HF-US imaging was performed under aseptic conditions in a laminar flow hood in the nude rat housing room at the DLAR facility. Before each imaging session, rats were anesthetized with a ketamine/xylazine mixture (40/4 mg/kg administered intraperitoneally) and were placed on a sterile pad over a heating blanket. After topical application of 0.5% proparacaine HCl, the superior surface of the right eye was partially exposed using a stainless steel speculum. The hand-held, high-resolution HF-US system for small animal research (model MHF-1; E-Technologies, Inc.) was used to obtain B-scan ultrasound images of the right eye bearing the human choroidal melanoma xenograft. The MHF-1 features a lightweight probe with a 35-MHz transducer (P60-G water path probe; Paradigm Medical, Gardena, CA) and includes a probe attachment (nosepiece) that permits direct contact of a sterile membrane with the tissue. Different nosepieces are available to provide different offsets and allow the scan to be focused at a given depth. The scan angle (sweep) can be changed using the accompanying software. The probe was attached to a manipulator that permitted three-dimensional orthogonal motion of the probe head (E-Tech Manual Baseplate System; ALA Scientific Instruments, Inc., Westbury, NY). In preliminary studies, we found that a sweep angle of 10° and a probe nosepiece with a 6-mm offset yielded the best images of the rat eye. Before each rat was imaged, the membrane was soaked in an activated dialdehyde solution (Cidex OPA; ASP, Irvine, CA) for at least 10 minutes so that it could be disinfected. As suggested by the manufacturer, a value of 1532 m/s was used as the speed of sound in axial dimension calculations. This is in the range of standard values for the speed of sound in biological tissues.
To image the ocular tumor, sterile ultrasound gel (Aquasonic 100; Parker Laboratories, Fairfield, NJ) was placed on the eye, and the membrane-covered probe was positioned over the gel
(Fig. 1) . The probe was oriented so that the cornea was to the left of the B-scan image, and the probe was moved vertically until the sclera and lens were visible. The probe was then moved to the temporal edge of the eye, and an image was captured using the accompanying software. The probe was moved nasally in 250-μm steps using the micromanipulator, and images were captured at each step. We attempted to record at least three such image series across the eye during each imaging session. All images were stored for future analysis.
To evaluate the ability of the high-resolution HF-US system to image tumors in vivo, tumor-bearing right eyes were imaged on day 14 after implantation in 14 rats. The ability to serially image tumors in the same rat using HF-US was tested in another 12 rats, in which tumors were imaged on days 5, 9, and 14 after tumor implantation. At the end of all studies, the rats were humanely killed, and the tumor-bearing right eyes were enucleated, fixed in 10% formalin, and embedded in paraffin for possible future histologic evaluation.