Intensive Chemotherapy (IC) is More Effective than Hypomethylating Agents (HMA) for the Treatment of Younger Patients with Myelodysplastic Syndrome (MDS) and Elevated Bone Marrow Blasts

Retrospective studies of frontline IC in patients with MDS showed a lower complete remission (CR) rate and duration, and a higher incidence of prolonged myelosuppression as compared to patients with acute myeloid leukemia (AML) treated with similar regimens.¹ However, the outcome of younger patients (<60 years) with MDS and AML treated with IC was analyzed, and shown to be independent of bone marrow blast percentage at diagnosis.² The AZA-001 study has prospectively compared the use of HMA to conventional care regimens, including IC, in patients with high-risk MDS. While the use of azacitidine was associated with prolonged overall survival as compared to other treatment strategies, only 19 patients were treated with IC and of these only 12 were younger than 64, precluding any significant statistical comparison.³

We performed a retrospective analysis of 1063 patients with MDS or MDS/MPN, who had ≥ 10% bone marrow blasts and were followed at the University of Texas - MD Anderson Cancer Center between 10/1993 and 12/2017. Analysis was further restricted to 106 (10%) patients who were younger than 60 and received frontline treatment at our institution with either HMA or IC, the latter defined as a combination of anthracyclines and cytarabine. This study was approved by the Institutional Review Board and conducted in accordance with the principles of the Declaration of Helsinki.

Response duration (RD) was defined as time from achievement of response until progression or relapse. Overall survival (OS) was defined as time from initial therapy to date of death or last follow-up.

Exon 4, 5, 6, 7, 8 and 9 of TP53, and Exon 12 of NPM1 were amplified by polymerase chain reaction, and products processed by Sanger Sequence.

Categorical variables were compared using the χ2 or Fisher exact tests, as appropriate. Logistic regression was used for multivariate analysis (MVA) of categorical variables. Comparisons of survivals were made using the log-rank test, and a Cox proportional hazard regression model used for MVA.

Fifty-seven (54%) patients were treated with frontline HMA and 49 (46%) treated with frontline IC. Specific regimens are outlined in the Supplementary Table 1, and baseline characteristics in Supplementary Table 2.

On univariate analysis, patient treated with frontline IC were significantly younger (45 vs 55 years, p=0.03) had a higher percentage of bone marrow blasts (15% vs 13%, p<0.001), and had less frequently MDS/MPN as compared to MDS (2% vs 21%, p=0.003) or unfavorable cytogenetics (27% vs 46%, p=0.05).

Patients treated with frontline IC had a significantly higher overall response rate (ORR)(82% vs 60%, p=0.02), a significantly higher complete remission (CR) rate (63% vs 30%, p<0.001), and a significantly shorter median time to response (1 vs 2 months, p<0.001).

The association between all baseline characteristics shown in Supplementary Table 2 and achievement of ORR was evaluated; on univariate analysis, factors associated with achievement of higher ORR were age younger than 50 (82% vs 61%, p=0.03), presence of NPM1 mutation (100% vs 66%, p=0.05) and use of IC as compared to HMA (82% vs 60%, p=0.02). On MVA, only use of IC maintained its association (odds ratio [OR] 3.3, 95% confidence interval [CI] 1.4–7.9; p=0.009).

Forty-seven (96%) patients treated with IC and 56 (98%) treated with HMA discontinued treatment (p=0.60), after a median time of 4 and 5 months, respectively (p=0.74). Patients treated with frontline IC were more likely to discontinue treatment as a consequence of therapy completion (12% vs 0%, p=0.05) and less likely as a consequence of primary refractoriness or progression (39% vs 58%, p=0.08). A comparable percentage of patients treated with IC as compared to HMA discontinued treatment to proceed to stem cell transplant (SCT) in first remission (29% vs 34%, p=0.68). Of the 19 patients treated with HMA who proceeded to SCT, 13 were in CR, 1 in CRi and 5 had hematological improvement (HI), whereas of the 14 patients treated with IC who proceeded to SCT, 12 were in CR and 2 in CRi. A comparable percentage of patients treated with IC as compared to HMA discontinued treatment as a consequence of toxicity (8% vs 2%, p=0.18) or because of patient’s preference (8% vs 4%, p=0.41). Early deaths, defined as 30 day mortality, occurred at comparable rates during frontline IC (10%) and HMA (4%)(p=0.25).

After a median follow-up of 15 months (range, 1–178 months), 38 (51%) out of 74 patients who had achieved a response relapsed or progressed, and median RD was 19 months (range, 1–166 months). Eight (8%) patients transformed to AML during frontline treatment, 7 (12%) among patients receiving HMA, 1 (2%) among patients receiving IC (p=0.07). The association between all baseline characteristics, treatment strategies and RD was evaluated. On univariate analysis, factors associated with prolonged RD were absence of unfavorable cytogenetics (26 vs 7 months, p=0.04), SCT in first remission (not reached vs 14 months, p=0.009) and use of frontline IC (26 vs 11 months, p=0.03)(Figure 1A). On multivariate analysis, use of frontline IC maintained its association (hazard ratio [HR] 2.9, 95% CI 1.4–5.8, p=0.003). Of interest, among patients treated HMA, a trend for longer RD was observed for patients achieving CR as compared to patients achieving CRi or HI (68 months vs 14 months vs 5 months, p=0.06).

Figure 1.

A. Response duration (RD) by treatment group. B. Overall survival (OS) by treatment group.

At most recent follow-up, 65 (61%) patients died and median OS was 21 months (range, 1–178 months).

Factors associated with longer OS on univariate analysis were: non-therapy related MDS (31 vs 9 months, p<0.001), <15% bone marrow blasts (43 vs 16 months, p=0.01), absence of unfavorable cytogenetics (43 vs 9 months, p<0.001), intermediate IPSS score (27 vs 14 months, p=0.03), SCT in first remission (not reached vs 6 months, p=0.01), presence of NPM1 mutation (not reached vs 19 months, p=0.009), and use of frontline IC (43 vs 15 months, p=0.05)(Figure 1B). The association between use of frontline IC and prolonged OS was not maintained on MVA (p=0.74) and the only 3 factors which maintained their association were non-therapy related MDS (HR 0.49, 95% CI 0.25–0.92, p=0.03), absence of unfavorable cytogenetics (HR 0.1, 95% CI 0.03–0.4; p=0.01) and SCT in CR1 (HR 0.1, 95% CI 0.01–0.5; p=0.006). To this regard, among patients receiving SCT in CR1, no significant difference in OS was observed based on use of frontline IC as compared to HMA (not reached vs 21 months, p=0.35).

Among the 39 patients with baseline unfavorable cytogenetics, use of IC as compared to HMA did not significantly affect ORR (61% vs 58%, p=1), median RD (6 months vs 9 months, p=0.68) or median OS (8 months vs 11 months, p=0.24). However, SCT in CR1 associated with a significantly prolonged OS in this group (17 months vs 8 months, p=0.001). Among the 43 patients with baseline TP53 mutation (19 with concurrent complex karyotype, none with monosomal karyotype), use of IC as compared to HMA was associated with a trend for higher ORR (80% vs 48%, p=0.06), but did not significantly affect median RD (19 months vs 9 months, p=0.57) or median OS (28 months vs 11 months, p=0.15).

In the AZA-001 trial, patients with unfavorable karyotype abnormalities had a significant benefit from HMA as compared to other conventional regimens.³ In addition, prior studies have shown that among patients with MDS or AML the presence of TP53 mutations predicts better response to decitabine,⁴ and does not impair response to azacitidine. However, in our series (which includes a larger population of patients), these findings were not confirmed. Larger prospective and randomized trials, stratifying patients by these risk factors, are needed to shed light on this discrepancy.

It is also important to notice that the association between NPM1 mutation and higher ORR observed in our study was lost after adjusting for treatment arm, confirming our previous observation that this mutation may identify a subgroup of patients with MDS who are more sensitive to IC than HMA.⁵

Allogeneic SCT remains the only curative options for patients with high-risk MDS, its feasibility being greater in younger patients.⁶ Patients with unfavorable karyotype relapsed, though, independently from transplant status, suggesting that high-risk patients may benefit from post-transplant maintenance therapy.

In conclusion, IC is more effective than HMA as frontline treatment of younger patients with MDS and elevated bone marrow blasts, independently from baseline karyotype and TP53 status. We acknowledge the limits of our analysis, due to its retrospective nature and the unequal distribution of baseline characteristics between the 2 arms. Future strategies, combining targeted agents with either HMA or IC, should be investigated to determine whether this advantage will be maintained for IC.