Research News

Cleveland Clinic researchers are advancing genetic testing for cancer treatment with a new tool called the chemogram. The method is published in PLoS Computational Biology and uses the tumor’s gene expression to predict a patient’s chemosensitivity, or how well their cancer is likely to respond to different chemotherapy drugs.  

The research team, consisting of graduate student and lead author Kristi Lin-Rahardja; former graduate student and current medical resident Jessica Scarborough, MD, PhD; and principal investigator Jacob Scott, MD, DPhil, aimed to design and demonstrate a simple and effective method that tailors chemotherapy to individual patients.  

“Standard chemotherapy benefits many cancer patients, but its effectiveness can’t be guaranteed on an individual basis given its one-size-fits-all approach,” Lin-Rahardja says. “As such, some people may experience potentially debilitating side effects of chemo with limited therapeutic benefit.”  

Physicians can’t predict whether a patient will benefit from a standard chemotherapy regimen without testing the tumor before treatment. This process is much easier said than done. Scientists are starting to develop tests that would determine which drugs might work best. These tests are moving the field in the right direction, but currently aren’t practical, Lin-Rahardja says. 

“Testing different chemotherapies on a tumor sample is just too expensive and time-consuming. Even if a patient can afford the potential six-figure price tag, they may not have the months it can take to get results,” she explains. “Because of this, cancer drug screens are rarely used to inform treatment in any scenario.”  

Genetic tests for cancer treatment are faster, cheaper and easier, but only one test is currently recommended for clinical use. Other tests aiming for approval also focus on one cancer type, or one drug at a time. Dr. Scott tasked Lin-Rahardja with developing a genetic test that could be used across cancer types, across multiple drugs and in real-world clinical settings. 

Using genetic testing for cancer treatment  

The chemogram uses a computer algorithm to analyze a tumor’s RNA, or gene expression, and produce a personalized list of existing, approved chemotherapy drugs, ranked by how effective they would be at treating the tumor. Ranking drugs based on how likely they are to work for that specific patient avoids confusion and gives physicians clear and straightforward framework for making clinical decisions. The team got the idea and name for the chemogram from a method called the antibiogram, which gives physicians a simple chart showing which antibiotics work and which don’t for a given infection.  

“Many other drug response prediction models can also predict how well a specific drug would work in a tumor, but those methods don’t let you compare the different drugs or generate ranked lists tailored to individual patients,” Lin-Rahardja says. 

Lin-Rahardja used a method previously developed by Dr. Scarborough to begin developing the chemogram. Using Dr. Scarborough’s method, she first identified 72 new gene signatures that predict a tumor’s response to 72 different cancer drugs, picked out the top 10 strongest signatures, and then made a framework that uses those signatures to rank which of those 10 drugs would be most to least effective against individual tumors. . She is now working to validate her method against clinical trial data. 

“Kristi has developed a wonderful method and demonstrated the proof-of-concept needed to justify its use in a clinical setting,” Dr. Scott says. “A paper describing an improved version of the chemogram is already on the way, and we have filed IP disclosures with Cleveland Clinic Innovations.”