions. In the long term, this work seeks to establish a new standard for assessing bone disease in multiple myelomamoving from qualitative visual assessment to quantitative, data-driven analysis. The results could improve diagnosis, treatment monitoring, and prognosis, ultimately contributing to better, more individualized care for patients living with this challenging disease.