Drs. Fakih and Padamanabhan have comprehensively and succinctly reviewed the role of carcinoembryonic antigen (CEA) monitoring in the management of colorectal cancer. The prognostic value of CEA in colorectal cancer is clear, but its role in detecting recurrence and monitoring advanced diseae is more controversial. With improvements in clinical outcome associated with the complete resection of hepatic and pulmonary recurrences in colorectal cancer, we feel patients should be on some form of surveillance program following primary surgical therapy, despite the lack of definitive prospective data.[1-4] However, the optimal frequency and composition of such a surveillance program remains debatable.
CEA Monitoring in Post-Curative-Intent Surgery Surveillance
Improvement in the resectability of recurrences and clinical outcome has been associated with surveillance programs that include CEA monitoring. This is apparent from a review of multiple sources of data, including three meta-analyses, six randomized trials, and other retrospective studies.[3-11] Only two of the six randomized trials were designed with and without CEA monitoring in the intensive surveillance and control arms, respectively (see Table 1 of Fakih and Padamanabhan's article), and the results of these two trials were conflicting.[7,11] Also, as the authors allude, the direct benefit of CEA monitoring in these studies remains nonquantifiable.
The 2005 update of the American Society of Clinical Oncology (ASCO) Colorectal Cancer Surveillance Practice Guideline recommends intensive surveillance following primary therapy for colorectal cancer. An absolute risk reduction of 7% in death from all causes from three meta-analyses supports the recommendation. ASCO also modified its 1997 recommendation to include annual chest and abdominal computed tomography (CT) scans for 3 years following primary therapy for colorectal cancer in patients who are at risk of recurrence and could be candidates for curative-intent surgery later. Every-3-month CEA monitoring for at least 3 years was also recommended for this group of patients.
One of the corroborating studies was by Chau et al, who performed a secondary analysis of a randomized trial involving stage II and III colorectal cancer patients assigned to bolus fluorouracil(Drug information on fluorouracil) (5-FU)/leucovorin or protracted venous-infusion 5-FU. The main aim of the study was to evaluate the value of routine CEA monitoring and CT scanning. The result was intriguing and hypothesis-generating.
Both arms were followed similarly after chemotherapy, according to a protocol-specific follow-up plan, and were grouped together during analysis. They underwent follow-up monitoring every 3 months for the first year, every 6 months for the second year, and annually thereafter. Serum CEA was measured at baseline and at each clinic visit. CT scans of the thorax, abdomen, and pelvis were performed at baseline, 12 months, and 24 months.
A total of 530 patients were followed for a median of 5.6 years, and 154 patients relapsed; 65 (42%) of those relapses were symptomatic. Among the asymptomatic relapses (n = 89), 45 were detected by CEA and 49 by CT. In 14 patients with relapse detected by CT, concomitant CEA elevations were discovered, and these patients were included in both CT and CEA groups when analyzed. The remaining 9 (6%) cases of relapse were detected by other methods.
The analysis found better survival associated with relapsed patients who proceeded to curative-intent surgery, confirming findings from other studies. Compared to the symptomatic group, a higher proportion of patients in the CT- and CEA-detected groups went on to curative-intent surgery (26.5% and 17.8%, respectively, vs 3.1% in the symptomatic group). However, overall survival from the time of randomization for relapsed patients did not differ significantly between the different detection groups (P = .313). Five-year survival rates were 30.6% among symptomatic patients (95% confidence interval [CI] = 19.9%-42.0%), 22.9% in the CEA-detected group (95% CI = 11.7%-36.3%), and 30.8% in the CT-detected group (95% CI = 17.9%-44.7%).
Several large randomized trials are currently investigating the benefit of more-intensive surveillance programs following curative colorectal cancer surgery. The Gruppo Italiano di Lavoro per la Diagnosi Anticipata (GILDA) trial aims to enroll 1,500 Duke's B or C colorectal cancer patients into the two arms.[14,15] Both arms of the study include 4-monthly CEA monitoring, and therefore, the role of CEA monitoring in surveillance is unlikely to be directly defined.
The UK Follow-up After Colorectal Surgery (FACS) study has a 2 × 2 factorial design that assesses survival, quality of life, and costs of symptomatic follow-up in the primary care setting vs intensive follow-up in a hospital setting. The trial aims to enroll 4,760 Duke's A to C colorectal patients who have undergone primary curative treatment. The COLOFOL study is a multicenter randomized trial being conducted in Denmark, Sweden, Poland, UK, and the Netherlands. This trial will evaluate the frequency of surveillance tests after curative resection in stage II and III colorectal cancer. Enrollment was scheduled to start by the end of 2005.
CEA Monitoring in Advanced Colorectal Cancer
In managing metastatic colorectal cancer, changes in chemotherapy regimens are often considered when there is evidence of progressive disease. The usefulness of CEA monitoring in this setting was examined by two studies reviewed by Drs. Fakih and Padamanabhan. The positive-predictive value of CEA monitoring in detecting progressive disease during fluoropyrimidine-based chemotherapy reportedly ranges from 85% to 100%.[18,19] However, in some studies, the positive-predictive value for radiographic response is only as good as the flip of a coin.
In a follow-up study of 91 patients with metastatic colorectal carcinoma who were receiving first- and second-line chemotherapy, the sensitivity and specificity of a CEA increase ≥ 200% between baseline and 8-week evaluation for radiologic tumor response were 19.5% and 98.0%, respectively. The positive-predictive value for progression of disease with a CEA increase ≥ 200% was 88.9%. The authors concluded that CEA monitoring should be used cautiously with CT for evaluating tumor response to chemotherapy. A CEA surge following chemotherapy further complicates the matter and renders the test even more unreliable.
As such, due to insufficient evidence, ASCO does not recommend the routine use of CEA alone in monitoring treatment response in patients with colorectal cancer. CEA monitoring should be performed only when no other simple test to indicate response is available, to be measured at the start of treatment for metastatic disease and every 2 to 3 months during active treatment. Progressive disease is defined as two successive values above baseline.
Given the survival improvement associated with curative-intent salvage surgery for recurrences, there is an increasing need to define a scientifically and economically sound surveillance strategy following primary colorectal therapy. Current evidence supports a combination of imaging, such as CT scans or ultrasound, and tumor markers, such as CEA. The optimal frequency of these tests remains debatable.
CEA monitoring has become an accepted part of surveillance programs in many trials despite poor-quality evidence. The question of whether colorectal recurrence detected by CEA monitoring, alone or with imaging, is clinically significant remains unknown. A well-designed randomized, controlled trial enrolling thousands of patients is required to answer the question. Unfortunately, patient accrual for such a trial would likely be difficult in this country. Some might even argue that it would not be worth the resources.
—Wen Wee Ma, MD
—Wells Messersmith, MD
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.