High-Resolution Anatomic and Elastographic Transrectal Ultrasound for Improved Diagnosis of Prostate Cancer

Improved imaging can detect prostate cancer earlier and more reliably.

Prostate cancer screening generally uses the Pro state Spe cific Antigen (PSA) blood test, free-PSA testing, and Digital Rectal Exam ination (DRE). When the PSA is used, there exists a significant gray area in which cancers may be missed. Addition ally, DRE is practically limited to the detection of shallow (subcapsular) palpable abnormalities. Even systematic multi-core biopsy fails to detect clinically detectable cancers in up to 34% of men. Thus, there is compelling clinical interest in finding improved detection methods.

Elastographic Experiments show: a) the irregular-shaped lesion mimicking a malignant lesion reconstructed from the elastograms, and b) the irregularshaped lesion inclusion reconstructed from sonograms. The shape of the irregular shaped inclusion, when reconstructed from the elastograms, has blunt edges and appears larger in size than that reconstructed from the sonogram or B-mode image.
The work conducted as part of this program can be divided into the following key goals:

  1. Research, design, development, and prototype testing of a new transrectal ultrasound transducer, syringe pump, and ultrasound instrumentation to facilitate a Synthetic Digital Rectal Examination (SDRE). A high-frequency (8-14 MHz) transducer array was designed. This specialized ultrasound transducer has two tracking arrays, each with 32 elements, a central imaging array with 192 elements, and the elements are spaced on a 0.2 mm pitch. This transducer is operable at up to 14 MHz, whereas the previously available transducer was only operable up to 8 MHz.
  2. Research, development, and prototype testing of techniques to enable quantitative (dimensionally accurate) 3D reconstructions of the prostate. An apparatus was assembled based on strain imaging. Using internally made phantoms, the design was iterated efficiently, and the replacement phantoms were fabricated quickly.
  3. Research, development, and test of techniques to improve ultrasound image quality and to facilitate automated (or semi-automated) border detection of lesions. An accurate 3D surface rendering was developed from 2D slices by implementing a 3D gradient vector flow (GVF) snake algorithm. The method is a stochastically driven compression filter called the “squeeze box filter” (SBF).

After the 3D surface of the prostate is segmented, the volume is determined by applying a novel blobbing technique, called the multi-directional connected component analysis (MDCCA). The volume in units of voxels of the object enclosed by the 3D surface is attained by summing the binary 3D data of MDCCA of non-intersecting slices. The volume is converted to units of cubic centimeter (or millimeter or other units) by multiplying with the voxel resolution. Since volumes can be measured directly, rather than extrapolating volume from a length dimension or cross-sectional area, the image contributions are well-matched and complement contributions in 3D and elastographic imaging.

This work was done by Dr. John A. Hossack of the University of Virginia for the U.S. Army Medical Research and Materiel Command. ARL-0065



This Brief includes a Technical Support Package (TSP).
Document cover
High Res Anatomic & Elastographic Transrectal Ultrasound for Improved Diagnosis of Prostate Cancer

(reference ARL-0065) is currently available for download from the TSP library.

Don't have an account? Sign up here.