A Visit to the fMRI

I’ve heard that for some the experience of undergoing an MRI scan is claustrophobic, but I find it oddly comfy and cocoon-like. OK, there are those gear-grinding screeches and thumps interrupting the music in the earphones. And the cumbersome set-up for imaging breasts, along with the usual admonition to keep perfectly still, does not leave me in a position I would choose for a nap. Still, I’m on the verge of dropping off when the whole table starts shaking under me. They had warned me about this before I climbed onto the apparatus, but for a second I’m not sure whether to laugh (without moving) or jump off.

The fMRI machine I’m visiting is located in the center of campus. I’m taking part in an experiment on breast imaging for cancer detection. This experiment couldn’t be more timely: One scholarly article after another is calling into question the blanket use of current screening methods, as well as for better methods for early breast cancer detection. Indeed, one of the goals of the experiment was to develop a method that could image the breasts of women of all ages and at various hormonal stages, with a completely safe, non-invasive method, while avoiding some of the frequent false positives that plague the usual types of screening.

Dr. Noam Nissan

PhD research student Noam Nissan, MD, who led this study, talked with me afterwards. Noam is supervised by Prof. Hadassa Degani of the Biological Regulation Department and Prof. Lucio Frydman of the Chemical Physics Department. Those two groups converge on the MRI equipment: Degani developed a highly sensitive, MRI-based method of breast cancer diagnosis called 3TP, which is in clinical use, and more recently she also developed the new, completely non-invasive method based on diffusion measurements; while Frydman is a pioneer in developing novel NMR and MRI techniques.

“I knew I wanted to research cancer, for personal reasons as well as scientific ones,” says Noam. “I had contacted Hadassa almost by mistake, but minutes into our first talk I was introduced to her work in breast cancer research, and I immediately recognized that she would be my supervisor, even though, at the time, I knew almost nothing about the methodology of MRI. ”

Prof. Hadassa Degani

For the experiment, volunteers from three different groups were recruited: Young, premenopausal women; older, postmenopausal women; and lactating women. Detection of cancer in each of these groups, Noam explains, presents  a different diagnostic challenge. For the young, premenopausal women, hormonal cycles can get in the way when the current standard MRI method that uses injection of a contrast material is used, thus restricting the scanning to a limited time of the month. Hormone replacement therapy in older women is also problematic for the standard MRI method and it is generally recommended to discontinue it a few months prior the scan, which, again, may lead to delayed diagnosis. There are also problems with using this method in lactating women and pregnant women. In addition, mammography – the standard screening method – is less effective in young women and women treated with hormonal replacement therapy who have unusually dense breast tissue. Although rare, pregnancy-associated breast cancer is not unknown, says Noam and because of the diagnostic challenge with conventional imaging methods, it is often diagnosed late and thus carries a less favorable outcome.

Degani’s new method, Diffusion-Tension Imaging (DTI), does not require any injection of contrast materials, and the experiment showed that it is not sensitive to hormonal changes, so it can potentially be used for all of the above. Part of the method relies on using diffusion gradients in multiple directions, which causes the feeling of shaking in the MRI bed. The other part is computerized analysis of the data. Call it extreme photoshopping: The algorithms they developed begin working immediately after the scanning – while the subject may still be in the fMRI scanner – and provides very fast answers. It consists of analyzing each pixel in the some 60 thin slices of each breast produced in the scan to find any in which the normal, relatively fast diffusion of water is reduced and becomes disordered – a sign of cancer, especially if the neighboring pixels show a similar reduction and disorder. The result is a 3-D, “searchable” image of the breasts that in normal breasts appear in cold colors (purple and blue), while any cancer is highlighted in living color (red and yellow).

The results were published in Radiology, a clinical journal, and the researchers hope that breast DTI will make a relatively quick transition from lab to clinic. According to Noam, DTI can complement the other breast screening methods, giving clinicians a valuable choice in non-standard situations. In the meantime, the lab team is working to apply the method to screening for other forms of cancer, particularly pancreatic cancer, which is often discovered only in the later stages of the disease.

From Radiology Nissan, et al. [0033-8419] 2014

When I asked Noam about the experiment, itself, he told me that I was not alone in my keenness to volunteer. The lab was flooded with calls, including quite a few lactating women. (The experimental setup was not suitable for pregnant women – otherwise they would likely have volunteered as well.) Clearly, he says, there is quite a bit of awareness of the problem as well as willingness to help find a solution.

- Judy Halper