A predictive algorithm using clinical and laboratory parameters may assist in ruling out and in diagnosing MDS

Abstract

We present a noninvasive Web-based app to help exclude or diagnose myelodysplastic syndrome (MDS), a bone marrow (BM) disorder with cytopenias and leukemic risk, diagnosed by BM examination. A sample of 502 MDS patients from the European MDS (EUMDS) registry (n > 2600) was combined with 502 controls (all BM proven). Gradient-boosted models (GBMs) were used to predict/exclude MDS using demographic, clinical, and laboratory variables. Area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were used to evaluate the models, and performance was validated using 100 times fivefold cross-validation. Model stability was assessed by repeating its fit using different randomly chosen groups of 502 EUMDS cases. AUC was 0.96 (95% confidence interval, 0.95-0.97). MDS is predicted/excluded accurately in 86% of patients with unexplained anemia. A GBM score (range, 0-1) of less than 0.68 (GBM < 0.68) resulted in a negative predictive value of 0.94, that is, MDS was excluded. GBM ≥ 0.82 provided a positive predictive value of 0.88, that is, MDS. The diagnosis of the remaining patients (0.68 ≤ GBM < 0.82) is indeterminate. The discriminating variables: age, sex, hemoglobin, white blood cells, platelets, mean corpuscular volume, neutrophils, monocytes, glucose, and creatinine. A Web-based app was developed; physicians could use it to exclude or predict MDS noninvasively in most patients without a BM examination. Future work will add peripheral blood cytogenetics/genetics, EUMDS-based prospective validation, and prognostication.

Graphical abstract

Figure 1.

GBM probability scores stratified by case (red) and control (green) status. The lavender region represents overlap between case and control patients. Threshold values of 0.68 (green vertical line) and 0.82 (red line) are indicated; above the red threshold value a patient is predicted to have MDS, below the green threshold, the patient is predicted not to have MDS. Between these 2 threshold values, no prediction is made. In this figure, case and control prevalence is assumed equal to illustrate the score distributions most clearly; in practice, case prevalence is likely to be much lower (we have taken 20% as indicative in our calculations).

Figure 2.

Receiver operating characteristic curve for the fitted GBM. The AUC is 0.96 (95% CI, 0.95-0.97).

Figure 3.

Relative influence values of variables in the GBM. Creat, creatinine; Gluc, glucose; Mono, monocyte; Neut, neutrophil.

Figure 4.

The Web-based app for the noninvasive diagnostic tool. (A) The quick response (QR) code and the full Web address allow entrance to the Web site. (B) Once in the site, the window opens for entering the values of the 10 variables and calculating the probability of having MDS. The variables: age, sex, Hb, MCV, WBC, neutrophil count, monocyte count, platelet count (Plt), serum creatinine, and serum glucose. F, female; M, male.

Figure 5.

Examples of the predictive app in practice. Values for a given patient are entered into the appropriate spaces, and the calculate button is pressed. A blue line indicates the probability of the patient having MDS. (A) Values are entered for a patient with pMDS. Note the position of the blue line in the red region. (B) Values for a patient who probably does not have MDS (pnMDS). (C) Patient with an indeterminate diagnosis. In this figure, a case prevalence of 20% is assumed (as opposed to Figure 1 where 50% was assumed).