In this multicenter phase II trial, protracted 5-fluorouracil (5FU) infusion administered in combination with LAR octreotide was an active regimen in the upfront treatment of advanced neuroendocrine carcinomas. This regimen was associated with an overall objective response rate of 24%, a biochemical response rate of 48%, and a symptomatic response rate of 60%. To our knowledge, this is the first trial to test the activity of the combination of octreotide and a chemotherapy drug in somatostatin-naïve patients. Due to the rarity of neuroendocrine tumors, this trial was not randomized and this represents a limitation. Nonetheless, the objective response rate obtained here seems to be higher than that expected with somatostatin alone (0-6%) [4
]. The recently presented results of the first randomized study testing the efficacy of octreotide versus placebo in metastatic well-differentiated neuroendocrine carcinoma showed that octreotide was superior to placebo in prolonging time to progression, despite a comparable response rate [26
]. These data further confirm that the antineoplastic activity of somatastatin analogues is not due to tumor shrinkage but rather to a prolonging of disease stabilization. The potentiality of continuous 5FU in inducing tumor shrinkage could improve the efficacy of systemic treatment, a finding that needs to be assessed in a randomized study.
Due to the potential dedifferentiation of neuroendocrine carcinomas as a consequence of progression, it is expected that response to treatment in metastatic sites may differ from that of primary lesions. Our series was homogeneous in this respect, since 27 out of 29 patients had metastatic disease. Tumor shrinkage (partial response or minimal response) obtained by our regimen made surgery for residual disease (with/without liver chemoembolization or liver radiofrequency ablation) feasible in 3 patients. Surgery for residual disease seemed to be efficacious, since 2 patients were disease free at the last follow-up examination after 27 and 35 months, respectively.
In order to verify study population homogeneity, the tumor samples were centrally reviewed by two pathologists. Central review also allowed the pathologists to assess the biological prognostic parameters. Significant outcome predictors were mitotic count and Ki67 proliferation index, both parameters being reciprocally correlated. Only Ki67, however, maintained an independent prognostic role in the multivariate analysis, confirming the importance of assessing this marker in these tumors [27
]. Ki67 is also an essential parameter in the proposed grading system for foregut neuroendocrine tumors of the stomach, duodenum, and pancreas [29
]. The proposed cutoff of 2% to discriminate G1 from G2 carcinomas was associated with better outcome, confirming a recent report [30
]. Ki67 is also used as predictive marker of treatment response in clinical decision-making algorithms [7
]. It has been suggested that neuroendocrine carcinomas with high Ki67 expression may be potentially sensitive to conventional chemotherapy, whereas tumors with lower Ki67 expression may not [31
]. This trend was confirmed in our series but failed to attain statistical significance.
It has been suggested that metronomic chemotherapy has antiangiogenetic activity [10
]. The activity of metronomic chemotherapy in reducing circulating VEGF levels was observed in a previous study in patients with advanced breast cancer [32
]. In the present study, plasma VEGF levels monitored every 3 months in 11 patients showed a stepwise reduction over time that failed to attain statistical significance due to the small sample size. On an individual basis, however, 7 patients were noted to have a decrease in VEGF and only 2 an increase. Taken together, these data evince an antiangiogenic effect of continuous 5FU plus octreotide treatment.
In a recent Italian multicenter study, a combination of oxaliplatin and metronomic capecitabine (an oral 5FU analogue) was administered to 40 patients with malignant neuroendocrine tumors. Interestingly, a disease response was observed in 23% of patients with high-grade and in 30% of those with low-grade neuroendocrine tumors, respectively [33
]. The introduction of metronomic capecitabine could account for the high response rate obtained in this well-differentiated subset.
Continuous 5FU infusion in this patient subset was well tolerated as most toxicities observed were mild (grade I or II). As expected, dose-limiting toxicities were non hematolological (mucositis, diarrhea, hand-foot syndrome) and delayed treatment in 40% of patients.