Revised international prognostic scoring system for myelodysplastic syndromes

Abstract

The International Prognostic Scoring System (IPSS) is an important standard for assessing prognosis of primary untreated adult patients with myelodysplastic syndromes (MDS). To refine the IPSS, MDS patient databases from international institutions were coalesced to assemble a much larger combined database (Revised-IPSS [IPSS-R], n = 7012, IPSS, n = 816) for analysis. Multiple statistically weighted clinical features were used to generate a prognostic categorization model. Bone marrow cytogenetics, marrow blast percentage, and cytopenias remained the basis of the new system. Novel components of the current analysis included: 5 rather than 3 cytogenetic prognostic subgroups with specific and new classifications of a number of less common cytogenetic subsets, splitting the low marrow blast percentage value, and depth of cytopenias. This model defined 5 rather than the 4 major prognostic categories that are present in the IPSS. Patient age, performance status, serum ferritin, and lactate dehydrogenase were significant additive features for survival but not for acute myeloid leukemia transformation. This system comprehensively integrated the numerous known clinical features into a method analyzing MDS patient prognosis more precisely than the initial IPSS. As such, this IPSS-R should prove beneficial for predicting the clinical outcomes of untreated MDS patients and aiding design and analysis of clinical trials in this disease.

Figure 1

IWG-PM patients marrow blast subgroups. Impact on survival. Survival related to MDS patients' individual marrow blast percent categories (Kaplan-Meier curves, Dxy 0.3, P < .001). The number of patients in each category and their proportional representation are shown in Table 1.

Figure 2

IWG-PM patients marrow blast subgroups: Impact on AML evolution. Progression to AML related to MDS patients' individual marrow blast percent categories (Kaplan-Meier curves, Dxy 0.47, P < .001). The number of patients in each category and their proportional representation are shown in Table 1.

Figure 3

Survival based on IPSS-R prognostic risk-based categories. Survival related to MDS patients' prognostic risk categories (Kaplan-Meier curves, n = 7012; Dxy 0.43, P < .001). The number of patients in each category and their proportional representation are shown in Table 1.

Figure 4

AML evolution based on IPSS-R prognostic risk-based categories. Progression to AML related to MDS patients' prognostic risk categories (Kaplan-Meier curves, n = 6485; Dxy 0.52, P < .001). The number of patients in each category and their proportional representation are shown in Table 1.

Figure 5

Survival based on patient ages > 60 years vs ≤ 60 years related to their IPSS-R prognostic risk-based categories (Kaplan-Meier curves). Age-related differential survivals are shown for patients in all groups, particularly for those in lower risk categories.

Figure 6

Age-adjusted IPSS-R risk categories. The nomogram describes predicted survival based on patient age and IPSS-R risk status (IPSS-RA). To determine an age-adjusted risk categorization, for example, follow the horizontal line, starting at the IPSS-R risk score 3.5 on the vertical axis (Int [Intermediate] risk category per Table 4) to the age of the patient and record the color at that point. If the patient is 45 years, the 3.5′-line and the vertical 45-year line cross in the gray field, placing the patient in the Low risk category, whereas if the patient is 95 years the 3.5′-line and the 95-year line cross in the yellow field, placing the patient in the Intermediate risk category. As indicated, for most patients in the Very high risk category there is no change of risk group, whereas for most patients in the lower risk categories there is greater possibility of category change. Note the “dotted” vertical line at 70 years, which is at the median age of the IWG-PM patient cohort from which the basic risk category scores were calculated (ie, without need for age correction for these patients). The formula to generate the age-adjusted risk score in the figure: (years − 70) × [0.05 − (IPSS-R risk score × 0.005)]. Example: For the 45-year-old patient with an IPSS-R risk score of 3.5 (Intermediate risk): (45-70) × [0.05 − (3.5 × 0.005)] = −0.81. Thus, 3.5-0.8 = 2.7 [age-adjusted IPSS-R score, IPSS-RA: Low risk].

Figure 7

Comparison of IPSS-R and IPSS subgroups within the IWG-PM database patient cohort. Vertical axis represents IPSS-R categories' and horizontal axis, IPSS categories. The proportion of patients in each category is shown in Table 9. Kendall τ = 0.73.