A prospective study of false-positive diagnosis of micrometastatic cells in the sentinel lymph nodes in colorectal cancer.

Abstract

Sentinel lymph node mapping (SLNM) with multilevel sections (MLS) and cytokeratin immunohistochemistry (CK-IHC) of sentinel lymph nodes (SLNs) upstages 15-20% of patients (pts). False-positive SLNs occur in breast cancer due to mechanical transport of cells during mapping procedures, or to pre-existing benign cellular inclusions. Our prospective study evaluated whether colorectal mapping procedures alone caused false positives. A total of 314 pts underwent SLNM with blue dye. Ninety of the pts underwent a second mapping in normal bowel away from the primary tumor. The first 1-5 blue nodes near the primary tumor were marked as SLNs; those near the second injection site were marked as nontumor SLNs (nt-SLNs). All SLNs and nt-SLNs were evaluated by MLS and CK-IHC. Of 314 pts, 30 had benign tumor and 284 had invasive cancer. SLNM was successful in 274/284 (96.5%) invasive cancer pts, with 728 SLNs identified. Forty-six of the 274 pts (16.8%) had low-volume metastasis in 57 SLNs: 31 pts (11.3%) had 38 SLNs with micrometastasis (>0.2 mm, <or=2 mm), while 15 pts (5.5%) had 19 SLNs with isolated tumor cells (<or=0.2 mm). For 100 pts with second SLNM (70/90 pts successfully mapped with 102 nt-SLNs), or with SLNM of benign pathology (30/30 pts successfully mapped with 88 SLNs), there were no false positives in any of 190 nodes (P < 0.001). No false positives due to mechanical transport of cells or to benign cellular inclusions were identified in 190 lymph nodes from 100 patients with SLNM in benign bowel.