Abstract:
|
A common phenomenon in cancer is that for the same individual, multiple sites may present primary cancer at different times in life. In Li-Fraumeni syndrome(LFS), a rare pediatric cancer syndrome, TP53 mutation carriers are known to have a higher probability of developing second primary cancer than the non-carriers. Modeling the development of multiple primary cancers is therefore desired for better clinical management of LFS. To this end, we have developed a non-homogeneous Poisson process model with multiplicative rate function in order to account for primary cancer events occurred over time. We constructed a novel familywise likelihood under the Poisson process, which allows for inclusion of all family history information even when the individual genotype is missing. We used Markov chain Monte Carlo algorithm to estimate model parameters, and derived age-at-onset penetrance for single and multiple primary cancers given mutation status, respectively. We applied our method to a pediatric sarcoma cohort collected at MD Anderson Cancer Center from 1944 to 1982. Our penetrance estimates are consistent with previous studies on LFS and reasonably predict future cancer risk.
|