Improving the identification of clinical risk through a new computational model that reconstructs the tumour’s development over time using information contained in patients’ DNA. The development of this technology is at the heart of the study “Prognostic Score for Myelodysplastic Syndromes Based on Molecular Evolution” and is the result of collaboration between research groups from Milano-Bicocca and Humanitas, Yale University and Moffitt Cancer Center in the United States, together with other leading European centres in the field of haematology.
The research group from the University of Milano-Bicocca, which led the work, is composed of Daniele Ramazzotti from Medicine and Surgery, Alex Graudenzi from Computer Science, Systems and Communication, and Ivan Civettini from Milano-Bicocca and Vita-Salute San Raffaele University.
The study, recently published in NEJM Evidence (New England Journal of Medicine Evidence), presents the development and validation of a new tool for predicting disease evolution in myelodysplastic syndromes (MDS), a group of different blood disorders with varying probabilities of progressing to more aggressive forms such as leukaemia. The study introduces ProgEvo, a platform developed by the researchers that reconstructs the molecular evolutionary trajectories of tumours using patients’ genomic data. By combining genetic data with clinical information, the new model makes it possible to classify patients more accurately according to their level of risk, improving predictions about the course of the disease.
This approach led to the development of IPSS-M-Evo, a new tool that updates the system currently used by clinicians to assess patient risk by adding information on how the disease may evolve over time on the basis of the genetic profile. The model was trained and validated on independent cohorts comprising more than 6,000 patients in total, showing a consistent improvement in predicting survival and the risk of disease progression.
A particularly significant result is the reclassification of approximately 40 per cent of patients into different clinical risk groups, representing a significant improvement compared with the tools currently in use. The model is also accompanied by freely accessible web tools for risk calculation and clinical decision support.
«With ProgEvo, we sought to move beyond the simple analysis of the mutations present in a tumour,» said Daniele Ramazzotti. «We do not limit ourselves to observing which genetic alterations are present, but reconstruct the pathway through which the disease may have developed. This evolutionary information can help identify patients at greater risk and support personalised clinical decisions.»