Objectives: Sonography has proven to be a reliable tool in early detection of lymph node and in-transit cutaneous-subcutaneous metastases. Those metastases normally appear as hypoechoic or even anechoic lesions on sonography. It has been assumed that this appearance is due to necrosis of the lesions, but so far, that assumption has never been proven. The purpose of this retrospective study was to evaluate whether the hypoechoic appearance of melanoma metastasis is really due to tumor necrosis.
Methods: From a radiographic database, we retrieved 212 melanoma cases imaged with sonography over a 2-year period for disease staging or follow-up. We selected 37 positive cases with 84 nodal and extranodal (satellite and in-transit) metastatic lesions and reviewed the sonograms and pathologic slides (slides available for 40 of 84 lesions). We retrospectively assessed the vascularization pattern (color Doppler images available for 78 of 84 lesions), categorizing it as poor, intermediate, or consistent. We also looked for necrosis on the histopathologic material, categorizing it into scores of 0, 1, 2, and 3 for absence of necrosis, less than 20% necrosis, 20% to 40% necrosis, and greater than 40% necrosis, respectively.
Results: Despite their gray scale appearance, most melanoma lesions were vascularized on color Doppler imaging and showed limited necrosis at histopathologic analysis. Consistent vascularization on Doppler imaging, excluding substantial necrosis, was found in 44 of 78 lesions (56.4%). Poor vascularization on Doppler imaging, suggesting necrosis, was present in only 14% of the lesions. Substantial necrosis (scores of 2 and 3) was found pathologically in only 10% of the lesions.
Conclusions: Necrosis seems to be an uncommon event in melanoma metastasis and is probably not the basis for its low-level echo pattern on sonography. The hypoechoic appearance is very typical of melanoma metastasis and is likely due to massive melanomatous infiltration (with the poor echo reflectivity of melanin). However, confirmation in larger pathologically proven series is required.