Magnetic Resonance Imaging of Intracranial Vasculopathies

Magnetic Resonance Imaging (MRI) is a non-invasive diagnostic techniques which allow investigators to literally "see inside" the body without surgery or ionizing radiation (i.e., x-rays). The MRI scanner uses a magnetic field and radio waves to produce a picture of the human anatomy. Magnetic resonance imaging techniques are constantly being developed and optimized to increase the speed of acquisition and expand the information content of the resulting images and spectra. Such developments are typically made possible by advances in the machine hardware and computer software. Improved software options of the MR system necessitate evaluation to determine how they can best be applied to research studies employing the equipment. MR software are initially evaluated using inanimate test phantoms. Trials involving human subjects are a necessary step in the development of new MR software since test phantoms are often inadequate for simulating real life conditions. For example, test phantoms cannot be used to visualize subtle changes in tissue contrast or evaluate the effects of voluntary patient movement, breathing, or blood flow. Therefore, limited human subject testing is necessary to evaluate and optimize the performance of the MR software. This will entail varying the MR pulse sequence parameters and assessing specific image quality measures such as tissue contrast, image artifact levels and signal to noise ratio. Investigational software may include pulse sequences and reconstruction software. Pulse sequences are software modules, which control the timing of current through the various coils of the imager. After reception and digitization of the radiofrequency signal, reconstruction software creates digital images, which are displayed on a video monitor. This software adheres to the FDA's definition of non-significant risk.

The aims of this study are to optimize MR imaging and MR angiography sequences and image reconstruction for the 3T magnetic resonance imaging system for subjects with or with suspected CNS vasculitis.

Fifty (50) adult subjects with or with suspected intracranial vasculopathy will be enrolled. These patients will be recruited by subspecialty physicians.

Eligibility will be determined by a screening interview. The treating physician will only introduce the study to the potential subject, and will not try to obtain any consent from them in order to avoid any possible coercion. The treating physician will also try to obtain their approval for being contacted by phone to discuss the study in more detail.

A researcher will later call the potential subjects and explain what the study is about. An IRB approved telephone script will be used in the case that a non-physician makes the telephone call.

Most subjects are expected to be recruited from the clinics. And as described above, these treating doctors from the clinics will only introduce the study and get the subjects approval to be contacted by phone for more detailed discussion. No subjects will be called if prior approval or interest was not obtained.

Since this is a study of the actual MRI images and spectra produced, and not of the intracranial pathology of the subjects, it is sometimes optimal to scan the same subject more than once for direct image comparison. This allows for the best comparison of the image, and eliminates the variability between subjects. Therefore, the investigators will ask subjects if they would like to return for one additional MRI session. They have no obligation to return or be scanned a second time, but will be offered the choice. If the investigators scan the same subject a second times, the investigators will obtain new consent for the second scanning session. Although this study is considered low risk, adult subjects will be limited two MRI scans in total for this study.

MR Imaging - MR images will be acquired using on the 3 Tesla scanners. Prior to scanning the subjects will be asked to remove metal (including jewelry, etc.) from their clothing and person. Subjects will wear earplugs as hearing protection during scanning. Subjects will lie still in the scanner with their head positioned comfortably on a pillow for the duration of the study. Subjects will be in communication with the investigators via two way intercom at all times during the imaging study. Subjects will be encouraged to communicate any feelings of discomfort to the investigator immediately. Periodically during the duration of the MRI exam the investigators will confirm that the subject is comfortable and wishes to continue. The entire imaging session will last 1-2 hours. Multiple anatomical or functional imaging sequences will be run to assess specific image quality measures.

The same anatomical scans will be repeated varying a limited range of pulse sequence parameters to assess their effect on the image contrast, image artifacts and signal to noise ratio. Parameters that will be varied may include the repetition time of the MR acquisition (TR), the echo time of the acquisition (TE), the presence of pre-acquisition inversion pulses, the time interval between the pre-acquisition inversion pulse and the excitation pulse, the spatial profile of the radiofrequency excitation pulse, the spectral profile of the pre-excitation fat saturation pulse, the timing of when the spatial encoding occurs within the repetition time of the sequence, the bandwidth of the signal digitizer, the amplitude of the readout encoding gradient, The shape of the readout gradient waveform, the duration and amplitude of diffusion weighting gradients, the spatial resolution of the MR images, the number of spin-echoes acquired per excitation and the tip angle of the RF excitation. The above parameters will also be evaluated in pulse sequences, which acquire parameters used in the reconstruction of the MR images. The range of these parameters will be maintained within the FDA's non-significant risk criteria as ensured by the software and hardware interlocks that are put in place by the scanner manufacturer. Anatomical imaging sequence may be repeated with different MR coil configurations to compare the performance of these coils. These coils will include volume head coils, transmit and receive surface coil and receive-only phased array coils.

Image Evaluation - Images will be evaluated to determine the effect of pulse sequence parameters and/or hardware configuration on image contrast to noise ratio, signal to noise ratio, and specific image artifacts such as the Nyquist ghost level. The images acquired with different pulse sequence parameters or MR hardware configurations will be compared. Comparison of the MR images entails comparing the result of specific image quality measures such as Signal to Noise Ratio (SNR) and Contrast to Noise Ratio (CNR) as well as the level of image artifacts such as Nyquist Ghost levels. The measurement of several of these quantities has been standardized by the National Electronics Manufacturers Association (NEMA). The NEMA standards have been adopted by FDA as standards for use during the evaluation of premarket submissions for medical devices (see "FDA Modernization Act of 1997: Guidance for the Recognition and Use of Consensus Standards; Availability", Federal Register Vol. 63 No. 37 page 9561). The investigators will use these standard measurements where appropriate.

In cases where a standard measurement is not available or appropriate, the investigators will use standard methods in use in the field. For example, the measurement of Nyquist Ghost levels will entail measuring the mean image intensity in a region of interest containing the ghost artifact but not the image and expressing it as a percent of the mean object intensity across a large Region of Interest (ROI) in the object.

The contrast-to-noise ratio measurement will compare the image intensity difference determined from Regions of Interest drawn in the two types of tissue regions of interest (gray matter and white matter for example) to the mean of the noise in the magnitude image corrected for the magnitude nature of the image using the correction factor described in the NEMA Signal-to-Noise Ratio measurement standard. While there is some subjectivity in this measurements incurred when the researcher places the ROIs, it is not felt to be subjective enough in the eyes of most MR scientists to require a blinded study. Therefore the results of the measurements will be compared by the investigators of the study in an unblinded manner.

In this study, the investigators will vary an MR image acquisition parameter and plot the value of an MR metric (such as SNR, CNR, and image artifact measures described above) as a function of the MR image acquisition parameter value. The image quality metric will be measured at four to ten values of the MR acquisition parameter. For studies such as RF coil comparisons, only a single measurement is required for each RF coil and the comparison requires only the comparison of the image quality metrics or image artifact metrics to determine which coil is more desirable. In cases where a continuous parameter is to be optimized, the range of over which the parameter to be varied will be first determined from what is possible to perform on the scanner by hardware timing and/or amplitude limits and by the safety limits as defined by the FDA's non significant risk criteria. These limits are determined in phantom studies prior to the human studies. The range limits of the human studies will also be narrowed using the information available from literature describing similar studies and knowledge based on models in the literature and the investigators previous studies and experience with the dependency of the MR signal on the sequence parameter in question.

The graph of MR quality metric vs. parameter value will be reviewed by the investigators to determine whether a subset of the parameter values merits a finer search. If the range of the graph exceeds a factor of 5% times the number of points, a finer search will be performed in an attempt to determine the maximum within the 5% accuracy criteria. A standard binary search strategy will be used. If the highest two points are adjacent to one-another, the region between them will be probed with an additional 3-10 measurements. The procedure will deemed to have converged when the range of the data points is less than 5% times the number of data points. If the graph shows evidence of multiple maxima (the highest points are not adjacent to one another), the multiple regions will be explored with additional measurements. Experience indicates that more than 2 maximum are unlikely when optimizing MR acquisition parameters over a range determined with prior knowledge (such as phantom studies or theoretical analysis). If the graph is monotonically increasing or decreasing, the limits will be expanded in the appropriate direction if possible. When the limits can no longer be expanded or a maximum is not revealed upon expansion, then the region near the range limit will be explored in 3-10 additional measurements to verify that the boundary is at the constraint and to determine the potential value of efforts to increase the available limit through hardware development or other tradeoffs within the pulse sequence.