Skip Ribbon Commands
Skip to main content

AACR Annual Meeting 2014: Why Cancer Drugs Stop Working in Some Patients

The AACR's immediate past-president Charles Sawyers believes finding the answers to drug resistance requires more collaboration by researchers and patients alike. By Marci A. Landsmann

As genomic analysis becomes increasingly common, Charles Sawyers, a physician-scientist at Memorial Sloan Kettering Cancer Center in New York City and a presenter at the Scientist↔Survivor Program at the American Association for Cancer Research (AACR) Annual Meeting 2014, feels confident that scientists may soon discover why drugs stop working in some cancer patients. 

“We’ve become extremely good at making drugs that inhibit many of the cancer-driving genes," 
Sawyers, the past-president of the AACR, told advocates during his presentation April 7. “We are killing more tumor cells with new drugs, but there are some residual cells that are looking for a back door. [The cancer then says,] ‘I’m going to change from a prostate cancer into something else. It’s changed its identity.’ ”

Once this happens, drugs that previously worked stop working—posing new problems and harder-to-treat cancers. By analyzing changes in the genes, however, Sawyers believes scientists can pinpoint more therapeutic targets. “We have unbelievable technology to define the genomic complexity of the tumor within days with minimal costs. We are doing this at unbelievable scale at many cancer centers,” he says.

The Cancer Genome Atlas (TCGA), which set out to sequence 500 cases of several types of cancer, mapped some of the common genetic mutations that occur in about 10 to 20 percent of all the samples, Sawyers says. However, approximately 2,000 tumor samples from each tumor would allow scientists to determine 98 percent of all the mutations in the cancer cells, according to a recent analysis that Sawyers recounted to advocates at the meeting. Using this statistical analysis, Sawyers estimates that approximately 100,000 tumor samples would be needed to map mutations in 50 different types of cancer. 

Many of these tumors have already been sequenced, but, unfortunately, they exist in disparate cancer centers across the world. “We [need to] learn how to pool the data that is going on collectively,” says Sawyers. Most U.S. cancer centers have enough samples sequenced already to come to close to these numbers, he says. 

Sawyers estimates that Memorial Sloan Kettering Cancer Center will analyze 10,000 tumors this year alone, but these samples are kept in what he calls “lead-shield silos.” Hospitals and facilities are not often able to share data because of patient privacy. To tackle this problem, an international organization, the Global Alliance for Genomics & Health, aims to confront some of these issues and to facilitate ways to manage and share genomic information between health care facilities and industries within an ethical and secure framework.

Patients can play a role in increasing this knowledge base by agreeing to release their medical information. To help give researchers the tools to overcome drug resistance, patients also can consent to matched biopsies, which means having a biopsy both before and after treatment. The genetic information obtained from these samples allows researchers to isolate new mutations as the cancer cells assume new genetic identities. Using matched biopsies, scientists would need approximately 250 to 1,000 tumor samples to isolate a mutation that occurs in 1 percent of a given cancer type. Without matched biopsies, Sawyers estimates scientists would require tens of thousands of tumor samples to find the same aberration, he says.

“One approach [to ensuring genetic information is available for research] is to go straight to consumers,” says Sawyers. “Instead of dealing with doctors and medical facilities, [you can] load up your clinical data, but there are people who worry about releasing their medical information. Just from a small stretch of DNA, people can identify you. It is almost impossible to share this information without people putting themselves at some risk that they will be identifiable.”

To this, Virginia Hetrick, a cancer advocate in the audience and an inflammatory breast and endometrial cancer survivor, shared a recent conversation she had with three of her doctors. “All of us agreed that we weren’t concerned about [patients] being identifiable as having this or that, as a consequence of our data being used [in this fashion]. I’ve had enough things go wrong with me to know that there’s not much else to lose,” Hetrick says. “It’s just dumb not to share information that can help people.”


MARCI A LANDSMANN is the editor of Cancer Today.


Receive monthly updates, including information about web exclusives, events, resources, articles and highlights from new issues—direct to your email inbox. Be among the first to hear the latest news from Cancer Today! Click here to sign up!