Machine Learning Improves Risk Stratification in Myelofibrosis: An Analysis of the Spanish Registry of Myelofibrosis

Abstract

Myelofibrosis (MF) is a myeloproliferative neoplasm (MPN) with heterogeneous clinical course. Allogeneic hematopoietic cell transplantation remains the only curative therapy, but its morbidity and mortality require careful candidate selection. Therefore, accurate disease risk prognostication is critical for treatment decision-making. We obtained registry data from patients diagnosed with MF in 60 Spanish institutions (N = 1386). These were randomly divided into a training set (80%) and a test set (20%). A machine learning (ML) technique (random forest) was used to model overall survival (OS) and leukemia-free survival (LFS) in the training set, and the results were validated in the test set. We derived the AIPSS-MF (Artificial Intelligence Prognostic Scoring System for Myelofibrosis) model, which was based on 8 clinical variables at diagnosis and achieved high accuracy in predicting OS (training set c-index, 0.750; test set c-index, 0.744) and LFS (training set c-index, 0.697; test set c-index, 0.703). No improvement was obtained with the inclusion of MPN driver mutations in the model. We were unable to adequately assess the potential benefit of including adverse cytogenetics or high-risk mutations due to the lack of these data in many patients. AIPSS-MF was superior to the IPSS regardless of MF subtype and age range and outperformed the MYSEC-PM in patients with secondary MF. In conclusion, we have developed a prediction model based exclusively on clinical variables that provides individualized prognostic estimates in patients with primary and secondary MF. The use of AIPSS-MF in combination with predictive models that incorporate genetic information may improve disease risk stratification.

Myelofibrosis and Survival Prognostic Models: A Journey between Past and Future

“…ML potentially plays a role yet to be fully exploited in medicine, opening new scenarios. In particular, in the context of MF, registry data of 1386 patients followed up in 60 different Spanish centers were analyzed by this computational approach [ 10 ] to create a new score called AIPSS. The training set data modeled OS and leukemia-free survival (LFS) based on standard clinical features collected at diagnosis, establishing an individualized prediction for every single patient.…”

“…Hemoglobin < 10 g/dL, white blood cell count < 4 or >30 × 10 9 /L, absolute monocyte count > 1 × 10 9 /L, presence of constitutional symptoms, platelet count < 100 × 10 9 /L, and blast ≥ 1% were early prognostic hematological parameters on which the first models were based [ 6 , 7 , 8 , 9 ]. Then, many other models were developed, the most recent ones with the help of molecular biology and artificial intelligence (AI) [ 10 ].…”

Refining risk prediction in pediatric acute lymphoblastic leukemia through DNA methylation profiling

“…An optimization problem simplifies the analysis by incorporating clinical outcomes and baseline prognostic information to derive a risk predictor for predicting outcomes in new patients [ 29 ]. This approach has paved the way for the development of prognostic and predictive tools in various onco-hematological fields, including myelofibrosis, myelodysplastic neoplasms, and multiple myeloma [ 30 – 32 ].…”

AIPSS‐MF machine learning prognostic score validation in a cohort of myelofibrosis patients treated with ruxolitinib

“…Beyond these, several working groups propose new scores to improve the ability to better identify patients with the worst outcome. In this context, the AIPSS‐MF (Artificial Intelligence Prognostic Scoring System for Myelofibrosis, based on machine learning) could play a pivotal role as it is based exclusively on clinical and easily accessible variables at diagnosis and has been shown to outperform the prognostic accuracy of the IPSS in PMF and the MYSEC‐PM in SMF 4 . Over the years, various drugs have been authorized or are ongoing tested for treating these conditions 5 .…”