Conquering Cancers with AI

We develop clinical decision support systems (CDS) in the following precision oncology domains:

• Chemotherapies 

Chemotherapy drugs  are the backbone of contemporary "standard-of-care" regimens for treating cancers.  Such regimens are usually not personalized  due to the lack of biomarkers and methods to guide their application.  Our AI methods can predict responses to common chemotherapy regimens currently used in treating different cancers.  Our recent work on predicting efficacy of oxaliplatin and bevacizumab in colon cancer provides a strong evidence that our AI method is clinically applicable and likely will change practice.  We believe that precision application of majority of chemotherapies can be achieved through AI, which would improve outcomes of the majority of cancer patients. 

• Molecular targeted anticancer drugs 

Conventionally, application of molecularly targeted drugs is guided by mutation status of targeted genes.  We have developed AI solutions to infer the activation state of targeted pathways in cancer cells, which lead to improved accuracy for predicting sensitivities to molecularly targeted drugs.

• Immune therapies  

Individual tumors exploit distinct immune evasion mechanisms.  We have developed an AI framework for modeling cell-cell communication within each tumor to gain insights of immune evasion mechanisms of individual tumors.  Such information is highly valuable in predicting responses to contemporary immune therapies and for drug discovery.