This study assessed the accuracy of volumetric measurements acquired with ultrasound tomography (UST). UST imaging generates 3D speed-of-sound maps that can identify tissue types and measure lesion volumes. Since tumor volume doubling time is associated with growth rate and tumor biology, accurate measurement of tumor volume is critical for oncologic diagnosis, staging, and treatment.

Six cylindrical agar phantoms were imaged using UST and hand-held ultrasound (HHUS). Each phantom contained 4 embedded "lesions" composed of irregular-shaped chicken breast with known volumes ranging from 1.3 cm3 to 7.4 cm3. Two board-certified breast imaging fellowship-trained radiologists independently performed blind interpretations of the UST and HHUS phantom images and calculated 24 lesion volumes. UST volumes were calculated with automated segmentation software (QT Ultrasound, Novato, CA). HHUS lesions were measured in 3 dimensions (a, b, c) and their volumes were calculated using 2 volume formulas: (1) (4/3)πr3 (with r = average of a, b, and c); and (2) (π/6)abc. These calculations were then statistically analyzed to determine the volumetric measurement accuracy of both UST and HHUS as compared to known true volumes calculated by water displacement methods.

The average lesion volume calculated from UST images was 3.93 1.55 cm3 and from HHUS was 6.39 2.61 cm3 (sphere) or 5.65 2.39 cm3 (ellipsoid), compared to the true average of 3.98 1.47 cm3. HHUS volumes were significantly larger than the true volumes with a mean over-estimation of 62.8% 36.9% cm3 (sphere) or 43.3% 32.6% cm3 (ellipsoid), whereas UST volumes agreed with the truth within measurement errors. Interobserver agreement was substantial (ICC = 0.95).

This study demonstrates that UST can accurately measure the volume of irregular-shaped masses, with superior accuracy than HHUS.

Ultrasound tomography can accurately measure tumor volume, demonstrating its potential utility in guiding oncologic management and treatment.