Purpose: To reassess traditional ultrasound descriptors of sentinel lymph node (SLN) metastases, to determine the minimum cross-sectional area (CSA) of an SLN metastasis detectable by ultrasound (US), and to establish whether targeted, high-resolution US of SLNs identified by lymphoscintigraphy before initial melanoma surgery can be used as a substitute for excisional SLN biopsy.
Methods: US was performed on SLNs identified in 871 lymph node fields in 716 patients. SLN biopsy was performed within 24 hours of lymphoscintigraphy and US examination. The CSA of each SLN metastatic deposit was determined sonographically and histologically.
Results: The sensitivity of targeted US in the detection of positive SLNs was 24.3% (95% CI, 19.5% to 28.7%), and the specificity was 96.8% (95% CI, 95.9% to 97.7%). The sensitivity was highest for neck SLNs (45.8%) and improved with greater Breslow thickness. The median histologic CSA of the SLN metastatic deposits was 0.39 mm(2) (12.75 mm(2) for US true-positive results and 0.22 mm(2) for US false-negative results). True-positive, US-detected SLNs had significantly greater CSAs (t test P < .001) than undetected SLN metastases and were more likely to be spherical in cross-section. More than two sonographic descriptors of SLN metastases or rounding of the node alone were factors highly suggestive of a melanoma deposit.
Conclusion: US is not an appropriate substitute for SLN biopsy, but it is of value in preoperative SLN assessment and postoperative monitoring.