SUMMARY: It is estimated that in the US approximately 13,000 people are diagnosed with MyeloDysplastic Syndromes (MDS) each year. The prevalence has been estimated to be from 60,000 to 170,000 in the US. MyeloDysplastic Syndromes are a heterogenous group of stem cell disorders characterized by marrow failure resulting in cytopenias, mainly symptomatic anemia, with associated cytogenetic abnormalities, and abnormal cellular maturation with morphologic changes in clonal cells. Majority of the individuals diagnosed with MDS are 65 years or older and die as a result of infection and/or bleeding, consequent to bone marrow failure. About a third of patients with MDS develop Acute Myeloid Leukemia (AML).
Elderly patients with mildly symptomatic anemia (macrocytic anemia) or pancytopenia are initially evaluated for B12, folate and iron deficiency, as well as hypothyroidism and hemolysis. The next recommended test for unexplained anemia is bone marrow examination, which is the current gold standard for diagnosis of MDS. However, this procedure is invasive and can be painful, and is occasionally associated with infectious and bleeding complications. For these reasons, many patients and their physicians may avoid this procedure, potentially delaying access to effective treatments.
The researchers in this study developed a noninvasive algorithm, to help diagnose or exclude MDS, without bone marrow evaluation. To develop this web-based app, 502 patients diagnosed with MDS based on bone marrow evaluation were randomly selected from the European MDS registry and this sample was combined with 502 controls with unexplained anemia, aged 50 years and older, who had normal findings on bone marrow evaluation. Patients with bone marrow involvement as a part of a hematological or other disease or with any degree of bone marrow dysplasia could not serve as controls. The authors using a logistic regression model were able to classify patients into 1 of 3 categories: probable MDS, probably not MDS, and indeterminate. The initial model that was developed by the researchers was further improved using the new Gradient-Boosted Models (GBMs), adding more variables based on their known association with MDS. They included 10 routinely measured and readily available demographic, clinical, and laboratory variables such as Age, Sex, Hemoglobin, White Blood Cell count, Platelet count, Mean Corpuscular Volume, Neutrophils, Monocytes, Glucose, and Creatinine, as well as including more patients. A web app was developed that would help clinicians diagnose, and more importantly rule out MDS noninvasively, without bone marrow examination.
The researchers then calculated Positive Predictive Values and Negative Predictive Values assuming a 20% prevalence of MDS within the population of patients to which the model would be applied in practice (patients with unexplained anemia, in whom other causes of anemia have been excluded, who would likely undergo bone marrow examination in clinical practice). Approximately 90% of the MDS patients had anemia, about 35-40% of them had neutropenia, thrombocytopenia or bicytopenia, and about 15% had pancytopenia, all according to WHO criteria. Using the more severe cytopenia criteria as would be used for the International Prognostic Scoring System (IPSS) score, about 50% of MDS patients were severely anemic, about 20-25% of patients were neutropenic, thrombocytopenic, or bicytopenic, and 5% were pancytopenic.
It was noted that this tool was reliably able to separate patients with and without MDS. This model had a sensitivity of 88% and specificity of 95%. In this patient population with unexplained anemia, probable MDS and probably not MDS could be determined in 86% of patients, leaving it to the patient and the physician to discuss whether the bone marrow evaluation should be performed in the indeterminate group, to make the definitive diagnosis. The researchers also determined the robustness of this model in patients with neutropenia, thrombocytopenia, as well as in those with bicytopenia and pancytopenia. It was noted that the predictive model continued to be reliable, especially in its ability to rule out MDS in almost all of these categories, with Negative Predictive Values all above 90% and relatively narrow 95% Confidence Intervals (CIs). Moreover, the lower boundaries of the 95% CI were all above 90%. However, when it came to making a diagnosis of MDS, the accuracy was somewhat diminished. The researchers attributed this to smaller number of patients in these groups, and further added that patients with multiple cytopenias should have bone marrow evaluation, irrespective of the model prediction. Based on this algorithm, the researchers developed a web-based predictor calculator which would serve as a practical tool for clinicians. The limitations of this algorithm are that morphology, blast percentage, genetics, and cytogenetics have not yet been integrated into the model.
It was concluded that based on this MDS model, a web-based computer app has been developed, to help the physician community to primarily exclude MDS in a cytopenic individual and also predict the possibility of MDS, without performing an invasive bone marrow evaluation. The authors plan to not only improve the predictive power of the model by increasing the number of measured variables, but also validate this model with independent prospective patient data, and develop a predictive prognostic tool, in addition to diagnosis.
A predictive algorithm using clinical and laboratory parameters may assist in ruling out and in diagnosing MDS. Oster HS, Crouch S, Smith A, et al. Blood Adv. 2021;5:3066-3075.
