Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging

The overall goal of the following experiment is to track the microstructure of the breast to detect breast cancer using a completely non-invasive and safe imaging method. This is achieved by taking MRI images of both breasts and measuring in the entire breast fibro glandular tissue. The water diffusion tensor to reveal the normal and pathological breast micro structural and cellular features.

As a second step, the diffusion tensor imaging data sets are processed using a symmetric tensor model in order to calculate the directional diffusion coefficients and the anisotropy indices pixel by pixel. Next vector and color coded parametric maps of the directional diffusion coefficients. Their average apparent diffusion coefficient and the fractional and maximal anti isotropy indices are produced to identify the best parameters for detecting malignancy and examining pathology.

The results show that the diffusion tensor parameters can reveal the breast architectural features and detect malignant growth by finding map regions with reduced diffusion coefficient values and reduced maximal anisotropy values compared to normal tissue. The new diffusion tensor imaging method showed a breast cancer diagnostic accuracy equivalent to the standard dynamic contrast enhanced three time point method. The main advantage of this technique over existing MRI methods like dynamic contrast enhanced MRI is that DTI does not require injections of an external contrast agent, and therefore it is completely safe and non-invasive.

It is also a relatively fast method, and his clinical performance is not affected by hormonal changes. This method detects increased cellularity due to cancer cells proliferation, and can help answer key questions in breast imaging, including detection, diagnosis and therapy, monitoring of breast cancer, as well as imaging of other organ with dactyl glandular components such as the prostate and the kidney. Generally, individuals new to this method will struggle because breast DTI requires optimization of the scanning protocol through fine tuning of all the experimental parameters, as well as advanced processing software of the raw data sets that yield clear parametric maps of the diffusion tensor parameters We felt had the idea for this method when we came across the drawing of the breast ductal system book on the anatomy of breast from 1840 and the recent computer resistance with dimensional reconstruction of the Mamm Reductor system in the reports of Atkin, colleagues from 2001 and going in moat from 2004.

Visual demonstration of this method is critical as the use and implication of this method is based on color coded parametric maps that enable evaluation of breast tissue and detection of cancer. The procedure will be demonstrated by Dr.Edna Fran, by Dr.Mara Shapiro Feinberg, Dr.Noam Nisan, Mr.Dove Gogel, and our two excellent MRI technician, Ms.Fania Tao and Mr.Holster. The new breast diffusion tensor imaging method is evaluated in comparison to the current standard dynamic contrast enhanced breast MRI method and therefore begin with inserting an intravenous catheter in either arm of the patient to deliver the contrast agent then position the patient for the breast scan.

The patient must be lying in the prone position with both breasts hanging freely in the bilateral openings of the breast coil. The patient should also be resting their head and neck on a pillow for comfort. Now have the patient extend both arms overhead and check the breast positioning.

Again, ensure that each breast is centrally positioned, hanging loose and as deeply as possible within the coil opening. Now, connect the automated injection pump to the intravenous catheter, move the patient into the scanner and proceed with scanning. Using a pilot image, localize both breasts and determine the field of view followed by the number of slices needed and slice thickness needed to completely cover both breasts with extension to the axilla and to the chest wall.

Keep these three experimental parameters consistent in all the scanning sequences. Next, using the interface computer of the scanner, locate a region that includes both breasts and the axilla to define a shimming box. Then apply an iterative shimming strategy to optimize the magnetic field while observing the signal in the proton spectrum.

Adjust the shimming by centering the frequency on the water resonance and then on the fat resonance. By separating the fat and water signals, optimize the signal intensity and shape the irradiation frequency must be centered on the water resonance frequency. For this to work now apply a 2D transversal T two weighted turbo spin echo multis slice sequence at high spatial resolution without fat saturation using grappa with an acceleration factor of two.

Set the variable parameters according to table one. Next, perform fat suppressed spin echo DTI. Using the twice refocused echo planar imaging sequence, again, apply grpa with an acceleration factor of two.

For the spin echo DTI use the following parameters, the most critical are highlighted in purple. They include the echo time te, the spatial resolution, the number of diffusion gradient directions, and the gradients B values. Now apply a field mapping sequence to correct geometric distortions in echo planer imaging.

Obtain phase differential images as described by desert and balaban, ensure that the imaging sequence includes 2D transversal gradient echo images with two different in phase echo times and ensure that the phase and coding direction is the same as in the DTI sequence. Table one lists the values to use for this imaging sequence as well. The next step is to apply a dynamic contrast enhanced protocol using a 3D fast gradient echo sequence without fat suppression and with parameters optimized according to the three time point method.

Again, consult table one for the parameter values. 15 seconds before the end of the second pre contrast image acquisition. Inject with the computer controlled automatic injector, the contrast agent at two milliliters per second.

Follow this injection with automatic injection of 20 milliliters of saline as a flush delivered at the same speed. Now continue recording seven sequential 3D data sets at seven fixed time points post-injection. Next, create a 2D transversal T two weighted fat suppressed turbo spin echo multis slice sequence using experimental details similar to those used for T two weighted imaging.

The overall examination time is 30.5 minutes with the duration of the DTI sequence being six minutes, begin processing the images by first transferring the entire dataset to a remote workstation or personal computer. Then open the data with programs devoted to analyzing breast diffusion, tensor imaging, and D-C-E-M-R-I. Here a home-built software package is put to use with all the steps for DTI image processing.

First, evaluate the noise level outside of the breast and the remaining tissue in three or four regions of interest. Each around 100 pixels. Find the maximum noise level for the pixels above the noise level.

Use the 30 gradient directions to calculate the six diffusion coefficients of the symmetric tensor D in all 60 slices. Use the staal tanner equation and a nonlinear regression fitting program. Then diagonalize the symmetric diffusion tensor in each pixel using principle component analysis.

This process yields three iGen vectors per pixel, which define the diffusion direction in the three orthogonal AEs of an ellipsoid shape. These values coincide with the diffusion frame of the tissue and the corresponding diffusion eigen values that determine the directional diffusion coefficients. Next, for each pixel, calculate the apparent diffusion coefficient, which is the average of the three egen values.

Then calculate the maximal absolute anti isotropy index, which is the difference in the egen values of Lambda one and Lambda three. Now determine the fractional anti isotropy index. For each pixel.

The values range from zero indicating isotropic diffusion to one indicating free diffusion in one direction. With all these calculations performed, use the graphing abilities of the software to make a vector map of the prime or first eigen vector V one. Also create a map with three colors showing the main directions of V one and overlay them on a T two weighted image of the same slice back in the DTI software construct diffusion tensor imaging, parametric maps that display the values of all the diffusion tensor parameters for every pixel in every slice.

Then overlay those values on the T two weighted image of the same slice. The method was tested and demonstrated a young healthy volunteer. The relatively high fraction of fibro glandular tissue can be clearly seen on the T two weighted image as gray areas.

The bright areas are fat. The direction of the prime diffusion coefficient Lambda one shows a large fraction of pixels pointing towards the nipple. As expected, the values of the diffusion tensor coefficients declined from Lambda one to Lambda two to Lambda three.

Using these three diffusion coefficients enabled calculation of the average diffusivity, A DC, the fractional anti atropy, and the maximal anti atropy, to which the fractional anti atropy was highly congruent. 68 patients with breast malignancies were scanned using the DTI protocol. The parametric maps of the highest diffusion coefficient, Lambda one and of maximal anisotropy Lambda one minus Lambda three were found to be significantly lower in the cancer tissue as compared to the normal breast tissue and the most effective for the detection of breast cancer.

In these patients, the regions of low values of Lambda one and Lambda one minus Lambda three indicating can paralleled the regions that were identified as cancer with dynamic contrast enhanced MRI visualized by the color coding of the three time point method, the second pair of Lambda one and Lambda one minus Lambda three images shows a cancer in a patient who subsequently underwent neoadjuvant chemotherapy treatment. After successful neoadjuvant chemotherapy before surgery DTI was able to characterize the tissue's response. A significant increase had occurred in the diffusion coefficients reflecting the presence of reparative connective tissue replacing the cancer cells Once mastered the sequencing and imaging processing of the diffusion tensor imaging data sets can be performed in minutes yielding diffusion parametric maps throughout the entire breast tissue, which enables the detection of malignancy with a high sensitivity.

While attempting this procedure, it's important to remember to optimize the MRI scanner, field homogeneity, and adhere to the DTI protocol and imaging processing specifications. Following this procedure, other MRI methods like dynamic contrast enhancement can be performed in order to verify the DTI results and increase the accuracy of breast cancer diagnosis After its development. This technique may pave the way for researchers in exploring developmental aspects and the hormonal regulation of the normal breast tissue and characterizing early changes in the breast leading to the development of malignancy.

After watching this video, you should have a good understanding of how to adapt and implement to your advanced MRI scanner, the acquisition and image processing tools we have described, as well as to evaluate breast diffuse tenal parametric maps and identify breast malignancy.