The Annals of Applied Statistics

Bootstrap aggregating continual reassessment method for dose finding in drug-combination trials

Ruitao Lin and Guosheng Yin

Full-text: Access denied (no subscription detected)

We're sorry, but we are unable to provide you with the full text of this article because we are not able to identify you as a subscriber. If you have a personal subscription to this journal, then please login. If you are already logged in, then you may need to update your profile to register your subscription. Read more about accessing full-text


Phase I drug-combination trials are becoming commonplace in oncology. Most of the current dose-finding designs aim to quantify the toxicity probability space using certain prespecified yet complicated models. These models need to characterize not only each individual drug’s toxicity profile, but also their interaction effects, which often leads to multi-parameter models. We propose a novel Bayesian adaptive design for drug-combination trials based on a robust dimension-reduction method. We continuously update the order of dose combinations and reduce the two-dimensional searching space to a one-dimensional line based on the estimated order. As a result, the common approaches to single-agent dose finding, such as the continual reassessment method (CRM), can be applied to drug-combination trials. We further utilize the ensemble technique in machine learning, the so-called bootstrap aggregating (bagging) in conjunction with Bayesian model averaging, to enhance the efficiency and reduce the variability of the proposed method. We conduct extensive simulation studies to examine the operating characteristics of the proposed method under various scenarios. Compared with existing competitive designs, the bagging CRM demonstrates its precision and robustness in terms of pinning down the correct dose combination. We apply the proposed bagging CRM to two recent cancer clinical trials with combined drugs for dose finding.

Article information

Ann. Appl. Stat., Volume 10, Number 4 (2016), 2349-2376.

Received: April 2016
Revised: August 2016
First available in Project Euclid: 5 January 2017

Permanent link to this document

Digital Object Identifier

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

Bagging Bayesian model averaging dimension reduction drug combination maximum tolerated dose


Lin, Ruitao; Yin, Guosheng. Bootstrap aggregating continual reassessment method for dose finding in drug-combination trials. Ann. Appl. Stat. 10 (2016), no. 4, 2349--2376. doi:10.1214/16-AOAS982.

Export citation


  • Ahn, C. (1998). An evaluation of phase I cancer clinical trial designs. Stat. Med. 17 1537–1549.
  • Bendell, J. C., Jones, S. F., Hart, L., Spigel, D. R., Lane, C. M., Earwood, C., Infante, J. R., Barton, J. and Burris, H. A. (2015). A phase Ib study of linsitinib (OSI-906), a dual inhibitor of IGF-1R and IR tyrosine kinase, in combination with everolimus as treatment for patients with refractory metastatic colorectal cancer. Invest. New Drugs 33 187–193.
  • Breiman, L. (1996). Bagging predictors. Mach. Learn. 24 123–140.
  • Breiman, L. (2001). Random forests. Mach. Learn. 45 5–32.
  • Bril, G., Dykstra, R., Pillers, C. and Robertson, T. (1984). Algorithm AS 206: Isotonic regression in two independent variables. J. R. Stat. Soc. Ser. C. Appl. Stat. 33 352–357.
  • Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. Chapman & Hall/CRC, Boca Raton, FL.
  • Clyde, M. A. and Lee, H. K. H. (2001). Bagging and the Bayesian bootstrap. In Artificial Intelligence and Statistics (T. Richardson and T. Jaakkola, eds.) 169–174. Elsevier, New York.
  • Conaway, M. R., Dunbar, S. and Peddada, S. D. (2004). Designs for single- or multiple-agent phase I trials. Biometrics 60 661–669.
  • Fan, S. K., Venook, A. P. and Lu, Y. (2009). Design issues in dose-finding phase I trials for combinations of two agents. J. Biopharm. Statist. 19 509–523.
  • Freund, Y. and Schapire, R. E. (1997). A decision-theoretic generalization of on-line learning and an application to boosting. J. Comput. System Sci. 55 119–139.
  • Friedman, J., Hastie, T. and Tibshirani, R. (2000). Additive logistic regression: A statistical view of boosting. Ann. Statist. 28 337–407. With discussion and a rejoinder by the authors.
  • Gandhi, L., Bahleda, R., Tolaney, S. M., Kwak, E. L., Cleary, J. M., Pandya, S. S., Hollebecque, A., Abbas, R., Ananthakrishnan, R., Berkenblit, A., Krygowski, M., Liang, Y., Turnbull, K. W., Shapiro, G. I. and Soria, J.-C. (2014). Phase I study of neratinib in combination with temsirolimus in patients with human epidermal growth factor receptor 2-dependent and other solid tumors. J. Clin. Oncol. 32 68–75.
  • Harrington, J. A., Wheeler, G. M., Sweeting, M. J., Mander, A. P. and Jodrell, D. I. (2013). Adaptive designs for dual-agent phase I dose-escalation studies. Nat. Rev. Clin. Oncol. 10 277–288.
  • Hastie, T., Tibshirani, R. and Friedman, J. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd ed. Springer Series in Statistics. Springer, New York.
  • Heyd, J. M. and Carlin, P. B. (1999). Adaptive design improvements in the continual reassessment method for phase I studies. Stat. Med. 18 1307–1321.
  • Hirakawa, A., Hamada, C. and Matsui, S. (2013). A dose-finding approach based on shrunken predictive probability for combinations of two agents in phase I trials. Stat. Med. 32 4515–4525.
  • Hirakawa, A., Wages, N. A., Sato, H. and Matsui, S. (2015). A comparative study of adaptive dose-finding designs for phase I oncology trials of combination therapies. Stat. Med. 34 3194–3213.
  • Hoeting, J. A., Madigan, D., Raftery, A. E. and Volinsky, C. T. (1999). Bayesian model averaging: A tutorial. Statist. Sci. 14 382–417. With comments by M. Clyde, David Draper and E. I. George, and a rejoinder by the authors.
  • Houede, N., Thall, P. F., Nguyen, H., Paoletti, X. and Kramar, A. (2010). Utility-based optimization of combination therapy using ordinal toxicity and efficacy in phase I/II trials. Biometrics 66 532–540.
  • Huang, X., Biswas, S., Oki, Y., Issa, J.-P. and Berry, D. A. (2007). A parallel phase I/II clinical trial design for combination therapies. Biometrics 63 429–436.
  • Iasonos, A. and O’Quigley, J. (2014). Adaptive dose-finding studies: A review of model-guided phase I clinical trials. J. Clin. Oncol. 32 2505–2511.
  • Isakoff, S. J., Wang, D., Campone, M., Calles, A., Leip, E., Turnbull, K., Bardy-Bouxin, N., Duvillié, L. and Calvo, E. (2014). Bosutinib plus capecitabine for selected advanced solid tumours: Results of a phase 1 dose-escalation study. Br. J. Cancer 111 2058–2066.
  • Ivanova, A. and Wang, K. (2004). A non-parametric approach to the design and analysis of two-dimensional dose-finding trials. Stat. Med. 23 1861–1870.
  • Korn, E. L. and Simon, R. (1991). Selecting dose-intense drug combinations: Metastatic breast cancer. Breast Cancer Res. Treat. 20 155–166.
  • Korn, E. L. and Simon, R. (1993). Using the tolerable-dose diagram in the design of phase I combination chemotherapy trials. J. Clin. Oncol. 11 794–801.
  • Korn, E. L., Midthune, D., Chen, T. T., Rubinstein, L. V., Christian, M. C. and Simon, R. M. (1994). A comparison of two phase I trial designs. Stat. Med. 13 1799–1806.
  • Kramar, A., Lebecq, A. and Candalh, E. (1999). Continual reassessment methods in phase I trials of the combination of two drugs in oncology. Stat. Med. 18 1849–1864.
  • Lin, R. and Yin, G. (2016). Bayesian optimal interval design for dose finding in drug-combination trials. Stat. Methods Med. Res., DOI:10.1177/0962280215594494.
  • Mander, A. P. and Sweeting, M. J. (2015). A product of independent beta probabilities dose escalation design for dual-agent phase I trials. Stat. Med. 34 1261–1276.
  • Mandrekar, S. J. (2014). Dose-finding trial designs for combination therapies in oncology. J. Clin. Oncol. 32 65–67.
  • O’Quigley, J. and Conaway, M. (2010). Continual reassessment and related dose-finding designs. Statist. Sci. 25 202–216.
  • O’Quigley, J., Pepe, M. and Fisher, L. (1990). Continual reassessment method: A practical design for phase 1 clinical trials in cancer. Biometrics 46 33–48.
  • Oron, A. P. and Hoff, P. D. (2013). Small-sample behavior of novel phase I cancer trial designs. Clin. Trials 10 63–80.
  • Papadatos-Pastos, D., Luken, M. D. M. and Yap, T. A. (2015). Combining targeted therapeutics in the era of precision medicine. Br. J. Cancer 112 1–3.
  • Raftery, A. E., Madigan, D. and Hoeting, J. A. (1997). Bayesian model averaging for linear regression models. J. Amer. Statist. Assoc. 92 179–191.
  • Riviere, M.-K., Dubois, F. and Zohar, S. (2015a). Competing designs for drug combination in phase I dose-finding clinical trials. Stat. Med. 34 1–12.
  • Riviere, M.-K., Yuan, Y., Dubois, F. and Zohar, S. (2014). A Bayesian dose-finding design for drug combination clinical trials based on the logistic model. Pharm. Stat. 13 247–257.
  • Riviere, M. K., Le Tourneau, C., Paoletti, X., Dubois, F. and Zohar, S. (2015b). Designs of drug-combination phase I trials in oncology: A systematic review of the literature. Ann. Oncol. 26 669–674.
  • Rubin, D. B. (1981). The Bayesian bootstrap. Ann. Statist. 9 130–134.
  • Saura, C., Garcia-Saenz, J. A., Xu, B., Harb, W., Moroose, R., Pluard, T., Cortes, J., Kiger, C., Germa, C., Wang, K., Martin, M., Baselga, J. and Kim, S. B. (2014). Safety and efficacy of neratinib in combination with capecitabine in patients with metastatic human epidermal growth factor receptor 2-positive breast cancer. J. Clin. Oncol. 32 3626–3633.
  • Shen, L. Z. and O’Quigley, J. (1996). Consistency of continual reassessment method under model misspecification. Biometrika 83 395–405.
  • Siegel, D. S., Richardson, P., Dimopoulos, M., Moreau, P., Mitsiades, C., Weber, D., Houp, J., Gause, C., Vuocolo, S., Eid, J., Graef, T. and Anderson, K. C. (2014). Vorinostat in combination with lenalidomide and dexamethasone in patients with relapsed or refractory multiple myeloma. Blood Cancer J. 4 e182.
  • Thall, P. F. (2010). Bayesian models and decision algorithms for complex early phase clinical trials. Statist. Sci. 25 227–244.
  • Thall, P. F., Millikan, R. E., Mueller, P. and Lee, S.-J. (2003). Dose-finding with two agents in phase I oncology trials. Biometrics 59 487–496.
  • Tighiouart, M., Piantadosi, S. and Rogatko, A. (2014). Dose finding with drug combinations in cancer phase I clinical trials using conditional escalation with overdose control. Stat. Med. 33 3815–3829.
  • Ullenhag, G. J., Rossmann, E. and Liljefors, M. (2015). A phase I dose-escalation study of lenalidomide in combination with gemcitabine in patients with advanced pancreatic cancer. PLoS ONE 10 e0121197.
  • Wages, N. A., Conaway, M. R. and O’Quigley, J. (2011). Continual reassessment method for partial ordering. Biometrics 67 1555–1563.
  • Wang, K. and Ivanova, A. (2005). Two-dimensional dose finding in discrete dose space. Biometrics 61 217–222.
  • Wilky, B. A., Rudek, M. A., Ahmed, S., Laheru, D. A., Cosgrove, D., Donehower, R. C., Nelkin, B., Ball, D., Doyle, L. A., Chen, H., Ye, X., Bigley, G., Womack, C. and Azad, N. S. (2015). A phase I trial of vertical inhibition of IGF signalling using cixutumumab, an anti-IGF-1R antibody, and selumetinib, an MEK 1/2 inhibitor, in advanced solid tumours. Br. J. Cancer 112 24–31.
  • Yin, G. (2012). Clinical Trial Design: Bayesian and Frequentist Adaptive Methods. John Wiley & Sons, Hoboken, New Jersey.
  • Yin, G. and Lin, R. (2015). Comments on ‘Competing designs for drug combination in phase I dose-finding clinical trials’ by M.-K. Riviere, F. Dubois, and S. Zohar [MR3286233]. Stat. Med. 34 13–17.
  • Yin, G. and Yuan, Y. (2009a). Bayesian dose finding in oncology for drug combinations by copula regression. J. R. Stat. Soc. Ser. C. Appl. Stat. 58 211–224.
  • Yin, G. and Yuan, Y. (2009b). Bayesian model averaging continual reassessment method in phase I clinical trials. J. Amer. Statist. Assoc. 104 954–968.
  • Yuan, Z. and Chappell, R. (2004). Isotonic designs for phase I cancer clinical trials with multiple risk groups. Clin. Trials 1 499–508.
  • Yuan, Y. and Yin, G. (2008). Sequential continual reassessment method for two-dimensional dose finding. Stat. Med. 27 5664–5678.