The MALMO Project combines artificial intelligence (AI) with mechanistic modeling to improve our understanding of melanoma progression and treatment resistance. Melanoma is one of the leading causes of cancer-related deaths globally, accounting for an annual death toll of 57,000. Late-stage melanoma sufferers typically have poor clinical prognosis (i.e., median survival rate of 8 months), even when surgical interventions are coupled with chemotherapeutic treatments. Treatment resistance has been suggested to stem from a phenomenon known as metabolic rewiring. As cancer cells infiltrate the host tissue cells, they create tumor microenvironments which are depleted of nutrients and oxygen. While this impacts the host, the cancer cells are able to adapt to these newly created conditions through metabolic rewiring, ensuring their survival even under the most harshest of conditions (even those of its own doing). In the MALMO Project, it is our assumption that capturing these metabolic changes over time will help elucidate the mechanisms behind treatment resistance. For our study, we focused our efforts on whole slide images (WSI) - digitized pathology slides. Using these WSIs, we created a 2D- and 3D-based pipeline, in which our biomarker of interest - the blood vessel, given it is a network providing nutrients and oxygen - was isolated from WSIs and an animated 3D model was captured of the vasculature network. The pipeline has two purposes: to automate current pathology diagnostic processes (which are currently arduous and time consuming in its processes) and to extract blood vessel features which can then be used to create a predictive mechanistic model that captures oxygen diffusion in melanoma samples. Research output and demonstration videos can be accessed through the below links. For our study, we focused our efforts on whole slide images (WSI) - digitized pathology slides. Using these WSIs, we created a 2D- and 3D-based pipeline, in which our biomarker of interest - the blood vessel, given it is a network providing nutrients and oxygen - was isolated from WSIs and an animated 3D model was captured of the vasculature network. The pipeline has two purposes: to automate current pathology diagnostic processes (which are currently arduous and time consuming in its processes) and to extract blood vessel features which can then be used to create a predictive mechanistic model that captures oxygen diffusion in melanoma samples. Research output and demonstration videos can be accessed through the below links
- Project website: MALMO project
- GitHub: MALMO repository
- Research output: MALMO research
