Multi-syndrome, multi-gene risk modeling for individuals with a family history of cancer with the novel R package PanelPRO

Abstract

Identifying individuals who are at high risk of cancer due to inherited germline mutations is critical for effective implementation of personalized prevention strategies. Most existing models focus on a few specific syndromes; however recent evidence from multi-gene panel testing shows that many syndromes are overlapping, motivating the development of models that incorporate family history on several cancers and predict mutations for a comprehensive panel of genes. We present PanelPRO, a new, open-source R package providing a fast, flexible back-end for multi-gene, multi-cancer risk modeling with pedigree data. It includes a customizable database with default parameter values estimated from published studies and allows users to select any combinations of genes and cancers for their models, including well-established single syndrome BayesMendel models (BRCAPRO and MMRPRO). This leads to more accurate risk predictions and ultimately has a high impact on prevention strategies for cancer and clinical decision making. The package is available for download for research purposes at https://projects.iq.harvard.edu/bayesmendel/panelpro.

Statistical methods for Mendelian models with multiple genes and cancers

“…The contents of Sections 2.1-2.2 expand on and provide additional context for the framework described in Lee et al (2021), including a generalization of the genotype vectors to multiple variants/mutation states, as well as notation for net future risk and further context for net versus crude penetrance and future risk estimates. The following sections, Sections 2.3-2.4, contain entirely new elements that cover extensions for secondary cancers and other additional risk modifiers and describe conditions for model collapsibility.…”

“…However, it may be less effective for large pedigrees. Further discussion of genetic linkage algorithms and their computational considerations can be found in Lee et al (2021).…”

“…The framework introduced is highly flexible and can be used to obtain the future risk of cancer for all cancers in the model. Lee et al (2021) present the R software package implementation for PanelPRO. It covers the informatics aspects of PanelPRO; detailed statistical topics are not included in Lee et al (2021) and are the focus of this study.…”

“…We provide comprehensive notation for our statistical model, including the flexible incorporation of secondary cancers and additional risk modifiers, and derive conditions for model collapsibility. As previously mentioned in Lee et al (2021), we reduce the computational expense of the peeling algorithm to a feasible level for clinical use by making some reasonable simplifying assumptions (Madsen et al, 2018). We scrupulously verify the comparability of the prediction performance of PanelPRO against existing models in simulations and evaluate model performance using simulated data and a validation study using a high-risk multigene panel testing data set from the USC-Stanford Hereditary Cancer Panel (HCP) Testing Study.…”

Predicting cancer risk based on family history

“…Nevertheless, older risk modeling programs generally include only a small number of genes in their predictions. Now, in eLife, Danielle Braun and colleagues – including Gavin Lee and Jane Liang as joint first authors – report on a new software package that has the capacity to evolve alongside advances in cancer research ( Lee et al, 2021 ).…”