Antiangiogenic therapy is the basis of various cancer therapies. Bevacizumab, a recombinant humanized monoclonal antibody against vascular endothelial growth factor (VEGF), has been widely employed in the standard chemotherapy of many solid malignancies such as colorectal cancer (1). Despite extensive research for the detection of biomarkers such as angiogenesis-related biochemical factors, genetic polymorphisms, gene and/or protein expression, and radiographical images, no practical biomarker predicting antitumor efficacy and toxicity has been identified yet (2).

Age-related macular degeneration (AMD) is a common eye disease that usually occurs in people aged 50 and older causing irreversible central vision loss. Diagnostic retinal examination using optical coherence tomography or fluorescein angiography reveals choroidal neovascularization that develop under the retina scarring of the macula. Direct intravitreal injection of anti-VEGF agents, such as ranibizumab, aflibercept, and brolucizumab, is a treatment option for AMD (3). Bevacizumab, commonly administered at a dose of 1.25 mg, is an alternative option causing similar effects on both vision and retinal thickness, despite its off -label use. According to a prospective study of 11 patients with AMD, the retinal arteriolar diameter was decreased after each intravitreal injection of ranibizumab, a smaller fragment (Fab) of bevacizumab probably due to vasoconstriction (4).

Retinal observation is routine practice in ophthalmology. Besides its use in the diagnosis of ophthalmic diseases, it is also known that aberrant alternation of the retinal microvessels reflects systemic disorders, such as atherosclerosis (5), hypertension (6), and diabetes (diabetic retinopathy) (7,8). We, thus, hypothesized that retinal microvessel diameters might decrease after systemic administration of bevacizumab in cancer patients receiving systemic chemotherapy, indicating tumor response to systemic treatment. Direct observation of patient’s retinal microvessels could thus be a biomarker predicting antitumor efficacy and toxicity of angiogenesis inhibitors.

This research was performed in compliance with the Declaration of Helsinki and was approved from the institutional Review Board (No. 1050). Each patient provided written informed consent for participating in the study.

Patients. A total of 70 patients with colorectal cancer that received chemotherapy with or without bevacizumab in an outpatient setting between April 2011 and August 2012 were evaluated. Among these patients, 10 patients received scheduled preoperative chemotherapy consisting of intravenous bevacizumab 7.5 mg/kg and oxaliplatin 130 mg/m² on day one in combination with oral capecitabine 1,000 mg/m² twice daily for two weeks, followed by a one week break, repeated every three weeks.

As a routine practice, patients who had a record of thrombotic diseases, interstitial pneumonia, peptic ulcer, hemoptysis, or other bleeding or coagulation disorders, as well as patients receiving anticoagulants or aspirin were excluded from the bevacizumab receiving arm.

Measurement of retinal microvessel diameters. Retinal photographs, left or right, were obtained using a nonmydriatic retinal camera (AFC-230, Nidek, Gamagori, Japan) before (baseline) and after each intravenous infusion of bevacizumab and oxaliplatin, throughout the therapy period in each patient. Measurements were commonly collected with a total of three treatments from the second to the fourth preoperative courses as they were enrolled before the second course. Retinal photographs were usually collected within one hour after the end of infusion. Based on our preliminary study of six healthy volunteer subjects (three males and three females), there were no significant modifications in retinal microvessel diameters due to daily activities such as walking and eating.

As retinal arteries and veins demonstrated similar decrease in diameters in another preliminary study, retinal veins were examined for the following analyses: diameters of retinal veins were measured at any eight locations within ×1.5-fold of the diameter of the optic disc from the outer edge of the optic disc on retinal photograph using Adobe Photoshop™ version 9 (Adobe Systems Incorporated, San Jose, CA, USA). The retinal vessel diameter on each photograph was standardized based on the diameter of the optic disc. The measurement was repeated by five different investigators to minimize inter-investigator variation.

R software, version 2.15.0 (R Foundation for Statistical Computing, Vienna, Austria) was used for statistical analysis.

No significant difference was observed in the decrease of microvessel diameters between right and left fundi in five patients who received chemotherapy with bevacizumab (15.1±7.2% in right fundus; 16.0±5.7% in left fundus, mean± standard deviation, total 24 records). It was thus decided for each patient whether the left or the right fundus would be examined, but the same fundi were examined before and after chemotherapy. Furthermore, a decrease of retinal microvessel diameter persisted for several hours after the end of bevacizumab and oxaliplatin infusions, and it usually recovered by the initiation of the next chemotherapy.

Retinal microvessels diameters of 10 patients (nine males and one female) with a median age of 49 years (range=29-67 years) decreased 9.9±2.6% (total 54 records) from baseline after the end of bevacizumab and oxaliplatin infusions, and then recovered by the initiation of the subsequent chemotherapy three weeks later, as presented in the preliminary studies (Table I, Figure 1). Among the 10 patients, five achieved a tumor shrinkage of the primary lesion of more than 30% on radiological images, equivalent to partial response defined with RECIST criteria, while the other five did not. Retinal microvessel diameters of the five responders decreased temporarily from baseline by 12.3±5.4% (in a total of 26 records), while that of the five nonresponders decreased by 8.2±5.6% (in a total 28 records, p=0.008, Wilcoxon rank sum test, Figure 2). In contrast, the decrease of retinal microvessel diameters of four patients who obtained moderate histopathological regression (Grade 2, necrosis more than two-third of the lesion, 11) did not differ from that of the other nonresponders (9.1±1.7%, total 22 records vs. 10.4±2.9%, in a total of 32 records). Among the 10 patients, three experienced bevacizumab-associated grade 2 hypertension (increased blood pressure) according to CTCAE version 4.0 during treatment. No significant difference in the change of the retinal microvessel diameters was observed among the three patients (10.6±1.0%, in a total of 16 records) and the other seven patients (9.63±2.9%, total 38 records).

In addition, the findings from another 68-year-old female patient who continued bevacizumab-based chemotherapy for more than one year revealed an association between the decrease of retinal microvessel diameters and the efficacy of chemotherapy (Figure 3). The decrease of retinal microvessel diameters was relatively consistent (2.7±4.0%, mean±standard error, in a total of eight records) while the chemotherapy was effective. On the other hand, the decrease was inconsistent (1.6±9.6%, total four records) while the patient continued receiving the same chemotherapeutic scheme despite clinical disease progression. After switching to another chemotherapy, the decrease on microvessel diameters was superior and consistent (14.4±5.8%, in a total 18 records).

With the direct observation of retina in patients who received bevacizumab-based chemotherapy, the retinal microvessel diameters were observed to be decreased from baseline temporarily, shortly after a systemic infusion of bevacizumab. Although the underlying mechanism of this interesting phenomenon remains unclear (9,10), bevacizumab causes vasoconstriction by decreasing blood flow in the retinal microvessels, thus reducing the diameter of microvessels. As it is believed that angiogenesis inhibitors exert their therapeutic effects by reducing blood flow to the tumor, it is expected that a decrease of retinal microvessel diameter could reflect a response to systemic treatment with bevacizumab-based chemotherapy. In particular, in the present study it was observed that a decrease in microvessel diameters was associated with greater tumor reduction.

Apart from the limited number of patients, this preliminary study contains some further limitations, and the results should thus be interpreted with caution. Initially, the decrease of retinal vessel diameters was not related with histopathological tumor response or toxicity (hypertension) in the present study. If retinal microvessel diameter can accurately reflect response to systemic treatment, it should also be correlated with the histopathological tumor response as well as with toxicity. Moreover, the possibility that this phenomenon is specific to bevacizumab or other angiogenesis inhibitors should not be excluded as oxaliplatin in combination with bevacizumab may also decrease the retinal microvessel diameters. The present study did not examine a sufficient number of patients receiving monotherapy of oxaliplatin or bevacizumab. The ratio of oxaliplatin and bevacizumab contribution to vascular modifications thus remains unclear.

Nevertheless, this study demonstrated that direct observation of retinal microvessels is feasible in patients receiving systemic chemotherapy. As fundus examination is a noninvasive technique in ophthalmology and retinal microvessels are the only vessels in the body that can be directly observed, ophthalmologic assessment of retinal microvessels could be a potential biomarker candidate for the assessment of systemic treatment response upon angiogenesis inhibitors treatment like bevacizumab.

The UK Biobank, which includes 68,550 participants, uses fully-automated image analysis to assess non-mydriatic retinal imaging, blood pressure, and arteriosclerosis index (11). Recently retinal capillary density and caliber have been associated with the severity of diabetic retinopathy derived by optical coherent tomography angiography (OCTA) from a cross-sectional study involving about 600 eyes in the United States (12). Development of easy analytical systems enables assessment of blood pressure or diabetic retinopathy in large numbers in the real world clinics. Such technologies may be utilized to evaluate retinal microvessel diameters in large numbers of cancer patients, which may play an important role in detecting therapeutic effects and side effects of new angiogenesis inhibitors.

Dr. Ando reports grants and personal fees from Chugai Pharmaceutical Co., Ltd., grants and personal fees from Kyowa Kirin Co., Ltd., grants and personal fees from Nippon Kayaku Co., Ltd., grants and personal fees from Yakult Honsha Co., Ltd., grants and personal fees from Ono Pharmaceutical Co., Ltd., grants and personal fees from Taiho Pharmaceutical Co., Ltd., grants and personal fees from Daiichi-Sankyo Co., Ltd., grants and personal fees from Eisai Co., Ltd., personal fees from Eli Lilly Japan K.K., personal fees from Novartis Pharma K.K., personal fees from Bayer Holding Ltd., personal fees from Bristol-Myers Squibb, personal fees from Sawai Pharmaceutical Co., Ltd, personal fees from Roche Diagnostics K.K., personal fees from MSD K.K, personal fees from Astellas Pharma Inc., personal fees from Otsuka Holdings Co., Ltd., personal fees from Sanwa Kagaku Kenkyusho Co., Ltd., personal fees from Hisamitsu Pharmaceutical Co., Inc., and grants from Mochida Pharmaceutical Co., Ltd.

Dr. Nakayama reports personal fees from Chugai Pharmaceutical Co., Ltd., personal fees from Yakult Honsha Co., Ltd., personal fees from Taiho Pharmaceutical Co., Ltd., personal fees from Eli Lilly Japan K.K., personal fees from Miyarinsan Pharmaceutical Co., Ltd., personal fees from Takeda Pharmaceutical Co., Ltd., and research funding to his institution from Eli Lilly Japan K.K. as outside the submitted work. All remaining Authors have declared no conflicts of interest.

Conceptualization, A.M., Y.I., Y.A.; Methodology, A.M., Y.I., T.S., Y.A.; Software, A.M., Y.I.; Validation, A.M., T.S., Y.I.; Formal Analysis, A.M.; Investigation, A.M., Y.I., T.S., C.I., K.U., G.N., Y.A.; Writing – Original Draft Preparation, A.M.; Writing – Review & Editing, A.M., Y.A.; Visualization, A.M., Y.I., Y.A.; Supervision, H.T., Y.A; Project Administration, A.M., Y.A.; Funding Acquisition, Y.A. All Authors have read and agreed to the published version of the manuscript.

This work was supported by JSPS KAKENHI Grant Number 23590638. We would like to thank Keiko Yoshida, Yukiko Kato, Kiyoko Miura, and Yuka Murasaki for their support of this study.