Machine learning can identify newly diagnosed patients with CLL at high risk of infection.
Journal
Nature communications
ISSN: 2041-1723
Titre abrégé: Nat Commun
Pays: England
ID NLM: 101528555
Informations de publication
Date de publication:
17 01 2020
17 01 2020
Historique:
received:
04
06
2019
accepted:
11
12
2019
entrez:
19
1
2020
pubmed:
19
1
2020
medline:
10
4
2020
Statut:
epublish
Résumé
Infections have become the major cause of morbidity and mortality among patients with chronic lymphocytic leukemia (CLL) due to immune dysfunction and cytotoxic CLL treatment. Yet, predictive models for infection are missing. In this work, we develop the CLL Treatment-Infection Model (CLL-TIM) that identifies patients at risk of infection or CLL treatment within 2 years of diagnosis as validated on both internal and external cohorts. CLL-TIM is an ensemble algorithm composed of 28 machine learning algorithms based on data from 4,149 patients with CLL. The model is capable of dealing with heterogeneous data, including the high rates of missing data to be expected in the real-world setting, with a precision of 72% and a recall of 75%. To address concerns regarding the use of complex machine learning algorithms in the clinic, for each patient with CLL, CLL-TIM provides explainable predictions through uncertainty estimates and personalized risk factors.
Identifiants
pubmed: 31953409
doi: 10.1038/s41467-019-14225-8
pii: 10.1038/s41467-019-14225-8
pmc: PMC6969150
doi:
Substances chimiques
Antineoplastic Agents
0
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
363Subventions
Organisme : NIDCR NIH HHS
ID : K08 DE026500
Pays : United States
Références
da Cunha-Bang, C. et al. Improved survival for patients diagnosed with chronic lymphocytic leukemia in the era of chemo-immunotherapy: a Danish population-based study of 10455 patients. Blood Cancer J. 6, e499 (2016).
pubmed: 27834937
pmcid: 5148052
doi: 10.1038/bcj.2016.105
Hallek, M. et al. Addition of rituximab to fludarabine and cyclophosphamide in patients with chronic lymphocytic leukaemia: a randomised, open-label, phase 3 trial. Lancet 376, 1164–1174 (2010).
doi: 10.1016/S0140-6736(10)61381-5
Goede, V. et al. Obinutuzumab plus chlorambucil in patients with CLL and coexisting conditions. N. Engl. J. Med. 370, 1101–1110 (2014).
pubmed: 24401022
doi: 10.1056/NEJMoa1313984
pmcid: 24401022
Farooqui, M. Z. H. et al. Ibrutinib for previously untreated and relapsed or refractory chronic lymphocytic leukaemia with TP53 aberrations: a phase 2, single-arm trial. Lancet Oncol. 16, 169–176 (2015).
pubmed: 25555420
doi: 10.1016/S1470-2045(14)71182-9
pmcid: 25555420
Hallek, M. et al. iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL. Blood 131, 2745–2760 (2018).
doi: 10.1182/blood-2017-09-806398
Andersen, M. A. & Niemann, C. U. Immune failure, infection and survival in chronic lymphocytic leukemia in Denmark. Haematologica 103, e330 (2018).
pubmed: 29970495
pmcid: 6029532
doi: 10.3324/haematol.2018.197889
Crassini, K. R., Best, O. G. & Mulligan, S. P. Immune failure, infection and survival in chronic lymphocytic leukemia. Haematologica 103, e329 (2018).
pubmed: 29970494
pmcid: 6029549
doi: 10.3324/haematol.2018.196543
Eichhorst, B. et al. First-line chemoimmunotherapy with bendamustine and rituximab versus fludarabine, cyclophosphamide, and rituximab in patients with advanced chronic lymphocytic leukaemia (CLL10): an international, open-label, randomised, phase 3, non-inferiority trial. Lancet Oncol. 17, 928–942 (2016).
pubmed: 27216274
doi: 10.1016/S1470-2045(16)30051-1
pmcid: 27216274
Wierda, W. G. et al. Prognostic nomogram and index for overall survival in previously untreated patients with chronic lymphocytic leukemia. Blood 109, 4679–4685 (2007).
pubmed: 17299097
doi: 10.1182/blood-2005-12-051458
pmcid: 17299097
Pepper, C. et al. Defining the prognosis of early stage chronic lymphocytic leukaemia patients. Br. J. Haematol. 156, 499–507 (2012).
pubmed: 22171799
doi: 10.1111/j.1365-2141.2011.08974.x
pmcid: 22171799
International CLL-IPI working group. An international prognostic index for patients with chronic lymphocytic leukaemia (CLL-IPI): a meta-analysis of individual patient data. Lancet Oncol. 17, 779–790 (2016).
doi: 10.1016/S1470-2045(16)30029-8
Riches, J. C. et al. T cells from CLL patients exhibit features of T-cell exhaustion but retain capacity for cytokine production. Blood 121, 1612–1621 (2013).
pubmed: 23247726
pmcid: 3587324
doi: 10.1182/blood-2012-09-457531
Burger, J. A. et al. Blood-derived nurse-like cells protect chronic lymphocytic leukemia B cells from spontaneous apoptosis through stromal cell-derived factor-1. Blood 96, 2655–2663 (2000).
pubmed: 11023495
doi: 10.1182/blood.V96.8.2655
pmcid: 11023495
Herishanu, Y. et al. The lymph node microenvironment promotes B-cell receptor signaling, NF-kappaB activation, and tumor proliferation in chronic lymphocytic leukemia. Blood 117, 563–574 (2011).
pubmed: 20940416
pmcid: 3031480
doi: 10.1182/blood-2010-05-284984
Herman, S. E. M. et al. Modeling tumor-host interactions of chronic lymphocytic leukemia in xenografted mice to study tumor biology and evaluate targeted therapy. Leukemia 27, 2311–2321 (2013).
pubmed: 23619564
pmcid: 4126654
doi: 10.1038/leu.2013.131
Bagnara, D. et al. A novel adoptive transfer model of chronic lymphocytic leukemia suggests a key role for T lymphocytes in the disease. Blood 117, 5463–5472 (2011).
pubmed: 21385850
pmcid: 3109718
doi: 10.1182/blood-2010-12-324210
Niemann, C. U. et al. Disruption of in vivo chronic lymphocytic leukemia tumor-microenvironment interactions by ibrutinib–findings from an investigator-initiated phase II Study. Clin. Cancer Res. 22, 1572–1582 (2016).
pubmed: 26660519
doi: 10.1158/1078-0432.CCR-15-1965
pmcid: 26660519
Bachireddy, P. & Wu, C. J. Arresting the inflammatory drive of chronic lymphocytic leukemia with Ibrutinib. Clin. Cancer Res. 22, 1547–1549 (2016).
pubmed: 26847060
pmcid: 4818665
doi: 10.1158/1078-0432.CCR-15-3106
Pleyer, C. et al. Partial reconstitution of humoral and cellular immunity in patients with chronic lymphocytic leukemia treated with acalabrutinib. 132, 1874–1874 (2018).
Rossi, D. et al. Integrated mutational and cytogenetic analysis identifies new prognostic subgroups in chronic lymphocytic leukemia. Blood 121, 1403–1412 (2013).
pubmed: 23243274
pmcid: 3578955
doi: 10.1182/blood-2012-09-458265
Baliakas, P. et al. Refining prognosis of chronic lymphocyticleukemia with somatically hypermutated B-cell receptors: a novel prognosticindex on behalf of the european research initiative on CLL (ERIC). Haematologica 100, 52–52 (2015).
Delgado, J. et al. Chronic lymphocytic leukemia: A prognostic model comprising only two biomarkers (IGHV mutational status and FISH cytogenetics) separates patients with different outcome and simplifies the CLL-IPI. Am. J. Hematol. 92, 375–380 (2017).
pubmed: 28120419
doi: 10.1002/ajh.24660
pmcid: 28120419
Danish National Research Foundation. Precision medicine and bioinformatics with 320 mph: when data sharing increases the understanding of diseases. DNRF. https://dg.dk/en/precision-medicine-and-bioinformatics-with-320-mph-when-data-sharing-increases-the-understanding-of-diseases/ (2017).
Parikh, S. A. & Shanafelt, T. D. Prognostic factors and risk stratification in chronic lymphocytic leukemia. Semin. Oncol. 43, 233–240 (2016).
pubmed: 27040701
doi: 10.1053/j.seminoncol.2016.02.009
pmcid: 27040701
Fournier, P.-E. et al. Blood culture-negative endocarditis: improving the diagnostic yield using new diagnostic tools. Medicine 96, e8392 (2017).
pubmed: 29381916
pmcid: 5708915
doi: 10.1097/MD.0000000000008392
Salimnia, H. et al. Evaluation of the filmarray blood culture identification panel: results of a multicenter controlled trial. J. Clin. Microbiol. 54, 687–698 (2016).
pubmed: 26739158
pmcid: 4767991
doi: 10.1128/JCM.01679-15
Iroh Tam, P.-Y. et al. Detection of Streptococcus pneumoniae from culture-negative dried blood spots by real-time PCR in Nigerian children with acute febrile illness. BMC Res. Notes 11, 657 (2018).
pubmed: 30201041
pmcid: 6131910
doi: 10.1186/s13104-018-3770-2
Sivic & Zisserman. Video Google: a text retrieval approach to object matching in videos. Proc. Ninth IEEE Int. Conf. Comput. Vision 2, 1470–1477 (2003).
Raghavan, V., Bollmann, P. & Jung, G. S. A critical investigation of recall and precision as measures of retrieval system performance. ACM Trans. Inf. Syst. 7, 205–229 (1989).
doi: 10.1145/65943.65945
Saito, T. & Rehmsmeier, M. The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE 10, e0118432 (2015).
pubmed: 25738806
pmcid: 4349800
doi: 10.1371/journal.pone.0118432
Lundberg, S. M. & Lee, S.-I. A Unified approach to interpreting model predictions. Adv. Neural Inf. Process. Syst. 4765–4774 (2017).
Damle, R. N. et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood 94, 1840–1847 (1999).
doi: 10.1182/blood.V94.6.1840
Van Der Maaten, L. & Hinton, G. Viualizing data using t-SNE. J. Machin. Learning Res. 9, 2579–2605 (2008).
McInnes, L., Healy, J., Saul, N. & Großberger, L. UMAP: uniform manifold approximation and projection. JOSS 3, 861 (2018).
doi: 10.21105/joss.00861
Becht, E. et al. Dimensionality reduction for visualizing single-cell data using UMAP. Nat. Biotechnol. 37, 38–44 (2018).
doi: 10.1038/nbt.4314
Andersen, M. A. et al. Incidence and predictors of infection among patients prior to treatment of chronic lymphocytic leukemia: a Danish nationwide cohort study. Haematologica 103, e300–e303 (2018).
pubmed: 29519862
pmcid: 6029543
doi: 10.3324/haematol.2017.182006
da Cunha-Bang, C. et al. The danish national chronic lymphocytic leukemia registry. Clin. Epidemiol. 8, 561–565 (2016).
pubmed: 27822100
pmcid: 5094649
doi: 10.2147/CLEP.S99486
da Cunha-Bang, C., Christiansen, I. & Niemann, C. U. The CLL-IPI applied in a population-based cohort. Blood 128, 2181–2183 (2016).
pubmed: 27605511
doi: 10.1182/blood-2016-07-724740
pmcid: 27605511
Bishop, C. M. Pattern Recognition And Machine Learning (Springer, 2006).
Polikar, R. Bootstrap - inspired techniques in computation intelligence. IEEE Signal Process. Mag. 24, 59–72 (2007).
doi: 10.1109/MSP.2007.4286565
DiCiccio, T. J. & Efron, B. Bootstrap confidence intervals. Stat. Sci. 11, 189–212 (1996).
doi: 10.1214/ss/1032280214
Crassini, K. R. et al. Humoral immune failure defined by immunoglobulin class and immunoglobulin G subclass deficiency is associated with shorter treatment-free and overall survival in Chronic Lymphocytic Leukaemia. Br. J. Haematol. 181, 97–101 (2018).
pubmed: 29468645
doi: 10.1111/bjh.15146
pmcid: 29468645
Towards trustable machine learning. Nat. Biomed. Eng. 2, 709–710. https://doi.org/10.1038/s41551-018-0315-x (2018).
Burger, J. A. Targeting the microenvironment in chronic lymphocytic leukemia is changing the therapeutic landscape. Curr. Opin. Oncol. 24, 643–649 (2012).
pubmed: 22960555
doi: 10.1097/CCO.0b013e3283589950
pmcid: 22960555
Niemann, C. U. & Wiestner, A. B-cell receptor signaling as a driver of lymphoma development and evolution. Semin. Cancer Biol. 23, 410–421 (2013).
pubmed: 24060900
pmcid: 4208312
doi: 10.1016/j.semcancer.2013.09.001
Perrone, M. P. & Cooper, L. N. in How We Learn; How We Remember: Toward an Understanding of Brain and Neural Systems 342–358 (World Scientific, 1995).
Brown, G., Wyatt, J., Harris, R. & Yao, X. Diversity creation methods: a survey and categorisation. Inf. Fusion 6, 5–20 (2005).
doi: 10.1016/j.inffus.2004.04.004
Hsu, K.-W. A theoretical analysis of why hybrid ensembles work. Comput. Intell. Neurosci. 2017, 1930702 (2017).
pubmed: 28255296
pmcid: 5307253
doi: 10.1155/2017/1930702
Melville, P., Shah, N., Mihalkova, L. & Mooney, R. Experiments on ensembles with missing and noisy data. Lecture Notes Comput. Sci. 3077, 293–302 (2004).
doi: 10.1007/978-3-540-25966-4_29
Shui, C., Mozafari, A., Marek, J., Hedhli, I. & Gagné, C. Diversity regularization in deep ensembles. Preprint at arXiv https://arxiv.org/abs/1802.07881 (2018).
Wang, S. & Yao, X. Diversity analysis on imbalanced data sets by using ensemble models. in Proc. IEEE Symp. Comput. Intell. Data Mining (IEEE, 2009).
Baraldi, P., Razavi-Far, R. & Zio, E. A method for estimating the confidence in the identification of nuclear transients by a bagged ensemble of FCM classifiers. in Proc. NPIC&HMIT. 283–293 (2010).
Muhlbaier, M., Topalis, A. & Polikar, R. in Multiple Classifier Systems Vol. 3541 (eds Oza, N. C., Polikar, R., Kittler, J. & Roli, F.) 326–335 (Springer, 2005).
Molica, S., Giannarelli, D., Mirabelli, R., Levato, L. & Shanafelt, T. D. Chronic lymphocytic leukemia international prognostic index (CLL-IPI) in patients receiving chemoimmuno or targeted therapy: a systematic review and meta-analysis. Ann. Hematol. 97, 2005–2008 (2018).
pubmed: 29740683
doi: 10.1007/s00277-018-3350-5
pmcid: 29740683
Persimune. Danish National Foundation for Research’s (DNRF) centre of excellence (COE) for personalised medicine of infectious complications in immune deficiency (PERSIMUNE). http://www.persimune.dk (2019).
Lundberg, S. M. et al. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nat. Biomed. Eng. 2, 749–760 (2018).
pubmed: 31001455
pmcid: 31001455
doi: 10.1038/s41551-018-0304-0
Topol, E. J. High-performance medicine: the convergence of human and artificial intelligence. Nat. Med. 25, 44–56 (2019).
doi: 10.1038/s41591-018-0300-7
Dietterich & Thomas, G. Ensemble Methods in Machine Learning (Springer-Verlag, 2000).
Hansen, L. K. & Salamon, P. Neural network ensembles. IEEE Trans. Pattern Anal. Mach. Intell. 12, 993–1001 (1990).
doi: 10.1109/34.58871
Breiman, L. Random Forests. Mach. Learn. 45, 5–32 (2001).
Chen, T. & Guestrin, C. XGBoost: a scalable tree boosting system. in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discov. Data Mining (ACM Press, 2016).
Pedregosa, Fabian et al. Scikit-learn: machine learning in Python %J. J. Mach. Learn. Res. 12, 2825–2830 (2011).
Breiman, Leo. Bagging predictors. Mach. Learn. 24, 123–140 (1996).
Holland & John, H. Adaptation in Natural and Artificial Systems. (MIT Press, 1992).
Bahlo, J. et al. Evaluation of the international prognostic index for chronic lymphocytic leukemia (CLL-IPI) in elderly patients with comorbidities: analysis of the CLL11 Study Population. Blood 128, 4401–4401 (2016).
doi: 10.1182/blood.V128.22.4401.4401
Rajkomar, A. et al. Scalable and accurate deep learning with electronic health records. npj Digital Med. 1, 18 (2018).
doi: 10.1038/s41746-018-0029-1
Lundberg, S. M., Erion, G. G. & Lee, S.-I. Consistent individualized feature attribution for tree ensembles. Preprint at arXiv https://arxiv.org/abs/1802.03888 (2018).
Ribeiro, M. T., Singh, S. & Guestrin, C. Why should I trust you?: Explaining the predictions of any classifier. in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discov. Data Mining (ACM Press, 2016).