Cancer Avatars for Personalized Medicine

They’re not blue and they’re not science fiction.  They are avatars, at least of the cancer kind.  From the hands of Dr. Sandeep Pingle, MD PhD from Moore’s Cancer Center in UC San Diego, CA comes computer representations of tumor cells.  Defined as an icon or figure representing a person, avatar is an appropriate term to describe the computer simulation modeling the genomic signatures of individual cancers and how they will respond to specific therapeutic interventions. 

In light of the recent discussions on how each tumor is unique and the need for personalized medicine, this newest discovery provides a tool for a comprehensive analysis of individual cancers.  Most importantly, it can be customized for each patient, allowing clinicians to treat patients selectively with drugs believed to have the greatest chance of success.

Published recently in the Journal of Translational Medicine, the generation of this virtual cell combined intensive analysis of the literature with basic cell survival assays in response to targeted therapies.  Cell lines with different genetic backgrounds, like patients, will respond differently to specific therapeutics.  Using this knowledge, the virtual cell can predict how cells, or individual tumors, will respond to these different drug treatments.

To validate their model, they generated cell lines from glioblastoma tumors.  They used their cancer avatar to predict survival and tested this against actual survival data generated in the laboratory.  In this situation, their models saw a high degree of agreement with greater than 75% corroboration between in vitro cell line models and the cancer avatar.  Ultimately such a model should be tested in clinical trials with real patient data. 

Where do we go from here?

You can imagine that such a model could provide enormous benefit for clinicians, providing an effective platform for testing and developing personalized therapeutic regimens.  However, the future development of this model relies on its accuracy.  With new technologies, including next generation sequencing, we have access to large amounts of data including genomic, transcriptomic, metabolomic, and proteomic data, among others.  Integration of these datasets is necessary for accurate predictions.  The incorporation of extracellular factors such as the tumor microenvironment and tumor-associated inflammation is also critical. 

I’ll leave you with this thought:  Are we placing too much faith in computer simulations.  Cancer is, by definition, unpredictable and ever changing.  Can we accurately model that disease?

References:

Pingle SC, Sultana Z, Pastorino S, et al.  In silico modeling predicts drug sensitivity of patient-derived cancer cells.  J. Transl. Med.  2014. 12:128.