Purpose: Reverse-transcriptase PCR (RT-PCR) assays for carcinoembryonic antigen (CEA) have been described to identify lymph node micrometastases. These assays are not quantitative and can be confounded by false-positive results. The purpose of this study was to determine whether quantification of CEA in lymph nodes could more readily identify clinically relevant groups. Experimental Design: Specimens included 400 lymph nodes from 64 patients undergoing colon resections. Specimens were tested by immunohistochemistry and by RT-PCR using nested primers for CEA. Specimens from 59 patients that were positive by nested RT-PCR were further quantified by detection of CEA mRNA fluorescence increase at a threshold PCR cycle. Results: CEA was detected by nested RT-PCR analysis in 4 of 34 (12%) nodes of nonneoplastic disease, 2 of 13 (15%) nodes from T 1N0 patients, 32 of 81 (40%) nodes of T2N 0 patients, 49 of 109 (45%) nodes from T3N0 patients, and 92 of 163 (56%) nodes from T1-4N1-2 patients. The overall presence of any RT-PCR-detectable CEA in nodes did not differentiate patient groups. Immunohistochemistry was positive in nodes from 7% of T 3N0 patients and 100% of T1-3N1-2 patients. CEA quantification revealed that 0 of 7 patients with nonneoplastic disease and 2 of 17 (12%) patients with stage I T1-2N0 cancers had one or more lymph nodes with ≥1.0 × 102 CEA transcripts per sample. In contrast, 4 of 13 (31%) patients with stage II T 3N0 cancer and 10 of 22 (45%) stage III patients with known metastases had lymph nodes with ≥1.0 × 102 CEA transcripts. Conclusions: These data suggest that quantification of CEA levels in lymph nodes may more accurately identify patients at risk for cancer recurrence than does routine nested RT-PCR or immunohistochemistry.