MD Anderson Scoring System (MDACC) Predicts Outcomes After Hematopoietic Stem Cell Transplantation (HSCT) Better Than Other Prognostic Classifications In MDS

MDS is a heterogeneous group of hematopoietic stem cell disorders. Various prognostic models have been established to categorize patients with MDS including the International Prognostic Scoring System (IPSS), the Revised-IPSS (r-IPSS) and MDACC Scoring System. In this analysis, we compared those three classification schemas for their outcome predictability after HSCT.

We analyzed 291 MDS patients with a median age of 55 (interquartile range (IQR) 47-60.7 years) who underwent HSCT between January 2001 and December 2011. Histology by WHO classification included RA/RARS 48 (16.5%), RCMD 28 (9.6%), RAEB-1 59 (20.2%), RAEB-2 63 (21.7%), MDS unclassified 67 (23%), and CMML 26 (9%). Of 291, 117 patients (40.2%) had therapy related MDS (t-MDS). Conditioning regimen was myeloablative in 201 patients (69.1%) and reduced intensity in 90 patients (30.9%). Donors were matched related (MRD), matched unrelated (MUD), mismatched (MMD) in 131 (45%), 114 (39.2%) and 46 (15.8%) patients respectively. Risk categorization was performed by IPSS, r-IPSS and MDACC scoring systems at the time of diagnosis. IPSS, r-IPSS and MDACC scoring systems could be assessed in 239 (82.1%), 241 (82.8%) and 231 (79.4%) patients respectively.

The median follow up time of 109 survivors was 45 months. The median time from diagnosis to HSCT was 7.3 months (IQR 4.6-12.4 months). Three-year overall survival (OS) was 38.1% (95%CI 32.3-43.9) with 3-year event free survival (EFS) of 34.2% (95%CI 28.4-40). Cumulative relapse incidence (RI) at 3-year was 28.8% (95%CI 23.3-34.5). Cumulative incidence of treatment related mortality (TRM) at 3 year post-transplant was 27.9% (95%CI 22.6-33.6).

In univariate analysis, IPSS and r-IPSS were able to differentiate 2 risk groups for OS and EFS. High risk group per IPSS and very high risk group per r-IPSS had lower OS with hazard ratio (HR) of 2.4 to 3.1, lower EFS with HR of 2.2 to 2.7. While IPSS could not predict RI, very high risk group by r-IPSS had higher RI with HR of 3.6 compared with lower risk groups. Both IPSS and r-IPSS did not identify different risk groups for TRM.

On the other hand, MDACC scoring system was able to identify 4 different risk groups for EFS and OS in univariate analysis. Three-year OS was 68%, 46.1%, 30.3% and 11.4% for patients with MDACC risk score of 0-4, 5-6, 7-8 and ≥9 respectively (p<0.001) (figure1). Three-year EFS with MDACC risk score of 0-4, 5-6, 7-8 and ≥9 was 61.7%, 40.8%, 28.1% and 7.4% respectively (p<0.001). For RI and TRM, only MDACC risk scores of ≥9 was associated with poor outcomes with 3-year RI of 38.9% and 3-year TRM of 41.7% compared with 13.3% and 15.5% in risk scores of 0-4 (p=0.01 and p=0.01 respectively).

In multivariate analysis, MDACC score, matched unrelated and mismatched donors were associated with inferior OS (table1).

As a summary, MDACC risk scoring system for MDS better differentiates prognostic groups than IPSS or r-IPSS. Considering the high frequency of t-MDS among transplanted MDS patients, we propose that MDACC scoring system should be used for prognostic classification for hematopoietic transplantation.