To the editor, We read with great interest the study by She-Juan An et al., which showed that posttreatment plasma VEGF levels associated with the overall survival of patients with advanced non-small-cell lung cancer treated with bevacizumab plus chemotherapy. In this study, before and 6 weeks after treatment, vascular endothelial growth factor (VEGF) levels measured. Also, this study showed that low posttreatment VEGF levels associated with better overall survival (OS) [1]. We have some insights for this decrease in serum VEGF levels. VEGF circulating in the blood of patients with cancer may be derived from various sources, as depicted in Fig. 1. Circulating levels of VEGF are dependent on the amount released from tumor cells (Fig. 1A) and/or platelets activated at the tumor-vessel endothelium. Activated platelets release VEGF (Fig. 1B), and platelet consumption causes increased levels of thrombopoietin, which in turn stimulates megakaryopoiesis and platelet production in the bone marrow. Newly produced platelets are larger in size than those produced during steady-state thrombopoiesis, and therefore contain an elevated amount of VEGF (Fig. 1C). Finally, host immune cells as macrophages that infiltrate tumor tissues can be an additional source of VEGF (Fig. 1 D) [2]. Serum VEGF levels have also been found to correlate with platelet count in a mixed population of metastatic cancers and renal cancer, and some investigators have also stated that VEGF should be corrected for platelet counts. To correct for variation in platelet counts between patients, it has been proposed that the concentration of VEGF per platelet (pg/10) can be calculated by dividing the serum VEGF concentration (pg/ml) by the platelet count (910/ml) [3]. In the study, She-Juan An et al., mean VEGF levels decreased from 81.5 pg/ml to 50.5 pg/ml after 6 weeks of treatment (P = 0.08), with a mean reduction of 31 pg/ml in overall. The relationship of the plasma VEGF levels with OS and progression-free survival (PFS), the patients were divided in the two groups based on their median VEGF value. The OS differed significantly between the low and high posttreatment VEGF levels (25.6 months vs. 13.4 months, P = 0.0284). No other relationship was found with plasma VEGF levels. It is known that most cytotoxic chemotherapeutics decrease the platelet count. In this study, thrombocytopenia was developed in 27.3% patients treated with bevacizumab plus chemotherapy. The association between VEGF levels and thrombocytopenia was not reported. Also, it is possible that platelet counts decreased with chemotherapy, but remained within normal limits. Because platelet counts before and after therapy has not been reported, the possible contribution of platelets on the decreasing VEGF levels remains obscure. We believe that the decrease in serum VEGF levels may at least in part be due to decreased platelet counts secondary to chemotherapy, and plasma VEGF levels may not accurately reflect the truth all the time before correcting the VEGF levels for platelet counts. M. A. N. Şendur S. Aksoy (&) Ş. Yaman Z. Arik N. Y. Ozdemir N. Zengin Department of Medical Oncology, Ankara Numune Education and Research Hospital, 06100 Sihhiye, Ankara, Turkey e-mail: saksoy07@yahoo.com