Purpose: The aims of the present work are to assess the diagnostic accuracy of high resolution color Doppler ultrasound (US) of the sentinel node (SN) in patients with cutaneous melanoma skin at stage I. The US findings of nodal involvement could spare the patient a surgical step (selective lymphaderectomy) allowing them to undergo radical lymphadenectomy directly.
Material and methods: From November 1998 to November 2000 94 patients (mean age 52.7 years) underwent lymphoscintigraphy in order to mark the SN site on their skin. An US scan (112 lymphatic basins) was performed within 24 hours with a 10-13 MHz electronic linear probe with color-power-Doppler (Esaote AU5 Idea Scanner, Genoa, Italy). The sonographic features we analysed were: shape (roundness index), hilum displacement, intranodal heterogenicity, eccentric cortical thickness, extranodal invasion, vessel irregularity.
Results: 26 nodes showed US findings consistent with malignant involvement, 86 were negative. All the nodes were surgically removed and controlled by histology. Sensitivity and specificity of US scanning were 89.4% and 90.3%, the positive and negative predictive values 65.3% and 97.6%, respectively. US correctly identified the involved SN in 15,1% cases, so that 17 patients could have avoided the selective lymphadenectomy
Conclusion: Preoperative lymphoscintigraphy and high-resolution color-Doppler US scanning constitute a useful diagnostic tool in identifying the metastatic SN, with a low margin of error. False negatives were technically induced, even using the more recent scanners, by the low US probe resolution, unable to recognise metastatic microdeposits. The two most reliable parameters in identifying involved lymphnodes were the roundness index and the absence of hilar echo. The advent of technologically more advanced probes should allow better spatial resolution and assessment of lymph node vascularization, enabling diagnosis of metastasis measuring less than 2 mm in diameter.