Thank you for the opportunity to respond to the letter entitled “Amaurosis fugax secondary to imaging methods of MRI scan.” We would agree with the authors of this letter, that their method of trying to physically compress the eyes by tightly taping them with underlying pads and overlying tape, places unnecessary and undo pressure on the eyes.
Gently taping the eye tends to put the lids at rest, minimizing lid movement and the motion artifact this induces. Adding the moistened gauze shifts the air-water interface away from the ocular surface, improving image quality for anterior structures in the front of the eye. Neither of these procedures can or should be used in attempt to reduce the movement of the eye via compression of orbital tissues. Tightly taping the eye, or compressing the globe with tightly tapped stacks of gauze, would be expected to produce precisely the visual consequences reported by the letter’s authors We assumed that such a consequence would be obvious to any practitioner, and now emphasize the importance of tapping the lids gently and vertically, only tight enough to obtain closure, and adding the gauze pads only to serve as a means for reducing susceptibility artifact.
To ensure that the eye is not compressed, we routinely examine the scout images for the presence of any visible distortion of the corneal curvature. If the regular curvature of the cornea was altered at all, we remove the subject from the magnet and further loosen both the tape and the gauze. Normal intraocular pressure is only slightly higher than the pressure in the central retinal vein and modest pressure on the eye can begin to reduce arterial flow as well. Misuse of the technique we reported should not negate the value of the technique. It was not the soaked gauze or the tape that caused the transient ischemic accident in this patient, it was the inappropriate application of pressure-something that was absolutely not a part of the technique we have described. To suggest that because the authors compressed the eye too hard is reason to abandon a valuable method is simply inappropriate.
Our technique was designed for high-resolution examination of one eye, with attention to the anterior chamber. The key concept was to use conjugate gaze to steady the eye, not pressure . At least one eye was kept open to fixate on a target. This was the basis for controlling motion, not applying pressure to both eyes by trying to tape them in place.
We will take this opportunity to review the differences in the two techniques. We will include key excerpts from our article and further clarifications.
Our goal in this work was both to define a simple and practical means for reducing the artifacts present in traditional eye MRI [magnetic resonance imaging] and to carefully evaluate this method. Our purpose was to investigate low-cost means of controlling eye and eyelid motion, as well as susceptibility effects based on simple physiologic principles. Motion in the eye occurs from both voluntary and involuntary movements. Fixing one’s gaze on a stationary target controls voluntary movements. However, this requires that the eyes remain open, again increasing susceptibility effects and motion resulting from blinking. If the eyes are closed, or a patient is sleeping, gross involuntary movements can occur.
These two artifacts can be simultaneously controlled, however, if use is made of the fact that gaze is conjugate in the normal population. That is, if one eye is shut while the open eye is fixed on a target, the closed eye will pseudo-track the target. This allows the imaged eye to remain closed, while still allowing the nonimaged eye to fixate on a stationary target.
We did loosely place a piece of tape vertically on the imaged eyelid in two of our four techniques to help reduce blinking artifacts. This loose piece of tape was placed directly on the lid. We simply laid the tape in place from just below the eyebrow to the cheek. The tape that we used was low adhesive in nature so that the subjects could easily peel it off at the end of the experiment.
Visible tonic movements of the lids occur when subjects try to hold their lids closed. These movements are visually imperceptible for the eyelid that is taped shut.
Gauze was placed over the eye and loosely taped into place in a similar fashion, in one of our techniques; the technique that we use. We simply placed a single piece of 1-inch, low-adhesive tape in an oblique fashion from the forehead to the cheek. We were always careful never to place direct pressure on the globe. Many of our subjects stated that the warm gauze pads felt soothing and comfortable, similar to the experience obtained in salons when warm pads are laid loosely on the eyes.
Additionally, to reduce T2* dephasing of the MRI signal from magnetic field susceptibility differences, we added water-soaked gauze pads on top of the eyelid to move the air interface further from the surface of the cornea. Finally, during fixation on a target, involuntary fine saccadic movements still occur and ocular pulsations exist. There are no obvious means of compensating for the latter.
The images in our studies were obtained with surface coils from different manufacturers. We did use a 1-cm loose piece of tape, obliquely placed from the forehead to the cheek, to steady the coil in some cases. The surface coil was braced by the orbit, with no pressure placed on the globe. In other cases (Siemens unit), the temporomandibular joint/eye coils were connected to the head coil unit. We found that the head coil unit did help in providing head, not eye, stability. Some manufacturers have Velcro head straps or pads that are commercially used to help control head motion. We used these commercially available devices to control head, not ocular motion. When securing the head within a head coil, care was always taken never to place straps, tape, or any other support in a position that would place pressure on the globe.
To study simple techniques for reducing the composite effects of motion, susceptibility, and B field drop off, two subjects were imaged on one of the MRI systems… with (a) both eyes open and fixed on a target, (b) one eye closed with the other fixed on a target, (c) one eye closed and taped with the other fixed on a target, and (d) one eye closed, taped, and water-soaked gauze placed over the eye, with the open eye again fixed on a target. Each patient was positioned in a head coil with pads and straps to minimize head motion.
Occasionally there were motion artifacts in our data, from the involuntary motion of the eye. On the few occasions that these motion artifacts were observed, we assumed that the involuntary motion of the eye was primarily along the phase encode direction and attempted to rectify this by switching the phase encode/readout directions. This strategy worked about 80% of the time in reducing motion artifacts from the involuntary motion of the orbit itself. The phase encode direction that we used initially was anteroposterior with the readout along the right-left direction. This configuration of the gradient directions resulted in minimal motion artifacts a majority of the time, implying that, for most cases, involuntary eye motions were primarily along the right-left (readout) direction. This type of motion is consistent with orbit rotation or saccades. For the few cases in which motion artifacts were minimized with the phase encode and readout gradients swapped, we would argue that the primary direction of involuntary eye motion was along the anteroposterior direction, which is consistent with radial pulsations within the eye.