Intraoperative detection of occult colon cancer micrometastases using ¹²⁵I-radiolabeled monoclonal antibody CC49

BACKGROUND. The detection of locally-disseminated disease is one of the principal goals of oncologic surgery. For this study, a hand-held, gamma-detecting probe was used intraoperatively to assess the extent of colorectal carcinoma in patients previously injected with radiolabeled antibody to the TAG-72 antigen (CC49); this technique is known as Radioimmunoguided Surgery (RIGS) (Neoprobe Corporation, Dublin, OH). RIGS-positive areas (i.e. those with increased signal over background) have previously been shown to contain carcinoma in a high proportion of cases. However, some RIGS-positive areas had no tumor detectable by clinical examination or routine histopathologic analysis. This study was undertaken to determine if the presence of occult metastases might account for this disparity. METHODS. A total of 57 regional lymph nodes (LN), resected from 16 patients with primary (9) or recurrent (7) colorectal carcinoma, were studied. The patients were injected with ¹²⁵I labeled CC49 murine monoclonal antibody approximately 3 weeks prior to surgery. After routine histologic evaluation, the LN were analyzed for occult metastases; paraffin sections were cut at 5 levels (50 μm apart) and were examined by histology (hematoxylin and eosin stain [H & E]) and by immunohistochemistry (IHC) with a cocktail of monoclonal antibodies to cytokeratins. RESULTS. Fifty-seven LN were included in this study; 17 were H & E-positive (i.e., contained tumor by routine histologic examination [overt tumor]), while 40 LN were H & E-negative (i.e., no evidence of tumor after routine histologic examination). Thirty-nine LN were RIGS-positive, but only 14 of these were H & E-positive. Of the 25 RIGS-positive/H & E-negative LN, 10 (40%) demonstrated the presence of occult metastases after serial section/IHC analysis. Thus, a total of 27 LN contained metastatic carcinoma (17 overt, 10 occult); routine histologic analysis was able to identify tumor in only 17 of these 27 LN (63%), while the probe signaled the presence of tumor in 24 of these LN (89%). None of the RIGS-negative/H & E-negative LN were found to have occult metastases (0/15). Specific immunoreactivity with CC49 antibody was observed in 5 of 15 RIGS-positive/H & E-negative LN in which no tumor could be identified by any method (histopathology or IHC). CC49 immunoreactivity was not observed in 15 RIGS-negative/H & E-negative LN. CONCLUSIONS. The finding of a RIGS-positive LN had a significant association with the presence of tumor cells (P < 0.05). In this study, the RIGS procedure was more sensitive than clinical or histopathologic examination in detecting the regional spread of a tumor. Furthermore, in LN that showed no evidence of tumor by routine histopathologic examination, a positive RIGS reading was significantly associated with the presence of occult LN metastases (P < 0.01). This study is the first to demonstrate the detection of histologically occult tumor by a remote imaging device. RIGS assessment is a highly sensitive method for detecting occult tumor deposits, and may guide therapeutic intervention in patients with colorectal carcinoma.