Skip to main content
Log in

Likelihood inference for the destructive exponentially weighted Poisson cure rate model with Weibull lifetime and an application to melanoma data

  • Original Paper
  • Published:
Computational Statistics Aims and scope Submit manuscript

Abstract

In this paper, we develop the steps of the expectation maximization algorithm (EM algorithm) for the determination of the maximum likelihood estimates (MLEs) of the parameters of the destructive exponentially weighted Poisson cure rate model in which the lifetimes are assumed to be Weibull. This model is more flexible than the promotion time cure rate model as it provides an interesting and realistic interpretation of the biological mechanism of the occurrence of an event of interest by including a destructive process of the initial number of causes in a competitive scenario. The standard errors of the MLEs are obtained by inverting the observed information matrix. An extensive Monte Carlo simulation study is carried out to evaluate the performance of the developed method of estimation. Finally, a known melanoma data are analyzed to illustrate the method of inference developed here. With these data, a comparison is also made with the scenario when the destructive mechanism is not included in the analysis.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3

Similar content being viewed by others

References

  • Balakrishnan N, Koutras MV, Milienos FS, Pal S (2016) Piecewise linear approximations for cure rate models and associated inferential issues. Methodol Comput Appl Probab (to appear). doi:10.1007/s11009-015-9477-0

    Google Scholar 

  • Balakrishnan N, Pal S (2013a) Lognormal lifetimes and likelihood-based inference for flexible cure rate models based on COM-Poisson family. Comput Stat Data Anal 67:41–67

    Article  MathSciNet  Google Scholar 

  • Balakrishnan N, Pal S (2013b) Expectation maximization-based likelihood inference for flexible cure rate models with Weibull lifetimes. Stat Methods Med Res (to appear). doi:10.1177/0962280213491641

    Google Scholar 

  • Balakrishnan N, Pal S (2015) An EM algorithm for the estimation of parameters of a flexible cure rate model with generalized gamma lifetime and model discrimination using likelihood- and information-based methods. Comput Stat 30:151–189

    Article  MathSciNet  MATH  Google Scholar 

  • Balakrishnan N, Peng Y (2006) Generalized gamma frailty model. Stat Med 25:2797–2816

    Article  MathSciNet  Google Scholar 

  • Berkson J, Gage RP (1952) Survival curve for cancer patients following treatment. J Am Stat Assoc 47:501–515

    Article  Google Scholar 

  • Boag JW (1949) Maximum likelihood estimates of the proportion of patients cured by cancer therapy. J R Stat Soc Ser B 11:15–53

    MATH  Google Scholar 

  • Chen M-H, Ibrahim JG, Sinha D (1999) A new Bayesian model for survival data with a surviving fraction. J Am Stat Assoc 94:909–919

    Article  MathSciNet  MATH  Google Scholar 

  • Cooner F, Banerjee S, Carlin BP, Sinha D (2007) Flexible cure rate modeling under latent activation schemes. J Am Stat Assoc 102:560–572

    Article  MathSciNet  MATH  Google Scholar 

  • Cox D, Oakes D (1984) Analysis of survival data. Chapman & Hall, London

    Google Scholar 

  • Dempster AP, Laird NM, Rubin DB (1977) Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser B 39:1–38

    MathSciNet  MATH  Google Scholar 

  • Farewell VT (1982) The use of mixture models for the analysis of survival data with long-term survivors. Biometrics 38:1041–1046

    Article  Google Scholar 

  • Hoggart CJ, Griffin JE (2001) A Bayesian partition model for customer attrition. In: George EI (ed) Bayesian methods with applications to science, policy, and official statistics (Selected Papers from ISBA 2000), Proceedings of the sixth world meeting of the international society for Bayesian analysis. International Society for Bayesian Analysis, Creta, pp 61–70

  • Ibrahim JG, Chen M-H, Sinha D (2001) Bayesian survival analysis. Springer, New York

    Book  MATH  Google Scholar 

  • Lange K (1995) A gradient algorithm locally equivalent to the EM algorithm. J R Stat Soc Ser B 57:425–437

    MathSciNet  MATH  Google Scholar 

  • Li CS, Taylor JMG, Sy JP (2001) Identifiability of cure models. Stat Probab Lett 54:389–395

    Article  MathSciNet  MATH  Google Scholar 

  • Louis TA (1982) Finding the observed information matrix when using the EM algorithm. J R Stat Soc Ser B 44:226–233

    MathSciNet  MATH  Google Scholar 

  • Maller RA, Zhou X (1996) Survival analysis with long-term survivors. Wiley, New York

    MATH  Google Scholar 

  • McLachlan GJ, Krishnan T (2008) The EM algorithm and extensions, 2nd edn. Wiley, Hoboken, NJ

    Book  MATH  Google Scholar 

  • Meeker WQ, Escobar LA (1998) Statistical methods for reliability data. Wiley, New York

    MATH  Google Scholar 

  • Pal S, Balakrishnan N (2015) Likelihood inference based on EM algorithm for the destructive length-biased Poisson cure rate model with Weibull lifetime. Commun Stat Simul Comput (to appear). doi:10.1080/03610918.2015.1053918

    Google Scholar 

  • Peng Y, Xu J (2012) An extended cure model and model selection. Lifetime Data Anal 18:215–233

    Article  MathSciNet  MATH  Google Scholar 

  • R Core Team (2014) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. http://www.R-project.org/

  • Rodrigues J, Cancho VG, de Castro M, Balakrishnan N (2012) A Bayesian destructive weighted Poisson cure rate model and an application to a cutaneous melanoma data. Stat Methods Med Res 21:585–597

    Article  MathSciNet  Google Scholar 

  • Rodrigues J, de Castro M, Balakrishnan N, Cancho VG (2011) Destructive weighted Poisson cure rate models. Lifetime Data Anal 17:333–346

    Article  MathSciNet  MATH  Google Scholar 

  • Rodrigues J, de Castro M, Cancho VG, Balakrishnan N (2009) COM-Poisson cure rate survival models and an application to a cutaneous melanoma data. J Stat Plan Inference 139:3605–3611

    Article  MathSciNet  MATH  Google Scholar 

  • Scheike TH, Martinussen T (2006) Dynamic regression models for survival data. Springer, New York

    MATH  Google Scholar 

  • Sy JP, Taylor JMG (2000) Estimation in a Cox proportional hazards cure model. Biometrics 56:227–236

    Article  MathSciNet  MATH  Google Scholar 

  • Yakovlev AY, Tsodikov AD (1996) Stochastic models of tumor latency and their biostatistical applications. World Scientific Publishers, Singapore

    Book  MATH  Google Scholar 

  • Yakovlev AY, Tsodikov AD, Bass L (1993) A stochastic-model of hormesis. Mathematical Biosciences 116:197–219

    Article  MATH  Google Scholar 

  • Yin G, Ibrahim JG (2005) Cure rate models: a unified approach. Can J Stat 33:559–570

    Article  MathSciNet  MATH  Google Scholar 

Download references

Acknowledgments

The second author expresses his thanks to the Natural Sciences and Engineering Research Council of Canada for funding this research through an individual discovery grant. The authors are also thankful to the editor and anonymous reviewers for their useful comments and suggestions on an earlier version of this manuscript which led to this improved one.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Suvra Pal.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (pdf 275 KB)

Appendix: Expressions useful in M-step

Appendix: Expressions useful in M-step

$$\begin{aligned} \frac{\partial Q}{\partial \beta _{1l}}\!= & {} \!\sum _{J_1}x_{il}\left( 1\!-\!p_i\right) \left\{ 1\!-\!\eta _ip_ie^{\phi }F\left( t_i\right) \right\} -\sum _{J_0}x_{il}\left( 1-\pi _i^{\left( k\right) }\right) \eta _ip_i\left( 1\!-\!p_i\right) e^{\phi }\\&+\,\sum _{J_0}x_{il}\pi _i^{\left( k\right) }\eta _ip_i\left( 1-p_i\right) e^{\phi }\frac{\left\{ q_{0i}-S_{ pop }\left( t_i\right) F\left( t_i\right) \right\} }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} },\\ \frac{\partial Q}{\partial \beta _{2k}}= & {} \sum _{J_1}z_{ik}\left\{ 1-\eta _ip_ie^{\phi }F\left( t_i\right) \right\} -\sum _{J_0}z_{ik}\left( 1-\pi _i^{\left( k\right) }\right) \eta _ip_ie^{\phi }\\&+\,\sum _{J_0}z_{ik}\pi _i^{\left( k\right) }\eta _ip_ie^{\phi }\frac{\left\{ q_{0i}-S_{ pop }\left( t_i\right) F\left( t_i\right) \right\} }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} },\\ \frac{\partial Q}{\partial \gamma _1}= & {} \frac{1}{\gamma _1}\sum _{J_1}\left\{ \frac{\left( \gamma _2t_i\right) ^{\frac{1}{\gamma _1}}}{\gamma _1}\log \left( \gamma _2t_i\right) -\frac{\log \left( \gamma _2t_i\right) }{\gamma _1}-1\right\} -\sum _{J_1}\eta _ip_ie^{\phi }F_1\left( t_i\right) \\&-\,\sum _{J_0}\pi _i^{\left( k\right) }\frac{\eta _ip_ie^{\phi }S_{ pop }\left( t_i\right) F_1\left( t_i\right) }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} },\\ \frac{\partial Q}{\partial \gamma _2}= & {} \frac{1}{\gamma _1\gamma _2}\sum _{J_1}\left\{ 1-\left( \gamma _2 t_i\right) ^{\frac{1}{\gamma _1}}\right\} \\&-\,\sum _{J_1}\eta _ip_ie^{\phi }F_2\left( t_i\right) -\sum _{J_0}\pi _i^{\left( k\right) }\frac{\eta _ip_ie^{\phi }S_{ pop }\left( t_i\right) F_2\left( t_i\right) }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} },\\ \frac{\partial ^2Q}{\partial \beta _{1l}\partial \beta _{1l^{\prime }}}= & {} -\sum _{J_1}x_{il}x_{il^\prime }p_i\left( 1-p_i\right) \left\{ 1+\eta _i\left( 1-2p_i\right) e^{\phi }F\left( t_i\right) \right\} \\ \end{aligned}$$
$$\begin{aligned}&-\,\sum _{J_0}x_{il}x_{il^\prime }\left( 1-\pi _i^{\left( k\right) }\right) \eta _ip_i\left( 1-p_i\right) \left( 1-2p_i\right) e^{\phi }\\&+\,\sum _{J_0}x_{il}x_{il^\prime }\pi _i^{\left( k\right) }\frac{\eta _ip_i\left( 1-p_i\right) e^{\phi }}{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} }\left[ \left( 1-2p_i\right) \left( q_{0i}-S_{ pop }\left( t_i\right) F\left( t_i\right) \right) \right. \\&\left. -\frac{\eta _ip_i\left( 1-p_i\right) e^{\phi }q_{0i}S_{ pop }\left( t_i\right) \left( 1-F\left( t_i\right) \right) ^2}{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} }\right] ,\\ \frac{\partial ^2Q}{\partial \beta _{1l}\partial \beta _{2k}}\!= & {} \!-\sum _{J_1}x_{il}z_{ik}\eta _ip_i\left( 1\!-\!p_i\right) e^{\phi }F\left( t_i\right) \!-\!\sum _{J_0}x_{il}z_{ik}\left( 1\!-\!\pi _i^{\left( k\right) }\right) \eta _ip_i\left( 1-p_i\right) e^{\phi }\\&+\,\sum _{J_0}x_{il}z_{ik}\pi _i^{\left( k\right) }\frac{\eta _ip_i\left( 1-p_i\right) e^{\phi }}{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} }\\&\times \left[ q_{0i}-S_{ pop }\left( t_i\right) F\left( t_i\right) -\frac{\eta _ip_ie^{\phi }q_{0i}S_{ pop }\left( t_i\right) \left( 1-F\left( t_i\right) \right) ^2}{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} }\right] ,\\ \frac{\partial ^2Q}{\partial \beta _{2k}\partial \beta _{2k^{\prime }}}= & {} -\sum _{J_1}z_{ik}z_{ik^\prime }\eta _ip_ie^{\phi }F\left( t_i\right) -\sum _{J_0}z_{ik}z_{ik^\prime }\left( 1\!-\!\pi _i^{\left( k\right) }\right) \eta _ip_ie^{\phi }\\&+\,\sum _{J_0}z_{ik}z_{ik^\prime }\pi _i^{\left( k\right) }\frac{\eta _ip_ie^{\phi }}{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} }\\&\times \left[ q_{0i}-S_{ pop }\left( t_i\right) F\left( t_i\right) -\frac{\eta _ip_ie^{\phi }q_{0i}S_{ pop }\left( t_i\right) \left( 1-F\left( t_i\right) \right) ^2}{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} }\right] , \end{aligned}$$
$$\begin{aligned} \frac{\partial ^2Q}{\partial \beta _{1l}\partial \gamma _1}= & {} -\sum _{J_1}x_{il}\eta _ip_i\left( 1-p_i\right) e^{\phi }F_1\left( t_i\right) +\sum _{J_0}x_{il}\pi _i^{\left( k\right) }\eta _ip_i\left( 1-p_i\right) e^{\phi }\\&\times \,\left[ \frac{S_{ pop }\left( t_i\right) F_1\left( t_i\right) \left\{ q_{0i}-S_{ pop }\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}\left( 1-F\left( t_i\right) \right) \right\} }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\right] ,\\ \frac{\partial ^2Q}{\partial \beta _{1l}\partial \gamma _2}= & {} -\sum _{J_1}x_{il}\eta _ip_i\left( 1-p_i\right) e^{\phi }F_2\left( t_i\right) +\sum _{J_0}x_{il}\pi _i^{\left( k\right) }\eta _ip_i\left( 1-p_i\right) e^{\phi }\\&\times \,\left[ \frac{S_{ pop }\left( t_i\right) F_2\left( t_i\right) \left\{ q_{0i}-S_{ pop }\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}\left( 1-F\left( t_i\right) \right) \right\} }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\right] ,\\ \frac{\partial ^2Q}{\partial \beta _{2k}\partial \gamma _1}= & {} -\sum _{J_1}z_{ik}\eta _ip_ie^{\phi }F_1\left( t_i\right) +\sum _{J_0}z_{ik}\pi _i^{\left( k\right) }\eta _ip_ie^{\phi }\\&\times \,\left[ \frac{S_{ pop }\left( t_i\right) F_1\left( t_i\right) \left\{ q_{0i}-S_{ pop }\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}\left( 1-F\left( t_i\right) \right) \right\} }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\right] ,\\ \frac{\partial ^2Q}{\partial \beta _{2k}\partial \gamma _2}= & {} -\sum _{J_1}z_{ik}\eta _ip_ie^{\phi }F_2\left( t_i\right) +\sum _{J_0}z_{ik}\pi _i^{\left( k\right) }\eta _ip_ie^{\phi }\\&\times \,\left[ \frac{S_{ pop }\left( t_i\right) F_2\left( t_i\right) \left\{ q_{0i}-S_{ pop }\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}\left( 1-F\left( t_i\right) \right) \right\} }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\right] ,\\ \frac{\partial ^2 Q}{\partial {\gamma _1^2}}= & {} \frac{1}{\gamma _1^2}\sum _{J_1}\left[ 1+\frac{2\log \left( \gamma _2t_i\right) }{\gamma _1}-\frac{\log \left( \gamma _2t_i\right) \left( \gamma _2t_i\right) ^{\frac{1}{\gamma _1}}}{\gamma _1}\left\{ 2+\frac{\log \left( \gamma _2t_i\right) }{\gamma _1}\right\} \right] \\ \end{aligned}$$
$$\begin{aligned}&\quad -\,\sum _{J_1}\eta _ip_ie^{\phi }F_{11}\left( t_i\right) \\&-\,\sum _{J_0}\pi _i^{\left( k\right) }\frac{\eta _ip_ie^{\phi }S_{ pop }\left( t_i\right) }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\\&\times \left[ \left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} F_{11}\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}F_1^2\left( t_i\right) \right] ,\\ \frac{\partial ^2 Q}{\partial {\gamma _2^2}}= & {} -\frac{1}{\gamma _1\gamma _2^2}\sum _{J_1}\left\{ 1+\left( \frac{1}{\gamma _1}-1\right) \left( \gamma _2t_i\right) ^{\frac{1}{\gamma _1}}\right\} -\sum _{J_1}\eta _ip_ie^{\phi }F_{22}\left( t_i\right) \\&-\,\sum _{J_0}\pi _i^{\left( k\right) }\frac{\eta _ip_ie^{\phi }S_{ pop }\left( t_i\right) }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\\&\times \left[ \left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} F_{22}\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}F_2^2\left( t_i\right) \right] ,\\ \frac{\partial ^2 Q}{\partial {\gamma _2}\partial {\gamma _1}}= & {} \frac{1}{\gamma _1^2\gamma _2}\sum _{J_1}\left[ \left( \gamma _2t_i\right) ^{\frac{1}{\gamma _1}}\left\{ 1+\frac{\log \left( \gamma _2t_i\right) }{\gamma _1}\right\} -1\right] -\sum _{J_1}\eta _ip_ie^{\phi }F_{12}\left( t_i\right) \\&-\,\sum _{J_0}\pi _i^{\left( k\right) }\frac{\eta _ip_ie^{\phi }S_{ pop }\left( t_i\right) }{\left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} ^2}\\&\times \left[ \left\{ S_{ pop }\left( t_i\right) -q_{0i}\right\} F_{12}\left( t_i\right) +\eta _ip_ie^{\phi }q_{0i}F_1\left( t_i\right) F_2\left( t_i\right) \right] , \end{aligned}$$

where

$$\begin{aligned} F_1\left( t_i\right)= & {} \frac{\partial F\left( t_i\right) }{\partial \gamma _1}=-\frac{1}{\gamma _1^2}S\left( t_i\right) \left( \gamma _2 t_i\right) ^{\frac{1}{\gamma _1}}\log \left( \gamma _2 t_i\right) , F_2\left( t_i\right) =\frac{\partial F\left( t_i\right) }{\partial \gamma _2}\\= & {} \frac{1}{\gamma _1\gamma _2}S\left( t_i\right) \left( \gamma _2 t_i\right) ^{\frac{1}{\gamma _1}},\\ F_{11}\left( t_i\right)= & {} \frac{\partial ^2 F\left( t_i\right) }{\partial \gamma _1^2}=-F_1\left( t_i\right) \left\{ \frac{1}{\gamma _1^2}\left\{ \log \left( \gamma _2t_i\right) +2\gamma _1\right\} +\frac{F_1\left( t_i\right) }{S\left( t_i\right) }\right\} ,\\ F_{22}\left( t_i\right)= & {} \frac{\partial ^2 F\left( t_i\right) }{\partial \gamma _2^2}=F_2\left( t_i\right) \left\{ \frac{1}{\gamma _2}\left( \frac{1}{\gamma _1}-1\right) -\frac{F_2\left( t_i\right) }{S\left( t_i\right) }\right\} ,\\ F_{12}\left( t_i\right)= & {} \frac{\partial ^2 F\left( t_i\right) }{\partial \gamma _2\partial \gamma _1}=-F_2\left( t_i\right) \left\{ \frac{1}{\gamma _1}\left( 1+\frac{\log \left( \gamma _2t_i\right) }{\gamma _1}\right) +\frac{F_1\left( t_i\right) }{S\left( t_i\right) }\right\} . \end{aligned}$$

The above quantities are all defined for \(l,l^{\prime }=0,1,\ldots ,s,\,x_{i0}=1 \forall i=1,2,\ldots ,n,\) and \(k,k^{\prime }=1,2,\ldots ,s^\prime .\)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Pal, S., Balakrishnan, N. Likelihood inference for the destructive exponentially weighted Poisson cure rate model with Weibull lifetime and an application to melanoma data. Comput Stat 32, 429–449 (2017). https://doi.org/10.1007/s00180-016-0660-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00180-016-0660-8

Keywords

Navigation