CARCINOEMBRYONIC ANTIGEN C 1 ini cal Ajip 1 i cat io n

I'AY LIANC W. Go, M D

Carcinoembryionic antigen (CEA) a glycoprotein extracted from colonic cancer tissue (P-globulin electrophoretic mobility, sedimentation coefficient 7 to 8S, and mol wt -200,000) can be detected and measured by radioimmunoassay. Clinical evaluations of CEA determination have given the following results: In health: (1) Serum CEA level is not influenced by sex, age, blood type, time of blood sampling, or family history of cancer; (2) serum CEA level is influenced by a his- tory of smoking or inflammatory disease of the bowel, lung, pancreas, and other organs (occasionally, a CEA level as high as 10 ng/ml is noted); and (3) currently, CEA positivity is defined as >2.5 ngiml, however, 5 ng/ml may be more realistic. I n cancer: (1) CEA level may be increased in primary cancer of the gastroin- testinal (GI) tract as well as in non-GI neoplasia; (2) the CEA test is not recom- mended for screening to detect early cancer; (3) serum CEA level depends on the stage of the neoplasia and usually is not influenced by the grade of differen- tiation; and (4) markedly increased (>25 ng/ml) serum CEA values are highly suggestive of metastatic cancer, particularly hepatic metastasis. I n biological fluid: The CEA or CEA-like activity can be measured in gastrointestinal secre- tions. Quantitative studies of CEA levels in such fluids may yield more informa- tion than is obtainable from studies of serum. However, this possibility needs more study at present. Therefore, the currently available CEA tests cannot re- place any of the now standard diagnostic methods for cancer detection. This use for assessment of therapy in selected patients or for following those known to be at high risk for cancer appears promising in preliminary studies, but clinical value, if any, remains to be determined.

Cancer 37:562-566, 1976.

N 1965, GOLD AND FREEDMAN REPORTED THAT I carcinomas of the colon contained an anti- gen that was absent from normal adult intesti- nal mucosa but was present in primitive ento- derm.1g.2" The authors named this material carcinoembryonic antigen (CEA) of the diges- tive system. In 1969, Thomson et al. described a radioimmunoassay for CEA in serumBo and reported positive findings in 97yo of a series of patients with cancer of the colon. So CEA was acclaimed in the medical and lay litera- ture as a tumor-specific antigen whose detec- tion in the serum may be helpful in the diag-

Presented at the meeting of the ACS-NCI National Conference on Advances in Cancel Management, Part 11: Detection and Diagnosis, Denver, CO, May 1-3, 1975.

From the Gastroenteiology Unit, Mayo Clinic and Mayo Foundation, Rochester, Minnesota.

Supported by research giant CB 23854 from the National Cancer Institute, National Institutes of Health.

Address for reprints: V. L. MI. Go, MD, Mayo Clinic, Rochester, MN 55901.

nosis of colonic cancer. Subsequently, immu- nochemicnl tests showed that this material could be detected in the serum of patients with gastrointestinal or nongastrointestinal malignancy, in extracts from gastrointestinal a n d nongastrointestinal tumors, and in tissue extract? of the fetal gastrointestinal t r a ~ t . ~ J ~ $ ~ 1 4 4 Recently several laboratories have re- ported CEA immunoreactivity in urine,7,13.21 stool,l2 14 and intestinal secretionsl~l"~'~3"45 of cancer patients and of normals.

T w o coopeiative trials of the CEA test in blood have been carried out in a number of centers in the United States and Canada. T h e National Cancer Institute of Canada and the American Cancer Society sponsored a col- laborative study of CEA tests employinq the Thomson technique,25 while the Hoffmann- LaRoche Company organized a trial utilizing tlie Hansen technique.22 Through a contract awarded by the National Cancer Institute, we have carried out a prospective evaluation of CEA test5 in a clinical setting, using reagents

562

No. 1 CARCINOEMBRYONIC ANTIGEN - Go 563

kindly supplied by Dr. Hans Hansen, Hoff- mann-LaRoche Company. Over the last 30 months, 6,700 patients have been entered into the study. From our experience and that of others in such studies and in applying the CEA test to b10od3.2"30.31,~~.*3,~~ and to gastro- iritestinal1,15,17.",4~ and urinary secretions,13.21 I shall review the present status of this test.

PHYSICOCHEMICAL PROPERTIES OF CEA

Recent progress in elucidation of the physi- cal and chemical properties of purified CEA has been reviewed extensively by Terry et al.38 CEA is a glycoprotein, extracted, and purified from primary adenocarcinoma of the colon or from colonic carcinoma metastasized to the liver. Purified CEA has a molecular weight of about 200,000, sedimentation coefficient of 7.0 to 8.OS, and beta-globulin electrophoretic mo- hili ty. Electron microscopy has shown the molecule to be of uniform size with a complex secondary structure.6 It is soluble in I hl perchloric acid and in half-saturated ammo- nium sulfate.

Chemical analysis of purified CEA shows approximately 50 to 65y0 carbohydrate, the remainder being protein. The carbohydrate portion of CEA contained fucose, mannose, galactose, N-acetyl-glucosamine, sialic acid, and tracer of N-acetyl-galactosamine. The charge heterogeneity of various CEA preparations have been attributed, in part, to variations in sialic acid content of the molecule. Amino acid sequencing of the protein backbone of CEA is not yet complete. However, the first 24 residues from the N-terminal of purified CEA, obtained from liver metastasis or isolated from serum of patients with advanced meta- static colonic cancer, are identical.*'

CE,\ RADIOIMMLINOASSAY METHOD

Several techniques for radioimmunoassay to measure CEA have been described. The m a y measures inhibition of the binding of known amounts of radiolabeled purified CEA to the specific antibody. The techniques used most often are those originally described by Thornson et al.,5O by Hansen,31 and by Egan.11 Several modifications of these methods have been validated.30e34~42 To date, the CEA-Roche test is the only one approved by the FDA for clinical use. The assay techniques differ mostly in their ways of extracting CEA from plasma

or serum and their ways of separating free from antibody-bound antigen. Some tech- niques use either serum or plasma and some use inta.ct serum or plasma (direct assay), whereas for others the test serum must be ex- tracted first with perchloric acid (indirect as- say). The reliability and comparability of the various methodologies were demonstrated in a study organized by the National Cancer In- stitute in which CEA levels of 94 blood speci- mens ohtained from different patients were tested simultaneously by different 1a.borato- ries, including those of Das, Go, Gold, Han- sen, LoCerfo, Mach, Martin, Reynoso, Todd, and Zarncheck:"4 Correlations of 82 to 907; were obtained.

CIRCULATING CEA LEVEL I N HEALTH A N D BENIGN DISORDERS

The normal values for the different assays vary among laboratories. The Thomson assay and the CEA-Roche assay used the cut-off point of 2.5 ng/ml; the Egan-Todd assay is approximately 12.5 ng/ml; and other labora- tories, including Mach and our laboratory, use a cut-off point of 5 ng/ml. Therefore, the CEA result obtained by each laboratory should always be interpreted in comparison with tlie value obtained from normal controls using the same assay procedure.

The serum or plasma CEA level is influ- enced by the history of cigarette smoking, tending to he higher in heavy smokers than in nonsmokers.2') Elevated 'CEA levels between 2.6 and 10 ng/ml and occasionally as high as 10 ng/ml occur in many patients with a va- riety of benign conditions including bron- chitis, inflammatory bowel disease, pancrea- titis, cirrhosis, and pulmonary empliysetna.2~~~ x,21.2*.33.37 CEA positivity defined as greater than 2.5 ng/ml is not helpful in the diagnosis of cancer, particularly if there is a history of heavy smoking or any of the listed underlying henign diseases.

We have found that other factors-sex, age, time of the clay when tlie blood sample was drawn, and the presence or absence of family history of cancer-do not influence the CEA level.";

CEA IN MALIGNANT DISEASES

CEA tests have limited sensitivity and speci- ficity for cancer. The initial studies utilized cases of frank gastrointestinal cancer, many

564 CANCER Janirory Supplement 197G 1'01. 97

with metastasis; the CEA levels were higher than 2.5 ng/ml in more than 75% of them. Subsequent prospective evaluations of CEA tests in gastrointestinal malignancy, particu- larly in cancer of the colon, have shown that the CEA level varies with the stage of the can- cer. Large proportions of patients with early surgical and resectable colonic cancer have normal concentrations of circulating CEA. As the cancer spreads to adjacent organs or me- tastasi7es distantly, the amount of circulating CEA and the incidence of CEA positivity in- crease. Greatly increased serum CEA values (> 25 ng/ml) are highly suggestive of meta- static cancer, and particularly hepatic metas- tasis. Similar patterns have been observed in cancer of the pancreas, stomach, lung, breast, prostate, and other solid tumors. Therefore, the use of CEA test alone for screening pur- poses is not recommended. At present, CEA tests are useful as a diagnostic adjunct when cancer is suspected, but they cannot replace any of the currently standard methods for can- cer detection. A normal CEA level does not preclude the presence of primary, metastatic, or recurrent cancer.

Comparison of concentrations of circulating CEA and other enzymes such as serum alkaline phosphatase, gamma-glutamyl transpeptidase, lactate dehydrogenase, and other serum en- zymes suggest that measurements of CEA and enzyme concentrations were complementary to each other in patients with primary and meta- static mal ignan~ies . l8 .~0~~~ Similar findings were obtained from simultaneous measurements of CEA and other tumor-associated antigens such as alpha-fetoprotein. In our experience, when alpha-fetoprotein and CEA were measured si- multaneously, 54y0 of patients with gastric carcinoma had elevations of one or both, whereas only 27% had elevated alplia-fetopro- tein alone and 2 2 7 , had elevated CEA al0ne.3~ This suggests the possibility that measure- ments of CEA and other tumor-associated an- tigens such as alpha-fetoprotein, human cho- rionic gonadotropin, or Regan isoenzymes and other enzymes might increase significantly our ability to detect advanced gastrointestinal cancer.

CEA AS BIOLOGIC MARKER FOR CANCER THERAPY

Thomson, Gold et al. first reported five cases of colonic cancer with elevated CEA levels that declined to normal in three cases after complete resection of the cancer. Such a response was confirmed by Dhar et al.,'" Lo

Gerfo et al.,31J2 Reynoso et al.,23s43 Laurence et > I I . , ~ O Mach et al.,3-' and Khoo and Macky.41 However, a drop of the CEA level to normal after surgery does not always provide a good index of completeness of surgical excision, for in some cases the cancer recurs. In a small number of cases the CEA level has fallen to normal after surgical therapy but risen again during recurrence, perhaps before there was clinical evidence of recurrence. Our experi- ence showed that during recurrence after 110-

tentially curative resection the level of CEA depends on the stage of the recurrent disease. T h e CEA level is normal in localized recur- rent disease, and an elevated CEA level sug- gests regional and distant metastasis. Our data suggest that serial determinations of serum CEA could indicate advances of recurrent can- cer in long-term follow-up of patients who had undergone curative surgery and who have a normal postoperative CEA. However, i t must be made clear that in those cases where CEA increased, the tumor recurrence was readily detectable by other means.42

Serial CEA assays have been used for moni- toring cancer chemotherapy in patients with cancer of the colon, pancreas, stomach, biliary tract,*4-.J"4G breast,5?48 cervix, body of the uterus, ovary,"fi.2' and lung,51 and in patients with neuroblastoma.4~ These studies have shown a fair correlation between effectiveness of cancer chemotherapy and serum CEA con- centration, and our findings seem to confirm these general results. However, serial determi- nations of serum CEA must be interpreted in the clinical context, for we have noted that preterminally the CEA concentration may fall rather than rise and that patients whose orig- inal cancer produced no CEA may deteriorate without producing any. Furthermore, various cancer chemotherapy agents may alter the syn- thesis and secretion of CEA without affecting the progression of the neoplasia. It is in this area of research that CEA assay as a biologic marker for following cancer chemotherapy may prove fruitful.

CEA-LIKE ACTIVITY I N SECRETIONS

CEA or CEA-like activities have been de- tected by various immunologic tests in saliva, gastrointestinal secretion, feces, urine, bron- chial washings, and other biologic fluids and secretions of different organ sy~tems.l,7.1"1~,~~. 15~7.21 .35 .36~45 With use of several intestinal per- fusion techniques, we recently have quantified the secretion of CEA-like activity in the gastro- intestinal tract of healthy human subjects.

The secretory rate was highest in colonic se- cretion, arid next highest in pancreatico-biliary secretion.17 T h e CEA-like material extracted from normal human saliva, normal gastroin- testinal secretion, and feces has chromato- graphic and immunologic properties similar to those of purified CEA extracted from liver tissue containing metastatic cancer of the colon. Whether this immunoreactive CEA ma- terial is chemically the same as purified CEA needs further study.

It has been suggested that studies of CEA- like activity in fluid secreted by hollow vis- cera such as the gastrointestinal tract, lung, and urinary bladder might lead to improve- ment in early detection of cancer. Winawer found that the CEA concentration in colonic lavage specimens from 38 patients ranged from 0 to 4450 ng/mg protein.5' Patients with cancer and polyps were at the higher end of the range, patients with ulcerative colitis in- termediate, and normal subjects at the lower

end.5' In pancreatic secretion, we found that CEA output does not differentiate patients with pancreatic disease from healthy persons; yet one-third of patients with pancreatic can- cer and 20% of those with chronic pancreatitis had increased outputs of CEA in response to cholecystokinin stimulation.9 Hall and co- workers found abnormal urinary CEA values in two-thirds of their patients with active bladder cancer.tl We have confirmed their ob- servations and further noted that the fre- quency of abnormal CEA values correlated with the extent and grade of the tumor. How- ever, urinary tract infection can produce ab- normal CEA values in patients without can- cer.

These and 0therf5-36,~g various preliminary results suggest that measurement of CEA-like activities may have some diagnostic value. However, i ts ultimate clinical application re- mains to be determined.

No. 1 CARCINOEMBRYONIC ANTIGEN - Go 565

REFERENCES 1. Bhargava, .4. K., Chen, J. F., Vincent, R. G., and

Chu, T. M.: Characterization of a glycoprotein from ascites of pancreatic carcinoma and its comparisons with colonic carcinoembryonic antigen (CFA). Pro- ceedings of the American Association for Cancer R e - search of the American Society of Clinical Oncology 16: 131(#521), 1975.

2. Booth, S. N., King, J. P. G., Leonard, J. C., and Dykes, P. W.: Serum carcinoembryonic antigen in clinical disorders. Gut 14:794-799, 1973.

3. Booth, S. N., King, J . P. G., Leonard, J. C., and Dykes, P. W.: The significance of elevation of serum carcinoembryonic antigen (CEA) levels in inflamma- tory disease of the intestine. Scand. J. Gastroenterol.

4. Burtin, P., Sahine, M. C., and Chavanel, G.: Presence of carcinoemhryonic antigen in children's colonic mucosa. Znt. J. Cancer 1072, 1972.

5 . Chu, T. M., and Nemoto, T.: Evaluation of carcinoembryonic antigen in human mammary car- cinoma. ./. Notl . Cancer Ins t . 51:1119-1122, 1973.

6. Coligan, J. E., and Todd, C. W.: Structural studies on carcinoemhryonic antigen periodate oxidation. Bio- rhemology 14:805-810, 1975.

7. Coombes, G . B., Hall, R. R., Laurence, D. J. R., and Neville, A. M.: Urinary carcinoemhryonic antigen (CE.A)-like molecules and urothelial malignancy--i\ clinical appraisal. Br . J. Cancer 31:135-142, 19175. 8. Delwiche, R., Zamcheck, N., and Marcon, N.:

Positive serum carcinoembryonic antigen (CEA) in alcoholic pancreatitis. Cancer 31:328-330, 1973. 9. DiMagno, E. P., Moertel, C. G.. and Go, V. L. W.:

Pancreatic CEA-like activity and serum CEA in health and pancreatic disease. Clin . Res. 23:248A3, 1975.

10. Dhar, P., Moore, T. I>,, Zamcheck, N., and Kup- chik, H. Z.: Carcinoemhryonic antigen (CEA) in colonic cancer: Use in pre- and postoperative diagnosis and prognosis. J A M A 221:31-35, 1972.

11. Egan, M. I., Lautenschleger, J. T., Coligan. J . E., and Todd, C. W.: Radioimmunoassay of car- cinoemhryonic antigen. Zmn~u?zochemistry 9:289-299, 1972.

9 :651-656, 1974.

12. Elias, E. G., Holyoke, E. D., and Chu, T. M.: Carcinoemhryonic antigen (CEA) in feces and plasma of normal subjects and patients with colorectal car- cinoma. Dis. Colon Rectum 17:38:41, 1974.

13. Fraser, R. A., Ravry, M. J., Segura, J. W., and Go, V. L. W.: Clinical evaluation of urine and serum urinary bladder cancer. J. Urol., in press, 1975.

14. Freed, D. L. J., and Taylor, G.: Carcinoemhry- onic antigen in faeces. Br. M e d . J. 85-87, 1972.

15. Kawanishi, H., Sell, J. E., and Pollard, H. M.: Carcinoembryonic antigen and cytology of pancreatic fluid. Gastroenterology 68: A-66/923, 1975.

16. Go, V. L. W., Ravry, M., Taylor, W. IF., Schutt, A. J.. and Moertel, C. G.: Prospective evaluation of carcinoembryonic antigen (CEA) in clinical setting. Proceedings of the American Association for Cancer Research of the American Society of Clinical On- cology 15:194(#842), 1974.

17. Go, V. L. W., Ammon, H. K., Holtermuller, K. H., Krag, E., and Phillips, S. F.: Quantification of carcinoembryonic antigen-likc activities in normal hu- man gastrointestinal secretions. Clin. Res. 22:489A, 1974.

18. Go, V. L. W.: Unpublished data. 19. Gold, P., and Freedman, S. 0.: Demonstration

of tumor-specific antigens in human colonic carcino- mata by immunological tolerance and absorption techniques. .J . E x p . Med. 121:439-462, 1965.

20. Gold, P., and Freedman, S. 0.: Specific carcino- embryonic antigens of the human digestivc system. J . E x p . M e d . 122:467-481, 1965.

21. Hall, R. R., Laurence, D. J. R., Darcy, D., Stevens, U., James, R., Roberts, S., and Munro Neville, A.: Carcinoemhryonic antigen in the urine OE patients with urothelial carcinoma. Rr. Med. J. 3:G09-611, 1972.

22. Hansen, H. J.. Snyder, J. J., Miller, E., Vande- voorde, J. P., Milelr, 0. N.. Hines, L. R., and Burns, J. J.: Carcinoemhryonic antigen (CEA) assay. A lahora- tory adjunct in the diagnosis and management of cancer. J . Hum. Pathol. 5:139-147, 1974.

23. Holyoke, D., Reynosa, G., and Chu, T. M.:

566 CANCER Jnnzinry Supplement 1976 Vol. 37

Carcinoemhryonic antigen (CEA) in patients with carcinoma of the digestive tract. A n n . Szirg. 175:559- 563, 1972.

24. Holyoke, E. D., Reynoso, G., and Chu, T. M.: Carcinoemhryonic antigen in patients with carcinoma of the digestive tract. I n Embryonic and Fetal Antigens in Cancer, vol. 2, IJSAEC Report CONF-72020R. Spring- field, VA, 1J.S. Dept. of Commerce, 1972; pp. 215-219.

25. Joint National Cancer Institute of Canada/ American Cancer Society Investigation. A Collahora- tive Study of a Test for Carcinoenihryonic Antigen (CEA) in the Sera of Patients with Carcinoma of the Colon and Rectum. Con. hled. Assor. J . 107:25-33, 1972.

26. Khoo, S. K.. and Mackay, E. V.: Carcinoemhry- onic antigen by radioimmunoassay in the detection of recurrence during long-tcrm follow-up of female genital cancer. Cnnrer 34:542-548, 1974.

27. Khoo, S. K., and Mackay, E. V.: Carcinoembry- onic antigen in cancer of the female reproductive system: Sequential levels and effects of treatment. Aust. N.Z. J. Ohstet. Gynnecol. 13:l-7, 1973.

28. Khoo, S. K. . Warner, N. L., Lie, J. T., Mackay. I . R.: Carcinoembryonic antigenic activity of tissue extracts: A quantitative study of malignant and benign neoplasms. Cirrhotic liver, normal adult and fetal organs. I n t . .J . Cancer 11:681-687, 1973.

29. Kupchik. H. Z. , and Zamcheck. N.: Carcino- embryonic antigens(s) in liver disease. 11. Isolation from human cirrhotic liver and serum and from nor- mal liver. Gnstroenterology 63:95-101, 1972.

30. Laurence, D. J. R., Stevens, U., Bettelheim. R., et al.: Evaluation of the role of plasma rarcinoemhry- onic antigen (CEA) in the diagnosis of gastrointestinal, mammary and bronchial carcinoma. Br. Med . J. 3:605- 609. 1972.

31. LoGerfo, P., Krupey, J.. and Hansen, H. J.: Demonstration of an antigen common to several varie- ties of neoplasia. Assay using zirconyl phosphate gel. N . En@. J. Aled. 285:138-141, 1971.

32. LoGcrfo, P., Krupey. J . . Herter. F., Barker, H. G . , and Hansen, H. J.: Tumor-associatcd antigen in patients with carcinoma of the colon. A m . J . Sitrg.

33. Lurie. R. R., Loewenstcin. M. S., antl Zam- check. N.: Elevated carcinoemhryonic antigen (CEA) in benign biliary ohstrnction. C l i n . Reg. 22:363A. 1974.

34. Mach, J. P., Jaeger, P. H.. Bertholet, M. M., Ruegsegger, C. H., Loosli, R. M., and Pettavel, J.: Detection of recurrence of large-bowel carcinoma hy radioimmunoassay of circulating carcinoemhryonic antigen (CEA). Lnncrt 2:535-540, 1974.

35. Martin, F., and Devant, J.: Carcinoembryonic antigen in normal human saliva. J. Nnt l . Cancer Ins t .

36. McCabe. R. P.. Knpchik, H. Z., and Zam- check, H.: Identification of carcinoemhryonic (CEA) activity in the pancreatic juice of patients with car- cinoma of the pancrras. Fed. Proc. 33:637, 1974.

37. Moore, T. L., Dhar, P., Zamcheck. N.. Keeley, A.. Gottlieh. L.. and Kupchik, H. 2.: Carcinoemhryonic an- tigen(s) in liver disease. Gastroenterology 63:88-94, 1972.

38. Moore, 1’. L., Kupchik, H. Z., Marcon, N., and %amcheck, N.: Carcinoemhryonic antigen assay in can- cer of the colon and pancreas and other digestive tract

123: 127-131, 1972.

50:1375-1371), 1973.

disorders. A m . f . Dig. Dis. 16:l-7, 1971. 39. Mulcarr, R., and LoGerfo, P.: Tumor-associated

antigen in the chemotherapy of solid tumors. 1. S w g . O n d . 4:407-417, 1972.

40. Mnnjal, D., Chawla, P. I-.. I.okich, J . J . , and Zamcheck, N.: Circulating antl tissue levels of car- cinoernhryonic antigen (CEA), phosphohexose isomcrase (PHI), gamma-gliitamyl transpeptidase (y-GTP) and lactatr dehytlrogenasc (LDH) in cancer patients. C l i n , Res. 23:341A, 1975.

41. Ravry, M., McIntire, K. R., Moertel, C. G , , Waldmann, T . .4., Schutt. A. J., and Go, V. L. W.: Brief communication-Carcinoenihryonic antigen (CE.4) and alpha-fetoprotein (aFP) in the diagnosis of gastric and colonic cancer. A comparative clinical evaluation. J. N n f l . Cnncer I n s t i t . 52:1019-1021, 1974.

42. Ravry. M., Moertel, C. G., Schutt. A. J., antl Go, V. 1.. W.: Usefulness of serial serum carcinoemhryonic antigen (CEA) determinations during anticancer ther- apy or long-term followup of gastrointestinal car- cinoma. Cnncer 34: 1230-1234, 1974.

43. Reynow, G., Chu, T. M., Holyoke, D., et al.: Carcinoemhryonic antigen in patients with different cancers. , J A M A 220:361-365, 1972.

44. Rosai. J., Tillack, T. W.. and Marchesi. V. T.: Membrane antigens of human colonic carcinoma antl non-tumoral colonic mucosa-Results obtained with a new isolation method. I t i t . J . C m r e r 10357-367, 1972.

45. Sharma, M. P., Grcgg, J . A,, McCahe. R. P., Locwenstein, M. S., Lurie, R . B., and Zamcheck, N.: Carcinoembryonic antigen, (CEA)-like activity in pan- creatic juice of patients with pancreatic cancer antl pancreatitis. Gnstroenterology 66:A-122/776. 1974.

46. Skarin, A. T.. Delwiche. R.. Zamcheck, N., Lokich, J. I., and Frei, E.. 111. Carcinoembryonic an- tigen-Clinical correlaion with chemotherapy for metastatic gastrointestinal cancer. Cnncer 33: 1239- 1245, 1974.

47. Steele. I,.. Cooper, E. H., Mackay, A. M., Losow- sky, M. S., and Golighyr, J . C.: Comhination of carcinoemhryonic antigen and gamma glntamyl trans- peptidasc in the study of the cvolution of colorectal cancer. Br. J . Cnncer 30:319-324, 1974.

48. Steward, A . M., Nixon, D. W., Zamcheck, N., and Airenherg, A,: Carcinoemhryonic antigen in breast cancer patients-Serum levels and disease progress. Cnnrer 33:1246-1252, 1974.

49. Terry. W. D., Henkart, P. A.. Coligan, J. E., and Todd, C. W.: Carcinocmhryonic antigen-Char- acterirstion and clinical application. Transplant. Rev . 20:100-129, 1974.

50. Thomson, D. M. P., Krupey, J., Freedman, S. O., antl Gold, P.: The radioimmunoassay of circulating carcinoemhryonic antigen of the human digestive sys- tem. Proc. N a f l . A m d . Sri. IJSA 64:161-167, 1969.

51. Vincent, R. G., and Chu. T. M.: Carcinoemhry- onic antigen in patients with carcinoma of the lung. J. Thornc. Cordiovnsc. S w g . 66:320-328, 1973.

52. Winawer, S. J.: Detection of early colon cancer. Nationnl 1.arge Bowel Cancer Project Newsletter 3:7-9, 1975.

53. Zamcheck, N.: Carcinocmhryonic antigen-@an- titative variations in circulating levels in benign and malignant digestive tract diseases. In Advances in In- ternal Medicine, vol. 19. Chicago, Year Rook Medical, 1974; pp. 413--433.