The efficacy of several predictive markers previously thought to determine which patients might respond to temozolomide-based therapy could not be validated for patients with pancreatic neuroendocrine tumors.
Mauro Cives, MD
Treatment with capecitabine and temozolomide (CAPTEM) is an effective therapy for patients with metastatic pancreatic neuroendocrine tumors (pNETs), but the efficacy of several predictive markers previously thought to determine which patients might respond to the regimen could not be validated, according to a presentation at the 2015 NANETS Symposium.
“In our study we tried to assess the role of potential predictors of response to temozolomide-based regimens in metastatic pNETs,” according to lead author Mauro Cives, MD, who presented the findings at the 2015 North American Neuroendocrine Tumor Society symposium. The markers the investigators evaluated were expression of MGMT, proliferative activity, and ALT activation.
Investigators looked at 143 patients with metastatic pNETs treated at Moffitt who had undergone therapy with CAPTEM and retrospectively evaluated them for radiographic response. “This is the largest reported cohort of pNET patients treated with temozolomide-based chemotherapy,” said Cives, a research associate in the Department of GI Oncology at the Moffitt Cancer Center and Research Institute in Tampa, Florida.
The investigators observed a partial response in 54% of patients, with 35% experiencing stable disease. The median progression-free survival time was 17 months (95% CI, 15-25 months), and 78% of patients had some degree of tumor shrinkage on waterfall plot analysis. Cives reported that median overall survival was 73.2 months (95% CI, 51.9-81.1 months). Thrombocytopenia was the most common grade 3/4 toxicity.
Researchers also aimed to determine whether biomarkers might be used to determine which patients with pNETs were more likely to respond to treatment with CAPTEM. Temozolomide is an alkylating agent that was developed as an oral alternative to dacarbazine. It is cytotoxic even to slow-growing cells such as neuroendocrine tumors, Cives noted, and acts by inducing DNA breaks.
Some small retrospective studies have linked the MGMT protein, which repairs DNA mismatch errors, to the therapeutic success of temozolomide treatment in cancer, with low levels corresponding to a better response. In addition, in one study, about half of all pNETs were found to be deficient in MGMT, which suggested that MGMT status might be a good biomarker for these tumors.
In the current study, immunohistochemistry (IHC) was used to retrospectively evaluate MGMT expression. Tissue was available for staining in 65 patients. The results of MGMT expression were correlated to patient response to treatment and Cives said that they failed to predict responses for CAPTEM for metastatic pNETS.
The investigators also evaluated tumor proliferation as a predictor of response, as measured by mitotic count, expression of Ki-67 based on IHC, and tumor grade. Here, too, the investigators found that proliferative activity did not predict response to treatment with CAPTEM, although they noted that proliferative activity did vary by time and location.
Finally, the investigators evaluated ALT activation as a prognostic factor for response to CAPTEM. ALT has also been evaluated as a prognostic marker in the past, but its predictive value is unknown. FISH was used to evaluate ALT activation in 128 specimens. When the research team correlated the results of FISH with the patient outcomes, they did not find it predictive of response to CAPTEM. However, they did find a high degree of correlation between loss of DAXX and ALT activation, suggesting that it may be a positive prognostic factor in the metastatic setting.
The investigators also evaluated the status of frequently altered genes in pNETS and correlated them to response to CAPTEM, but have not yet reported those findings.
Cives M, Ghayouri M, Brelsford, et al. Identification of response predictors to capecitabine/temozolomide in metastatic pancreatic neuroendocrine tumors. Presented at: 2015 NANETS Symposium; October 15-17; Austin, TX. Abstract C13.