I. Introduction
Accurate detection of metastases in excised human lymph nodes (LNs) plays a crucial role in proper staging, patient management, and treatment planning. Current histological methods may fail to detect small but clinically significant metastases because only the sampled cross-sectional slices of LNs are microscopically evaluated. In particular, for colorectal and gastric cancers, only the central histological section of each LN is evaluated for metastases regardless of LN size [1]. No clinically available screening method exists for detecting metastases throughout the entire LN volume. Our previous work [1] –[4] has demonstrated the potential of a high-frequency (HF), three-dimensional (3-D), quantitative ultrasound (QUS) method for reliably evaluating the entire LN volume to detect metastases. In these previous studies, radio-frequency (RF) echo signals were acquired from entire LNs using HF ultrasound (HFU) (i.e., 15 MHz) so that tissue microstructures could be assessed. The QUS method consists of two major steps: 3-D segmentation and 3-D QUS-parameter estimation.