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.

  • Analysis
  • Published:

Precision oncology for acute myeloid leukemia using a knowledge bank approach

Nature Genetics volume 49, pages 332–340 (2017)Cite this article

Subjects

Abstract

Underpinning the vision of precision medicine is the concept that causative mutations in a patient's cancer drive its biology and, by extension, its clinical features and treatment response. However, considerable between-patient heterogeneity in driver mutations complicates evidence-based personalization of cancer care. Here, by reanalyzing data from 1,540 patients with acute myeloid leukemia (AML), we explore how large knowledge banks of matched genomic–clinical data can support clinical decision-making. Inclusive, multistage statistical models accurately predicted likelihoods of remission, relapse and mortality, which were validated using data from independent patients in The Cancer Genome Atlas. Comparison of long-term survival probabilities under different treatments enables therapeutic decision support, which is available in exploratory form online. Personally tailored management decisions could reduce the number of hematopoietic cell transplants in patients with AML by 20–25% while maintaining overall survival rates. Power calculations show that databases require information from thousands of patients for accurate decision support. Knowledge banks facilitate personally tailored therapeutic decisions but require sustainable updating, inclusive cohorts and large sample sizes.

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

Figure 1: Systematic comparison of models.
Figure 2: Multistage modeling of patient fate.
Figure 3: Multistage outcome predictions for 1,024 patients.
Figure 4: Individualized risk exemplified for two patients.
Figure 5: Benefit of allograft in CR1 versus allograft after relapse.
Figure 6: Extrapolations and power calculations.

Similar content being viewed by others

References

  1. Van Allen, E.M. et al. Whole-exome sequencing and clinical interpretation of formalin-fixed, paraffin-embedded tumor samples to guide precision cancer medicine. Nat. Med. 20, 682–688 (2014).

    Article  CAS  Google Scholar 

  2. Mardis, E.R. Genome sequencing and cancer. Curr. Opin. Genet. Dev. 22, 245–250 (2012).

    Article  CAS  Google Scholar 

  3. Garraway, L.A. & Lander, E.S. Lessons from the cancer genome. Cell 153, 17–37 (2013).

    Article  CAS  Google Scholar 

  4. McDermott, U., Downing, J.R. & Stratton, M.R. Genomics and the continuum of cancer care. N. Engl. J. Med. 364, 340–350 (2011).

    Article  CAS  Google Scholar 

  5. Papaemmanuil, E. et al. Clinical and biological implications of driver mutations in myelodysplastic syndromes. Blood 122, 3616–3627, quiz 3699 (2013).

    Article  CAS  Google Scholar 

  6. Cancer Genome Atlas Network. Comprehensive molecular portraits of human breast tumors. Nature 490, 61–70 (2012).

  7. Imielinski, M. et al. Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing. Cell 150, 1107–1120 (2012).

    Article  CAS  Google Scholar 

  8. Manolio, T.A. et al. Global implementation of genomic medicine: we are not alone. Sci. Transl. Med. 7, 290ps13 (2015).

    Article  Google Scholar 

  9. Collins, F.S. & Varmus, H. A new initiative on precision medicine. N. Engl. J. Med. 372, 793–795 (2015).

    Article  CAS  Google Scholar 

  10. Papaemmanuil, E. et al. Genomic classification and prognosis in acute myeloid leukemia. N. Engl. J. Med. 374, 2209–2221 (2016).

    Article  CAS  Google Scholar 

  11. Cancer Genome Atlas Research Network. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N. Engl. J. Med. 368, 2059–2074 (2013).

  12. Gale, R.P., Wiernik, P.H. & Lazarus, H.M. Should persons with acute myeloid leukemia have a transplant in first remission? Leukemia 28, 1949–1952 (2014).

    Article  CAS  Google Scholar 

  13. Döhner, H. et al. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood 115, 453–474 (2010).

    Article  Google Scholar 

  14. Koreth, J. et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and meta-analysis of prospective clinical trials. J. Am. Med. Assoc. 301, 2349–2361 (2009).

    Article  CAS  Google Scholar 

  15. Flowers, M.E.D. & Martin, P.J. How we treat chronic graft-versus-host disease. Blood 125, 606–615 (2015).

    Article  CAS  Google Scholar 

  16. Burnett, A.K. et al. Curability of patients with acute myeloid leukemia who did not undergo transplantation in first remission. J. Clin. Oncol. 31, 1293–1301 (2013).

    Article  Google Scholar 

  17. Schlenk, R.F. et al. The value of allogeneic and autologous hematopoietic stem cell transplantation in prognostically favorable acute myeloid leukemia with double-mutant CEBPA. Blood 122, 1576–1582 (2013).

    Article  CAS  Google Scholar 

  18. Doria-Rose, V.P., Harlan, L.C., Stevens, J. & Little, R.F. Treatment of de novo acute myeloid leukemia in the United States: a report from the Patterns of Care program. Leuk. Lymphoma 55, 2549–2555 (2014).

    Article  CAS  Google Scholar 

  19. Cressman, S. et al. Economic impact of genomic diagnostics for intermediate-risk acute myeloid leukemia. Br. J. Haematol. 174, 526–535 (2016).

    Article  Google Scholar 

  20. Khera, N., Zeliadt, S.B. & Lee, S.J. Economics of hematopoietic cell transplantation. Blood 120, 1545–1551 (2012).

    Article  CAS  Google Scholar 

  21. Forbes, S.A. et al. COSMIC: exploring the world's knowledge of somatic mutations in human cancer. Nucleic Acids Res. 43, D805–D811 (2015).

    Article  CAS  Google Scholar 

  22. Schoenfeld, D.A. Sample-size formula for the proportional-hazards regression model. Biometrics 39, 499–503 (1983).

    Article  CAS  Google Scholar 

  23. Schmoor, C., Sauerbrei, W. & Schumacher, M. Sample-size considerations for the evaluation of prognostic factors in survival analysis. Stat. Med. 19, 441–452 (2000).

    Article  CAS  Google Scholar 

  24. Jameson, J.L. & Longo, D.L. Precision medicine—personalized, problematic and promising. N. Engl. J. Med. 372, 2229–2234 (2015).

    Article  CAS  Google Scholar 

  25. Stone, R.M. et al. The multi-kinase inhibitor midostaurin (M) prolongs survival compared with placebo (P) in combination with daunorubicin (D)-cytarabine (C) induction (ind), high-dose C consolidation (consol) and as maintenance (maint) therapy in newly diagnosed acute myeloid leukemia (AML) patients (pts) age 18–60 with FLT3 mutations (muts): an international prospective randomized (rand) P-controlled double-blind trial. Blood 126, 6 (2015).

    Article  Google Scholar 

  26. Ishwaran, H., Kogalur, U.B., Blackstone, E.H. & Lauer, M.S. Random survival forests. Ann. Appl. Stat. 2, 841–860 (2008).

    Article  Google Scholar 

  27. Ley, T.J. et al. DNMT3A mutations in acute myeloid leukemia. N. Engl. J. Med. 363, 2424–2433 (2010).

    Article  CAS  Google Scholar 

  28. Welch, J.S. et al. The origin and evolution of mutations in acute myeloid leukemia. Cell 150, 264–278 (2012).

    Article  CAS  Google Scholar 

  29. Delhommeau, F. et al. Mutation in TET2 in myeloid cancers. N. Engl. J. Med. 360, 2289–2301 (2009).

    Article  Google Scholar 

  30. Schlenk, R.F. et al.; German-Austrian Acute Myeloid Leukemia Study Group. All-trans retinoic acid as adjunct to intensive treatment in younger adult patients with acute myeloid leukemia: results of the randomized AMLSG 07-04 study. Ann. Hematol. 95, 1931–1942 (2016).

    Article  CAS  Google Scholar 

  31. Schlenk, R.F. et al. Phase 3 study of all-trans retinoic acid in previously untreated patients 61 years or older with acute myeloid leukemia. Leukemia 18, 1798–1803 (2004).

    Article  CAS  Google Scholar 

  32. Schlenk, R.F. et al. Prospective evaluation of allogeneic hematopoietic stem cell transplantation from matched related and matched unrelated donors in younger adults with high-risk acute myeloid leukemia: German–Austrian trial AMLHD98A. J. Clin. Oncol. 28, 4642–4648 (2010).

    Article  Google Scholar 

  33. Therneau, T.M., Grambsch, P.M. & Pankratz, V.S. Penalized survival models and frailty. J. Comput. Graph. Stat. 12, 156–175 (2003).

    Article  Google Scholar 

  34. Shah, R.D. & Samworth, R.J. Variable selection with error control: another look at stability selection. J. R. Stat. Soc. Series B Stat. Methodol. 75, 55–80 (2013).

    Article  Google Scholar 

Download references

Acknowledgements

We thank C. Holmes for stimulating discussions. We gratefully acknowledge D. Weber for clinical data managing, V. Teleanu for assistance in cytogenetics data classification and S. Kayser for assistance in morphological evaluation. This work was supported by grants from the Wellcome Trust (077012/Z/05/Z; P.J.C.), the Bloodwise charity (P.J.C.), the Leukemia and Lymphoma Society (P.J.C.) and the Deutsche Krebshilfe (109675; K.D.), in part by grants from the German Bundesministerium für Bildung und Forschung (BMBF) (01GI9981 (H.D.) and 01KG0605 (R.F.S. and H.D.)), by a Wellcome Trust Senior Clinical Research Fellowship (WT088340MA; P.J.C.), by an EHA early career fellowship (E.P.), and by the Deutsche Forschungsgemeinschaft (project B3, Sonderforschungsbereich (SFB) 1074; K.D. and L.B.); H.D. is coordinating investigator. L.B. is a Heisenberg Professor of the DFG (BU 1339/3-1). We are grateful to all members of the German–Austrian AML Study Group (AMLSG) for their participation in this study and for providing patient samples; a list of participating institutions and investigators appears in the Appendix of Papaemmanuil et al.5. AMLSG treatment trials were in part supported by Amgen and DKH grant 109675.

Author information

Author notes

  1. Moritz Gerstung and Elli Papaemmanuil: These authors contributed equally to this work.

  2. Hartmut Döhner and Peter J Campbell: These authors jointly directed this work.

Authors and Affiliations

  1. Cancer Genome Project, Wellcome Trust Sanger Institute, Hinxton, UK

    Moritz Gerstung, Elli Papaemmanuil, Inigo Martincorena, Niccolo Bolli, Ultan McDermott & Peter J Campbell

  2. European Bioinformatics Institute EMBL-EBI, Hinxton, UK

    Moritz Gerstung

  3. Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, New York, USA

    Elli Papaemmanuil

  4. Department of Internal Medicine III, Ulm University, Ulm, Germany

    Lars Bullinger, Verena I Gaidzik, Peter Paschka, Konstanze Döhner, Richard F Schlenk & Hartmut Döhner

  5. Department of Hematology, Hemostasis, Oncology and Stem Cell Transplantation, Hannover Medical School, Hannover, Germany

    Michael Heuser, Felicitas Thol & Arnold Ganser

  6. Division of Hematology, Fondazione IRCCS Istituto Nazionale dei Tumori, University of Milan, Milan, Italy

    Niccolo Bolli

  7. Department of Pathology (UOC), University of Otago, Christchurch, New Zealand

    Peter Ganly

  8. Department of Haematology, University of Cambridge, Cambridge, UK

    Peter J Campbell

Authors
  1. Moritz Gerstung
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Elli Papaemmanuil
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Inigo Martincorena
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Lars Bullinger
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Verena I Gaidzik
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Peter Paschka
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Michael Heuser
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Felicitas Thol
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Niccolo Bolli
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Peter Ganly
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Arnold Ganser
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Ultan McDermott
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Konstanze Döhner
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Richard F Schlenk
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Hartmut Döhner
    View author publications

    You can also search for this author in PubMed Google Scholar

  16. Peter J Campbell
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

M.G. developed the statistical methods, analyzed data and wrote the manuscript and supporting information, with input from E.P. and P.J.C. E.P. prepared and curated the genetic and clinical data. I.M. analyzed TCGA data. R.F.S., H.D., K.D., L.B., V.I.G., P.P., M.H., F.T. and A.G., along with all of the institutions contributing to the study group (AMLSG), recruited patients in this study, and collated and contributed clinical data. N.B., P.G. and U.M. provided input into analyses and interpretation of results. E.P., K.D., H.D., R.F.S. and P.J.C. initiated the study. P.J.C. and H.D. wrote the manuscript and are joint corresponding authors.

Corresponding authors

Correspondence to Hartmut Döhner or Peter J Campbell.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Constellations of risk factors for overall survival.

(a) Predicted log-hazard for overall survival versus patient outcome (color) of 1,540 patients. The black curve denotes the censored fraction, averaged over 50 adjacent patients. (b) Radar plots showing the distribution of aggregated risk in seven categories for five selected patients. Color indicates the observed survival, with solid lines around each symbol denoting an observed death and censored outcome otherwise. NK, normal karyotype; HR, hazard ratio. (c) Constellations of risk factors for 1,024 randomly selected patients, arranged by hierarchical clustering according to their risk profiles and laid out in 2D according to a Hilbert curve to preserve the proximity of the clustering. Related to Figure 2.

Supplementary Figure 2 Predicted effects of allograft in CR1 or after relapse.

(a–d) Predicted outcome for four patients, under a therapy regimen with standard chemotherapy in CR1 and allograft after relapse (left) or with allograft in CR1 (right). The height of each colored segment corresponds to the probability of being alive in remission (purple), alive after relapse (yellow), dead after relapse (green) or dead without relapse (blue). The bars at the bottom illustrate the actual outcome and treatment the patient received. Predictions are based on a model where the given patient was excluded. Related to Figure 3.

Supplementary Figure 3 Effects of allogeneic transplants.


(a–e) Kaplan–Meier curves for patients from different ELN risk groups. In each individual panel, patients were stratified by their predicted benefit from allografts in CR1 and by whether they had received allogeneic transplant in CR1.
There was a clear benefit to allogeneic transplants over chemotherapy in CR1 in the 12% of patients we predicted to have >10 percentage points improvement in absolute survival (HR = 0.5, 95% CI = 0.3–0.81, P = 0.003, ELN-stratified Cox proportional-hazards model). In contrast, for the other 88% of patients, the stratified hazard ratio was found to be HR = 0.74 with 95% CI = 0.55–0.89 (P = 0.01), matching our predictions of lower benefit from transplant of these patients in CR1. Related to Figure 4c.

Supplementary Figure 4 Predicted population-level survival gains for different estimates of overall benefit of early allograft.


(a) Scenario with low benefit of allografts in CR1 (5% quantile of all predictions). This scenario is based on the high toxicity of allografts, especially in elderly patients. Overall, only about one-third of patients (in CR, younger than 60 years) would benefit from an early allograft, as indicated by the decline of the red curve, beyond 75% total allografts (CR1 + relapse combined). The added survival of the knowledge bank would be about 1%, and the same survival could be achieved using much fewer allografts. (b) Scenario with intermediate benefit (50% quantile of all predictions), similar to Figure 4d. (c) Scenario with high benefit (95% quantile of all predictions). In this limit case, all patients are predicted to benefit from early allografts and there is no decline of the curves. Related to Figure 5.

Supplementary Figure 5 Screenshot of the multistage calculator.

The calculator is available online at http://cancer.sanger.ac.uk/aml-multistage. A docker image can be installed locally from https://hub.docker.com/r/mg14/aml-multistage. All source code can be found at http://github.com/mg14/aml-multistage.

Supplementary Figure 6 Implications of the number of genes sequenced.


(a) Subsampling of genes shows that the predicted variation in risk increases linearly with the average number of mutations per patient. This allows extrapolation from this cohort (111 genes sequenced; average of 2.3 driver mutations per patient) to an exome study (TCGA: ~20,000 genes sequenced; ~3.7 driver mutations predicted per patient.
(b) Algorithmic imputation of missing genes based on the multistage model, as implemented in the web portal. Shown is the concordance if only information on a subset of genes is present. With seven genes, the concordance is about 72% and plateaus at about 30 genes entered into the portal.

Supplementary Figure 7 Extrapolated power for detecting novel associations.

(a) Power calculations showing the expected ability to detect significant genomic features as a function of effect and mutation frequency for cohort sizes of 100, 1,540 and 10,000 patients. Shown are selected terms and the distribution of effect sizes in the random-effects model. (b) Distribution of P values (y axis) in comprehensive simulations using 100 (+), 1,000 (Δ) and 10,000 (o) simulated patients, respectively. P values are shown as a function of the product of N (sample size), ψ (uncensored fraction), p (mutation frequency) and β2 (squared log-hazard), as indicated on the x axis. The solid line indicates an analytical approximation by Schoenfeld, described in Schmoor (2000).

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–7 and Supplementary Note (PDF 15110 kb)

Supplementary Tables 1–6

Supplementary Tables 1–6 (XLSX 149 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Gerstung, M., Papaemmanuil, E., Martincorena, I. et al. Precision oncology for acute myeloid leukemia using a knowledge bank approach. Nat Genet 49, 332–340 (2017). https://doi.org/10.1038/ng.3756

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/ng.3756

This article is cited by

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing