Despite significant gains in our understanding of the molecular biology and genetics of cancer, mortality rates for North American women from breast cancer have remained disturbingly constant over the past fifty years. While X-ray mammography is the most effective breast screening method for older women, it is not clearly beneficial for younger women or women with dense breasts. The ability of mammography to diagnose is also particularly unreliable, which is evidenced by the fact that the majority of mammographically indeterminate lesions that progress to surgical biopsy are benign. The development of new non-invasive methods for characterizing indeterminate lesions is an important area of research in medical imaging.

Recent studies have shown that the random motion, or diffusion, of water molecules in tissue is different in benign and malignant tumors, but these studies vary greatly in the measured values of the diffusion characteristics. This lack of agreement may be due to the fact that non-uniformity in the tissue and the flow of blood in microvessels has not been taken into account. The ability to measure the density of blood vessels in tumors may provide an accurate diagnostic method since it is known that tumor progression depends on the ability to stimulate the development of new blood vessels. This process is known as tumor angiogenesis and is required to supply nutrients and oxygen to the expanding tumor. Further, the density of microvessels in "hotspots" in the periphery of invasive breast tumors has been shown to outperform lymph-node positivity as a predictor for the development of distant metastatic disease, and microvessel density is the only prognostic indicator that is a statistically significant predictor of overall survival for node-negative women.

We are investigating the potential for MR diffusion imaging (MRDI) to provide non-invasive, high-resolution maps of the microstructures in and around breast lesions. We predict that this information will allow diagnosis of benign versus malignant tumors based on high resolution maps of the cellular microarchitecture and the amount of vascularization of the tumor. In the case of benign tumors, the maps will provide information regarding the malignant potential. This information will allow improved treatment of benign tumors. Conversely, tumors with high potential to switch to a malignant phenotype may be removed, thus providing an early method of detection of malignant tumors and prevention of metastases. Finally, the information obtained may lead to a better understanding of tumor structure, growth, and angiogenesis, as well as the ability to measure the effects of treatment methods such as anti-angiogenic pharmaceuticals.

To date, we have implemented a spontaneous breast tumour model and have developed a MRDI method on our 1.5T clinical MRI system that measures the diffusion properties of tissue. An MR image of a typical tumour is shown in Figure 1. We have found that there are three unique rates of diffusion in these breast tumours Æ a low diffusion rate, D1, an intermediate rate, D2, and a high rate, D3. Figure 2 shows these diffusion rates pre-cessation (solid symbols) and post-cessation (open symbols) of blood flow versus the factional volume of the tumour that contributes to each diffusion rate. Note that, in most cases, when blood flow has been stopped, the high diffusion rate disappears leaving only two diffusion rates. This suggests that D3 is the rate of blood flow in the tumours.

Figure 1:

Figure 2: