Angiogenesis is a dynamic process essential for tumor growth. Maeda et al. first reported its association with poor prognosis in gastric cancer in 1996 and, since then, there have been numerous similar studies examining various factors that mediate angiogenesis such as vascular endothelial growth factor (VEGF) and platelet-derived growth factor [1]. It is therefore surprising that a decade and a half later there is still no clear consensus on the role of antiangiogenic agents in the treatment of gastroesophageal cancer and, more importantly, recognition of the patient subset who are most likely to benefit from these expensive agents. In this issue of the Journal, Liu et al. [2] report that VEGF expression is an independent prognostic factor for tumor recurrence and decreased survival in patients with node-negative gastric cancer. Unlike prior studies, the authors excluded all node-positive patients in the hopes of stratifying node-negative patients into lowand high-risk groups for tumor recurrence. This is an important step in identifying patients most likely to benefit from adjuvant therapy, but the manner in which patients were deemed node-negative is particularly relevant. What was the median number of nodes resected, and how were the nodes examined? It seems unusual that 11 patients with pT4 tumors were node-negative. If we presume all nodes were examined using a single section and conventional [hematoxylin-eosin (H&E)] staining, then this means that only about 1% of the lymph node tissue was examined, and some node-positive patients almost certainly were missed. In a retrospective study last year, we found that almost one-third of esophageal cancer patients who were node-negative by conventional histology proved to have occult tumor deposits in their lymph nodes after serial sectioning and immunohistochemical evaluation [3]. In other words, until standardized and more stringent histopathology is performed on lymph nodes, it is difficult for the reader to be confident that all patients were truly nodenegative. Also, in Liu et al.’s study, it is possible that 24 of 55 (44%) patients who were VEGF-positive and had a worse survival were those who might have been nodepositive with more stringent histopathologic testing. Nevertheless, their study raises an interesting question. Should patients with early gastric cancer and positive VEGF staining be given an antiangiogenic agent to reduce their risk of tumor recurrence? Bevacizumab (Avastin), a recombinant humanized monoclonal antibody directed against VEGF-A, was approved for patients with metastatic colorectal cancer in 2004, in combination with 5-fluorouracil chemotherapy, after a Phase III (randomized controlled multicenter) study showed an increase in survival of 4.7 months with the inclusion of this agent in the treatment regimen [4]. This study paved the way for a myriad of other clinical trials evaluating more than 40 antiangiogenesis agents in various solid tumors. In gastric cancer, a Phase III trial called AVAGAST (Avastin in Gastric Cancer) has just been published in the Journal of Clinical Oncology on the back of three Phase II (nonrandomized) trials evaluating the addition of bevacizumab to chemotherapy in patients with unresectable or metastatic gastroesophageal cancer [5]. Although the smaller Phase II trials reported response rates of 65 to 68% with encouraging progression-free survival [6], the Phase III trial found only a 2 month survival benefit between groups, which was not significant [5]. Subsequently, bevacizumab has not been S. K. Thompson (&) Department of Surgery, Royal Adelaide Hospital, University of Adelaide, Level 5, Eleanor Harrald Building, Adelaide, South Australia 5000, Australia e-mail: sarah.thompson@adelaide.edu.au