Abstract

Abstract 4944

Background:

Despite its common occurrence, there are few large population-based studies on low-grade myelodysplastic syndromes (MDS), which became available to United States central cancer registries only in 2001. We evaluated the outcomes for low-grade MDS according to histological subtype and cause of death.

Methods:

The Surveillance, Epidemiology and End Results (SEER) database was searched for patients with refractory anemia (RA), RA with ringed sideroblasts (RARS), refractory cytopenia with multilineage dysplasia (RCMD) and MDS with 5q deletion (5q-), diagnosed between 2001 and 2009 and with complete demographic information. Overall survival (OS) and disease-specific survival (DSS) were calculated from diagnosis to death from any cause and death from MDS or acute leukemia, respectively. Univariate survival was estimated by the Kaplan-Meier method and curves compared by log rank. Univariate and multivariable Hazard ratios were calculated by Cox Proportional Hazards. Incidence rates were calculated from SEERStat 7. 1.

Results:

8606 patients met inclusion criteria, including 3866 (45%) RA, 2639 (31%) RARS, 1498 (17%) RCMD and 603 (7%) 5q-. The percentage of females was highest in 5q- (61%) and lowest in RCMD (36. 5%) (p < 0. 001). Incidence rates per 100, 000 population were 0. 55, 0. 44, 0. 26 and 0. 09 for RA, RARS, RCMD and 5q-, respectively. Deaths from MDS or acute leukemia compared to other causes were 15. 9% vs. 38. 4% in RA, 17. 4% vs. 33. 6% in RARS, 29. 9% vs. 28. 9% in RCMD and 33% vs 24. 8% in 5q-. Both OS and DSS were significantly higher for RA and RARS compared to RCMD and 5q- in both univariate and multivariable analyses (Table).

Conclusions:

The survival for 5q- was similar to RCMD and significantly inferior to RA and RARS. This surprisingly shorter survival for 5q-, particularly with the highest incidence of death from MDS or acute leukemia has not been previously reported. Since 5q- is the least common subset of MDS and studies usually have a small number of patients, it is possible that a larger database such as SEER, which includes 21% of all cases diagnosed in US, may have unveiled a more accurate prognosis for this MDS subset. Although we cannot rule out the presence of additional genetic abnormalities to the 5q- subgroup, the poor outcomes observed in our study warrant further evaluation.

Table
5-year OS5-year DSS
N5-year OSHRu (95% CI)HRm (95% CI)5-year DSSHRu (95% CI)HRm (95% CI)
RA 3866 37.7% 1.00 1.00 76.3% 1.00 1.00 
RARS 2639 41.3% 0.86 (0.81–0.92) 0.82 (0.77–0.88) 74.2% 1.01 (0.90–1.14) 0.96 (0.85–1.08) 
RCMD 1498 24.9% 1.48 (1.37–1.60) 1.5 (1.39–1.63) 48.8% 2.56 (2.26–2.89) 2.52 (2.23–2.86) 
5q- 603 26.3% 1.31 (1.17–1.47) 1.35 (1.17–1.47) 45.7% 2.56 (2.18–3.01) 2.68 (2.28–3.15) 
5-year OS5-year DSS
N5-year OSHRu (95% CI)HRm (95% CI)5-year DSSHRu (95% CI)HRm (95% CI)
RA 3866 37.7% 1.00 1.00 76.3% 1.00 1.00 
RARS 2639 41.3% 0.86 (0.81–0.92) 0.82 (0.77–0.88) 74.2% 1.01 (0.90–1.14) 0.96 (0.85–1.08) 
RCMD 1498 24.9% 1.48 (1.37–1.60) 1.5 (1.39–1.63) 48.8% 2.56 (2.26–2.89) 2.52 (2.23–2.86) 
5q- 603 26.3% 1.31 (1.17–1.47) 1.35 (1.17–1.47) 45.7% 2.56 (2.18–3.01) 2.68 (2.28–3.15) 

HRu – Hazard ratio in univariate analysis.

HRm – Hazard ratio in multivariable analysis after adjusting for age at diagnosis, sex, race and year of diagnosis.

Disclosures:

Dhodapkar:Celgene: Research Funding; KHK: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.