Measuring Medicine: How New Technologies Could Help Doctors Predict Patient Outcomes

SAN DIEGO - As potential cancer therapies proliferate, researchers and clinicians are striving to measure their effectiveness and to more accurately predict which patients will receive the most benefit. At the American Association for Cancer Research 2008 Annual Meeting, researchers present data on a new role for MRI in brain cancer, how doctors can more effectively measure response to commonly used cancer drugs, and a unique method for predicting the risk of breast cancer spread.

A phase II study of the efficacy and tolerability of lapatinib in patients with advanced hepatocellular carcinomas: Abstract LB306

Results of a phase II trial suggest that lapatinib, currently approved for breast cancer treatment, shows promise for stabilizing disease in patients with liver cancer.

“Lapatinib is well-tolerated and may have some activity in hepatocellular carcinoma (HCC),” said Joseph Markowitz, M.D., Ph.D., a researcher at The Ohio State University, who works with Tanios Bekaii-Saab, MD, the principal investigator on the study. “More work is needed to understand the underlying molecular mechanisms of this cancer.”

HCC rates are rising in the United States, which correlates with the increase in hepatitis C-related liver disease, a known risk factor for HCC, Markowitz says. “There is also a link to an increased incidence of what we call ‘fatty liver’ as a result of the increasing rates of obesity and diabetes mellitus in the U.S. population.”

Lapatinib blocks the activity of the tyrosine kinase of both epidermal growth factor (EGFR) and HER2/neu, Markowitz says. “A dual inhibitor such as lapatinib should be effective in patients who express one or both receptors. Given the lack of curative or even modestly effective treatment options for patients with advanced hepatocellular carcinomas, new therapies are desperately needed,” he said.

Markowitz, Bekaii-Saab and colleagues assessed the efficacy of lapatinib as an HCC treatment in a phase II trial with 26 participants.

Patients took a 1,500 mg oral lapatinib dose daily throughout a 28-day cycle. The median number of cycles during the trial was two with some patients receiving as few as one cycle and some receiving as many as 12 cycles. The researchers performed radiological assessments every eight weeks.

In this study, where 20 percent of all patients had previous treatment before receiving lapatinib, there were no objective responses. However, 31 percent of all patients receiving lapatinib had stable disease; 8 percent had stable disease lasting longer than six months.

The most common toxicities were diarrhea (69 percent) and nausea (54 percent).Three patients had more severe toxicities including diarrhea, rash and acute renal failure. Researchers found no evidence of cardiac dysfunction. Side effects were considered tolerable.

A single-institution prospective study evaluating the functional diffusion map (fDM) as an early imaging biomarker for overall survival in high-grade glioma: Abstract LB248

Researchers have found that functional diffusion mapping (fDM), which assesses early changes in tumors by diffusion MRI, can provide an earlier assessment of response for patients with glioma, a notoriously hard-to-treat cancer of the central nervous system.

“Diffusion MRI does not necessarily make a better picture of the tumor, but it may be able to assess the response of the tumor to therapy much faster than traditional MRI images,” said Brian D. Ross, Ph.D., co-director for the Center for Molecular Imaging at the University of Michigan. “In addition, by combining diffusion MRI and conventional MRI, one can get a better measurement of results than could be achieved with either test alone.”

Diffusion MRI is a type of magnetic resonance imaging scan that can be performed on any MRI system and which allows for the measurement of the movement of water within body tissue. It is commonly used to evaluate brain injury in stroke patients, and is gaining clinical popularity as a method for distinguishing different types of tissues or tumors.

In this study, researchers assessed brain tumor response by diffusion MRI in 60 patients with high grade glioma who were undergoing radiation therapy. The patients were assessed at one, three and 10 weeks after the start of radiation treatment. Researchers found measurable changes in diffusion MRI as assessed by fDM as early as the first week of treatment. Assessment at week three of therapy was a strong predictor of survival at one year. The strongest predictor was a combination of diffusion MRI and conventional MRI at week 10.

Ross says diffusion MRI is a better predictor of outcome at early stages of disease because it measures changes in cellular density. Researchers believe that a tumor responding to treatment will show decreased cell density, and, as a result, surrounding water will move more freely. Diffusion MRI enables physicians to see this water movement almost immediately instead of waiting the eight to 10 weeks traditionally needed to see if a particular treatment, such as radiation therapy, has shrunk a tumor, or if the tumor has grown.

This is a more accurate measure of tumor response than simply measuring the size of the tumor. “Tumors may appear to get larger or have more contrast enhancement as a response to therapy even if the tumor is not actually growing,” said Ross. “Diffusion MRI may help differentiate which patients are doing well even if the tumor grows.”

Presence of amphiregulin autocrine-loop predicts sensitivity of EGFR wild type cancers to gefitinib and cetuximab: Abstract 4958

Researchers have discovered a biomarker in epidermal growth factor receptor (EGFR) wild-type non-small cell lung cancer (NSCLC) and head and neck squamous cell carcinoma (HNSCC) cell lines. This so-called amphiregulin autocrine loop appears to predict response to the targeted therapies gefitinib and cetuximab.

“Amphiregulin expression could, in fact, be a suitable biomarker to select patients with EGFR wild-type NSCLC who would be likely to benefit from gefitinib or erlotinib,” said lead researcher Kimio Yonesaka, M.D., Ph.D., a researcher at the Dana Farber Cancer Institute in Boston.

EGFR is a therapeutic target in both NSCLC and HNSCC. Patients with NSCLC who present with EGFR mutations are more likely to respond to gefitinib, according to the researchers. However, about 30 percent of patients treated with gefitinib or erlotinib maintain stable disease, defined as no tumor regression and no tumor growth, even without EGFR mutations.

“Most NSCLC patients that develop stable disease with gefitinib or erlotinib therapy do not harbor EGFR mutations,” added co-researcher Pasi A. Jänne, M.D., Ph.D., Assistant professor of medicine at the Dana Farber Cancer Institute. “We have been interested in identifying biomarkers associated with stable disease as this is an important clinical benefit for non-small cell lung cancer patients.”

The presence of amphiregulin varied significantly among the cell lines, ranging from 4.6 to 1,625.8 pg/mL, according to Yonesaka and colleagues. All four EGFR mutant cell lines from the NSCLC samples had negligible levels of amphiregulin. Of note, the researchers detected greater than 250 mg/mL of amphiregulin in seven of the 14 cell lines with wild-type EGFR. These NSCLC and HNSCC cell lines were sensitive to gefitinib and cetuximab. In contrast, cell lines producing less than 250 pg/mL of amphiregulin were resistant to both gefitinib and cetuximab.

Immunohistochemistry of tissue samples revealed that eight of the 10 samples of patients with stable disease following treatment with gefitinib had high amphiregulin expression the researchers report. Only one of 14 samples with high amphiregulin expression was from a patient with progressive disease, suggesting that high amphiregulin expression was associated with the development of stable disease with gefitinib or erlotinib treatment, the researchers said.

The detection and prediction of circulating tumor cells in breast cancer patients: Abstract 3696A

Note: This researcher is not scheduled to participate in a press briefing at the meeting. Interviews can be arranged by contacting Staci Goldberg at 267-646-0616.

Researchers report a new, noninvasive method for measuring circulating tumor cells in patients with breast cancer, information which can be used to predict the likelihood that cancer will spread. The technique detects circulating tumor cells with 100 percent specificity and 88 percent sensitivity, researchers report.

“Metastasis, or the spread of cancer beyond the original site, is the main cause of death in breast cancer,” said Tim Molloy, Ph.D., a post-doctoral fellow at the Netherlands Cancer Institute. “If we can improve ways of measuring risk of metastasis, we can more effectively target therapy and manage these patients.”

Specificity is a statistical calculation that measures the likelihood that a negative result will be associated with the absence of disease. Sensitivity measures the likelihood that a positive result will be associated with disease.

Molloy and colleagues used a quantitative polymerase chain reaction-based detection platform that combined genetic information from four accepted tumor markers into a single score. The higher the score, the greater likelihood of circulating tumor cells.

When researchers applied this test to 131 individuals, an elevated score was observed in 88 percent of patients with metastatic breast cancer, 18 percent of patients with non-metastatic breast cancer and none of the healthy control participants.

After identifying patients whose tumors gave rise to high numbers of circulating tumor cells, a technique called microarray analysis was used to build a genetic profile of their tumors. Microarrays allow researchers to look at thousands of genes simultaneously in a particular cell or tissue to determine which are activated at the time of sampling. From these data Molloy and colleagues were able to build a specific ‘genetic fingerprint’ of breast tumors which may give rise to large numbers of circulating tumor cells.

With this knowledge it was possible to use microarray analysis to predict whether a tumor was likely to disseminate large numbers of tumor cells and therefore metastasize in the future. Testing this on a small independent patient group, researchers found those with a tumor having a genetic profile corresponding to lower levels of circulating tumor cells had a longer time to metastasis at 51.4 months, compared with 29.6 months among those who had a tumor with a genetic profile consistent with higher levels of circulating tumor cells.