A team of neuro-oncologists has developed a graded prognostic assessment (GPA) tool to help clinicians determine how long their patients will live following the diagnosis of brain metastases. Not long ago all patients were thought to have the same dire prognosis given the spread of the primary cancer. But a retrospective analysis of the medical records from thousands of patients has provided the scientists with a new perspective: Not all brain metastases are alike.
MRI OF PATIENTS BRAI...Image Tools
According to the new study, published in the Dec. 27 online edition of the Journal of Clinical Oncology, the prognostic factors will determine survival and vary depending on the type of cancer that ultimately traveled into the brain.
DR. PAUL W. SPERDUTO...Image Tools
“Historically, patients with brain metastases have been treated the same way,” said the lead study author Paul W. Sperduto, MD, a radiation oncologist and director of the Gamma Knife program at the University of Minnesota. “A lot of neuro-oncologists have seen variability among their patients but this scoring system allows them to better assess survival and use that information to help families make decisions about the available treatment options.”
The investigators initially created the GPA after studying the files of 1,980 patients from four prospective randomized Radiation Therapy Oncology Group (RTOG) clinical trials. They have recently expanded to 11 centers that pulled together records from 3,940 patients with newly diagnosed brain metastases, all of whom had been treated. They matched up prognostic factors with different cancers and tumor subtypes and discovered that specific prognostic factors could be used for each of the cancers to help determine a prognosis. A score of zero offers the worst prognosis.
They studied patients with small-cell lung cancer, non-small cell lung cancer, melanoma, breast cancer, renal cell carcinoma and gastrointestinal (GI) cancers. The overall median survival differed for each of the cancers and specific prognostic factors were stronger in determining survival. They went back into the medical records to look at the primary tumor type, the number of brain metastases, age at diagnosis, the identification of new tumor growth outside of the brain, tumor subtypes, and Karnofsky performance score, which allows patients to be classified according to their functional impairment (a 0-100 scale, where 0 is normal, 100 is dead).
The patients entered into the study were newly diagnosed of brain metastases within two months. They also had a range of treatment. Patients with recurrent brain metastases were excluded from the study.
Using the GPA, they found that the median survival time (MST) varied substantially. For instance, patients with non-small-cell lung cancer lived an average of 7 months; small-cell lung cancer patients about 4.90 months; melanoma patients 6.74 months; patients with renal cell carcinoma 9.63 months; breast cancer patients 13.8 months; and patients with GI cancers, an average of 5.36 months.
The variation within each group is small, said Dr. Sperduto, who noted that there was a narrow range of survival reflected in the 95% confidence interval and the separation between each group was significant at p< 0.001). (See the table, Median Survival Time for Patients with Brain Metastases.)
The prognostic factors varied by prognosis. For lung cancer, the Karnofsky performance score, age, presence of extracranial metastases, and number of brain metastases were important in determining survival. In melanoma and renal cell carcinoma the prognostic factors were Karnofsky performance score and the number of brain metastases. For breast cancer, the prognostic factors were tumor subtype, Karnofsky performance score, and age was a weak predictor. Finally, for GI cancer, the only prognostic factor was the Karnofsky performance score.
The number of brain metastases was important in determining prognosis in lung cancer, melanoma and renal cell carcinoma but not for breast cancer and GI cancers, Dr. Sperduto said. Extracranial metastases were only prognostic of the length of survival in lung cancer.
“The scale is specific for each diagnosis,” he added. “It was a form of discrimination to group patients into the same dismal outcome.” He said that patients with a high GPA could be treated more aggressively than those who have a low GPA score. “We should not treat all patients with brain metastases the same way,” the clinicians wrote in the paper. “Treatment should be individualized and the past philosophy of fatalistic futility should be abandoned.”
The GPA index has been adopted by the Radiation Therapy Oncology Group to stratify two new prospective randomized clinical trials, and is in the process of being prospectively validated. The design of the original study was based on four prospective randomized clinical trials.
EXPERTS WEIGH IN
“It is always hard to get a reasonable idea of a person's prognosis,” said Patrick Y. Wen, MD, a professor of neurology at Harvard Medical School and director of the Center for Neuro-oncology at the Dana Farber/Brigham and Women's Cancer Center. “This paper will help significantly.” In particular, he said, he was surprised that systemic disease wasn't as important to survival. He added that the worksheet in the paper will help clinicians get a rough idea of how their patients will do. “It will also be useful in stratifying patients for research trials.” He said that the assessment tool comes at a time when some of the first potential medicines for brain metastases are being studied.
Lynne Taylor, MD, a neuro-oncologist and director of palliative care at Tufts Medical Center, agrees that it will help “treat more of the appropriate cases and fewer of the inappropriate ones.” She said that she recently had a patient with breast cancer who had two brain metastases that grew to 100 in one month. “Everyone assumed her prognosis was poor but this study tells us that the number of tumor metastases for breast cancer doesn't really factor into the prognosis. It gives us new insights in how to treat these patients,” said Dr. Taylor.
Lisa Rogers, DO, head of the medical neuro-oncology program at University Hospital Case Medical Center in Cleveland, said that such a tool becomes more important as the number of treatment options for brain metastases increase.
But not everyone agrees that the new assessment tool will help guide treatment for cancer patients with brain metastases. David Larson, MD, professor of radio-oncology at the University of California, San Francisco, said that he doesn't know how clinicians will use this information. “Survival is just one outcome,” Dr. Larson said. “Outcome is determined by many things, including treatment. Also, the researchers may have looked at when a patient died but they did not take into account a person's quality of life during that time.”
He noted, as well, that the patient review began in the 1980s, which was the beginning of better treatments for both primary cancers and brain metastases. He said that he would never use a worksheet to determine how long someone will survive. While he may not use it in his clinical practice, he does agree that it is a good tool for research.
David Schiff, MD, a neuro-oncologist at the University of Virginia, agrees that it may not be practical to use in the clinic — “there is too much to remember” — but it will be important for research to correctly stratify patients before they deliver treatment. Still, he said: “The study is provocative because it challenges certain dogma in the field.”
MEDIAN SURVIVAL TIME...Image Tools
• Sperduto PW, Kased N, Mehta M, et al. Summary report on the graded prognostic assessment: An accurate and facile diagnosis-specific tool to estimate survival for patients with brain metastases. J Clin Oncol 2011; E-pub 2011 Dec 27.
©2012 American Academy of Neurology