Clinical Value of Diagnosis, Prognosis, J. David

Beatty, MD;

Carlos

and

Carcinoembryonic Antigen

Follow-up

Romero, MD; Peter

W.

of Patients With Cancer

Brown, MD; Walter Lawrence, Jr, MD;

\s=b\ Plasma carcinoembryonic antigen (CEA) in nanograms per milliliter was assayed in 149 patients with benign and 567 patients with malignant disease. Elevated CEA level ( > 5.0) was a good indicator of malignant disease but a poor screening test for cancer because of the high false-negative rate. Degree of elevation of plasma CEA level correlated with incidence of metastatic disease in patients with colorectal, gastric, and breast carcinomas, but no correlation was seen between CEA levels and status of lymph nodes in patients with localized disease. Patients with localized colorectal cancer, but elevated CEA levels before resection, had a 2.1-fold increase in the incidence of recurrence; however, this added to the prognostic value of Dukes' staging only when the CEA level remained elevated postoperatively. In 87% of patients with colorectal cancer, the CEA level was elevated at the time of recurrence, but a therapeutic value of reexploration for unexplained CEA level elevation was not confirmed.

(Arch Surg 114:563-567, 1979)

antigen (CEA) was identified in of the colon by Gold and Freedman1 in 1965, and Thomson et al2 in 1969 outlined a radioimmunoassay for CEA levels in serum. In this assay, the antigenantibody complexes were separated by ammonium sulfate. Other methods were soon developed, differing primarily in the method of isolation of the antigen-antibody complexes. Egan et al1 and MacSween et al' reported on the use of a second antibody to precipitate the complexes, whereas Hansen et al' used zirconyl phosphate gel for the same purpose. The method of Hansen is used in the commercial assay currently available. Initially, 97% of patients with colon cancer were found to have elevated serum CEA levels, but as patients with earlier stages of disease were examined, it was realized that fewer than 50% of patients with early colon cancer had abnormal serum or plasma CEA levels" and that the CEA level could be abnormal in the absence of carcinoma of the colon or rectum.7 The present study was instituted in 1972 and a prelimi¬ nary report was issued in 1974." Our objectives have been to assess (1) the value of CEA level in the diagnosis of malignant neoplasms of different sites of origin, (2) the correlation of CEA level with stage of disease, (3) the value

Carcinoembryonic cancer

Accepted for publication Oct 10, 1978. From the Division of Surgical Oncology, Medical College of Virginia, Richmond. Dr Beatty is now with the Department of Surgery, University of Manitoba, Winnipeg, Canada. Reprint requests to Box 11, MCV Station, Medical College of Virginia, Richmond, VA

23298

(Dr Terz).

Jose J.

Terz,

MD

of CEA level in estimating the prognosis of malignant neoplasms, and (4) the value of CEA level in the follow-up of patients treated for malignant neoplasms. METHODS Plasma CEA level (nanograms per milliliter) was assayed by the Hansen" method on the initial visit of most patients referred to the Division of Surgical Oncology at the Medical College of Virginia, Richmond. Patients who had received treatment previously were excluded from this study unless they were referred at the time they appeared with recurrent disease. Most patients were treated by surgical resection, and postoperative plasma CEA levels were obtained on most patients. Patients treated for colorectal carcinoma had plasma CEA levels assayed at two weeks, one month, and three months postoperatively, and then about every three months. Recurrence and survival data for patients with colorectal carcinoma were collected and correlated with plasma CEA levels. Patients were considered to have recur¬ rences on the basis of clinical, roentgenographic, radiological, laboratory, and biopsy data. It was not feasible or in the patients' interest to document all recurrences histologically by biopsy. The pathological extent of disease was recorded and stratifica¬ tion was done on the basis of patients with localized disease only and those with advanced disease. The patients with localized disease were categorized according to the existence of tumor involvement of the regional lymph nodes, and Dukes' classification was used for colorectal carcinoma.10 The pretreatment CEA levels were compared with the subsequent operative and pathological findings regarding extent of disease. All plasma CEA levels have been measured in nanograms per milliliter, but for simplicity, these units have not been included after each CEA value reported in the text.

RESULTS

Of the 716 patients assayed, 149 (21%) had benign disease and 567 (79%) had malignant disease (Table 1). An abnormal CEA level ( > 5.0) was found in 11% of patients with benign and 41% of patients with malignant disease. A total of 94% of patients with CEA levels > 5.0, 99% of patients with CEA levels > 10.0, and 100% of patients with CEA levels > 25.0 had malignant disease. The commonest cancers were gastrointestinal (GI) (65%), and patients with GI cancer accounted for 73% of the abnormal CEA values and 84% of the CEA levels > 10.0 (Table 2). Of the remaining 22 patients who had CEA levels > 10.0,16 had breast cancer, two had benign disease, and one each had uterine cancer, lung cancer, lymphoma, and pseudomyxoma peritonei. Patients with carcinoma of the stomach, pancreas, and colorectum had the highest

Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/09/2015

incidence of abnormal CEA levels (65%), but even in these categories, a substantial number of patients with cancer had normal CEA levels (35%). These same cancers also were associated with higher CEA levels; 46% of the patients had CEA levels > 10.0, 26% of the patients had CEA levels > 25.0, and 16% of the patients had CEA levels > 100.0. Little difference was seen in the distribu¬ tion of CEA levels between patients with non-GI cancer and patients with GI cancer other than carcinoma of the stomach, pancreas, or colorectum. All patients with cancer were classified according to operative and pathological extent of disease, and when more than 50 patients had the same type of cancer, we looked for a correlation between pretreatment CEA level and extent of disease. In carcinoma of the oropharynx, the CEA levels were elevated in only 23% of patients and no correlation was observed; but in patients with breast, gastric, and colorectal carcinoma, the level of preoperative CEA did reflect the extent of disease. The overall distribution of CEA values for patients with carcinoma of the breast was similar to that observed for patients with non-GI cancer; generally, however, only 15% of patients with localized disease had an elevated CEA level (>5.0); whereas in 53% of patients with metastatic disease, the CEA level was abnormal (Table 3). The inci¬ dence of metastatic disease was 72%, 88%, and 100% for patients with breast carcinomas who had CEA lev¬ els > 5.0, > 10.0, and > 25.0, respectively, but within the group of patients having localized disease, CEA level did not reflect pathological axillary lymph node status. A similar phenomenon was observed with gastric carci¬ noma (Table 4). Only 29% of the patients with localized disease had CEA levels > 5.0, whereas 85% of the patients with metastatic disease had an abnormal CEA level. Three of the four patients with localized disease and CEA levels > 25.0 and the one patient with a CEA level > 100.0 demonstrated distant metastasis shortly after the resec¬ tion. Thus, CEA levels from 0 to 5.0, > 5.0, and > 100.0 indicated metastatic disease in 29%, 94%, and 100% respec¬ tively, of the patients with carcinoma of the stomach. Two patients with pancreatic carcinoma had localized resectable disease and their CEA values were both < 2.5. The remaining 26 patients with pancreatic carcinomas had Table

1.—Range

CEA Level,

ng/mL 0-5.0 >5.0 >10.0 >25.0

>100.0

of Plasma CEA Levels Before

All Patients

Table CEA Level, ng/mL 0-5.0 >5.0

>10.0 >25.0 > 100.0

Malignant

133/149(89%) 16/149(11%) 2/149(1%)

334/567(59%) 233/567 (41 139/567(25%) 78/567(14%) 46/567(8%)

Disease

467/716(65%) 249/716(35%) 141/716(20%) 78/716(11%) 46/716(6%)

0/149 0/149

2.—Range

0/92

Disease

of Plasma CEA Levels Before

Oropharynx (77%) (23%) (5%) 1/92 (1%)

71/92 21/92 5/92

Therapy

Benign

Stomach 17/51

(33%) 34/51 (67%) 20/51 (39%) 15/51 (29%) 6/51 (12%)

locally

unresectable

disease

or

metastatic

disease

(Table 2). A total of 151 patients were seen primarily with colorec¬ tal neoplasia before any treatment (Table 5). Of patients with benign polyps or Dukes' A carcinoma, 85% had normal CEA levels; whereas in 85% of patients with metastatic or recurrent disease, the CEA level was abnormal ( > 5.0). As the pathological extent of disease increased, the incidence of higher CEA values ( > 10, > 25, and > 100) also increased. The distribution of CEA levels at recurrence for patients whose disease was originally considered localized was almost identical to that of the patients who originally appeared with métastases. Analyzing the data by CEA levels we observed that a normal CEA level was associated with localized disease in 81% of patients, whereas CEA levels > 5.0, > 25.0, and > 100.0 were associated with metastatic disease in 65%, 79% and 92% of patients,

respectively.

Of the 58 patients with Dukes' A, B, and C colorectal carcinoma who had been followed up for 18 months or longer (mean, 46.4 months; range, 18 to 72 months) recur¬ rent disease developed in 19 of them. Recurrence developed in nine of 38 patients (24%) with normal CEA levels before resection, whereas recurrence developed in ten of 20 patients (50%) with abnormal preoperative CEA levels. Thus, an abnormal preoperative CEA level was associated with a 2.1-fold increase in the incidence of recurrence in patients with resectable colorectal cancer. The disease-free survival rate at three years of all patients with resectable colorectal cancer was 33/48 (69%). Recurrence did not develop in any patients with Dukes' A disease, patients with Dukes' disease had a 70% disease-free survival at three years, and patients with Dukes' C disease had a 50% disease-free survival at three years. Patients with normal CEA levels had a three-year disease-free survival rate of 73%, and those with a CEA level > 5.0 had a three-year disease-free survival rate of 60%. However, consistent correlation between preoperative CEA level and diseasefree survival could not be demonstrated within a given Dukes' stage. Levels of CEA were followed serially after operation in most patients with colorectal carcinoma. A total of 15 of the patients in whom recurrence developed had CEA levels monitored and 13 of these patients had a CEA level > 5.0 at the time recurrence was reported. In most patients with Dukes' A, B, or C carcinoma of the colon, the CEA levels returned to normal within two weeks to one month. Levels of CEA in several patients were not normal until three months postoperatively. In three patients, the CEA levels remained elevated after this time and were observed to be increasing from one measurement to the next. Metastatic disease was eventually identified in all three of these patients (6,10, and 23 months postoperatively). Three other Therapy

for Patients With Gastrointestinal Cancer Pancreas 10/28 (36%) 18/28 (64%) 12/28 (43%)

4/28(14%) 2/28 (7%)

Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/09/2015

Colorectal 54/155

(35%) 101/155 (65%) 75/155 (48%) 41/155 (26%) 29/155 (19%)

Other 32/41 (78%) 9/41 (22%) 7/41 (17%) 5/41 (12%) 2/41 (5%)

Table

3.—Range

4.—Range of Plasma CEA Levels Before Therapy for Patients With Gastric Carcinoma

of Plasma CEA Levels Before

Table

Therapy for Patients With Breast Carcinomas CEA Level,

Localized, Total

ng/mL 0-5.0 >5.0 >10.0 >25.0 > 100.0



(69%) (31%) (14%) 8/115(7%) 5/115 (4%)

79/115

36/115 16/115

Table

Nodes + or 56/66 (85%) 10/66 (15%) 2/66 (3%) 0/66 0/66

5.—Range

Metastatic 23/49 (47%) 26/49 (53%) 14/49 (29%) 8/49 (16%) 5/49 (10%)

of Plasma CEA Levels Before

CEA Level, ng/mL 0-5.0 >5.0

Localized, Nodes 12/17(71%) 5/17 (29%)

+ or —

Metastatic 5/34 (15%) 29/34 (85%)

_>10.0_4/17 (24%)_16/34 (47%) 4/17(24%) 1/17 (6%)

>25.0

> 100.0

Therapy

For Patients With Colorectal

11/34(32%) 5/34 (15%)

Neoplasia

Carcinomas CEA Level, 0-5.0 >5.0 >10.0 >25.0 >100.0

ng/mL

Benign

Polyps

14/16(88%) 2/16(13%)

1/16(6%) 0/16

0/16

9/11

Dukes' B, C 33/60 (55%)

57ÏÏ

27/60(45%) 20/60 (33%) 8/60 (13%) 2/60 (3%)

Dukes' A

(82%) 2/11(18%) 2/11 (18%) 1/11 (9%)

had transient increases in CEA level that had returned to normal when the tests were repeated. No recurrent disease developed in any of these patients subse¬ quently. Ten of the patients with recurrent disease had normal CEA levels postresection and then elevated serum CEA levels at the time the recurrence was diagnosed based on other investigations, such as liver function tests, liver scan, liver biopsy, chest or bone roentgenography, and bone scan or laparotomy. In eight of these ten patients, the diagnosis of recurrence was preceded by up to 15 months (mean, 7.9 months) by an increase in CEA level. In all eight cases, the CEA level was not just elevated, but repeated tests confirmed that it was also increasing. Reexamination of the original pretreatment CEA levels in these ten patients revealed that half (five) had had normal and half (five) had had abnormal CEA levels. Three patients were subjected to laparotomy for persis¬ tent, unexplained abnormal CEA levels that were continu¬ ing to increase. In two patients, the CEA level had returned to normal after resection and then had increased to abnormal levels. Laparotomy had been performed four and nine months after the initial postoperative increase in CEA level. In one patient, the CEA level never returned to normal postoperatively and the laparotomy was performed ten months after the original resection. None of the three patients having laparotomy were candidates for resection of the disease identified by the procedure because of multiple hepatic métastases affecting both lobes of the liver in all of them and also because of multiple peritoneal implants in one patient.

patients

COMMENT

The ideal substance for the early detection of tumors should be cancer-specific and, if possible, organ-specific; its concentration in plasma and/or other body fluids should be proportional to the tumor cell mass; and it should be easy to assay. Our first objective in this study was to evaluate plasma CEA levels as a marker for the early detection of disease. Neville'1 had noted that only one of 335 healthy controls had a CEA value greater than 5.0 ng/ml. Troncone

Metastatic

Recurrent

(16%) 54/64(84%) 42/64 (66%) 27/64 (42%) 23/64 (36%)

2/15(13%) 13/15(87%) 11/15(73%) 7/15 (47%) 5/15 (33%)

10/64

et al1- found three of 57

plasma CEA CEA level

>

healthy controls with an elevated level ( > 5.0 ng/ml) and Hirai13 found the 5.0 ng/ml in 1.4% of 416 normal adults.

Hansen et al,e reporting on over 2,000 healthy subjects, noted that the CEA level was greater than 5.0 in 0.2% of nonsmokers, 4.0% of smokers, and 1.7% of former smokers. On the basis of these articles, we used a plasma CEA level of 0 to 5.0 ng/ml as the most practical normal range and considered values above this as nonspecific indicators of disease. We used the zirconyl phosphate gel technique of Hansen" because it was rapid, reproducible, and readily available. The patients screened for plasma CEA levels had all been referred to the Surgical Oncology Division and thus were a selected population. Almost 80% of the patients had malig¬ nant disease, primarily GI and breast cancer. Few patients were referred for pulmonary, genitourinary, hematological, or neurological malignant neoplasms. The CEA levels that we observed in patients with benign disease compared favorably with the report by Hansen et al9 that 8% of patients with benign disease had CEA levels > 5.0. The overall distribution of CEA values that we observed in patients with a variety of cancers (Table 1) was similar to the distributions reported by Hansen et al9 and Neville,11 thus supporting our conclusion that a plasma CEA level > 5.0 was a nonspecific indicator of disease, benign or malignant, and that a level > 10.0 was a nonspecific indicator of cancer. Although a plasma CEA level > 10.0 suggested GI cancer as the primary condition, it was not diagnostic of a specific tumor type. The plasma CEA level was of much less value in ruling out the existence of malignant neoplasms. In 59% of patients with cancer, the plasma CEA level would have given a false-negative screening test. Caution must there¬ fore be exercised to avoid interpreting a normal CEA level as indicating "no malignant disease." The prognostic value of pretreatment plasma CEA levels was examined in four different tumor types by comparing CEA level with pathological extent of disease. In carcino¬ ma of the oropharynx, no correlation could be made

Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/09/2015

between CEA level and pathological stage of disease. Our results in patients with breast cancer (Table 3) are in keeping with those reported by Steward et al14 and Neville.11 Steward considered a plasma CEA level greater than 2.5 ng/ml as elevated and noted that 27% of patients with localized disease and 79% of patients with metastatic disease had elevated CEA levels. In patients with localized disease, they also showed little differences in the CEA levels between those with and without lymph node involve¬ ment. Only 1/22 patients with localized disease had a CEA level > 10 and none of them had a CEA level > 25. Thus, pretreatment determination of CEA level in patients with proved carcinoma of the breast was of value in indicating the presence of metastatic disease. We consider a CEA level > 10 as suspicious and > 25 as highly suspicious (but not diagnostic) of métastases. Martin et al," commenting on their 62 patients with breast carcinomas, came to the same conclusions. Patients with gastric and colorectal carcinoma had elevated preoperative CEA levels in 66% of cases (Table 2) and CEA levels correlated with the gross extent of disease (Table 4 and 5), but just as with breast cancer within the localized disease group, the CEA level did not correlate well with lymph node involvement. The observations in local¬ ized colorectal carcinoma that few patients with Dukes' A lesions had abnormal CEA levels has been reported by many authors,11111'1"20 but the finding of little difference in CEA levels between patients with Dukes' and C disease has been more controversial. Our results agree with those of some authors11·13-1"·18 and conflict with the results of others.1719-20 Plasma CEA level was not, therefore, conclusive for the presence of cancer at an early stage. Levels of CEA > 5.0 were suspicious for cancer and levels > 10.0 were highly suspicious, but patients with these levels ( > 10.0) also tended to have more advanced disease. The test was not specific for a given tumor type, but elevations of CEA levels > 10.0 did suggest GI cancer. The increasing extent of disease observed with higher CEA levels did indicate that the CEA level was proportional to the tumor cell mass in those patients having elevated CEA levels. Presently, determination of CEA level is not conclusive for the early detection of cancer but may be a useful adjunct to other tests.2' As suggested by Gold and Freedman22 a persistent¬ ly abnormal plasma CEA level should be interpreted as suspicious of cancer, and in the absence of an adequate reason for the abnormal CEA level, such patients should be followed carefully for the possible development of a GI malignant neoplasm. Edgington et al21 suggested that part of the nonspecificity of CEA levels was related to the heterogeneity of the CEA preparations, and they reported greater specificity with an isomerie species of CEA (CEAS). Our results indicated that depending on the site of origin of malignancy, the degree of elevation of CEA level correlated with the incidence of metastatic disease (breast, stomach, and colorectum) and pretreatment CEA level assessments were of substantial value in identifying patients at greatest risk of having extensive disease before an

operative exploration.

Our results showed that abnormal preoperative CEA levels were subsequently associated with a 2.1-fold increase

in the incidence of recurrence after resection of colorectal carcinoma. Similar findings, but for shorter durations of follow-up, were noted by LoGerfo and Herter17 and by Koch et al.24 Our results also showed the customary corre¬ lation between Dukes' stage and recurrence. LoGerfo and Herter, in addition, have shown substantial differences in the CEA levels among all three Dukes' stages and were able to show within a given pathological stage that the plasma CEA level correlated in a consistent fashion with recurrence. We could not show this phenomenon, perhaps because of the small numbers in the subgroups, and we therefore support the conclusions of Chu et al25 that preresection plasma CEA levels and Dukes' stage each can be used to estimate a prognosis, that Dukes' staging is more discriminating, and that the combining of Dukes' stage and preresection CEA level does not enhance the value of Dukes' stage alone in estimating a prognosis. Although CEA level was of value before therapy in estimating the extent of colorectal carcinoma and in indicating prognosis of resectable lesions, it was most helpful after "curative" resection in identifying recur¬ rence. Recurrence developed in all three patients with persistently elevated CEA levels after "curative" resection of colorectal cancer. Mach et al1* described five patients with the same findings. We observed that a high propor¬ tion of patients (13/15) with recurrence had elevated CEA levels at the time recurrence was identified. This was previously noted by Herrera et al2" and Rieger and Wahr¬ en.20 Our observation that a transient rise in CEA level was not uncommon and that it did not indicate recurrent disease agrees with the results of Sorokin et al.27 Rittgers et al2" stressed the importance of this observation if one is using the "CEA nomogram" outlined by Martin et al29 as a guide for a second-look procedure. In 11 of the 13 patients with elevated CEA levels, the CEA level increase preceded other indications of disease by months (mean, > eight months). This is similar to the observations of many other authors.18·20·26"33 In ten patients, the CEA level was normal after resection and then became abnormal 0 to 15 months (mean, 6.3 months) before the diagnosis of recurrence. Interestingly, half of these patients had had normal prere¬ section CEA levels, which indicates that CEA level is of value in the follow-up of patients with normal preresection CEA values. On the basis of persistently elevated and increasing CEA levels, and in the absence of other evidence of persistent disease, we had three patients undergo laparotomy. All three of them had nonresectable disease. In 1974, Gunderson and Sosin14 described 75 patients who had planned reoperations at six- to nine-month intervals after resection of rectal carcinoma. In only four of the 52 patients in whom persistent rectal carcinoma was identified were the condi¬ tions converted to a disease-free status. Mackman et aJ3S treated 102 high-risk patients with colorectal cancer with 5-fluorouracil postoperatively for six to 12 months and second-look operations were performed nine to 12 months after primary resection. In only five patients in whom a tumor was identified was the condition converted to a tumor-free status by the surgery. Recent reports from Minton's29·30·32 group at Ohio State University, Columbus, suggested that these low yields of therapeutic benefit from

Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/09/2015

the second-look procedures may be improved by the use of CEA levels as an indicator of recurrence. They reported that 22 celiotomies were performed because of elevated CEA levels alone and found that the conditions of six (27%) of the patients were converted to a disease-free status. They noted that the patients with unresectable recurrences had a longer duration between elevation of the CEA level and celiotomy and thus instituted a policy of earlier celiotomy for unexplained CEA level elevation after resec¬ tion in patients with colorectal cancer. Their most recent report of 14 patients having early celiotomy indicated 11 patients (78%) with locally resectable disease. McPherson and Koch31 described 14 patients with high levels of CEA and no evidence of disease. They converted the conditions of five patients to a tumor-free status by surgical resec¬ tion. Moertel et al36 recently reported that patients with local recurrences in areas accessible to physical examina¬ tion (ie, pelvis, perineum) have a 33% incidence of elevated CEA levels ( > 5.0 ng/ml). They suggest that most intraabdominal recurrences identified by an elevated CEA level cannot be treated surgically with resection. We have described here only three patients who had laparotomies because of elevated CEA levels. As noted by Minton et al,32 the long duration between CEA level increase and laparotomy in these three patients is consis¬ tent with a low incidence of resectability. As well, we have considered the CEA level range of 0 to 5.0 as normal in the serial follow-up and this may be less sensitive than estab¬ lishing the patient's own baseline and defining "abnormal" as > 2 SDs above this level (CEA nomogram). However, in all the reports of surgical resection for recurrent disease initiated by high CEA levels, the follow-up reported has been short. Minton et al32 noted 17 resections from 19721977, but only six disease-free survivors at one year. Thus, it will be some time before it is established that patients with resectable recurrences are indeed tumor-free. A cooperative trial to examine the value of reexploration for elevated plasma CEA level is currently being implemented by the Society of Surgical Oncology, and we hope it will indicate the clinical value of analyzing CEA levels in follow-up of patients with colorectal cancer after "cura¬ tive" resection. Dr

Beatty is recipient of the McLaughlin Foundation Fellowship, Canada. was provided by Hoffmann-LaRoche Ine, Nutley, NJ.

CEA assay

References 1. Gold P, Freedman SO: Demonstration of tumor specific antigens in human colonic carcinomata by immunologic tolerance and absorption techniques. J Exp Med 121:439-462, 1965. 2. Thomson DMP, Krupey J, Freedman SO, et al: The radio-immunoassay of circulating carcinoembryonic antigen of the human digestive system. Proc Natl Acad Sci USA 64:161-167, 1969. 3. Egan ML, Lantenschleger JT, Coligan JE, et al: Radioimmune assay of

carcinoembryonic antigen. Immunochemistry 9:289-299, 1972. 4. MacSween JM, Warner NL, Bankhurst AD, et al: Carcinoembryonic antigen in whole serum. Br J Cancer 26:356-360, 1972. 5. Hansen HJ, Lance KP, Krupey J: Demonstration of an ion sensitive antigenic site on carcinoembryonic antigen using zirconyl phosphate gel,

abstracted. Clin Res 19:143, 1971. 6. Zamcheck N: Carcinoembryonic antigen: Quantitative variations in circulating levels in benign and malignant digestive tract disease, in Advances in Internal Medicine, Chicago, Year Book Medical Publisher Inc, 1974, vol 19, pp 413-433. 7. LoGerfo P, Krupey J, Hansen HJ: Demonstration of an antigen common to several varieties of neoplasia. N Engl J Med 285:138-141, 1971. 8. Papachristou D, Terz JJ, Lawrence W: Clinical evaluation of the assay

for carcinoembryonic antigen (CEA). Va Med Monthly 101:108-114, 1974. 9. Hansen HJ, Snyder JJ, Miller E, et al: Carcinoembryonic antigen (CEA) assay: A laboratory adjunct in the diagnosis and management of cancer. Hum Pathol 5:139-147, 1974. 10. Dukes CE: The classification of cancer of the rectum. J Pathol 35:323-332, 1932. 11. Neville AM: The role of the carcinoembryonic antigen (CEA) in the early detection of cancer, in Bostrom H, Larsson T, Ljungstedt N (eds): Health Control in Detection of Cancer. Stockholm, Almqvist & Wiksell, 1975, pp 44-60. 12. Troncone L, Salvo D, LaVecchia F, et al: CEA assay in the follow-up of patients with extra-gastrointestinal malignancies. Bull Cancer 63:495-504, 1976. 13. Hirai H: A collaborative clinical study of carcinoembryonic antigen in Japan. Cancer Res 37:2267-2274, 1977. 14. Steward AM, Nixon D, Zamcheck N, et al: Carcinoembryonic antigen in breast cancer patients: Serum levels and disease progress. Cancer 33:1246-1252, 1974. 15. Martin EW, Kibbey WE, DiVecchia L, et al: Carcinoembryonic antigen: Clinical and historical aspects. Cancer 37:62-81, 1976. 16. Livingstone AS, Hampson LG, Shuster J, et al: Carcinoembryonic antigen in the diagnosis and management of colorectal carcinoma: Current status. Arch Surg 109:259-264, 1974. 17. LoGerfo P, Herter FP: Carcinoembryonic antigen and prognosis in patients with colon cancer. Ann Surg 181:81-84, 1975. 18. Mach JP, Jaeger P, Bertholet MM, et al: Detection of recurrence of large-bowel carcinoma by radioimmunoassay of circulating carcinoembryonic antigen (C.E.A.). Lancet 2:535-540, 1974. 19. National Cancer Institute of Canada, American Cancer Society. A collaborative study of a test for carcinoembryonic antigen (CEA) in the sera of patients with carcinoma of the colon and rectum. Can Med Assoc J 107:25-33, 1972. 20. Rieger A, Wahren B: CEA levels at recurrence and metastases: Importance for detecting secondary disease. Scand J Gastroenterol 10:869\x=req-\ 874, 1975. 21. McCartney WH, Hoffer PB: The value of carcinoembryonic antigen (CEA) as an adjunct to the radiologic colon examination in the diagnosis of malignancy. Radiology 110:325-328, 1974. 22. Gold P, Freedman SO: Tests for carcinoembryonic antigen: Role in diagnosis and management of cancer. JAMA 234:190-192, 1975. 23. Edgington TS, Astarita RW, Plow EF: Association of an isomeric species of carcinoembryonic antigen with neoplasia of the gastrointestinal tract. N Engl J Med 293:103-107, 1975. 24. Koch M, McPherson TA, Morrish DW: Carcinoembryonic antigen: Three years' experience in a cancer clinic. Can Med Assoc J 116:769-771, 1977. 25. Chu TM, Holyoke ED, Cedermark B, et al: A long-term follow-up of CEA in resective colorectal cancer, in Fishman WH, Sell S (eds): Oncodevelopmental Gene Expression. New York, Academic Press Inc, 1976, pp 427-431. 26. Herrera MA, Chu TM, Holyoke ED: Carcinoembryonic antigen (CEA) as a prognostic and monitoring test in clinically complete resection of colorectal carcinoma. Ann Surg 181:5-9, 1976. 27. Sorokin JJ, Sugarbaker PH, Zamcheck N, et al: Serial carcinoembryonic antigen assays: Use in detection of cancer recurrence. JAMA 228:49-53, 1974. 28. Rittgers RA, Steele G, Zamcheck N, et al: Transient CEA elevations following resection of colorectal cancer: A limitation in the use of serial CEA levels as an indicator for second-look surgery, abstracted. Clin Res 25:666, 1977. 29. Martin EW, James KK, Hurtubise PE, et al: The use of CEA as an early indicator for gastrointestinal tumor recurrence and second-look procedures. Cancer 39:440-446, 1977. 30. Balz JB, Martin EW, Minton JP: CEA as an early indicator for second-look procedure in colorectal carcinoma, abstracted. Rev Surg 34:1-4, 1977. 31. McPherson TA, Koch M: Predictive value of CEA levels in 145 patients with colorectal cancer. Clin Res 25:710, 1977. 32. Minton JP, James KK, Hurtubise PE, et al: The use of serial carcinoembryonic antigen determinations to predict recurrence of carcinoma of the colon and the time for a second-look operation. Surg Gynecol Obstet 147:208-210, 1978. 33. Sugarbaker PH, Skarin AT, Zamcheck N: Patterns of serial CEA assays and their clinical use in management of colorectal cancer. J Surg Oncol 8:523-537, 1976. 34. Gunderson LL, Sosin H: Areas of failure found at reoperation (second or symptomatic look) following "curative surgery" for adenocarcinoma of the rectum. Cancer 34:1278-1292, 1974. 35. Mackman S, Ansfield FJ, Ramirez G, et al: A second look at the second-look operation in colonic cancer after the administration of fluorouracil. Am J Surg 128:763-766, 1974. 36. Moertel CG, Schutt AJ, Go VLW. Carcinoembryonic antigen test for recurrent colorectal carcinoma: Inadequacy for early detection. JAMA 239:1065-1066, 1978.

Downloaded From: http://archsurg.jamanetwork.com/ by a Michigan State University User on 06/09/2015

Clinical value of carcinoembryonic antigen: diagnosis, prognosis, and follow-up of patients with cancer.

Clinical Value of Diagnosis, Prognosis, J. David Beatty, MD; Carlos and Carcinoembryonic Antigen Follow-up Romero, MD; Peter W. of Patients Wi...
791KB Sizes 0 Downloads 0 Views