In vivo imaging of microvasculature using optical coherence tomography

In vivo imaging technologies drive the development of improved cancer therapies by revealing critical aspects of the complex pathophysiology of solid tumors in small animal models[1]. The abnormal vascular function, which predicts tumor malignant potential and presents broad barriers to effective treatment, has been studied at the subcellular size scale using multiphoton (MP) microscopy [2], and at significantly larger size scales using ultrasound, μCT and μMRI[3-5]. However, limited in vivo imaging approaches exist to study the vascular function at the network level, i.e., with sufficient resolution to discern smaller vessels while maintaining a field of view and penetration depth large enough to reveal interconnectivity and inhomogeneities across the tumor and surrounding tissue. One promising technology operating at this size scale is optical frequency domain imaging (OFDI) using Doppler-methods to detect blood flow. We have recently designed and constructed a Doppler OFDI system specifically for the application of vascular imaging in tumor models[6]. The technical aspects of this system that enable the required levels of flow sensitivity and imaging speed are described. Beam scanning patterns used for this system will be reviewed and analyzed. The construction of the imaging system including high-rate data acquisition with the ability to continuously stream data at rates of 400 MB/sec will be described. Finally, the algorithms used to process, filter, and display the acquired volumetric datasets as vascular projections will be described. To validate the developed Doppler OFDI instrument for this application, its capabilities and limitations were explored relative to those of multiphoton microscopy, the standard optical imaging approach applied to the study of tumor biology. We investigated both the resolution and penetration depth, as well as differences in vascular visibility resulting from the differing mechanisms of contrast (endogenous flow in Doppler OFDI, exogenous fluorescent agent in multiphoton microscopy). Figure 1 illustrates a comparison between Doppler and MP in vivo imaging of a region of normal brain in a mouse cranial window model. Semi-automated vessel tracing algorithms were applied to each dataset, allowing quantitative comparison of visualized vessel sizes. As expected, multiphoton microscopy provides higher resolution, but, as indicated in Fig. 1(e), each modality provides consistent sizing of vessels exceeding 10 microns in diameter. To compare the ability of each modality to image the abnormal vessels within and surrounding tumors, we performed imaging with each modality in a series of tumors in a dorsal skin chamber models. ©2010 IEEE. 59 60 59-60