Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Correspondence
  • Published:

Myelodysplastic Syndrome

Revised 15-item MDS-specific frailty scale maintains prognostic potential

Leukemia volume 34, pages 3434–3438 (2020)Cite this article

Subjects

Our recent publication in Leukemia described a 42-item frailty index (FI) based on cumulative deficits that improved risk stratification of existing clinical prognostication models and helped generate a composite prognostic score that comprised of IPSS, age, transfusion dependence, and frailty covariates [1]. However, the large number of measurements required to calculate this FI is cumbersome and reduces its clinical utility. Machine learning algorithms have been widely used to overcome some of the limitations of traditional statistical approaches [2]. One of the strengths of these algorithms is the inclusion of all variables prior to building the final model thus taking into account the interaction between these variables and the desired outcome without selecting only the significant variables in univariate analysis. This important step will assure that all variables are treated equally and their interactions with each other is accounted for. Herein, we applied similar methodology to identify the most critical elements of the FI and parsed it down to a more user-friendly form. We then evaluated the prognostic impact of this 15-item frailty scale (FS-15) compared with the original, and the additional value it contributed to the composite prognostic score [3, 4].

The MDS-CAN registry study received IRB approval at all sites, and all patients provided written informed consent prior to enrollment. The study is registered under clinicaltrials.gov number NCT02537990. A gradient Boosting Machine (GBMs) algorithm was used to build the model. GBMs are nonparametric decision tree algorithms that achieve state of the art performance compared to other algorithms in regards to structured data [5]. Shapley Additive Explanation (SHAP) values were used to identify the 15 most important variables that impacted the algorithm decision, which were subsequently used to calculate the new FS score [6]. These values are consistent and accurate estimations of the contributions of each variable to the final model prediction. More importantly, they account for feature interactions including instances where a given value may either increase or decrease risk. The variables that were identified by SHAP analysis were then used to calculate the new FS score. FS was calculated at enrollment in patients included in the national MDS registry (MDS-CAN) who were missing <20% of the selected 15 elements. The FS was calculated as a ratio of the number of weighted deficits divided by the total number considered for that patient. A log transformation was applied to the non-normally distributed FS scores, as per the original analysis. To identify the optimal FS cutoff scores, Cox proportional-hazard models of overall survival (OS) since enrollment were performed and compared between FS scores evaluated continuously and categorically. Hazard ratios and 95% confidence intervals (CI) of hazard ratios, and the generalized R2 statistic were also calculated for each FS cutoff score. The optimal cutoff score for the FS was determined based on highest R2 values whilst maintaining significant differences between all score categories.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Kaplan-Meier OS curves.

References

  1. Starkman R, Alibhai S, Wells R, Geddes M, Zhu N, Keating M, et al. An MDS-specific frailty index based on cumulative deficits adds independent prognostic information to clinical prognostic scoring. Leukemia. 2020;34:1394–406.

    Article  CAS  Google Scholar 

  2. Nazha A, Sekeres MA, Bejar R, Rauh M, Othus M, Komrokji R. et al. Genomic biomarkers to predict resistance to hypomethylating agents in patients with myelodysplastic syndromes using artificial intelligence. JCO Precis Ocol. 2019;3. https://doi.org/10.1200/po.19.00119.Licensed.

  3. Greenberg P, Cox C, LeBeau MM, Fenaux P, Morel P, Sanz G, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood. 1997;89:2079–88.

    Article  CAS  Google Scholar 

  4. Greenberg P, Tuechler H, Schanz J, Sanz G, Garcia-Manero G, Sole F, et al. Revised International Prognostic Scoring System for myelodysplastic syndromes. Blood. 2012;120:2454–4565.

    Article  CAS  Google Scholar 

  5. Hilton CB, Milinovich A, Felix C, Vakharia N, Crone T, Donovan C, et al. Personalized predictions of patient outcomes during and after hospitalization using artificial intelligence. npj Digit. Med. 2020;3:51. https://doi.org/10.1038/s41746-020-0249-z.

  6. Lundberg SM, Lee S-I. A unified approach to interpreting model predictions. In: Guyon I, Luxburg UV, Bengio S, Wallach H, Fergus R, Vishwanathan S, et al., editors. Advances in neural information processing systems 30. Curran Associates, Inc.; 2017, pp. 4765–4774.

  7. Chambless LE, Cummiskey CP, Cui G. Several methods to assess improvement in risk prediction models: extension to survival analysis. Stat Med. 2011;30:22–38.

    Article  Google Scholar 

  8. Uno H, Tian L, Cai T, Kohane IS, Wei LJ. A unified inference procedure for a class of measures to assess improvement in risk prediction systems with survival data. Stat Med. 2013;32:2430–42.

    Article  Google Scholar 

  9. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9:179–86.

    Article  CAS  Google Scholar 

  10. Bruera E, Kuehn N, Miller MJ, Selmser P, Macmillan K. The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment of palliative care patients. J Palliat Care. 1991;7:6–9.

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We thank Darren Kong and QxMD for their assistance in developing the MDS-Specific Frailty Scale online calculator.

Funding

The National Canadian Registry is financially supported by Celgene and Otsuka.

Author information

Authors and Affiliations

  1. Hematology/Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, ON, Canada

    Bo A. Wan, Rebecca Starkman, Richard. A. Wells, Lisa Chodirker, Liying Zhang & Rena Buckstein

  2. Hematology and Medical Oncology, Cleveland Clinic, Cleveland, OH, USA

    Aziz Nazha

  3. Geriatric Medicine/Oncology, University Health Network, Toronto, ON, Canada

    Shabbir Alibhai

  4. Hematology/Oncology, Tom Baker Cancer Centre, Calgary, AB, Canada

    Michelle Geddes

  5. Hematology, University of Alberta Hospital, Edmonton, AB, Canada

    Nancy Zhu

  6. Hematology/Oncology, Queen Elizabeth II Health Sciences Centre, Halifax, NS, Canada

    Mary-Margaret Keating

  7. Department of Medicine, McMaster University, Hamilton, ON, Canada

    Brian Leber

  8. Hematology/Oncology, University of Ottawa, Ottawa, ON, Canada

    Mitchell Sabloff & Grace Christou

  9. Hematology, St. Paul’s Hospital, University of British Columbia, Vancouver, BC, Canada

    Heather A. Leitch

  10. Hematology/Oncology, Dr. Georges-L-Dumont University Centre, Moncton, NB, Canada

    Eve St-Hilaire & Nicholas Finn

  11. Hematology/Oncology, Jewish General Hospital, Montreal, QC, Canada

    April Shamy

  12. Hematology/Oncology, Princess Margaret Hospital, Toronto, ON, Canada

    Karen W. L. Yee

  13. Hematology, McGill University Health Centre—Royal Victoria Hospital, Montreal, QC, Canada

    John Storring

  14. Hematology/Oncology, Vancouver General Hospital, Vancouver, BC, Canada

    Thomas J. Nevill

  15. Hematology/Oncology, Centre de Recherche du CHU de Quebec—Universite Laval, Quebec City, QC, Canada

    Robert Delage

  16. Hematology, Saskatoon Cancer Centre, Saskatoon, SK, Canada

    Mohamed Elemary

  17. Research Institute in Oncology and Hematology, CancerCare Manitoba & Departments of Internal Medicine and Biochemistry & Medical Genetics, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, MB, Canada

    Versha Banerji

  18. Hematology Clinical Trials, Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, ON, Canada

    Anne Parmentier, Mohammad Siddiqui, Aksharh Kirubananthaan & Alexandre Mamedov

Authors
  1. Bo A. Wan
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Aziz Nazha
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Rebecca Starkman
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Shabbir Alibhai
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Richard. A. Wells
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Michelle Geddes
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Nancy Zhu
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Mary-Margaret Keating
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Brian Leber
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Lisa Chodirker
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Mitchell Sabloff
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Grace Christou
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Heather A. Leitch
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Eve St-Hilaire
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Nicholas Finn
    View author publications

    You can also search for this author in PubMed Google Scholar

  16. April Shamy
    View author publications

    You can also search for this author in PubMed Google Scholar

  17. Karen W. L. Yee
    View author publications

    You can also search for this author in PubMed Google Scholar

  18. John Storring
    View author publications

    You can also search for this author in PubMed Google Scholar

  19. Thomas J. Nevill
    View author publications

    You can also search for this author in PubMed Google Scholar

  20. Robert Delage
    View author publications

    You can also search for this author in PubMed Google Scholar

  21. Mohamed Elemary
    View author publications

    You can also search for this author in PubMed Google Scholar

  22. Versha Banerji
    View author publications

    You can also search for this author in PubMed Google Scholar

  23. Anne Parmentier
    View author publications

    You can also search for this author in PubMed Google Scholar

  24. Mohammad Siddiqui
    View author publications

    You can also search for this author in PubMed Google Scholar

  25. Aksharh Kirubananthaan
    View author publications

    You can also search for this author in PubMed Google Scholar

  26. Alexandre Mamedov
    View author publications

    You can also search for this author in PubMed Google Scholar

  27. Liying Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  28. Rena Buckstein
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Rena Buckstein.

Ethics declarations

Conflict of interest

AN: Honoraria-DCI, Consulting or Advisory Role-Karyopharm Therapeutics, Tolero Pharmaceuticals, Speakers’ Bureau-Novartis, Incyte, Research Funding-Jazz Pharmaceuticals. RAW: Alexion: Honoraria, Research Funding; Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. MG: Celgene: Honoraria, Research Funding; Alexion: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. M-MK: Consultant—Seattle Genetics (brentuximab vedotin), Honoraria—Novartis, Scientific Advisory Board—Janssen, Shire Hoffman-La Roche, Sanofi Genzyme, Celgene. BL: Medical Advisory Boards and Speaker Bureau—Celgene, Pfizer, Novartis, AMGEN, Astellas, Otsuka, Jazz, Abbvie. MS: Advisory Boards—Jazz, Pfizer, Novartis, Astellas, Celgene. HAL: Advisory Boards, Honoraria, Research Funding—Abbvie, Alexion, Celgene, Novartis, Advisory Boards, Honoraria—Otsuka. ESt-H: Advisory Boards—BMS, Amgen, Celgene, Teva, Novartis, Speaker Honoratrium—BMS, Sanofi, Novartis. NF: Research Support—Boehringer Ingelheim, Roche, Pfizer, Merck, Honoraria—Roche, Novartis, Bristol Myers Squibb, Celgene, Pfizer, Scientific Advisory Board—Novartis, Pfizer, Lundbeck, Celgene, Janssen, Sanofi, Alexion, Roche, Ipsen, Takeda, Merck, Amgen, Bristol Myers Squibb, Astra Zeneca. AS: Advisory Boards: Amgen, Janssen, Abbvie, Celgene, Sponsored talks—Janssen. KWLY: Research Funding—Astex, Hoffman-La Roche, MedImmune, Merck, Millenium, Roche/Genentech, Honoraria—Novartis, Pfizer, Board of Directors or Advisory Committees: Astellas, Celgene, Novartis, Pfizer, Takeda. JS: Advisory Boards—Celgene, Novartis, Pfizer, Astellas, Abbvie, Amgen, speaking fees—Teva, Celgene, travel support—Pfizer. TJN: Ad boards and Sponsored talks—Celgene, Alexion, Novartis, Ad boards—Paladin Labs, Otsuka. RD: Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. VB: Industry Research Funding—Gilead, Janssen, Roche, Ad boards—Abbvie, Astra Zeneca, Janssen, Gilead, Roche, Sponsored talk—CAPhO Abbvie, Academic Research Support—CIHR, LLSC, Research Manitoba and CancerCare Manitoba Foundation, Licensing Fees—BIOGEN and The Dana-Farber Cancer Institute. RB: Research Funding support and Honoraria for Advisory Board panels—Celgene, Research Funding support—Otsuka, Takeda. The other authors declare that they have no conflict of interest.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wan, B.A., Nazha, A., Starkman, R. et al. Revised 15-item MDS-specific frailty scale maintains prognostic potential. Leukemia 34, 3434–3438 (2020). https://doi.org/10.1038/s41375-020-01026-y

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41375-020-01026-y

Search

Advanced search

Quick links