Prognosis in Myelodysplastic Syndrome
October 1, 1999 5 minutes read
Prognosis in Myelodysplastic Syndrome
Abstract & Commentary
Source: Maes B, et al. Ann Oncol 1999;10:825-829.
Since 1982, the french-american-british (fab) classification has been used to categorize patients with myelodysplasia into five clinical categories: refractory anemia (RA), refractory anemia with ringed sideroblasts (RARS), refractory anemia with excess blasts (RAEB), RAEB in transformation to acute leukemia (RAEB-t), and chronic myelomonocytic leukemia (CMML).1 These distinctions were made based solely on bone marrow and peripheral blood criteria and vaguely separated patients into distinct prognostic groups, though substantial overlap between groups was noted and individual prognosis within groups varied widely.
Through the intervening years, other morphological, clinical, and biological variables have been reported to influence survival and a number of ad hoc prognostic indices were developed to try to improve assessing prognosis in an individual patient. In 1997, an international workshop combined data derived from seven large studies of patients with myelodysplasia and from this meta-analysis, an International Prognostic Scoring System (IPSS) was developed.2 The multivariate analysis that produced the IPSS was not generated on one sample and tested on another to assess its validity. However, a recent paper applied the IPSS criteria to 184 patients treated at the University Hospital in Leuven, Belgium.
The patients had been accrued between 1980 and 1997; median follow-up was 6.5 years. The median age was 64 years. The distribution of FAB diagnoses was as follows: RA 32%, RARS 12%, RAEB 29%, RAEB-t 12%, CMMS 15%. This distribution is largely comparable to that used to develop the IPSS, with the exception that the RAEB-t category was a bit larger in Leuven.
The IPSS prognostic scoring system is a point system based upon the percentage of bone marrow blasts, cytogenetic abnormalities, and number of lineages affected by cytopenia. Thus, for bone marrow blasts, less than 5% is zero points, 5-10% is 0.5 points, 11-20% is 1.5 points, and 21-30% is 2 points. For cytogenetics, the abnormalities are rated as good (normal karyotype or isolated abnormalities of deleted Y, deletion of 5q, or deletion of 20q), poor (presence of 3 or more abnormalities or chromosome 7 abnormalities) or intermediate (all other abnormalities). Good cytogenetics is zero points, intermediate is 0.5 points, and poor is 1 point. For cytopenias, cell counts must be below a specific threshold to be considered decreased; hemoglobin less than 10 g/dL, neutrophil count less than 1500/microliter, and platelet count less than 100,000/microliter. If zero or 1 lineage is decreased, the point score is 0; if two or three lineages are decreased, the point score is 1. The prognostic risk is formed by adding up the points: low-risk, total score 0; low intermediate (or intermediate-1), score 0.5-1; high intermediate (or intermediate-2), score 1.5-2; high-risk, score 2.5 or greater.
Using the IPSS scoring system, 22% were in the low- risk group, 46% in intermediate-1 group, 25% in intermediate-2, and 7% in the high-risk group. The IPSS significantly separated patients into distinct natural histories. Median survival was 6.5 years for the low-risk group, 2.6 years for intermediate-1, 1.3 years for intermediate-2 and 0.7 years for the high-risk group. Age had a significant independent effect on prognosis only in the low-risk category. Low-risk patients younger than age 60 had a median survival of eight years while those older than 60 had a median survival of 3.7 years. Using the FAB system, patients with RA, RARS, and CMML were mainly in the lowest two categories (10% of RA and 22% of CMML were as high as intermediate-2), patients with RAEB were evenly distributed between intermediate-1 and intermediate-2, and RAEB-t were evenly distributed between intermediate-2 and high-risk groups.
IPSS risk group predicted likelihood to evolve into acute leukemia: low-risk, 16%; intermediate-1, 26%; intermediate-2, 30%; and high-risk, 62%. The time from diagnosis to evolution to acute leukemia was also reflected by the risk group. The mean time for progression to acute leukemia was 28 months in the intermediate-1 group, 10 months for the intermediate-2 group, and four months for the high-risk group.
The results of this large, single-institution study confirm the clinical use of the IPSS prognostic scoring system for patients with myelodysplasia.
Commentary
The patients in the Leuven series differed somewhat from the patient population from which the IPSS scheme was devised. The Leuven group had a higher percentage of patients with poor prognosis cytogenetic features and the fraction of patients younger than age 60 years was significantly higher. This is a good thing. It is reassuring that the IPSS scheme is predictive even in patients who differ from those from whom the scheme was developed.
The IPSS system is useful to clinicians because it provides somewhat clearer definitions of the likely clinical course and it does so largely independent of FAB diagnosis. Though heterogeneity is inherent in such a biologically diverse set of clinical problems, the variability has been dramatically reduced by the new system. Not only does the IPSS appear to provide more valuable information than the FAB regarding prognosis, some evidence suggests that the IPSS can provide meaningful prognostic information within a single FAB diagnosis. For example, a group from Japan has reported its success in predicting the course of patients with RA.3
The IPSS is likely to be only a temporary improvement in the way clinicians approach patients with myelodysplastic syndromes. Reports of other prognostic factors not considered in the IPSS system continue to appear and seem to have merit. For example, immunohistochemical characterization of CD34 expression on the bone marrow may influence survival, as may other variables that better reflect the molecular pathology of the defect.4 Nevertheless, the advantage of a scoring system such as IPSS is that one can evaluate the efficacy of an intervention in a more meaningful fashion before moving to phase III randomized comparisons. At the moment, we are left to infer that improvements in supportive care and perhaps the earlier application of anti-leukemia therapy may have improved survival based on comparing survival of patients diagnosed before 1990 (for example) to survival of those diagnosed after 1990. It is possible that the relative homogeneity in survival curves from multiple sites based on IPSS would give us another standard for comparison.
Of course, the IPSS also has its practical limitations. Few places in the world are providing karyotypic analysis in a clinically useful time frame. Use of the IPSS to stratify patients for therapeutic studies cannot be validated unless one can reliably ascertain the risk group shortly after the patient presents. It remains to be seen whether an IPSS score can predict response to treatment. A recent phase II study failed to show a relationship between the IPSS group and the response to granulocyte-colony-stimulating factor plus erythropoietin.5
References
1. Bennett JM, et al. Br J Haematol 1982;51:189-199.
2. Greenberg P, et al. Blood 1997;89:2079-2088.
3. Takahashi M, et al. Am J Hematol 1998;58:250-252.
4. Soligo DA, et al. Am J Hematol 1994;46:9-17.
5. Hellstrom-Lindberg E, et al. Blood 199892:68-75.
Which of the following features does not affect the IPSS prognostic score for myelodysplastic syndrome?
a. karyotype
b. bone marrow blast count
c. presence of thrombocytopenia
d. presence of anemia
e. age greater than 60 years
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content