PerspectiveIn the Journals Plus

Elevated biomarker level postoperatively increases colon cancer recurrence risk

Show Citation

December 22, 2017

Elevated postoperative serum carcinoembryonic antigen increased risk for recurrence, especially within the first 12 months after surgery, among patients with colon adenocarcinoma, according to a retrospective cohort analysis published in JAMA Oncology.

However, elevated preoperative carcinoembryonic antigen (CEA) that normalizes following resection did not predict poor prognosis, suggesting routine measurement of postoperative, and not preoperative CEA is warranted.

“Emphasis should be placed on postoperative CEA, and in the setting of modern high-quality imaging, we question the utility of measuring preoperative CEA,” Martin R. Weiser, MD, surgical oncologist in the colorectal service at Memorial Sloan Kettering Cancer Center, and colleagues wrote. “Patients with elevated postoperative CEA tend to experience recurrence early, which might justify a risk-adjusted and individualized surveillance strategy.”

Guidelines recommend clinicians measure CEA levels preoperatively among patients with colon cancer. Although CEA levels that are persistently high after resection are associated with increased recurrence risk, the effect of levels that normalize is unknown.

Weiser and colleagues evaluated data from 1,027 patients (50.4% men; median age, 64 years) with stage I to stage III colon adenocarcinoma who underwent curative resection between 2007 and 2014.

After excluding patients without postoperative measurements, researchers divided patients into three cohorts: those with normal preoperative CEA (n = 715), those with elevated preoperative but normalized postoperative CEA (n = 142), and those with elevated preoperative and postoperative CEA (n = 57).

Three-year RFS served as the study’s primary endpoint.

During a median follow-up of 38 months, 94 patients (10.3%) had recurrences and 42 patients (4.6%) died. Three-year RFS for all patients was 88.4% (95% CI, 85.9-90.5).

Patients with normal preoperative CEA had a 3-year RFS of 89.7%, which was 7.4 percentage points higher than the 82.3% rate observed among the combined cohorts of patients with high preoperative levels (P = .01).

However, 3-year RFS did not significantly differ between the normal preoperative group and the normalized postoperative group alone (87.9%).

Patients with persistently elevated CEA levels showed a 3-year RFS 14.9 percentage points lower than the combined cohorts with normal postoperative CEA (74.5% vs. 89.4%; P = .001).

Multivariate analyses showed elevated postoperative CEA increased risk for recurrence (HR = 2; 95% CI, 1.1-3.5), whereas normalized postoperative CEA did not (HR = 0.77; 95% CI, 0.45-1.3).

Analyses of the smoothed curve of the hazard function showed recurrence risk appeared higher, and peaked earlier, in the elevated postoperative CEA group.

These results suggest the value of CEA as a biomarker is greatest as an early indicator of tumor recurrence, Rebecca Anne Miksad, MD, MPH, and Neal J. Meropol, MD, both employees of Flatiron Health, wrote in an accompany editorial.


The observation by Konishi and colleagues that the timing of recurrence is earlier among patients with elevated postoperative CEA introduces the concept of risk-stratified surveillance that varies based on overall risk and the timing of that risk,” they wrote. “While the term precision medicine is quite familiar in the setting of cancer treatment selection, its application to surveillance after definitive primary treatment also holds great promise if early subclinical metastasis can be identified and effectively treated with curative potential.” – by Alexandra Todak


Disclosures: Weiser and other study authors report no relevant financial disclosures. Miksad and Meropol report employment with Flatiron Health.

itj+ Perspective

Wafik El-Deiry

Tumor marker CEA is a rapid inexpensive blood test that can detect gastrointestinal cancer, such as colorectal cancer, providing a number in the test result that can be checked again over time. The CEA test result can be elevated among patients with early-stage disease (stage II-III). After surgery, when most patients are cured, the CEA number would be expected to normalize. It is a red flag if the CEA remains elevated after surgery, possibly indicating residual disease, ie, cancer that was not removed. This is usually due to occult metastatic disease that was not detected prior to surgery. Preoperative CEA is not always checked before surgery and, when it is checked, it is not always elevated. When it is not elevated prior to surgery, most of the time it is not helpful to check CEA postoperatively as a marker of relapse, although it can still detectably rise with metastatic disease. In practice, when CEA is not elevated prior to surgery it can still be checked postoperatively; however, a negative test does not mean there is no relapse, and that’s a situation where imaging provides very useful information.

Another test called CA19.9 can sometimes be elevated in colorectal cancer when CEA is normal and may be useful to follow if CEA is not elevated. When both CEA and CA19.9 are normal for preoperative early-stage disease or postoperative relapsed local or metastatic disease, liquid biopsy may in the future offer a potential alternative as a tumor marker that can then be serially measured to detect disease clearance and relapse.

In this study, Konishi and colleagues retrospectively explored outcomes in a large cohort of patients with stage I to stage III colon cancer. The authors conclude that an elevated preoperative CEA is not a poor prognostic marker when it normalizes postoperatively, but it is a poor prognostic marker when it is elevated postoperatively. These conclusions from their data are certainly reasonable, expected and consistent with clinical experience.

The authors further advocate for routine measurement of postoperative, rather than preoperative, CEA. However, it is debatable whether preoperative CEA measurement is useful or not. The authors cite literature in which elevated preoperative CEA was infrequently associated with elevated CEA at relapse, and plenty of examples where elevated CEA at recurrence was associated with normal CEA preoperatively.

One of the limitations of the study, however, and reason clinicians may wish to keep checking preoperative CEA is that patients in the present study with stage I through stage III disease were grouped together. More data looking at the outcomes with CEA by stage in the three CEA groups is needed in the future with larger patient numbers to support the idea that preoperative CEA may not be warranted in stage II or stage III colorectal cancer. Otherwise such conclusions are premature. It would be expected that patients with stage I colon cancer may have undetectable CEA that may be elevated if the disease relapses with metastases. In practice, a normal preoperative CEA in stage III disease could signify that the tumor does not secrete CEA; so, with postoperative CEA measurements, clinicians should consider the possibility where measuring CEA postoperatively may not be sensitive to relapse. However, an elevated CEA in stage III preoperatively is useful to know, and it is useful for patients to know whether it normalized after surgery.

In advanced disease, CEA trends can be very helpful and can save patients precious time if progression is suspected so they can get scans earlier, or if trends are favorable, reduce the frequency of scans and reduce radiation exposure. Checking preoperative CEA in early stage II to stage III colorectal cancer is not useless and should not be discouraged or discontinued without more definitive evidence.

The choice is not so much preoperative vs. postoperative CEA because the test is inexpensive, but capturing the CEA value preoperatively, as well as postoperatively and subsequently during disease course, has value for disease status monitoring and early detection of relapse or progression.

Wafik S. El-Deiry, MD, PhD, FACP

HemOnc Today Editorial Board Member 
Fox Chase Cancer Center

Disclosure: El-Deiry reports no relevant financial disclosures.