Abstract Despite a detailed understanding of the genes mutated in myelodysplastic syndromes (MDS), diagnostic and treatment decisions for patients with MDS rely primarily on clinical and cytogenetic variables as considered by the Revised International Prognostic Scoring System (IPSS-R). Here we describe the recently developed Molecular IPSS (IPSS-M), a clinico-genomic risk stratification system that considers clinical, cytogenetic and genetic parameters; the implementation of a web portal to facilitate its adoption, a strategy to handle missing variables, and the worldwide utilization of the web calculator as a clinical support tool. The IPSS-M was trained on 2,957 clinically annotated diagnostic MDS samples profiled for mutations in 156 driver genes. To maximize the clinical applicability of the IPSS-M and account for missing genetic data (i.e genes missing from a sequencing panel), we implemented a strategy to calculate a risk score under three scenarios: best, worst and average. Last, we developed an online calculator as a standalone single-page web application using VueJs, and D3Js for the interactive visualizations, deployed through a CI/CD pipeline on AWS, where collection of anonymous usage analytics allows to track adoption and usability of the new proposed model. The model incorporates clinical, morphological, genetic variables informed by cytogenetics and constructed from the presence of oncogenic mutations in 31 genes. It delivers a unique risk score for each individual patient, as well as an assignment to one of six IPSS-M risk strata. Compared to the IPSS-R the IPSS-M re-stratified 46% of MDS patients. The model was validated in an external dataset of 754 MDS patients. We released an open-access IPSS-M web calculator available at https://mds-risk-model.com. By specifying the patient clinical and molecular profiles, the tool returns the patient-specific IPSS-M risk score and category, and the probability estimates over time for three clinical endpoints, i.e. leukemia free survival (LFS), overall survival, and incidence of leukemic transformation. Since its launch in June 2022, the calculator has been used by >6000 users in >75 countries, reaching a daily average of 100 users per day. Risks have been calculated for >45,000 patient profiles. 99.28% of the sessions initiated reach an IPSS-M score, suggesting that the calculator is intuitive and easy to use. We trained and validated the IPSS-M on 3,711 patients, a patient tailored risk stratification tool for patients with MDS that considers clinical, morphological and genetic variables inclusive of cytogenetics and mutations in one of 31 genes. The development of a web based tool was instrumental to the global dissemination of the model, enabling non-expert users to leverage the power of molecular biomarkers in risk stratification for patients with MDS. Citation Format: Elsa Bernard, Juan E. Arango Ossa, Heinz Tuechler, Peter L. Greenberg, Robert P. Hasserjian, Yasuhito Nannya, Sean M. Devlin, Maria Creignou, Philippe Pinel, Lily Monier, Juan S. Medina-Martinez, Dylan Domenico, Martin Jädersten, Ulrich Germing, Guillermo Sanz, Arjan A. van de Loosdrecht, Olivier Kosmider, Matilde Y. Follo, Felicitas Thol, Lurdes Zamora, Ronald F. Pinheiro, Andrea Pellagatti, Detlef Haase, Pierre Fenaux, Monika Belickova, Michael R. Savona, Virginia M. Klimek, Fabio P. Santos, Jacqueline Boultwood, Ioannis Kotsianidis, Valeria Santini, Francesc Solé, Uwe Platzbecker, Michael Heuser, Peter Valent, Kazuma Ohyashiki, Carlo Finelli, Maria Teresa Voso, Lee-Yung Shih, Michaela Fontenay, Joop H. Jansen, José Cervera, Norbert Gattermann, Benjamin L. Ebert, Rafael Bejar, Luca Malcovati, Mario Cazzola, Seishi Ogawa, Eva Hellström-Lindberg, Elli Papaemmanuil. Implementation and adoption of a web tool to support precision diagnostic and treatment decisions for patient with myelodysplastic syndromes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6168.
Read full abstract