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

Howard S. Oster; Simon Crouch; Alexandra Smith; Ge Yu; Bander Abu Shrkihe; Shoham Baruch; Albert Kolomansky; Jonathan Ben-Ezra; Shachar Naor; Pierre Fenaux; Argiris Symeonidis; Reinhard Stauder; Jaroslav Cermak; Guillermo Sanz; Eva Hellström-Lindberg; Luca Malcovati; Saskia Langemeijer; Ulrich Germing; Mette Skov Holm; Krzysztof Madry; Agnes Guerci-Bresler; Dominic Culligan; Laurence Sanhes; Juliet Mills; Ioannis Kotsianidis; Corine van Marrewijk; David Bowen; Theo de Witte; Moshe Mittelman

doi:10.1182/bloodadvances.2020004055

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

Howard S. Oster, Simon Crouch, Alexandra Smith, Ge Yu, Bander Abu Shrkihe, Shoham Baruch, Albert Kolomansky, Jonathan Ben-Ezra, Shachar Naor, Pierre Fenaux, Argiris Symeonidis, Reinhard Stauder, Jaroslav Cermak, Guillermo Sanz, Eva Hellström-Lindberg, Luca Malcovati, Saskia Langemeijer, Ulrich Germing, Mette Skov Holm, Krzysztof MadryAgnes Guerci-Bresler, Dominic Culligan, Laurence Sanhes, Juliet Mills, Ioannis Kotsianidis, Corine van Marrewijk, David Bowen, Theo de Witte, Moshe Mittelman

Health Sciences

Research output: Contribution to journal › Article › peer-review

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 \gt; 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\ 0.95-0.97). MDS is predicted/excluded accurately in 86\range, 0-1) of less than 0.68 (GBM \lt; 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 \lt; 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.

Original language	Undefined/Unknown
Pages (from-to)	3066-3075
Number of pages	10
Journal	Blood Advances
Volume	5
Issue number	16
DOIs	https://doi.org/10.1182/bloodadvances.2020004055
Publication status	Published - 13 Aug 2021

Bibliographical note

Access to Document

10.1182/bloodadvances.2020004055

A predictive algorithm using clinical and laboratory parameters may assist in ruling out and in diagnosing MDS
© 2021 by The American Society of Hematology. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details
Final published version, 1.49 MB

Cite this

Oster, H. S., Crouch, S., Smith, A., Yu, G., Abu Shrkihe, B., Baruch, S., Kolomansky, A., Ben-Ezra, J., Naor, S., Fenaux, P., Symeonidis, A., Stauder, R., Cermak, J., Sanz, G., Hellström-Lindberg, E., Malcovati, L., Langemeijer, S., Germing, U., Holm, M. S., ... Mittelman, M. (2021). A predictive algorithm using clinical and laboratory parameters may assist in ruling out and in diagnosing MDS. Blood Advances, 5(16), 3066-3075. https://doi.org/10.1182/bloodadvances.2020004055

@article{4b0a4cf31a9c4b139ea1b1a2e71deefe,

title = "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 \gt; 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\ 0.95-0.97). MDS is predicted/excluded accurately in 86\range, 0-1) of less than 0.68 (GBM \lt; 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 \lt; 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.",

author = "Oster, {Howard S.} and Simon Crouch and Alexandra Smith and Ge Yu and {Abu Shrkihe}, Bander and Shoham Baruch and Albert Kolomansky and Jonathan Ben-Ezra and Shachar Naor and Pierre Fenaux and Argiris Symeonidis and Reinhard Stauder and Jaroslav Cermak and Guillermo Sanz and Eva Hellstr{\"o}m-Lindberg and Luca Malcovati and Saskia Langemeijer and Ulrich Germing and Holm, {Mette Skov} and Krzysztof Madry and Agnes Guerci-Bresler and Dominic Culligan and Laurence Sanhes and Juliet Mills and Ioannis Kotsianidis and {van Marrewijk}, Corine and David Bowen and {de Witte}, Theo and Moshe Mittelman",

note = "{\textcopyright} 2021 by The American Society of Hematology. Uploaded in accordance with the publisher{\textquoteright}s self-archiving policy. Further copying may not be permitted; contact the publisher for details ",

year = "2021",

month = aug,

day = "13",

doi = "10.1182/bloodadvances.2020004055",

language = "Undefined/Unknown",

volume = "5",

pages = "3066--3075",

journal = "Blood Advances",

issn = "2473-9529",

publisher = "American Society of Hematology",

number = "16",

}

Oster, HS, Crouch, S , Smith, A , Yu, G, Abu Shrkihe, B, Baruch, S, Kolomansky, A, Ben-Ezra, J, Naor, S, Fenaux, P, Symeonidis, A, Stauder, R, Cermak, J, Sanz, G, Hellström-Lindberg, E, Malcovati, L, Langemeijer, S, Germing, U, Holm, MS, Madry, K, Guerci-Bresler, A, Culligan, D, Sanhes, L, Mills, J, Kotsianidis, I, van Marrewijk, C, Bowen, D, de Witte, T & Mittelman, M 2021, 'A predictive algorithm using clinical and laboratory parameters may assist in ruling out and in diagnosing MDS', Blood Advances, vol. 5, no. 16, pp. 3066-3075. https://doi.org/10.1182/bloodadvances.2020004055

TY - JOUR

T1 - A predictive algorithm using clinical and laboratory parameters may assist in ruling out and in diagnosing MDS

AU - Oster, Howard S.

AU - Crouch, Simon

AU - Smith, Alexandra

AU - Yu, Ge

AU - Abu Shrkihe, Bander

AU - Baruch, Shoham

AU - Kolomansky, Albert

AU - Ben-Ezra, Jonathan

AU - Naor, Shachar

AU - Fenaux, Pierre

AU - Symeonidis, Argiris

AU - Stauder, Reinhard

AU - Cermak, Jaroslav

AU - Sanz, Guillermo

AU - Hellström-Lindberg, Eva

AU - Malcovati, Luca

AU - Langemeijer, Saskia

AU - Germing, Ulrich

AU - Holm, Mette Skov

AU - Madry, Krzysztof

AU - Guerci-Bresler, Agnes

AU - Culligan, Dominic

AU - Sanhes, Laurence

AU - Mills, Juliet

AU - Kotsianidis, Ioannis

AU - van Marrewijk, Corine

AU - Bowen, David

AU - de Witte, Theo

AU - Mittelman, Moshe

PY - 2021/8/13

Y1 - 2021/8/13

N2 - 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 \gt; 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\ 0.95-0.97). MDS is predicted/excluded accurately in 86\range, 0-1) of less than 0.68 (GBM \lt; 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 \lt; 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.

AB - 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 \gt; 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\ 0.95-0.97). MDS is predicted/excluded accurately in 86\range, 0-1) of less than 0.68 (GBM \lt; 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 \lt; 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.

U2 - 10.1182/bloodadvances.2020004055

DO - 10.1182/bloodadvances.2020004055

M3 - Article

SN - 2473-9529

VL - 5

SP - 3066

EP - 3075

JO - Blood Advances

JF - Blood Advances

IS - 16

ER -