Artificial intelligence models in chronic lymphocytic leukemia – recommendations toward state-of-the-art

References

• Andersen MA, Eriksen CT, Brieghel C, et al. Incidence and predictors of infection among patients prior to treatment of chronic lymphocytic leukemia: a Danish Nationwide Cohort Study. Haematologica. 2018;103(7):e300–e303.
   PubMed Web of Science ®Google Scholar
• Crassini KR, Best OG, Mulligan SP. Immune failure, infection and survival in chronic lymphocytic leukemia. Haematologica. 2018;103(7):e329.
   PubMed Web of Science ®Google Scholar
• Andersen MA, Niemann CU. Immune failure, infection and survival in chronic lymphocytic leukemia in Denmark. Haematologica. 2018;103(7):e330.
   PubMed Web of Science ®Google Scholar
• Strati P, Shanafelt TD. Monoclonal B-cell lymphocytosis and early-stage chronic lymphocytic leukemia: diagnosis, natural history, and risk stratification. Blood. 2015;126(4):454–462.
   PubMed Web of Science ®Google Scholar
• Moreira J, Rabe KG, Cerhan JR, et al. Infectious complications among individuals with clinical monoclonal B-cell lymphocytosis (MBL): a cohort study of newly diagnosed cases compared to controls. Leukemia. 2013;27(1):136–141.
   PubMed Web of Science ®Google Scholar
• Andersen MA, Moser CE, Lundgren J, et al. Epidemiology of bloodstream infections in patients with chronic lymphocytic leukemia: a longitudinal nation-wide cohort study. Leukemia. 2019;33(3):662–670.
   PubMed Web of Science ®Google Scholar
• Andersen MA, Niemann CU, Rostgaard K, et al. Differences and temporal changes in risk of invasive pneumococcal disease in adults with hematological malignancies: results from a nationwide 16-year cohort study. Clin Infect Dis. 2021;72(3):463–471.
   PubMed Web of Science ®Google Scholar
• da Cunha-Bang C, Simonsen J, Rostgaard K, 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. 2016;6(11):e499.
   PubMed Web of Science ®Google Scholar
• Agius R, Brieghel C, Andersen MA, et al. Machine learning can identify newly diagnosed patients with CLL at high risk of infection. Nat. Commun. 2020;11:363.
   PubMed Web of Science ®Google Scholar
• Molica S. Chronic lymphocytic leukemia prognostic models in real life: still a long way off. Expert Rev Hematol. 2021;14(2):137–141.
   PubMed Web of Science ®Google Scholar
• Kreuzberger N, Damen JA, Trivella M, et al. Prognostic models for newly-diagnosed chronic lymphocytic leukaemia in adults: a systematic review and meta-analysis. Cochrane Database Syst Rev. 2020;7:CD012022.
   PubMedGoogle Scholar
• Steyerberg EW, Vickers AJ, Cook NR, et al. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21(1):128–138.
   PubMed Web of Science ®Google Scholar
• Collins GS, de Groot JA, Dutton S, et al. External validation of multivariable prediction models: a systematic review of methodological conduct and reporting. BMC Med Res Methodol. 2014;14:40.
   PubMed Web of Science ®Google Scholar
• Molica S, Giannarelli D. Prognostic models for chronic lymphocytic leukemia (CLL): a systematic review and meta-analysis. Leukemia. 2021;35(2):615–618.
   PubMed Web of Science ®Google Scholar
• Lee C, Zame W, Alaa A, et al. Temporal quilting for survival analysis; 2019.
   Google Scholar
• Delgado J, Doubek M, Baumann T, 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. 2017;92(4):375–380.
   PubMed Web of Science ®Google Scholar
• Bulian P, Rossi D, Forconi F, et al. IGHV gene mutational status and 17p deletion are independent molecular predictors in a comprehensive clinical-biological prognostic model for overall survival prediction in chronic lymphocytic leukemia. J Transl Med. 2012;10:18.
   PubMed Web of Science ®Google Scholar
• 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. 2016;17:779–790.
   PubMed Web of Science ®Google Scholar
• Wierda WG, O'Brien S, Wang X, et al. Prognostic nomogram and index for overall survival in previously untreated patients with chronic lymphocytic leukemia. Blood. 2007;109(11):4679–4685.
   PubMed Web of Science ®Google Scholar
• Soumerai JD, Ni A, Darif M, et al. Prognostic risk score for patients with relapsed or refractory chronic lymphocytic leukaemia treated with targeted therapies or chemoimmunotherapy: a retrospective, pooled cohort study with external validations. Lancet Haematol. 2019;6(7):e366–e374.
   PubMed Web of Science ®Google Scholar
• Rossi D, Rasi S, Spina V, et al. Integrated mutational and cytogenetic analysis identifies new prognostic subgroups in chronic lymphocytic leukemia. Blood. 2013;121(8):1403–1412.
   PubMed Web of Science ®Google Scholar
• Pflug N, Bahlo J, Shanafelt TD, et al. Development of a comprehensive prognostic index for patients with chronic lymphocytic leukemia. Blood. 2014;124(1):49–62.
   PubMed Web of Science ®Google Scholar
• Molica S, Giannarelli D, Mirabelli R, et al. Unavailability of thymidine kinase does not preclude the use of German Comprehensive Prognostic Index: results of an external validation analysis in early chronic lymphocytic leukemia and comparison with MD Anderson Cancer Center Model. Eur J Haematol. 2016;96(1):72–77.
   PubMed Web of Science ®Google Scholar
• Molica S, Shanafelt TD, Giannarelli D, et al. The chronic lymphocytic leukemia International Prognostic Index predicts time to first treatment in early CLL: independent validation in a prospective cohort of early stage patients. Am J Hematol. 2016;91(11):1090–1095.
   PubMed Web of Science ®Google Scholar
• Wolpert DH, Macready WG. No free lunch theorems for optimization. IEEE Trans Evol Comput. 1997;1(1):67–82.
   Google Scholar
• Katzman J, Shaham U, Cloninger A, et al. Deep survival: a deep Cox proportional hazards network; 2016.
   Google Scholar
• Deep multi-task Gaussian processes for survival analysis with competing risks. Proceedings of the International Conference on Neural Information Processing Systems; 2017.
   Google Scholar
• Faraggi D, Simon R. A neural network model for survival data. Stat Med. 1995;14(1):73–82.
   PubMed Web of Science ®Google Scholar
• Lee C, Zame W, Yoon J, et al. DeepHit: a deep learning approach to survival analysis with competing risks. Proceedings of the AAAI Conference on Artificial Intelligence; 2018.
   Google Scholar
• Luck M, Sylvain T, Cardinal H, et al. Deep learning for patient-specific kidney graft survival analysis. arXiv; 2017.
   Google Scholar
• Bellot A, van der Schaar M. Multitask boosting for survival analysis with competing risks. Proceedings of the International Conference on Neural Information Processing Systems; 2018.
   Google Scholar
• Spooner A, Chen E, Sowmya A, et al. A comparison of machine learning methods for survival analysis of high-dimensional clinical data for dementia prediction. Sci Rep. 2020;10(1):20410.
   PubMedGoogle Scholar
• Soave DM, Strug LJ. Testing calibration of Cox survival models at extremes of event risk. Front Genet. 2018;9:177.
   PubMed Web of Science ®Google Scholar
• Langerbeins P, Bahlo J, Rhein C, et al. The CLL12 trial protocol: a placebo-controlled double-blind phase III study of ibrutinib in the treatment of early-stage chronic lymphocytic leukemia patients with risk of early disease progression. Future Oncol. 2015;11(13):1895–1903.
   PubMed Web of Science ®Google Scholar
• Herling CD, Cymbalista F, Groß-Ophoff-Müller C, et al. Early treatment with FCR versus watch and wait in patients with stage Binet A high-risk chronic lymphocytic leukemia (CLL): a randomized phase 3 trial. Leukemia. 2020;34(8):2038–2050.
   PubMed Web of Science ®Google Scholar
• Wang P, Li Y, Reddy CK. Machine learning for survival analysis. ACM Comput Surv. 2019;51(6):1–36.
   Web of Science ®Google Scholar
• Ishdorj G, Streu E, Lambert P, et al. IgA levels at diagnosis predict for infections, time to treatment, and survival in chronic lymphocytic leukemia. Blood Adv. 2019;3(14):2188–2198.
   PubMed Web of Science ®Google Scholar
• Tissino E, Pozzo F, Benedetti D, et al. CD49d promotes disease progression in chronic lymphocytic leukemia: new insights from CD49d bimodal expression. Blood. 2020;135(15):1244–1254.
   PubMed Web of Science ®Google Scholar
• Rigolin GM, Saccenti E, Bassi C, et al. Extensive next-generation sequencing analysis in chronic lymphocytic leukemia at diagnosis: clinical and biological correlations. J Hematol Oncol. 2016;9(1):88.
   PubMedGoogle Scholar
• Böttcher S, Ritgen M, Fischer K, et al. Minimal residual disease quantification is an independent predictor of progression-free and overall survival in chronic lymphocytic leukemia: a multivariate analysis from the randomized GCLLSG CLL8 trial. J Clin Oncol. 2012;30(9):980–988.
   PubMed Web of Science ®Google Scholar
• Crassini KR, Zhang E, Balendran S, 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. 2018;181(1):97–101.
   PubMed Web of Science ®Google Scholar
• Andersen MA, Vojdeman FJ, Andersen MK, et al. Hypogammaglobulinemia in newly diagnosed chronic lymphocytic leukemia is a predictor of early death. Leuk Lymphoma. 2016;57(7):1592–1599.
   PubMed Web of Science ®Google Scholar
• Brieghel C, da Cunha-Bang C, Yde CW, et al. The number of signaling pathways altered by driver mutations in chronic lymphocytic leukemia impacts disease outcome. Clin Cancer Res. 2020;26(6):1507–1515.
   PubMed Web of Science ®Google Scholar
• Tissino E, Benedetti D, Herman SEM, et al. Functional and clinical relevance of VLA-4 (CD49d/CD29) in ibrutinib-treated chronic lymphocytic leukemia. J Exp Med. 2018;215(2):681–697.
   PubMed Web of Science ®Google Scholar
• Hernández JÁ, Hernández-Sánchez M, Rodríguez-Vicente AE, et al. A low frequency of losses in 11q chromosome is associated with better outcome and lower rate of genomic mutations in patients with chronic lymphocytic leukemia. PLOS One. 2015;10(11):e0143073.
   PubMedGoogle Scholar
• Huang SJT, Gillan TL, Gerrie AS, et al. Influence of clone and deletion size on outcome in chronic lymphocytic leukemia patients with an isolated deletion 13q in a population-based analysis in British Columbia, Canada. Genes Chromosomes Cancer. 2016;55(1):16–24.
   PubMed Web of Science ®Google Scholar
• Baliakas P, Jeromin S, Iskas M, et al. Cytogenetic complexity in chronic lymphocytic leukemia: definitions, associations, and clinical impact. Blood. 2019;133(11):1205–1216.
   PubMed Web of Science ®Google Scholar
• Shea LK, Honjo K, Redden DT, et al. Fc receptor-like 2 (FCRL2) is a novel marker of low-risk CLL and refines prognostication based on IGHV mutation status. Blood Cancer J. 2019;9(6):47.
   PubMedGoogle Scholar
• Hu B, Patel KP, Chen H-C, et al. Association of gene mutations with time-to-first treatment in 384 treatment-naive chronic lymphocytic leukaemia patients. Br J Haematol. 2019;187(3):307–318.
   PubMed Web of Science ®Google Scholar
• Else M, Blakemore SJ, Strefford JC, et al. The association between deaths from infection and mutations of the BRAF, FBXW7, NRAS and XPO1 genes: a report from the LRF CLL4 trial. Leukemia. 2021.
   Google Scholar
• Wang H-Q, Jia L, Li Y-T, et al. Increased autocrine interleukin-6 production is significantly associated with worse clinical outcome in patients with chronic lymphocytic leukemia. J Cell Physiol. 2019;234(8):13994–14006.
   PubMed Web of Science ®Google Scholar
• Herishanu Y, Polliack A, Shenhar-Tsarfaty S, et al. Increased serum C-reactive protein levels are associated with shorter survival and development of second cancers in chronic lymphocytic leukemia. Ann Med. 2017;49(1):75–82.
   PubMed Web of Science ®Google Scholar
• Roos-Weil D, Nguyen-Khac F, Chevret S, et al. Mutational and cytogenetic analyses of 188 CLL patients with trisomy 12: a retrospective study from the French Innovative Leukemia Organization (FILO) Working Group. Genes Chromosomes Cancer. 2018;57(11):533–540.
   PubMed Web of Science ®Google Scholar
• deAndrés-Galiana EJ, Fernández-Martínez JL, Luaces O, et al. Analysis of clinical prognostic variables for chronic lymphocytic leukemia decision-making problems. J Biomed Inform. 2016;60:342–351.
   PubMed Web of Science ®Google Scholar
• Tailoring the targeted treatment of chronic lymphocytic leukemia — ERA-LEARN [Internet]; 2021 [cited 2021 May 15]. Available from: https://www.era-learn.eu/network-information/networks/era-permed/multidisciplinary-research-projects-on-personalised-medicine-2013-pre-clinical-research-big-data-and-ict-implementation-and-user2019s-perspective/tailoring-the-targeted-treatment-of-chronic-lymphocytic-leukemia
   Google Scholar
• Haider H, Hoehn B, Davis S, et al. Effective ways to build and evaluate individual survival distributions. arXiv; 2018.
   Google Scholar
• Molica S, Mauro FR, Callea V, et al. The utility of a prognostic index for predicting time to first treatment in early chronic lymphocytic leukemia: the GIMEMA experience. Haematologica. 2010;95(3):464–469.
   PubMed Web of Science ®Google Scholar
• Moncada-Torres A, van Maaren MC, Hendriks MP, et al. Explainable machine learning can outperform Cox regression predictions and provide insights in breast cancer survival. Sci Rep. 2021;11(1):6968.
   PubMed Web of Science ®Google Scholar
• Lundberg SM, Nair B, Vavilala MS, et al. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nat Biomed Eng. 2018;2(10):749–760.
   PubMed Web of Science ®Google Scholar
• Lundberg SM, Lee S-I. A unified approach to interpreting model predictions; 2017.
   Google Scholar
• Al-Sawaf O, Zhang C, Tandon M, et al. Venetoclax plus obinutuzumab versus chlorambucil plus obinutuzumab for previously untreated chronic lymphocytic leukaemia (CLL14): follow-up results from a multicentre, open-label, randomised, phase 3 trial. Lancet Oncol. 2020;21(9):1188–1200.
   PubMed Web of Science ®Google Scholar
• Gentile M, Shanafelt TD, Rossi D, et al. Validation of the CLL-IPI and comparison with the MDACC prognostic index in newly diagnosed patients. Blood. 2016;128(16):2093–2095.
   PubMed Web of Science ®Google Scholar
• Isabel G-GyM, Maria P, Sanchez Maria H, et al. Validation of the new International Prognostic Index for chronic lymphocytic leukaemia (CLL-IPI) and modified score used in the MD Anderson Cancer Center (MDACC). 2016;101:338.
   Google Scholar
• Muñoz-Novas C, Poza-Santaella M, González-Gascón Y Marín I, et al. The International Prognostic Index for patients with chronic lymphocytic leukemia has the higher value in predicting overall outcome compared with the Barcelona-Brno biomarkers only prognostic model and the MD Anderson Cancer Center Prognostic Index. Biomed Res Int. 2018;2018:1–8.
   Web of Science ®Google Scholar
• Da Cunha-Bang C, Agius R, Kater AP, et al. PreVent-ACaLL short-term combined acalabrutinib and venetoclax treatment of newly diagnosed patients with CLL at high risk of infection and/or early treatment, who do not fulfil IWCLL treatment criteria for treatment. A randomized study with extensive immune phenotyping. Blood. 2019;134(Suppl. 1):4304.
   Google Scholar
• Klausen TW, Gregersen H, Abildgaard N, et al. The majority of newly diagnosed myeloma patients do not fulfill the inclusion criteria in clinical phase III trials. Leukemia. 2019;33(2):546–549.
   PubMed Web of Science ®Google Scholar
• da Cunha-Bang C, Christiansen I, Niemann CU. The CLL-IPI applied in a population-based cohort. Blood. 2016;128(17):2181–2183.
   PubMed Web of Science ®Google Scholar
• Baliakas P, Mattsson M, Stamatopoulos K, et al. Prognostic indices in chronic lymphocytic leukaemia: where do we stand how do we proceed? J Intern Med. 2016;279(4):347–357.
   PubMed Web of Science ®Google Scholar
• Niemann CU, Herman SEM, Maric I, 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. 2016;22(7):1572–1582.
   PubMed Web of Science ®Google Scholar
• Burger JA. Targeting the microenvironment in chronic lymphocytic leukemia is changing the therapeutic landscape. Curr Opin Oncol. 2012;24(6):643–649.
   PubMed Web of Science ®Google Scholar
• Niemann CU, Wiestner A. B-cell receptor signaling as a driver of lymphoma development and evolution. Semin Cancer Biol. 2013;23(6):410–421.
   PubMed Web of Science ®Google Scholar
• Paredes BR, Argyriou A, Berthouze N, et al. Exploiting unrelated tasks in multi-task learning; 2012.
   Google Scholar
• Fei N, Lu Z, Gao Y, et al. Meta-learning across meta-tasks for few-shot learning. arXiv; 2020.
   Google Scholar
• Vanschoren J. Meta-learning: a survey. arXiv; 2018.
   Google Scholar
• Zhang XS, Tang F, Dodge HH, et al. MetaPred: meta-learning for clinical risk prediction with limited patient electronic health records. KDD. 2019;2019:2487–2495.
   PubMedGoogle Scholar
• Liu L, Liu Z, Wu H, et al. Multi-task learning via adaptation to similar tasks for mortality prediction of diverse rare diseases. arXiv; 2020.
   Google Scholar
• Jain P, O’Brien S. Richter’s transformation in chronic lymphocytic leukemia. Oncology. 2012;26:1146–1152.
   PubMed Web of Science ®Google Scholar
• Strati P, Abruzzo LV, Wierda WG, et al. Second cancers and Richter transformation are the leading causes of death in patients with trisomy 12 chronic lymphocytic leukemia. Clin Lymphoma Myeloma Leuk. 2015;15(7):420–427.
   PubMed Web of Science ®Google Scholar
• Rossi D, Spina V, Deambrogi C, et al. The genetics of Richter syndrome reveals disease heterogeneity and predicts survival after transformation. Blood. 2011;117(12):3391–3401.
   PubMed Web of Science ®Google Scholar
• Brieghel C, Aarup K, Torp MH, et al. Clinical outcomes in patients with multi-hit TP53 chronic lymphocytic leukemia treated with ibrutinib. Clin Cancer Res. 2021.
   PubMedGoogle Scholar
• de Weerdt I, Hofland T, de Boer R, et al. Distinct immune composition in lymph node and peripheral blood of CLL patients is reshaped during venetoclax treatment. Blood Adv. 2019;3(17):2642–2652.
   PubMed Web of Science ®Google Scholar
• Konečný J, McMahan HB, Yu Fx Richtarik P, et al. Federated learning: strategies for improving communication efficiency; 2016.
   Google Scholar
• Xu L, Skoularidou M, Cuesta-Infante A, et al. Modeling tabular data using conditional GAN. Proceedings of the International Conference on Neural Information Processing Systems; 2019.
   Google Scholar
• Tucker A, Wang Z, Rotalinti Y, et al. Generating high-fidelity synthetic patient data for assessing machine learning healthcare software. Npj Digital Med. 2020;3:147.
   PubMed Web of Science ®Google Scholar
• Russakovsky O, Deng J, Su H, et al. ImageNet large scale visual recognition challenge. Int J Comput Vis. 2015;115(3):211–252.
   Web of Science ®Google Scholar
• Pham H, Dai Z, Xie Q, et al. Meta pseudo labels. arXiv; 2020.
   Google Scholar
• Lee C, Yoon J, van der Schaar M. Dynamic-DeepHit: a deep learning approach for dynamic survival analysis with competing risks based on longitudinal data. IEEE Trans Biomed Eng. 2020;67(1):122–133.
   PubMed Web of Science ®Google Scholar
• Vieira D, Gimenez G, Marmerola G, et al. XGBoost survival embeddings: improving statistical properties of XGBoost survival analysis implementation [Internet]; 2020 [cited 2021 Aug 9]. Available from: http://github.com/loft-br/xgboost-survival-embeddings
   Google Scholar
• Groenwold RHH, White IR, Donders ART, et al. Missing covariate data in clinical research: when and when not to use the missing-indicator method for analysis. Can Med Assoc J. 2012;184(11):1265–1269.
   PubMed Web of Science ®Google Scholar
• Feng D, Zhao L. BDNNSurv: Bayesian deep neural networks for survival analysis using pseudo values. CoRR. 2021.
   Google Scholar
• Lundberg SM, Erion GG, Lee S-I. Consistent individualized feature attribution for tree ensembles. arXiv; 2018.
   Google Scholar