Identification of Poor Risk Patients in Low and Intermediate-1 (Int-1) IPSS MDS with the New Ipssr Index and Comparison with Other Prognostic Indexes. A Study by the Spanish Group of MDS (GESMD)

Abstract 702

Despite that low and intermediate-1 (int-1) IPSS groups are commonly considered as low risk diseases with a median overall survival exceeding 60 months, some of these patients will evolve as higher risk myelodysplastic syndrome (MDS). Recently several new prognosis indexes (PI) have been proposed: The new IPSSr, WPSSr, MD Anderson for lower risk patients (MDA) Index, and the Spanish Group of MDS (GESMD) proposal that considers as high risk those patients with int-1 IPSS and at least one of the following: platelets <30×10⁹/L, granulocytes <0.5×10⁹/L, poor or very poor-risk karyotype or the presence of bone marrow (BM) fibrosis. The aim of the study was to compare the four PI and to analyze which of them was the best to identify patients with the poorest risk (defined as those with a median overall survival (OS) lower than 30 months) and to segregate different risk groups in a population of lower risk MDS patients. Indexes were compared using the Akaike analysis methodology.

A total of 2410 patients from the Spanish registry of MDS with low or int-1 IPSS were included. Median age was 74 years (42.6% female). The IPSS value was of: 0, 0.5 and 1 in 1314, 761 and 335 patients, respectively. The four poor risk variables defined by the GESMD confirmed its adverse predictive value for OS: granulocytes <0.5×10⁹/L (n=101, P<0.001), platelets <30×10⁹/L (n=94, P<0.001), poor or very poor risk karyotype (n=35, P=0.007), and BM fibrosis (n=109 of 698 evaluable patients, P<0.001). The presence of at least one of these was associated with adverse prognosis in the int-1 group but not in the low IPSS risk group, thus only the former was considered as high risk.

The distribution of patients across the four PI is detailed in the Table. These new PI identified between 16.9% and 46% of patients having a median OS of around 30 months within the int-1 patients (wide line in the table), but none of the PIs could identify such a poor prognosis patients in the low IPSS group. The PI that identified the highest number of patients with shorter OS was the new IPSSr, while MDA IP was the most discriminative in the Akaike analysis. In conclusion, IPSS is not discriminative enough in the int-1 group. In contrast, the application of the new PI can be employed to better identify poor prognosis patients within the int-1 group who could benefit from a high-risk approach.

Figure.

View largeDownload slide

Actuarial curves of overall survival according to the different PI.

Figure.

View largeDownload slide

Actuarial curves of overall survival according to the different PI.

Close modal