Cure models
in osteosarcoma

Background

Cure rate models are designed for handling survival data in the presence of long-term survivors or cured individuals. For example, for many cancer types treatment can lead to cure for a fraction of the patients. Even though these models have gained considerable attention in the statistical literature in the recent years, they have been hardly applied in childhood cancer. This is particularly important when the heterogeneity of the population (long and short term survivors) leads to a violation of the proportional hazards assumption. In such cases the use of the popular Cox model would give misleading results. Moreover, cure models provide additional information with respect to the Cox model since they are able to distinguish between a curative or a life-prolonging effect of a treatment or of other factors. Estimating the probability of being ‘cured’ on the basis of personal characteristics can in itself be of interest in order to assess the patient’s condition and to make better decisions in terms of further treatment strategies.

Aims

The project aims at developing new statistical methods for dealing with cancer survival data in the presence of cure (long term survivors) and promoting the use of cure rate models in the medical context. In particular, we will focus on childhood osteosarcoma and the following aspects:

• Can cure models help us to understand better the effect of chemotherapy intensification on survival and cure?

• Evaluating the risk of recurrence through a competing risk model with a cure fraction.

• Understanding whether the follow-up of a particular study is sufficient for enabling us to use cure models and to be confident about the results.

Relevance for cancer research

Through the use of cure models and the developed methodology, the project aims to provide new insights into the role of  therapy and dose intensification in curing or prolonging the life of children diagnosed with osteosarcoma. A distinction will be made between prognostic factors for the cure status and for the survival of the uncured patients, based on which personalized survival predictions can be made. The methods can be applied also to other cancer types with cure possibility and sufficient follow-up.

Project Outcomes

• Jacobs, T (2023). Identifiability of cure models in a competing risk framework. MSc Thesis; supervisors: prof.dr. M. Fiocco, dr. E. Musta

• Musta E, van Geloven N, Anninga JK, Gelderblom H & Fiocco M (2022). Short-term and long-term prognostic value of histological response and intensified chemotherapy in osteosarcoma: a retrospective reanalysis of the BO06 trial. BMJ Open, 12:e052941. doi: 10.1136/bmjopen-2021-052941

Team

• Prof.dr. M. Fiocco, Mathematical Institute Leiden University, Department of Biomedical Data Sciences Leiden University Medical Center & Princess Máxima Center for Pediatric Oncology 

• Prof.dr. H. Gelderblom, Department of Medical Oncology at the Leiden University Medical Centre 

• T. Jacobs, BSc, Mathematical Institute Leiden University 

• Prof.dr. J.H.M. Merks,  Princess Máxima Center for Pediatric Oncology

• Dr. E. Musta, Korteweg-de Vries Institute for Mathematics, University of Amsterdam