As discussed throughout this chapter, MDS is heterogeneous in both morphology and clinical course. Patients with MDS typically have at least one persistent otherwise unexplained cytopenia (hemoglobin less than 11 g/dL, ANC < 1500/mm3, or platelet count less than 100,000/mm3), plus one of the following: (1) ≥5% blasts in the bone marrow, (2) any of several chromosomal abnormalities that are found recurrently in MDS; (3) a meaningful degree of dysplasia, usually 10% or more, detectable in at least one cell lineage on bone mar row examination, or (4) other evidence of clonal hematopoiesis.20 In the past this latter category has usually comprised either karyotype or FISH results demonstrating chromosomal abnormalities missed by conventional karyotype, but evidence of clonal hematopoiesis based on detection of recurrent somatic mutations may need to be considered as well. However, given that at least 10% of the population aged 70 or older harbor point mutations in genes associated with myeloid malignancies, detection of such a mutation in the absence of other diagnostic criteria cannot be considered equivalent to MDS. Patients with unexplained cytopenias who have clonal mutations have a considerably higher likelihood of developing MDS meeting WHO criteria or another myeloid neoplasm compared to those who lack clonal mutations, especially if multiple mutations or splicing mutations are present, or the clonal burden is high.
As described above, the WHO criteria divide adult MDS into several general categories. Practitioners should note that these criteria have several limitations, the most obvious being the complexity of the criteria themselves, which can make them cumbersome to use. Second, as described above, there are a number of conditions that can mimic aspects of MDS, and the diagnostic criteria assume these diagnostic possibilities have been eliminated. Third, morphologic criteria are still based around arbitrary percentages of dysplasia and blasts, which can be interpreted differently by different pathologists or may be vulnerable to sampling variations. Finally, although WHO criteria incorporate cytogenetics to a limited extent, the criteria have not yet been updated to include types of newer genetic results that have been shown to have diagnostic value, and it is likely that incorporating some of these features will improve diagnostic clarity.
IPSS and IPSS-R
Although the WHO classification system (and the FAB system before it) is the accepted standard for diagnosing MDS, WHO criteria con fer limited prognostic information, so other systems have been developed expressly for this purpose. The most widely used has been the International Prognostic Scoring System (IPSS, Table1) and its newer revised iteration, the IPSS-R (Table 2). The original IPSS was first published in 1997 and was derived by analyzing baseline characteristics in over 800 newly diagnosed MDS patients from all diagnostic categories who received supportive care. Patients were scored based on three characteristics: the percentage of blasts in the marrow, the presence of specific cytogenetic abnormalities, and the number of cytopenias in the blood. Patients were then stratified based on total score into four risk groups (low, intermediate-1, intermediate-2, and high) that were shown to have prognostic value, both in estimating the chance of progressing to AML and in estimating the duration of overall survival. The original IPSS excluded patients with therapy-related MDS, patients with proliferative CMML, and patients who underwent AHCT.
Table1. 1997 International Prognostic Scoring System for Myelodysplastic Syndromes
Table2. 2012 Revised International Prognostic Scoring System for Myelodysplastic Syndromes
The original IPSS had several limitations that ultimately necessitated its revision. First, it was formulated four years prior to the adoption of the first iteration of WHO classification criteria, and some of the changes in those criteria (for instance, changing the cutoff for AML from 30% blasts to 20% blasts) invalidated certain aspects of the IPSS. Furthermore, broad application of the criteria to larger populations of MDS patients began to reveal that the original IPSS gave insufficient weight to some important clinical features, particularly the presence of severe cytopenias and complex karyotype. The IPSS was only validated in de novo disease, not t-MDS, and only in patients treated with supportive care alone. In response to some of these evident limitations, several alternative prognostic systems were proposed, but none were widely adopted.
The IPSS-R, introduced in 2012, revised the MDS blast cutoff to 20% and assigned points for severity of cytopenias. Furthermore, it incorporated a broader range of cytogenetic abnormalities than the original IPSS, rebalanced karyotype importance versus blast proportion, and stratified patients into five risk categories instead of four. The resulting system thus incorporates more informative biology and offers better refinement of prognosis than its predecessor.
The median survival times estimated by the IPSS and IPSS-R are largely reflective of natural history since they were originally derived in patients who did not receive any therapy for their MDS. However, IPSS-R stratifications have been shown to retain relevance when applied to treated populations, including risk stratification of patients receiving hypomethylating agents and those who undergo allogeneic transplant.
Nevertheless, the IPSS-R still has limitations. It does not incorporate some laboratory findings, such as the LDH and the serum ferritin, that have been shown to have independent prognostic value in dedicated studies. It does not capture the kinetics of disease or comorbid conditions. Additionally, similar to the WHO classification criteria, IPSS-R does not include any mutational data other than cytogenetics.
A molecular classification system, IPSS-RM, has been in development for more than 5 years. A European consortium evaluated more than 2,000 cases and identified 8 distinct mutation clusters with variant prognosis. Currently there is a high degree of variability in how clinicians use molecular profiling in evaluating patients with suspected MDS, and the availability of sequencing in different practice settings. The complexity of developing a molecular classification system appropriate for clinical use is underscored by the heterogeneity of co-mutation patterns, clonal architecture, and clone size associated with even a single allele.
Other Risk Stratification Systems
The limitations of the IPSS prompted attempts at more refined risk stratification using alternative models. Few of these have gained much traction in clinical practice but retain value for use in clinical trials and retrospective studies. For instance, a need for better stratification of lower-risk patients (i.e., IPSS-low or Int-1) led to the creation of a Lower-Risk Prognostic Scoring System (LR-PSS) by the MD Anderson group. The LR-PSS uses criteria similar to the IPSS but sets lower thresholds for point assignment (e.g., 2 points for a platelet count <50 × 109/L and 1 point for bone mar row blasts ≥4%). Subsequent studies have confirmed the LR-PSS’s ability to risk stratify patients, with those in the highest risk group having the greatest risk of developing AML, and the poorest overall survival.
Other attempts to improve prognostic systems have included attempts to incorporate formal measures of comorbidity into the IPSS, and an attempt to combine the WHO classification system with the IPSS that has been dubbed the WPSS. This latter system was formulated prior to the introduction of the IPSS-R.
