Radiology, Vol 190, 77-80, Copyright © 1994 by Radiological Society of North America
Performance testing of transrectal US scanners
EL Madsen, JA Zagzebski, IR Medina and GR Frank
Department of Medical Physics, University of Wisconsin, Madison 53706.
PURPOSE: To test the performance of transrectal ultrasound (US) scanners in
the detection of low-contrast focal lesions as a function of depth, size,
and contrast. MATERIALS AND METHODS: US images of two phantoms were
analyzed by three observers. One phantom was most suitable for US units
allowing scanning in the transverse or longitudinal planes, while the other
was suitable for any scanning situation. Both phantoms were made of
tissue-mimicking material and had six sections. Five sections contained
different sizes of low-contrast simulated lesions, which were of different
echogenicities and were randomly distributed to a depth of 8 cm. The sixth
section contained only background material and served as a reference. A
resolution zone, corresponding to the depth range over which a given type
of lesion was detectable, was determined for each of the five
lesion-containing sections and a figure of merit equal to a weighted sum of
resolution zone lengths computed. RESULTS: Data from 12 different scanner
configurations were analyzed; a broad range of figures of merit were found.
Agreement was found among the three observers regarding relative
performance of the scanner configurations. CONCLUSION: The phantoms are
useful for comparing performances of different transrectal US
configurations.
