Colorectal cancer (CRC) remains the second deadliest cancer in the United States, despite steady reduction in mortality rate over the last three decades. Colonoscopy is the gold-standard screening modality with high sensitivity and specificity to mature polyps. However, the miss rate for small (<5 mm) lesions is estimated to be as high as 26%. Because the five-year survival rate for CRC detected at the local stage is 90%, there is a clear need for a screening procedure that is sensitive to these small lesions. Optical coherence tomography (OCT) has become a major biomedical imaging modality since its invention in 1991. As the optical analog to ultrasound, OCT provides information in both lateral and depth dimensions with resolution < 10 µm and an imaging depth of about 1.5 mm in scattering tissue. In this dissertation, I describe my efforts to develop new uses of OCT for improved early detection of adenoma in the azoxymethane mouse model of CRC. In recent years, commercial OCT systems have reached imaging speeds sufficiently high for in vivo volumeric imaging while laterally sampling the tissue at the Nyquist limit. First, I describe the design of a miniature endoscope and the integration of this probe with a commercial OCT system. Then I describe the development of two OCT imaging methods, one structural and one functional, that could be used for future work in diagnostic or therapeutic studies. The structural method produces en face images of the colon surface showing the colonic crypts, the first such demonstration of crypt visualization in the mouse. Changes in the crypt pattern are correlated with adenoma and are one of the earliest morphological changes. The functional method uses a Doppler OCT algorithm and image processing to detect the colon microvasculature. This technique can be used for vessel counting and blood flow measurements. Angiogenesis occurs at the beginning of tumorigenesis, and the tumor-originated arterioles are incapable of regular vasodilation. This Doppler OCT technique could potentially detect tumors at the earliest stages by measuring the change in local blood flow velocity in response to vasodilatory stimuli.