Shave biopsy (SB) is used for the diagnosis of suspicious skin lesions, including melanoma. Its accuracy for melanoma has not been confirmed. We examined our experience with SB to determine its ability to predict true Breslow depth (BD). We performed a retrospective review of the tumor registry for all patients diagnosed with melanoma by SB from 1995 to 2004. Site and depth of lesion, tumor stage, correlation of BD between SB and wide local excision (WLE), and changes in surgical management due to discordance were examined. Melanoma-in-situ was defined as a depth of 0 for this analysis. One hundred thirty-nine patients were diagnosed with melanoma by SB. Pathology after WLE were as follows: 54 (39%) patients had no residual disease, 67 (48%) had a BD equal to or less than the SB, and 18 (13%) had a thicker BD compared with the SB. For these 18 patients, the median BD by SB and WLE was 1.1 mm (range 0-6.5) and 3.5 mm (range 0.5-20.5), respectively ( P = 0.0017). Upstaging of final BD from SB to WLE was significantly associated with increasing tumor depth and higher stage of melanoma ( P < 0.0001). Only seven of the 139 patients (5%) required further surgery because of the increased depth of the WLE. SB underestimated the final BD of melanoma in 13 per cent of patients, but changed the management of few patients. SB is a valuable tool for practitioners in the diagnosis of melanoma. Nevertheless, patients diagnosed with melanoma by SB should be counseled about the rare need for additional surgery.

The role of lymphadenectomy for gastric carcinoma has been debated for decades. Lymphatic mapping has revolutionized the treatment of melanoma and breast cancer. However, its potential utility in guiding lymphadenectomy for gastric carcinoma is unknown. Therefore we initiated a trial to investigate lymphatic mapping for gastric carcinoma at Wake Forest University Baptist Medical Center. Lymphatic mapping for gastric carcinoma was attempted in 14 cases of gastric carcinoma. Mapping was performed by perilesional injection of isosulfan blue and the first node in the draining basin was harvested and sent fresh to pathology. Sentinel lymph nodes (SLNs) were evaluated by hematoxylin and eosin (H&E) staining. Immunohistochemical analysis was performed on all SLNs that were found to be negative on initial histologic studies. Radical gastrectomy with celiac node dissection was performed in all cases. SLNs were identified in 14 cases. In one case the technique was abandoned because of bulk nodal disease. The average number of SLNs found in each case was 2.8 with a range of one to five. Eight of 14 patients were found to have SLNs positive for metastatic carcinoma. In seven of these patients pathologic analysis of the final resection specimen confirmed the presence of nodal disease. In one case carcinoma was found in a SLN on touch preparation and no nodal disease was noted in the resection specimen. Immunohistochemical studies performed on SLNs found to be negative on initial H&E histologic analysis failed to reveal the presence of carcinoma. The overall sensitivity and specificity were found to be 72.7 and 75 per cent, respectively. Lymphatic mapping is technically possible in the setting of gastric carcinoma and SLNs can be successfully identified in the majority of cases. Upstaging occurred in one case which may have ramifications for adjuvant therapy. SLN positivity accurately predicts the presence of additional nodal disease beyond the SLN in the final resection specimen (positive predictive value 89%). However, SLN negativity does not definitively prove that the remaining nodal basin is free of disease (negative predictive value 50%). Lymphatic mapping for gastric carcinoma is a promising technique worthy of further investigation.