Torben Martinussen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Torben Martinussen has a total of 105 co-authors. Here are the top co-authors by number of shared publications:
- Stijn Vansteelandt
Department of Mathematics, Computer Science and Statistics, Ghent University
6 shared publications - Lisbeth Høier Olsen
Department of Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Ridebanevej 9, 1870 Frederiksberg C, Denmark
3 shared publications Stijn Vansteelandt
Department of Applied Mathematics, Computer Science and Statistics, Ghent University , Krijgslaan 281 (S9), Ghent , Belgium
3 shared publications- Thomas Scheike
Section of Biostatistics, Department of Public Health, University of Copenhagen, Copenhagen K, Denmark
3 shared publications - David M. Zucker
Department of Statistics, Hebrew University, Mount Scopus, Jerusalem 9190501, Israel
2 shared publications Ditte Nørbo Sørensen
Section of Biostatistics, University of Copenhagen, Øster Farimagsgade 5B, 1014, Copenhagen K, Denmark
2 shared publications
Latest Publications
- Open Access
- 8 September 2026
- Published by Wiley in Animal Genetics
- Vol. 57 (5) , e70200
A Combination of Alleles in LMOD2 and a lncRNA is Strongly Associated With Myxomatous Mitral Valve Disease in Cavalier King Charles Spaniels
A previous genome‐wide association study identified regions on canine chromosome (cfa) 13 and 14 associated with early onset myxomatous mitral valve disease (MMVD) in Cavalier King Charles Spaniels (CKCS). In the present study, whole genome sequencing (WGS) of 9 CKCS cases (mitral regurgitation (MR) before 4.5 years or congestive heart failure (CHF) at any age due to MMVD) and 10 CKCS controls (no or mild MR after 8 years of age) identified > 2000 genetic variants in the MMVD associated cfa13 and cfa14 regions. Ensembl Variant Effect Predictor (VEP) identified a possible functional impact of 18 variants. These were genotyped in 250 CKCS; 117 cases and 133 controls. The most significantly associated variants were a splice‐site variant in a long noncoding RNA (lncRNA) on cfa13, a nonsynonymous variant in HYAL4 , a 39 base‐pair insertion in LMOD2 and a synonymous variant in ENSCAFG00000024436 ( p ‐values from 2.03E‐08 to 4.20E‐06). Concomitant homozygosity for risk alleles in LMOD2 and the lncRNA gave an odds‐ratio for MMVD of 52.5 compared to homozygosity for the nonrisk alleles ( p = 0.00034, 95% CI: 8.8–1023.8). Upon validation of our results in an independent cohort, this gene variant combination in CKCS is expected to enable targeted breeding programs to reduce MMVD prevalence in CKCS.
- 8 September 2026
- Published by Wiley in Animal Genetics
- Vol. 57 (5) , e70200
Open AccessA Combination of Alleles in LMOD2 and a lncRNA is Strongly Associated With Myxomatous Mitral Valve Disease in Cavalier King Charles Spaniels
A previous genome‐wide association study identified regions on canine chromosome (cfa) 13 and 14 associated with early onset myxomatous mitral valve disease (MMVD) in Cavalier King Charles Spaniels (CKCS). In the present study, whole genome sequencing (WGS) of 9 CKCS cases (mitral regurgitation (MR) before 4.5 years or congestive heart failure (CHF) at any age due to MMVD) and 10 CKCS controls (no or mild MR after 8 years of age) identified > 2000 genetic variants in the MMVD associated cfa13 and cfa14 regions. Ensembl Variant Effect Predictor (VEP) identified a possible functional impact of 18 variants. These were genotyped in 250 CKCS; 117 cases and 133 controls. The most significantly associated variants were a splice‐site variant in a long noncoding RNA (lncRNA) on cfa13, a nonsynonymous variant in HYAL4 , a 39 base‐pair insertion in LMOD2 and a synonymous variant in ENSCAFG00000024436 ( p ‐values from 2.03E‐08 to 4.20E‐06). Concomitant homozygosity for risk alleles in LMOD2 and the lncRNA gave an odds‐ratio for MMVD of 52.5 compared to homozygosity for the nonrisk alleles ( p = 0.00034, 95% CI: 8.8–1023.8). Upon validation of our results in an independent cohort, this gene variant combination in CKCS is expected to enable targeted breeding programs to reduce MMVD prevalence in CKCS.
- Open Access
- 30 May 2026
- Published by Wiley in Biometrical Journal
- Vol. 68 (3) , e70143
Nonparametric Estimation of the Patient‐Weighted While‐Alive Estimand
In clinical trials with recurrent events, such as repeated hospitalizations terminating with death, it is important to consider the patient events overall history for a thorough assessment of treatment effects. The occurrence of fewer events due to early deaths can lead to misinterpretation, emphasizing the importance of a while‐alive strategy as suggested in Schmidli et al. (2023, Statistics in Biopharmaceutical Research 15, no. 2: 238–248). In this study, we focus on the patient‐weighted while‐alive estimand, represented as the expected number of events divided by the time alive within a target window, and develop efficient estimation for this estimand. Specifically, we derive the corresponding efficient influence function and develop a one‐step estimator initially applied to the simpler irreversible illness‐death model. For the broader context of recurrent events, due to the increased complexity, this one‐step estimator is practically intractable due to likely misspecification of the needed conditional transition intensities that depend on a patient's unique history. Therefore, we suggest an alternative estimator that is expected to have high efficiency, focusing on the randomized treatment setting. In addition, we apply our proposed estimator to two real‐world case studies, demonstrating the practical applicability of this second estimator and benefits of this while‐alive approach over currently available alternatives.
- 30 May 2026
- Published by Wiley in Biometrical Journal
- Vol. 68 (3) , e70143
Open AccessNonparametric Estimation of the Patient‐Weighted While‐Alive Estimand
In clinical trials with recurrent events, such as repeated hospitalizations terminating with death, it is important to consider the patient events overall history for a thorough assessment of treatment effects. The occurrence of fewer events due to early deaths can lead to misinterpretation, emphasizing the importance of a while‐alive strategy as suggested in Schmidli et al. (2023, Statistics in Biopharmaceutical Research 15, no. 2: 238–248). In this study, we focus on the patient‐weighted while‐alive estimand, represented as the expected number of events divided by the time alive within a target window, and develop efficient estimation for this estimand. Specifically, we derive the corresponding efficient influence function and develop a one‐step estimator initially applied to the simpler irreversible illness‐death model. For the broader context of recurrent events, due to the increased complexity, this one‐step estimator is practically intractable due to likely misspecification of the needed conditional transition intensities that depend on a patient's unique history. Therefore, we suggest an alternative estimator that is expected to have high efficiency, focusing on the randomized treatment setting. In addition, we apply our proposed estimator to two real‐world case studies, demonstrating the practical applicability of this second estimator and benefits of this while‐alive approach over currently available alternatives.
- 18 July 2025
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 88 (1) , 120-142
Efficient nonparametric estimators of discrimination measures with censored survival data
Discrimination measures such as the concordance index and the cumulative-dynamic time-dependent area under the ROC-curve are widely used in the medical literature for evaluating the predictive accuracy of a scoring rule which relates a set of prognostic markers to the risk of experiencing a particular event. Often the scoring rule being evaluated in terms of discriminatory ability is the linear predictor of a survival regression model such as the Cox proportional hazards model. This has the undesirable feature that the scoring rule depends on the censoring distribution when the model is misspecified. In this work, we focus on linear scoring rules where the coefficient vector is a nonparametric estimand defined in the setting where there is no censoring. We propose so-called debiased estimators of the aforementioned discrimination measures for this class of scoring rules. The proposed estimators make efficient use of the data and minimize bias by allowing for the use of data-adaptive methods for model fitting. Moreover, the estimators do not rely on correct specification of the censoring model to produce consistent estimation. We compare the estimators to existing methods in a simulation study, and we illustrate the method by an application to a brain cancer study.
- 18 July 2025
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 88 (1) , 120-142
Efficient nonparametric estimators of discrimination measures with censored survival data
Discrimination measures such as the concordance index and the cumulative-dynamic time-dependent area under the ROC-curve are widely used in the medical literature for evaluating the predictive accuracy of a scoring rule which relates a set of prognostic markers to the risk of experiencing a particular event. Often the scoring rule being evaluated in terms of discriminatory ability is the linear predictor of a survival regression model such as the Cox proportional hazards model. This has the undesirable feature that the scoring rule depends on the censoring distribution when the model is misspecified. In this work, we focus on linear scoring rules where the coefficient vector is a nonparametric estimand defined in the setting where there is no censoring. We propose so-called debiased estimators of the aforementioned discrimination measures for this class of scoring rules. The proposed estimators make efficient use of the data and minimize bias by allowing for the use of data-adaptive methods for model fitting. Moreover, the estimators do not rely on correct specification of the censoring model to produce consistent estimation. We compare the estimators to existing methods in a simulation study, and we illustrate the method by an application to a brain cancer study.
- 3 July 2025
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 81 (3)
Doubly robust nonparametric estimators of the predictive value of covariates for survival data
The predictive value of a covariate is often of interest in studies with a survival endpoint. A common situation is that there are some well established predictors and a potential valuable new marker. The challenge is how to judge the potentially added predictive value of this new marker. We propose to use the positive predictive value (PPV) curve based on a nonparametric scoring rule. The estimand of interest is viewed as a single transformation of the underlying data generating probability measure, which allows us to develop a robust nonparametric estimator of the PPV by first calculating the corresponding efficient influence function. We provide asymptotic results and illustrate the approach with numerical studies and with 2 cancer data studies.
- 3 July 2025
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 81 (3)
Doubly robust nonparametric estimators of the predictive value of covariates for survival data
The predictive value of a covariate is often of interest in studies with a survival endpoint. A common situation is that there are some well established predictors and a potential valuable new marker. The challenge is how to judge the potentially added predictive value of this new marker. We propose to use the positive predictive value (PPV) curve based on a nonparametric scoring rule. The estimand of interest is viewed as a single transformation of the underlying data generating probability measure, which allows us to develop a robust nonparametric estimator of the PPV by first calculating the corresponding efficient influence function. We provide asymptotic results and illustrate the approach with numerical studies and with 2 cancer data studies.
- 25 June 2025
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 52 (4) , 1505-1555
Variable importance measures for heterogeneous treatment effects with survival outcome
Treatment effect heterogeneity plays an important role in many areas of causal inference, and in recent years, estimation of the conditional average treatment effect (CATE) has received much attention in the statistical community. While accurate estimation of the CATE‐function through flexible machine learning procedures provides a tool for prediction of the individual treatment effect, it does not provide further insight into the driving features of potential treatment effect heterogeneity. Recent papers have addressed this problem by providing variable importance measures for treatment effect heterogeneity. Most of the suggestions have been developed for continuous or binary outcome, while little attention has been given to censored time‐to‐event outcome. In this paper, we extend the treatment effect variable importance measure (TE‐VIM) proposed in Hines, Diaz‐Ordaz, and Vansteelandt (2022a) to the survival setting with censored outcome. We derive an estimator for the TE‐VIM for two different CATE functions based on the survival function and RMST, respectively. Along with the TE‐VIM, we propose a new measure of treatment effect heterogeneity based on the best partially linear projection of the CATE and suggest accompanying estimators for that projection. All estimators are based on semiparametric efficiency theory, and we give conditions under which they are asymptotically linear. The finite‐sample performance of the derived estimators is investigated through a simulation study. Finally, the estimators are applied and contrasted in two real data examples.
- 25 June 2025
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 52 (4) , 1505-1555
Variable importance measures for heterogeneous treatment effects with survival outcome
Treatment effect heterogeneity plays an important role in many areas of causal inference, and in recent years, estimation of the conditional average treatment effect (CATE) has received much attention in the statistical community. While accurate estimation of the CATE‐function through flexible machine learning procedures provides a tool for prediction of the individual treatment effect, it does not provide further insight into the driving features of potential treatment effect heterogeneity. Recent papers have addressed this problem by providing variable importance measures for treatment effect heterogeneity. Most of the suggestions have been developed for continuous or binary outcome, while little attention has been given to censored time‐to‐event outcome. In this paper, we extend the treatment effect variable importance measure (TE‐VIM) proposed in Hines, Diaz‐Ordaz, and Vansteelandt (2022a) to the survival setting with censored outcome. We derive an estimator for the TE‐VIM for two different CATE functions based on the survival function and RMST, respectively. Along with the TE‐VIM, we propose a new measure of treatment effect heterogeneity based on the best partially linear projection of the CATE and suggest accompanying estimators for that projection. All estimators are based on semiparametric efficiency theory, and we give conditions under which they are asymptotically linear. The finite‐sample performance of the derived estimators is investigated through a simulation study. Finally, the estimators are applied and contrasted in two real data examples.
- Open Access
- 3 February 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-13
18F-FDG PET/CT assessment of metabolic tumor burden predicts survival in patients with metastatic posterior uveal melanoma
The prognostic value of metabolic tumor burden parameters obtained from 18F-FDG PET/CT imaging was evaluated in this retrospective national multicenter study of patients with metastatic posterior uveal melanoma (PUM) and compared to the largest diameter of the largest metastatic lesion (LDLM) and the American Joint Committee on Cancer (AJCC) staging system. The Maximal Standard Uptake Value (SUVmax), Metabolic Tumor Volume (MTV), and Total Lesion Glycolysis (TLG) were obtained in 106 patients. Higher values of SUVmax (p = 0.007, log-rank), MTV (p < 0.001, log-rank), and TLG (p < 0.001, long-rank) were associated with shorter survival. The three parameters were also independent predictors in the multivariate Cox model, while the AJCC staging turned insignificant. Time-dependent positive predictive value (PPV) analysis and Receiver Operating Characteristics (ROC) curves showed that MTV (Area Under the Curve (AUC) = 0.78), TLG (AUC = 0.78), and LDLM (AUC = 0.76) were good predictors of 1-year survival. For the subset of 97 patients with liver metastases, the corresponding regional measurements in the liver tended to be even better predictors. In conclusion, MTV and TLG were found to be better predictors of survival in metastatic PUM than the AJCC staging system, but when LDLM was used as a continuous variable it showed an equally good prediction of 1-year survival.
- 3 February 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-13
Open Access18F-FDG PET/CT assessment of metabolic tumor burden predicts survival in patients with metastatic posterior uveal melanoma
The prognostic value of metabolic tumor burden parameters obtained from 18F-FDG PET/CT imaging was evaluated in this retrospective national multicenter study of patients with metastatic posterior uveal melanoma (PUM) and compared to the largest diameter of the largest metastatic lesion (LDLM) and the American Joint Committee on Cancer (AJCC) staging system. The Maximal Standard Uptake Value (SUVmax), Metabolic Tumor Volume (MTV), and Total Lesion Glycolysis (TLG) were obtained in 106 patients. Higher values of SUVmax (p = 0.007, log-rank), MTV (p < 0.001, log-rank), and TLG (p < 0.001, long-rank) were associated with shorter survival. The three parameters were also independent predictors in the multivariate Cox model, while the AJCC staging turned insignificant. Time-dependent positive predictive value (PPV) analysis and Receiver Operating Characteristics (ROC) curves showed that MTV (Area Under the Curve (AUC) = 0.78), TLG (AUC = 0.78), and LDLM (AUC = 0.76) were good predictors of 1-year survival. For the subset of 97 patients with liver metastases, the corresponding regional measurements in the liver tended to be even better predictors. In conclusion, MTV and TLG were found to be better predictors of survival in metastatic PUM than the AJCC staging system, but when LDLM was used as a continuous variable it showed an equally good prediction of 1-year survival.
- Open Access
- 15 February 2024
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 86 (3)
Torben Martinussen’s contribution to the Discussion of ‘Parameterizing and Simulating from Causal Models’ by Evans and Didelez’
Torben Martinussen; Torben Martinussen’s contribution to the Discussion of ‘Parameterizing and Simulating from Causal Models’ by Evans and Didelez’, Journal of
- 15 February 2024
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 86 (3)
Open AccessTorben Martinussen’s contribution to the Discussion of ‘Parameterizing and Simulating from Causal Models’ by Evans and Didelez’
Torben Martinussen; Torben Martinussen’s contribution to the Discussion of ‘Parameterizing and Simulating from Causal Models’ by Evans and Didelez’, Journal of
- 7 February 2024
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 51 (3) , 1161-1180
Estimation of treatment effect among treatment responders with a time‐to‐event endpoint
In a placebo‐controlled clinical study one may calculate the average treatment effect to convey the effect of the active treatment on some outcome. However, if it is speculated that the treatment only has an effect if the patient responds to the treatment defined by a certain biomarker response, then it is arguably more relevant to estimate the treatment effect among such responders. We present such a causal parameter that is based on principal stratification and is identified under the exclusion of a treatment effect among the nonresponders. We focus on time‐to‐event outcomes allowing for right censoring, and construct a doubly robust and efficient estimator based on the associated efficient influence function. The properties of the estimator are showcased in a simulation study and the methodology is applied to the Leader trial investigating the effect of liraglutide on the occurrence of cardiovascular events.
- 7 February 2024
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 51 (3) , 1161-1180
Estimation of treatment effect among treatment responders with a time‐to‐event endpoint
In a placebo‐controlled clinical study one may calculate the average treatment effect to convey the effect of the active treatment on some outcome. However, if it is speculated that the treatment only has an effect if the patient responds to the treatment defined by a certain biomarker response, then it is arguably more relevant to estimate the treatment effect among such responders. We present such a causal parameter that is based on principal stratification and is identified under the exclusion of a treatment effect among the nonresponders. We focus on time‐to‐event outcomes allowing for right censoring, and construct a doubly robust and efficient estimator based on the associated efficient influence function. The properties of the estimator are showcased in a simulation study and the methodology is applied to the Leader trial investigating the effect of liraglutide on the occurrence of cardiovascular events.
- Open Access
- 3 May 2023
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 50 (4) , 1919-1932
Efficient t0$ {t}_0 $‐year risk regression using the logistic model- 3 May 2023
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 50 (4) , 1919-1932
Open AccessEfficient t0$ {t}_0 $‐year risk regression using the logistic model - 16 April 2023
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 50 (3) , 1532-1551
Targeted estimation of state occupation probabilities for the non‐Markov illness‐death model
- 16 April 2023
- Published by Wiley in Scandinavian Journal of Statistics
- Vol. 50 (3) , 1532-1551
Targeted estimation of state occupation probabilities for the non‐Markov illness‐death model