Targeting Brain Cancer

SEVEN YEARS AFTER​ ​her brain cancer diagnosis, Liz Salmi decided she​ wanted to know more about the makeup of the tumor that changed her life.

Salmi had already undergone two surgeries to remove the grade II astrocytoma, an invasive but usually slow-growing brain tumor. After her second surgery, she was treated with Temodar (temozolomide), a chemotherapy drug, for two years. But Salmi, who was 29 when she was diagnosed in 2008, knew her cancer could not be cured. She wanted to be ready when it returned. “I’m living with a brain cancer, and I keep up with what’s going on,” says the 38-year-old communications specialist, who lives in Sacramento, California.

In 2015, Salmi read about powerful genomic sequencing tools that made it possible to analyze tumor tissue—peeking inside brain cancer cells and looking for the genetic drivers of the disease. Many experts were arguing that tumors should be classified not only by their appearance under a microscope—the traditional approach—but also according to the presence of certain genetic changes.

When Salmi first learned she had brain cancer after a series of debilitating seizures in the summer of 2008, this kind of molecular information wasn’t widely available to patients. “I almost feel like I was diagnosed in the Stone Age,” she says. But when the technology became more widespread, she wanted to know what molecular analysis could reveal about her tumor. She approached her neuro-oncologist, who cautioned her to think about her request for a few months.

“Once you know something, you can’t unknow it,” Salmi says, explaining her doctor’s concern. The results of tumor sequencing, for example, might suggest a favorable or an unfavorable prognosis. Or they could provide no meaningful information at all. Still, Salmi was curious. “The best way for me to cope was to know everything possible,” she says.

Understanding the molecular changes that drive brain cancer remains “potentially transformative,” says Paul Mischel, a cancer biologist and neuropathologist at the Ludwig Institute for Cancer Research in San Diego. Molecular biomarkers can help predict a brain tumor’s likely response to the chemotherapy drug Temodar. This testing can also help assess the risk that a tumor will grow and spread, and can reveal certain genetic variants in the tumor that may respond to experimental treatments. However, genomic testing of brain tumors has yet to uncover ways to extend survival beyond that provided by standard treatments.

Why the Head Is So Hard

More than 80,000 people in the U.S. are diagnosed with some type of brain tumor each year; of those tumors, about a third are malignant. Brain tumors are usually diagnosed following the onset of troubling symptoms, such as seizures, nausea, vomiting, drowsiness or memory problems.

Astrocytoma, Salmi’s tumor type, is a glioma. The most common type of brain cancer, gliomas originate in glial cells, which surround neurons in the brain. Gliomas are graded from I to IV. The most aggressive type is grade IV, called glioblastoma, which has the worst prognosis. The median survival after a glioblastoma diagnosis is around 15 months, a number that has barely budged in the past 50 years.

Compared to treatment advances for other types of cancer, progress in treating brain cancer has been frustratingly slow. The U.S. Food and Drug Administration (FDA) has approved 16 targeted therapies for lung cancer, 15 for breast cancer and seven for colon cancer, but only one—Avastin (bevacizumab)—for malignant brain tumors. And there’s no evidence that Avastin extends survival longer than chemotherapy and radiation, though it may delay symptoms and extend the time it takes for a tumor to grow.

Targeted drugs and immunotherapies that have generated excitement by extending survival in some patients with other cancer types have largely failed in clinical trials for brain cancer patients. “Brain tumors do not respond to a lot of the treatment strategies that we try to impose on them,” says Eric T. Wong, a neuro-oncologist at Beth Israel Deaconess Medical Center in Boston.

There are a number of reasons for this, he says. First, heterogeneity: Brain cancer tumors contain a staggering mix of cells with different mutations, not only from person to person but within a single tumor. That means a treatment that targets one type of mutation might be ineffective against cells in the tumor with a different set of mutations. A treatment-resistant tumor could be the result. “The cells that are most likely to be resistant are the ones most likely to thrive,” says Mischel.

Then there’s the hurdle of location. Even though brain cancer rarely spreads to other parts of the body, brain tumors such as gliomas can extend like tendrils into surrounding brain tissue. Thus, even the best surgical resections can’t remove all the cancerous cells from the delicate tissue of the brain.

In addition, “the barrier to the brain is tightly controlled,” says Mischel. This blood-brain barrier, which blocks foreign substances in the blood from entering the brain, can also impede cancer treatment. Chemotherapy, which is often given orally or intravenously after surgery to patients with fast-growing brain cancer, may not even get to a brain tumor because of the barrier. In the same way, targeted drugs that seemed promising in lab and animal studies on brain tumors may not improve survival in patients in clinical trials because they never reach the intended target, Mischel says.

In other cases, a clinical trial may benefit a small subset of patients and be declared a failure, halting development, because the drug didn’t benefit enough people. “Clinical trials [in brain cancer] usually fail,” says Mischel. “There’s a real emphasis in really thinking hard about an alternate way of doing clinical trials with cancers in the brain.”

Molecular analysis offers one way to rethink clinical trials and to help identify patients who might benefit from a given drug. The National Cancer Institute’s MATCH trial, for example, tests targeted drugs on patients based on their tumors’ genetic characteristics, rather than a cancer site.

Brain cancer has been the focus of intense genetic studies in Europe, Canada and the U.S. In 2006, glioblastomas were selected as one of the first tumor types to be studied as part of The Cancer Genome Atlas (TCGA), a 10-year sequencing effort to characterize the genomes of more than 20 tumor types.

“The TCGA really opened up the field,” says David Arons, CEO of the National Brain Tumor Society and chair of the National Cancer Institute’s Council of Research Advocates. “It allowed us to look under the hood of glioblastoma. If you can understand how the engine works and fits together, maybe you can understand how to stop the engine of tumorigenesis.”

The Impact on Patients

Investigations into brain tumors’ genetic machinery have produced some insights. A 2005 study by European and Canadian cancer research foundations, for example, found that the chemotherapy drug Temodar was more effective in treating glioblastoma patients with tumors whose MGMT gene had been switched off. In another advance, researchers showed that in low-grade gliomas, mutations in genes called IDH1 and IDH2 are associated with longer survival and less aggressive disease. Tumors without these mutations are more likely to grow and spread.

For that information alone, Arons recommends that brain cancer patients discuss molecular testing with their doctors early in their treatment. “I’d want to find out if my lower-grade brain tumor was more likely than not to transform into an aggressive brain tumor.”

Molecular research also spurred the classification changes that Salmi learned about. In May 2016, the World Health Organization, the branch of the United Nations that studies and promotes health worldwide, published a new system of classifying brain tumors based on molecular signatures, in addition to the usual evaluation of the cells under the microscope.

Given the lack of treatment options, however, having so much information can lead to frustration. Michael Yutkin, an asset manager in Irvine, California, was diagnosed in August 2012 with a glioblastoma during a business trip to Chicago. After surgery, he had radiation and received Temodar while also joining a clinical trial for the targeted therapy drug Velcade (bortezomib). The Temodar treatments ended after two years, but he continued to take Velcade until his white blood cell count dropped in February 2015. His tumor began to regrow in November 2015 and he underwent more radiation therapy and another year of Temodar.

In the spring of 2016, Yutkin had his tumor tissue analyzed. The report identified six genomic alterations, but five came with no recommended treatments or clinical trials. The sixth was an amplification, or production of multiple copies, of the EGFR gene. Patients with non–small cell lung cancer who have this alteration have responded to drugs called tyrosine kinase inhibitors. However, most patients with glioblastoma and an EGFR amplification have not responded to these drugs in clinical trials, so it wasn’t a treatment option for Yutkin.

He doesn’t regret getting the molecular tests done, “but I’m not sure how much it really told us,” he says. Scans taken in June 2017 showed that the tumor had started growing again, which led to a second surgery in October 2017, and he’s now hoping to enroll in a clinical trial for an experimental cancer vaccine or another immunotherapy. “The field isn’t moving fast enough,” he says.

In rare cases, genetic information can make a big difference—but these cases are not typical. Five years ago, Emil Lou, a medical oncologist at Masonic Cancer Center at the University of Minnesota in Minneapolis, had a 51-year-old patient with a brain cancer called medulloblastoma that is more commonly diagnosed in children. The patient, who was in a wheelchair as a result of symptoms from his condition, had multiple dime-sized tumors in his brain. He experienced constant nausea and vomiting, and Lou doubted he would survive chemotherapy. He landed upon a long-shot option: a drug that had been recently approved by the FDA for patients with basal cell carcinoma. The therapy blocks a molecular pathway involving a protein called sonic hedgehog (SHH), and Lou knew that SHH mutations were common in medulloblastoma patients. Genetic analysis of one of the patient’s tumors later revealed a mutation involved with SHH.​​

Lou prescribed the drug, Erivedge (vismodegib), for the patient. Four months later, his tumors had vanished. He continued the drug for 22 months and remains disease-free today. Lou cautions that his patient was not the norm; he is what is known as an exceptional responder. Still, the case offers a glimpse of the potential power of genetic information. Such cases depend on “whether or not we can identify a biological target and the drug it responds to,” says Lou. Without molecular analysis, these possibilities wouldn’t exist.​

Getting to Standard of Care

Molecular analysis of Salmi’s brain tumor fell in line with what she already knew from her treatment experience. Her tumor had an IDH mutation that has been associated with a better prognosis for her condition. It also showed methylated MGMT, which has been associated with a strong response to Temodar in patients with glioblastoma, though Salmi has a different type of glioma. She had her most recent scan in May 2017, and even though the scans showed no new tumor growth, she knows she hasn’t been cured.

“I still have cancer in my brain,” she says, “but it’s just not growing.” In the event the tumor does grow back, genomic information gained from being tested could point her toward clinical trials or new targeted treatments. She has also submitted saliva samples for a study following long-term survivors of low-grade brain tumors to look for markers that might be associated with survival.

Molecular analysis of brain cancer is not yet the universal standard of care, says Arons, although he predicts it will be in the next few years. Because these tests haven’t been associated with a clear patient benefit in all cases, they’re not universally covered by Medicare, and “there is resistance from insurance companies” to reimburse for them, says Lou.

This makes for a frustrating Catch-22: Information derived from molecular testing hasn’t been shown to improve survival, but researchers won’t be able to explore new treatments without more tumors being tested. In other tumor types—like lung and breast cancers—genomic testing has already influenced the development of new drugs and guided oncologists to use targeted therapies for patients whose tumors have specific mutations. Patients, researchers and oncologists all want to see that kind of shift in brain cancer. Lou says the rapidly growing knowledge of the molecular basis of the disease could make it possible. “The story is still being written,” he says.

Cancer Today magazine is free to cancer patients, survivors and caregivers who live in the U.S. Subscribe here to receive four issues per year.