Unrecognized Value of Carcinoembryonic Antigen in Recurrent Rectal and Sigmoid Colon Cancer: Case Series


Leonid L Yavorkovsky, MD, PhD1; Myron S Kwong, MD1; Sejal Jhatakia, MD1; Pilar Ivanov, MD1

Perm J 2019;23:18-022 [Full Citation]

E-pub: 05/31/2019


Introduction: Carcinoembryonic antigen (CEA) surveillance is recommended in patients with colorectal cancer for detection of potentially resectable metastases. In patients with appropriate symptoms, a highly increased CEA concentration (> 5 times the upper limit of normal) is considered strongly suggestive of cancer. Despite the recognized value, the test is neither absolutely sensitive nor specific for recurrent cancer. Generally, a greater diagnostic value has been assigned to elevated CEA levels, most commonly greater than 5 ng/mL. Fluctuations within the established normal CEA range are not customarily analyzed.
Case Presentations: We report here on 11 patients (8 women, 3 men) who, during the postoperative follow-up period, received a diagnosis of recurrent cancer despite their CEA levels exhibiting very subtle increases. Our cohort shared several similar characteristics such as a nonsmoking status, younger age (median, 52 years at initial diagnosis), and exclusive localization of the cancer to the rectosigmoid region.
Discussion: This important clinical observation may expand a prognostic value of CEA in a certain category of patients with colorectal cancer.


Carcinoembryonic antigen (CEA) surveillance is recommended in patients with colorectal cancer (CRC) for detection of potentially resectable metastases.1,2 In patients with appropriate symptoms, a highly increased CEA concentration (> 5 times the upper limit of normal) is considered strongly suggestive of cancer.3 Despite the recognized value, the test is neither absolutely sensitive nor specific for recurrent cancer. Generally, a greater diagnostic value has been assigned to elevated CEA levels, most often greater than 5 ng/mL. The fluctuations within the established normal CEA range are not customarily analyzed.

We describe 11 patients in whom recurrent CRC was diagnosed during postoperative follow-up despite their CEA levels exhibiting very subtle increases.


From 2000 through 2016, a total of 11 patients in our department who were undergoing standard CRC surveillance according to the National Comprehensive Cancer Network guidelines4 were identified and their medical records were reviewed. The cohort was selected on the basis of cancer recurrence in the setting of CEA levels that exhibited an incline but fell short of surpassing the upper limit of normal.

The relevant clinical data of the patients are presented in Table 1. Eight of the 11 patients were women, and the median age of the whole group was 52 years (range, 44-73 years; mean, 56 years). At the time of initial diagnosis, 1 patient had stage I, 4 patients had stage II, and 6 patients had stage III cancer. Only a single patient had a poorly differentiated adenocarcinoma, whereas most had moderately or well-differentiated tumors. The cancer was located in the rectum in 5 patients, in the sigmoid colon in 5 patients, and at the rectosigmoid junction in 1 patient. The recurrent cancer was confirmed by results of the biopsy of the suspicious lesion detected on computed tomography or positron emission tomography/computed tomography, or during the surgery. In 1 patient the biopsy was not technically feasible, so the recurrence was diagnosed on the basis of retroperitoneal fluorodeoxyglucose-avid lymph nodes (LNs) and CEA elevation, which was effectively treated with stereotactic radiosurgery. Seven patients had relapsed with a single metastasis: 4 in the lung, 1 in the liver, 1 in an abdominal LN, and 1 in a retroperitoneal LN. Four patients displayed multiple sites of recurrence, including the peritoneum (n = 1); the liver and colon (n = 1); the retroperitoneal, mediastinal, and supraclavicular LNs (n = 1); and the lung and brain (n = 1).

The serum CEA assay was performed according to an immunochemical automated analyzer (Abbott AxSym, Abbott Laboratories, North Chicago, IL) using a microparticle enzyme immunoassay with a normal range of equal to or below 5.0 ng/mL. The median preoperative CEA level (assessed in 9 of 11 patients) was 3.58 ng/mL (range, < 1.0 to 12.3 ng/mL), and the postoperative CEA was 1.03 ng/mL (range, < 1.0 ng/mL to 2.9 ng/mL). At the time of recurrence, the median CEA level was 2.97 ng/mL (range, 1.1 ng/mL to 4.8 ng/mL). In 1 patient (Table 1, Case 1), CEA demonstrated an abrupt, minuscule but perceptible increase above the level that had been remarkably stable over 5 years. Most patients displayed at least 3 steadily rising CEA values above the nadir over variable periods.

The median time to recurrence was 37.6 months (range, 19 to 65 months), with a median survival of 91.9 months (range, 41 to 162 months). Five of 11 patients died. The 6 patients who are alive have survived from 41 to 162 months (median, 83.8 months). The median time to recurrence in surviving patients varied from 19 to 66 months (median, 34.2 months).

18 022


Despite the lack of survival benefit, CEA surveillance has been recommended in the postoperative period to improve the detection of recurrent CRC, including those cancers that could be treated with curative intent. The accuracy of the test has been debated for years, prompting more sophisticated methods of CEA assessment such as CEA doubling time and CEA half-life.5 In general, the higher the level of CEA, the higher the specificity of the test. Therefore, the studies have traditionally focused on the high-risk patients with elevated preoperative CEA (> 5 ng/mL) and, especially, patients with persistent CEA elevation after surgery.6 An interpretation of the CEA levels is more problematic when the levels fluctuate in the normal or near-normal range, mostly for the following reasons: 1) the small variations are subject to oversight; 2) the minor fluctuations may be attributed to medical conditions other than cancer or to tobacco exposure, which may trigger minor but false CEA elevations3; and 3) CEA sensitivity and specificity can vary depending on the brand of assay used, so no uniform normal CEA values have been established. Even more important, up to 70% of CRC recurrences occur without CEA elevation (at normal CEA level).7 In many such cases the serum CEA has no biological connection to the cancer because approximately 30% of all CRC recurrences do not produce CEA.8 Despite all these challenges, a normal CEA level could still be a subtle marker of cancer recurrence, as was demonstrated in the study, with an upper limit of normal CEA of 10 ng/mL.9 

We identified 11 patients with recurrent cancer whose CEA levels before the diagnosis demonstrated small elevations that were confined to the established normal limits. The interpretation of such narrow-range CEA fluctuations requires careful distinction from nonspecific, potentially false-positive deviations. For example, previous studies have seldom made a distinction between smokers and nonsmokers,10 or men vs women, despite the proven differences in their CEA levels. According to our laboratory reference studies, the 90th percentile upper limit for smokers is 5.0 ng/mL and the 95th percentile for nonsmokers is 3.0 ng/mL. In a large reference study, the upper limits of the CEA assay for male smokers and nonsmokers were 6.2 and 3.4 ng/mL, respectively, and for female smokers and nonsmokers, 4.9 and 2.5 ng/mL, respectively.11 Thus, smoking appears to almost double the serum concentration of CEA and may certainly confound the interpretation of the results. All our patients were nonsmokers at the time of evaluation. Four of 11, however, showed CEA greater than 3.0 ng/mL (upper level for nonsmokers). Additionally, none of our patients had conditions that potentially may result in nonspecific CEA elevation, including cirrhosis, hepatitis, biliary obstruction, pulmonary emphysema, diverticulitis, inflammatory bowel disease, peptic ulcer disease, or collagen vascular disease.12

The CEA levels in patients with CRC have rarely been studied in regard to the cancer location.13 The most striking observation was that the cancers in our patients were localized exclusively to the rectal and sigmoid region. The cause of such a location clustering is unclear. It is plausible, however, that our cohort exemplifies a biologically distinct subcategory of CRC. For example, our patients have demonstrated more indolent than average cancer growth. Furthermore, 7 of 11 patients presented with oligometastatic recurrence, which is a recognized prerequisite of curability. The latter features could account for the lower, and stagnant, CEA levels in our patients (as well as a longer survival). The lower preoperative CEA levels have previously been linked to better outcomes. This was convincingly shown in both pulmonary14,15 and liver16 metastasectomy settings. Thus, the evidence suggests better outcomes when the cancer produces no or very small amounts of CEA.

Another distinguishing feature of our cohort was that it was nearly 10 years younger than a statistical median age for patients with rectal cancer.17 Also, the prevalence of the female sex, which is a reverse of average statistics, may provide an additional clue for this cluster (target) population identification.

Although more intensive surveillance with CEA screening consistently results in an increased rate of curative-intent surgical treatment, the cure rate remains suboptimal, as no survival benefit has been demonstrated yet. To improve the validity of CEA screening, an individualized and perhaps selective measurement approach may be necessary. Our study could be a step in this direction.

Our collective data on patients with CRC for more than 6 years (2010-2015) yielded an estimated incidence of the cohort of 0.7%. (On average, 91 new patients were seen annually during this period.) The fact that our cohort represents only a small subset of the CRC population highlights the importance of rigorous identification of such patients. Despite the rare occurrence, such cases could be easily overlooked and thus their incidence underestimated.


The upward trend of postoperative CEA level even within the normal range may be clinically relevant and, given appropriate attention, potentially lifesaving. Our observations suggest that a correctly defined subset of patients with rectal and sigmoid cancers may constitute a target population requiring maximum vigilance and cautious interpretation of CEA trends within an established normal range.

Our findings have also exposed an important but elusive relevance of the patient’s gender and tobacco use to the subtle postoperative CEA changes. Hopefully, the recognition of better-defined subcategories of patients with CRC as well as early intervention, before CEA levels become grossly abnormal, will expand a prognostic value of CEA and, ultimately, improve clinical outcomes. Although our study was small and included only 11 patients, it illustrates an important clinical observation that should be further evaluated in future studies.

Disclosure Statement

The author(s) have no conflicts of interest to disclose.


Kathleen Louden, ELS, of Louden Health Communications performed a primary copy edit.

How to Cite this Article

Yavorkovsky LL, Kwong MS, Jhatakia S, Ivanov P. Unrecognized value of carcinoembryonic antigen in recurrent rectal and sigmoid colon cancer: Case series. Perm J 2019;23:18-022. DOI: https://doi.org/10.7812/TPP/18-022

Author Affiliations

1 Oncology Department, San Jose Medical Center, CA

Corresponding Author

Leonid L Yavorkovsky, MD, PhD (leonid.yavorkovsky@kp.org)

1. Locker GY, Hamilton S, Harris J, et al; ASCO. ASCO 2006 update of recommendations for the use of tumor markers in gastrointestinal cancer. J Clin Oncol 2006 Nov 20;24(33):5313-27. DOI: https://doi.org/10.1200/JCO.2006.08.2644.
 2. Figueredo A, Rumble RB, Maroun J, et al. Follow-up of patients with curatively resected colorectal cancer: A practice guideline. BMC Cancer 2003 Oct 6;3:26. DOI: https://doi.org/10.1186/1471-2407-3-26.
 3. Duffy MJ. Carcinoembryonic antigen as a marker for colorectal cancer: Is it clinically useful? Clin Chem 2001 Apr; 47(4):624-30.
 4. Clinical Practice Guidelines in Oncology: Rectal Cancer. Version 3. Plymouth Meeting, PA: National Comprehensive Cancer Network. [Internet] Plymouth Meeting, PA: National Comprehensive Cancer Network; 2017 [cited 2017 Apr 6]. Available from: www.nccn.org/store/login/login.aspx?ReturnURL=https://www.nccn.org/professionals/physician_gls/pdf/rectal.pdf. [password protected].
 5. Konishi F. CEA doubling time and CEA half-life in the prediction of recurrences after colorectal cancer surgery. Jpn J Clin Oncol 2002 Feb;32(2):41-2. DOI: https://doi.org/10.1093/jjco/hyf018.
 6. Wanebo HJ, Rao B, Pinsky CM, et al. Preoperative carcinoembryonic antigen level as a prognostic indicator in colorectal cancer. N Engl J Med 1978 Aug 31;299(9):448-51. DOI: https://doi.org/10.1056/nejm197808312990904.
 7. Abir F, Alva S, Longo WE, Audiso R, Virgo KS, Johnson FE. The postoperative surveillance of patients with colon cancer and rectal cancer. Am J Surg 2006 Jul;192(1):100-8. DOI: https://doi.org/10.1016/j.amjsurg.2006.01.053.
 8. Carl J, Bentzen SM, Nørgaard-Pedersen B, Kronborg O. Modelling of serial carcinoembryonic antigen changes in colorectal cancer. Scand J Clin Lab Invest 1993 Nov;53(7):751-5. DOI: https://doi.org/10.3109/00365519309092581.
 9. Irvine T, Scott M, Clark CI. A small rise in CEA is sensitive for recurrence after surgery for colorectal cancer. Colorectal Dis 2007 Jul;9(6):527-31. DOI: https://doi.org/10.1111/j.1463-1318.2007.01176.x.
 10. Khan K, Athauda A, Aitken K, et al. Survival outcomes in asymptomatic patients with normal conventional imaging but raised carcinoembryonic antigen levels in colorectal cancer following positron emission tomography-computed tomography imaging. Oncologist 2016 Dec;21(12):1502-8. DOI: https://doi.org/10.1634/theoncologist.2016-0222.
 11. Wilson APM, Van Dalen A, Sibley PE, Kasper LA, Durham AP, el Shami AS. Multicenter tumour marker reference range study. Anticancer Res 1999 Jul-Aug;19(4A):2749-52.
 12. Clinical practice guidelines for the use of tumor markers in breast and colorectal cancer. Adopted on May 17, 1996 by the American Society of Clinical Oncology. J Clin Oncol 1996 Oct;14(10):2843-77. DOI: https://doi.org/10.1200/jco.1996.14.10.2843.
 13. Grossman I, de Bock GH, Meershoek-Klein Kranenbarg WM, van de Velde CJ, Wiggers T. Carcinoembryonic antigen (CEA) measurement during follow-up for rectal carcinoma is useful even if normal levels exist before surgery. A retrospective study of CEA values in the TME trial. Eur J Surg Oncol 2007 Mar;33(2):183-7. DOI: https://doi.org/10.1016/j.ejso.2006.10.035.
 14. Girard P, Grunenwald D, Baldeyrou P, Spaggiari L, Régnard JF, Levasseur P. Resectable lung metastases from colorectal cancer: Look at the serum CEA level! Ann Thorac Surg 1996 Dec;62(6):1888-9.
 15. Salah S, Watanabe K, Welter S, et al. Colorectal cancer pulmonary oligometastases: Pooled analysis and construction of a clinical lung metastasectomy prognostic model. Ann Oncol 2012 Oct;23(10):2649-55. DOI: https://doi.org/10.1093/annonc/mds100.
 16. Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases. World J Surg 1995 Jan-Feb;19(1):59-71. DOI: https://doi.org/10.1007/bf00316981.
 17. Howlader N, Noone AM, Krapcho M, et al. SEER Cancer Statistics Review. 1975-2010. Bethesda, MD: National Cancer Institute; 2013.

Keywords: cancer recurrence, cancer surveillance, carcinoembryonic antigen, colorectal cancer


Click here to join the eTOC list or text ETOC to 22828. You will receive an email notice with the Table of Contents of The Permanente Journal.


2 million page views of TPJ articles in PubMed from a broad international readership.


Indexed in MEDLINE, PubMed Central, EMBASE, EBSCO Academic Search Complete, and CrossRef.




ISSN 1552-5775 Copyright © 2021 thepermanentejournal.org

All Rights Reserved