A new machine-learning (ML) model based exclusively on clinical variables that are readily available in any health care facility demonstrated high prognostic accuracy for predicting survival and leukemic transformation in patients with myelodysplastic syndromes (MDS), according to a poster presentation at the 64th American Society of Hematology Annual Meeting and Exposition.
Adrián Mosquera Orgueira, MD, PhD, of the Hospital Clínico Universitario de Santiago de Compostela in Spain, and colleagues presented the results.
“This simple model has high prognostic accuracy for predicting survival and leukemic transformation in patients with MDS, outperforming the [Revised International Prognostic Scoring System (IPSS-R)],” the investigators said.
The new model, AIPSS-MDS (Artificial Intelligence Prognostic Scoring System for Myelodysplastic Syndromes), is based on eight readily available clinical variables determined at MDS diagnosis: age, sex, bone marrow blast, hemoglobin, leukocytes, platelet, percentage of neutrophils, and IPSS-R cytogenetic risk group.
To train and test the model, the investigators obtained registry data from 7,202 patients diagnosed with MDS between 2006 and 2022 in 90 Spanish institutions. The patients were randomly divided into a training set (80% of the cohort) and a test set (20% of the cohort). The random forest ML technique was used to model overall survival (OS) and leukemia-free survival (LFS) in the training set. The model’s performance was then validated in the test set.
In testing, AIPSS-MDS achieved high accuracy in predicting OS (c-indexes, 0.759 and 0.776 in the training and test sets, respectively) and LFS (c-indexes, 0.812 and 0.845 in the training and test sets, respectively). Both c-indexes and time-dependent areas under the curve confirmed that the model was superior to the IPSS-R and the age-adjusted IPSS-R in the global population. This difference persisted in different age ranges and in all MDS subgroups.
The investigators noted that AIPSS-MDS is based exclusively on clinical and cytogenetic variables available at diagnosis. This fact sets it apart from proposed models (such as recent versions of IPSS-R) that include molecular data, which are not available in all institutions.
The researchers said that a free, interactive online calculator for AIPSS-MDS is currently under development.
Reference
Orgueira AM, Encinas MP, Varela NAD, et al. Supervised machine learning improves risk stratification in newly diagnosed myelodysplastic syndromes: an analysis of the Spanish group of myelodysplastic syndromes. Abstract #468. Presented at the 64th ASH Annual Meeting and Exposition; December 10-13, 2022; New Orleans, Louisiana.