2-D and 3-D Ultrasound for Tumor Volume Analysis: A Prospective Study

Ultrasound (US) allows real-time tumor assessment. We evaluated the volumetric limits of 2-D and 3-D US, compared with magnetic resonance imaging (MRI), with a prospective institutional review board-approved clinical evaluation of US-to-MRI volumetric correlation. US images of pre- and post-neoadjuvant breast cancers were obtained. Volume discrepancy was evaluated with the non-parametric Wilcoxon signed-rank test. Expected inter-observer variability <14% was evaluated as relative paired difference (RPD); clinical relevance was gauged with the volumetric standard error of the mean (SEM). For 42 patients, 133 of 170 US examinations were evaluable. For tumors ≤20 cm3, both highly correlated to MRI with RPD within inter-observer variability and Pearson's correlation up to 0.86 (0.80 before and 0.86 after neoadjuvant chemotherapy, respectively). Lesions 20-40 cm3 had US-to-MRI discrepancy within inter-observer variability for 2-D (RPD: 13%), but not 3-D (RPD: 27%) US (SEM: 1.47 cm3 for 2-D, SEM: 2.28 cm3 for 3-D), suggesting clinical utility. Tumors >40 cm3 correlated poorly. Tumor volumes ≤20 cm3 exhibited a good correlation to MRI. Studies of clinical applications are warranted.