Leukemia: researchers develop a tool to identify the optimal time for a bone marrow transplant

An innovative tool has been developed by researchers to assist doctors in determining the most appropriate time to perform a stem cell transplant in patients with myelodysplastic syndromes, which are diseases that can be a prelude to leukemia.

This tool was the result of a collaboration between doctors, researchers and data scientists from the Center for Accelerating Leukemia/Lymphoma Research (CARL), a research center for accelerating clinical innovation in hematological diseases at Humanitas, the Humanitas AI Center and the Politecnico di Milano. The tool, which processes information on the genetic profile of the disease and the patient, is a step forward for precision medicine.

Stem cell transplant is the only potentially curative therapy for patients with myelodysplastic syndromes (MDS), blood diseases with a very heterogeneous clinical course that include extremely indolent forms alongside rapid evolution into acute leukemia. Identifying the patients at highest risk of disease progression and the best timing to perform a transplant is crucial to offer effective treatment strategies.

An international team of researchers led by Prof. Matteo Della Porta – Head of Leukemia and Myelodysplasia at Humanitas and Professor of Hematology at Humanitas University and Prof. Francesca Ieva – Professor of Statistics at the Politecnico di Milano, has created a clinical decision support tool (‘Clinical Decision Support System’) using innovative technologies. The aim is to help hematologists in the clinical process of deciding whether and when to perform a bone marrow transplant. Among the key parameters to guide the decision is a ‘molecular signature’ that includes genomic characteristics of the disease, which differ from patient to patient. The results of the study were published in the Journal of Clinical Oncology. The innovative methodology used was published jointly in the Journal of Clinical Oncology CCI.

‘The tool we have developed represents a significant step forward in the field of personalized medicine for hematological patients,’ stated Prof. Matteo Della Porta. To date, there is no clinical study that provides solid evidence to support the choice of the most appropriate time to perform a transplant procedure in the individual patient with MDS. With our model, it is possible to predict the evolution of each patient’s disease more accurately and suggest the best time to perform the transplant. This has only been possible thanks to a multidisciplinary research ecosystem, where doctors are in contact with researchers who are experts in technology, with the common goal of improving care for people with blood diseases. This ecosystem was formed within CALR, the Research Centre for Accelerating Clinical Innovation in Hematological Diseases of Humanitas, Humanitas AI Center and the Politecnico di Milano: an innovative reality to face the global challenge of creating precision medicine programs for all patients’.

‘We are very proud and satisfied with this result, which is an example of how a fruitful synergy between clinical practice and data analysis can enable research to make significant progress, and above all benefit patients,’ continues Prof. Francesca Ieva – This result, based on microsimulation and advanced statistical models, proves that today data analysis is an essential partner for clinicians. I am very grateful to the Politecnico for having invested significantly in Health Data Science in recent years’.

The study was made possible thanks to funding from two European projects – GenoMed4All and Synthema – and support from the AIRC Foundation for Cancer Research.

Myelodysplastic syndromes and prognostic markers

With more than 3000 diagnoses per year in Italy alone, myelodysplastic syndromes are blood diseases that occur most frequently in the elderly. To date, the only potentially curative therapy is hematopoietic stem cell transplantation (HSCT), a procedure which, however, presents major risks and criticalities, especially in the elderly.

Today, hematologists decide on treatment strategies, including the appropriateness and timing of donor transplantation, based on a so-called clinical ‘score’ called IPSS-R. This index is based on values such as the number of leukemic cells in the bone marrow, the levels of white, red and platelet cells in the blood, and the presence of abnormalities in the chromosomes of the blood cells. However, the score does not provide information on the genomic profile of the disease, i.e. on the presence of specific genetic mutations in the cancer cells that we now know are crucial for the evolution of the disease and consequently for the effectiveness of treatments.

In 2023, also under the leadership of Prof. Della Porta, the same group of researchers who signed the articles in the Journal of Clinical Oncology had demonstrated the usefulness of a new molecular score called IPSS-M. This indicator, also based on mutations in 31 genes, improves the ability to predict both the efficacy of stem cell transplantation and the life expectancy of patients with myelodysplastic syndromes (MDS).

The study

The current breakthrough started with a retrospective analysis of more than 7,000 patients with myelodysplastic syndromes in 26 institutions across Europe and the US. The researchers showed that the use of the new clinical decision support system based on the IPSS-M molecular score would change treatment choices in 17% of cases, compared to the current use of a score based on clinical parameters (IPSS-R). In particular, the new clinical decision support system showed that 15 per cent of the patients who would normally receive a transplant immediately, would instead benefit from a delayed procedure, while 19 per cent of patients who would normally receive a transplant delayed would instead benefit from an immediate procedure.

Della Porta explains: ‘This is because certain genetic mutations in the tumor provide key information about the true risk of disease progression, as well as the likelihood of recurrence. Including these mutations among the information on which the new decision-support system is based makes it possible to refine the clinical assessment, increasing the risk profile of some patients and reducing that of others, all other clinical indicators being equal’.

Today, many hospitals are still unable to perform routine molecular and genomic tests. However, the costs of these analyses are decreasing and the technologies required for these tests are becoming more accessible. This is why the switch to a molecular approach is getting closer.

‘The result was only possible thanks to the collaboration between doctors, data scientists and artificial intelligence experts, and to the possibility of accessing thousands of data shared within the consortium of the European projects GenoMed4All and Synthema. A fine example of how we can improve care by collaborating internationally and bringing together different disciplines and skills,’ Della Porta concludes.

Humanitas’ commitment to AI and digitalization

The Istituto Clinico Humanitas is the first hospital in Italy to have an integrated Artificial Intelligence Research Centre: the Humanitas AI Center. The center’s mission is to establish a space for sharing knowledge and skills between engineers, doctors and data scientists to achieve even higher standards of care through the construction of artificial intelligence algorithms. These algorithms should be able to process a large amount of clinical information in order to find combinations and define prediction models useful for the advancement of scientific research and innovation in areas such as predictive medicine, personalized treatments and diagnostic imaging. Humanitas is engaged in the application of Artificial Intelligence in various fields: from the prevention of colon and rectal cancer to diagnostic imaging and hematology.