Abstract

Patients with low-risk myelodysplastic syndromes (MDS), defined as refractory anemia (RA), RA with ringed sideroblasts (RARS), or International Prognostic Scoring System (IPSS) scores of 0–1.0, are treated with growth factors (GF) or chemotherapy (Chemo) in the setting of transfusion dependence. Predictive models (PM) for response to GF have been developed based on red blood cell (RBC) transfusion needs and erythropoietin (epo) levels, but report no difference in survival among good, intermediate, and poor PM groups (the PM good group has few transfusion needs and a low epo level, while the PM poor group has high transfusion needs and high epo levels:

Jadersten et al.
Blood
2005
;
106
:
803
). The optimal therapy (chemo vs. GF) based on response rates (RR) and survival has not been defined.

Methods: We performed a decision analysis to determine the optimal therapy for low-risk MDS patients. Patients receiving chemo (n=417) or GF (n=382) were identified from a MEDLINE search on the keywords MDS, RA, RARS, treatment, GF, and chemo. Original articles with individual patient characteristics, RA or RARS MDS subtypes, and documented effect of therapy were included. From 1985 to 2005, 137 papers representing 2,301 patients with low-risk MDS were reviewed. Individual data was available on 799 patients. Chemo included differentiating agents, immunomodulators, and non-ablative cytotoxic agents; GF included erythropoietin, granulocyte colony-stimulating factor, and granulocyte-macrophage colony-stimulating factor. IPSS was calculated when sufficient data were available. RR were standardized and recalibrated according to the International Working Group (IWG) criteria and included complete response, partial response, and hematologic improvement. Median survival in GF patients derived from the Jadersten et al. study, and was validated with data from GF patients in our database. Survival in chemo patients was determined from the database. Cut-points for which treatment strategy to choose, based on RR and survival, were calculated.

Results: Baseline variables for chemo patients included (median (range)): age 65(21–83) years; requiring 2(0–8) units RBC transfusions/month; having an epo level of 279(17–4590). IPSS score was 0–1.0 in 88.3% of patients. GF patients were older (72 years, range 43–87), but with similar transfusion needs, IPSS scores, and epo levels to chemo patients. Overall, patients receiving chemo had a RR of 41.5% and a median survival of 78 months, compared to 46% and 44 months, respectively, for GF patients. Survival and RR differences were not statistically significant. Using the decision model, patients in a PM good response group should receive GF unless chemo can yield a RR at a cut-point of >34.1%. Patients in the PM intermediate response group should receive chemo if the RR to chemo is >10.6%; and those in the PM poor response group should receive chemo if the RR to chemo is >3.2%. RR and median survivals for chemo patients did not differ when they were classified according to the three PM response groups.

Conclusions: Based on available survival data, low-risk MDS patients falling into a good response predictive group (low RBC transfusion needs and low epo levels) should be treated with GF unless chemo RR are >34.1%. Other low-risk MDS patients should probably be treated with chemo. Additional cost and quality of life data are needed to incorporate into future decision models.

Author notes

Corresponding author