Dynamic Assessment of RBC-Transfusion Dependency (RBC-TD) Improves the Molecular International Prognostic Scoring System (IPSS-M) Risk Stratification of MDS

Aim: RBC-transfusion remains an important cornerstone of MDS management and 40-50% patients require regular transfusions. We have previously shown that RBC-TD predicts poor survival and dynamic assessment of RBC-TD improves the Revised-International Prognostic Scoring System (IPSS-R) (Hiwase et al Am J Hematol 2017). Integrating genomic profiling with hematologic and cytogenetic parameters, the IPSS-M improves the risk stratification of MDS; however, it does not include RBC-TD. This study evaluates the prognostic value of dynamic assessment of RBC-TD along with IPSS-M in an independent cohort. Method: IPSS-M categories were calculated utilizing blood counts, bone marrow blast counts, cytogenetic and somatic mutation data, with mean values assigned to the missing 18.1% of genetic mutation data (Bernard et al NEJM Evidence 2022). RBC-TD was defined as ≥8 RBC units within sixteen weeks. In RBC-TD patients, pre-transfusion baseline blood counts were used. Overall survival (OS) was estimated with the Kaplan-Meier (KM) method. Cox proportional-hazards modelling was used to assess the effect of RBC-TD with covariates IPSS-M, age and sex. Transfusion and survival data were recorded for a minimum of one year from diagnosis to a censor date of April 9th 2024. Patients were censored at time of hematopoietic stem cell transplantation. Results: Of 461 patients, there were 398 (86.3%) with MDS and 63 (13.7%) with MDS/MPN. The median age at diagnosis was 71 years (IQR 64-78) and 65.1% were male. The median OS was 39.6 months (95% CI 32.6-46.0). The distribution of IPSS-M risk categories was: Very Low (11.1%, n=51), Low (31.2%, n=144), Moderate low (15.0%, n=69), Moderate high (8.7%, n=40), High (16.7%, n=77), and Very high (17.4%, n=80). There was a significant difference in OS between these groups, with a median OS of 151.3 months for Very Low, 66.2 months for Low, 41.1 months for Moderate low, 32.9 months for Moderate high, 21.9 months for High, and 9.7 months for Very high (P <0.0001). At diagnosis, 23.2% of patients were classified as RBC-TD and the OS of RBC-TD patients was significantly shorter compared to RBC transfusion independent (RBC-TI) (12.4 vs. 54.8 months; P <0.0001). Frequency of RBC-TD dependency progressively increases with higher IPSS-M risk groups and is associated with poor OS even in IPSS-M Moderate High, High and Very High IPSS-M (10.3 vs. 23.2 months; P <0.0001). Furthermore, multivariable Cox proportional-hazards analysis showed that RBC-TD status at diagnosis (hazard ratio, HR 2.98, 95% CI 2.23-3.98; P < 0.001) was an independent predictor of OS after adjusting for IPSS-M, age and gender. OS was poor in IPSS-M Low (HR 2.06, 95% CI 1.26-3.38), Moderate low (HR 2.72, 95% CI 1.60-4.64), Moderate high (HR 4.02, 95% CI 2.25-7.19), High (HR 4.43, 95% CI 2.61-7.52), and Very high (HR 18.56, 95% CI 10.67-32.3) risk groups compared to Very Low IPSS-M. Relative to age under 60, median OS was poorer in patients of ages 60-70 years (HR 1.47, 95% CI 0.86-2.50), 70-80 years (HR 2.29, 95% CI 1.37-3.83), and over 80 (HR 3.62, 95% CI 2.14-6.13) respectively. Finally, males were associated with poor OS compared to females (HR 1.44, 95% CI 1.13-1.85). RBC-transfusion requirement can change during the disease course; 25.0%, 19.0% and 17.5% of alive patients were RBC-TD at 6, 12 and 24 months after diagnosis, respectively. Multivariable Cox proportional-hazards analysis showed poor outcome with RBC-TD at landmarks of 6-months (HR 2.94, 95% CI 2.15-4.02), 12-months (HR 3.86, 95% CI 2.72-5.47) and 24-months (HR 2.95, 95% CI 1.92-4.54), after adjusting for IPSS-M category, age category and sex at diagnosis. Conclusion: RBC-TD status provides dynamic risk information independent of IPSS-M and should be considered for guiding treatment decisions including potentially curative allogenic stem cell transplantation.