September 2012, Vol 1, No 4

← Back to Issue

New Approach for Predicting Treatment-Related Side Effects

Phoebe Starr

Conference News

Researchers are evaluating the use of a cluster of single-nucleotide polymorphisms (SNPs) identified by a Bayesian method from an individual cancer patient to predict the occurrence of treatment-related side effects in that patient. Two unpublished studies have shown that SNPs identified by the Bayesian method have an accuracy of more than 90% in predicting treatment-related side effects, explained Stephen Sonis, MD, Brigham and Women’s Hospital, Boston, Massachusetts. Sonis presented preliminary data from the SNP project at the recent meeting of the Multi­national Association of Supportive Care in Cancer in New York City.

“We believe this [method] will…let us prospectively evaluate and understand who is at risk for side effects from cancer therapy. This can have a major impact on…how we optimize outcomes for our patients,” he stated.

SNPs are the most common variation in the genome and outnumber genes exponentially, with over 10 million SNPs and only 25,000 genes. SNPs are not necessarily associated with function, he explained.

Using the Bayesian method to analyze 1 million to 2 million individual SNPs and their interactions, it is possible to pick the ones that are relevant. There is no predetermined number of relevant SNPs, or “team players,” as he called them.

At Dana-Farber Cancer Center, a retrospective study used Bayesian networks to identify 82 SNPs that were able to predict patients who would develop mucositis with 90% accuracy. The investigators examined charts of myeloma patients slated for transplant from 2001-2006 and identified 153 subjects; classified them as oral mucositis (OM)-negative (n=102) and OM-positive (n=51), with positivity defined as having 2 consecutive days of WHO grade >2 OM; and extracted DNA from their saliva specimens.

The investigators then moved on to other cancer types in the OnPART study, which is currently being conducted at the West Clinic in Memphis, Tennessee. This study attempts to use the same method to identify relevant SNPs from individual patients with breast, co­lorectal, ovarian, and non–small cell lung cancer who are being treated with 3 or more cycles of chemotherapy.

Preliminary data in 30 breast cancer patients treated with anthracycline/taxane chemotherapy showed that the rates of side effects were nausea/vomiting, 42%; OM, 38%; diarrhea, 20%; fatigue, 58%; cognitive dysfunction, 23%; and peripheral neuropathy, 16%. The accuracy of prediction using a preselected network of SNPs was 96.7%.

Sonis said the data are consistent with the first study in transplant patients with myeloma conducted at Dana-Farber.

Uncategorized - September 19, 2012

Adaptive Clinical Trial Design: From Simple Dose-Finding Trials to Large-Scale Personalized Medicine Trials

While there’s great excitement about the potential of personalized medicine to improve care – particularly in oncology – there’s also a healthy dose of pessimism regarding the cost of clinical trials needed to bring optimal, targeted therapies to market. While many researchers believe that personalized medicine is the future of [ Read More ]

Conference News - September 19, 2012

Severe Diarrhea Associated With Molecularly Targeted Agents Can Impact Quality of Life and Healthcare Resource Utilization

A preliminary report of a meta-analysis of clinical trials of molecularly targeted therapies shows that they are not benign and can add to the toxicity of standard chemotherapy. In particular, increased rates of oral mucositis and diarrhea are reported with several FDA-approved agents. Increased mucositis seen with bevacizumab and erlotinib [ Read More ]