In partnership with international researchers, Sylvester Comprehensive Cancer Center researchers at the University of Miami Miller School of Medicine have created an advanced AI algorithm that can conduct sophisticated computational analysis to identify treatment targets for glioblastoma multiforme targets and different cancer forms.

Groundbreaking research in GBM treatment 

The research published in the Nature Cancer journal could have significant implications for GBM treatment alongside lung, breast, and pediatric cancers. GBM is an aggressive and fatal cancer. 

Sylvester Comprehensive Cancer Center deputy director and senior study author Antonio Iavarone said the work reflects translational science, offering immediate opportunities to transform clinical ways of managing GBM. Most importantly, the algorithm provides precision cancer treatment applications, offering oncologists the latest tool to fight GBM and other cancers. 

The algorithm is known as SPHINKS, which is an acronym for Substrate Phosphosite-based Inference for Network KinaseS that is used in machine learning to help in the identification and experimental validation of the PKCδ and DNAPKcs protein kinases. The two protein kinases are responsible for tumor progression in two subtypes of GBM and as possible therapeutic targets for more cancers. 

Usually, protein kinases are the main targets in precision cancer treatment, and they can be personalized to specific cancer properties. The researchers identified the most active kinases, labeling them “master kinases, ” which clinicians can target directly as a hallmark of cancer treatment.

Researchers used lab-grown tumor organoids for targeted treatment 

Additionally, besides the master kinases, Ivarone and his team employed tumor organoids cultured in the lab from patient samples to demonstrate that targeted medication that interferes with master kinases’ activity can prevent tumor progression. They labeled organoids as “patient-derived tumor avatars.  

The researchers reported a new glioblastoma category by capturing important tumor cell traits and categorizing GBM patients depending on their potential survival and their cancerous cells’ vulnerability to drugs. Various omics platforms confirmed the new classification.

SPHINKS uses machine learning to refine the omics data sets and establish an interactome. Findings show that multi-omic data can create new algorithms that show the best-targeted therapies for enhanced outcomes.