- Email: [email protected]
Carcinoembryonic antigen in intraepithelial neoplasia of the uterine cervix
Carkmembryonic antigen in intraepithelial of the uterine cervix J.
R.
W. J.
VAN
R. C.
NAGELL.
MEEKER, PARKER,
RAFIAH
JR.,
VIOLA
McCOLLUM. IYentwfry
M.D.
M.D. M.D.
JR.,
KASHMIRI.
Lt?xinglon,
neoplasia
M.S. B.A.
Carcinoembryonic antigen (CEA) was elevated (>2.5 ng. per milliliter) in 29 of IO0 with cervical intraepithelial neoplasia (CIN). CEA concentration was related to the amount of‘ intraepithelial neoplasia and to the presence of glandular extension. Lymphoplasmacytic infiltration of tumor cells was unrelated to CEA levels. CEA va1ue.s returned to normal within 8 weeks following surgery in 77 per cent of patients. A jwkstently elevated (>5.0 ng. per milliliter) plasma CEA value following conization u~as associated with residual GIN in the cervix. These results suggest that sequential CEA determinations may be of value in the follow-up of those cervical cancer patients who initially huvr hug-h plasma antigen leuels. patients
Material and methods
SINCE ITS original description by Gold and Freedman”‘ ’ in 1965. carcinoembryonic antigen (CEA) has been identified in the tumors and plasma of patients with a variety of invasive cancers.“* i3* I43 2o In a previous investigation from this institution,” 80 per cent of patients with invasive squamous cell carcinoma of the cervix were found to have elevated (>2.5 ng. per milliliter) plasma CEA levels. The incidence of elevated plasma CEA was directly related to the stage of disease and to the presence of vessel invasion by tumor cells. Very little has been reported, however, concerning the presence of this antigen in patients with noninvasive cervical lesions. The present prospective study was undertaken to determine (1) the incidence of elevated plasma CEA in patients with cervical intraepithelial neoplasia (CIN), (2) the relationship of CEA elevation to the extent and histomorphologic characteristics of cervical intraepithelial tumors, and (3) the effect of surgical removal of these tumors on plasma CEA. From the Depurtments Su~gryy, und P&olo~, Cmtw Rec&vd,for Accvptud
publication
Clinical material. From January, 1954, through April, 1975, 100 patients with CIN underwent cervical conization on the Gynecologic Oncology Service at the University of Kentucky Medical Center. Clinical cvaluation of each patient was based upon physical. radiologic, and laboratory findings prior to surgery. Data were collected concerning age, parity. height, weight, and history of cigarette smoking for all patients. Plasma samples for CEA were dravtn on the day prior to conization, 2 days fc>llo~~ing surgery. and X weeks thereafter. Eight to 20 sections (rut,an 11 .O) of each conization specimen were cut in orrlcr to define the extent of intraepithelial neoplasia. Eac,h histologic section was reviewed and classified accctrding to the following criteria: cell type, extension of moplasia to endocervical glands, benign adenomatoid or endocervital glandular hyperplasia (Fig. 1). d~keratosis or abnormal keratin production (Fig. 2). and lymphoplasmacytic infiltration. In addition, cervical epithelial abnormalities were further subdivided Into moderate dyspiasia, severe dysplasia, or carcinoma in situ according to the morphologic criteria report4 by Johnson and colleagues.x A group of 176 healthy vohmtccrs and 95 patients with benign gynecologic rliwasc studied previously at this institution by van Na;gcll and associates” served as control populations lirt- this investigation.
of Obsietrics and Gynecology, L+ziniwrsity of Kentucky Medical October
1, 1975.
Drcembe1- 12, 1975.
Hepint requests: Dr. John R. uan Nag&, Department of Obstetrics and Gynecology, Kentucky, Lexingioton, Kentucky 40506.
Jr., University
of 105
106
van Nagell et al.
Fig. 1. Benign glandular (adenomatoid) hyperplasia characterized by prominent small mucus-producing, back-to-back endocervical glands with scattered mononuclear inflammatory cells. (Hematoxylin and eosin: X 100.)
Methods Plasma CEA levels were determined as previously described by LoGerfo and associates”’ and Meeker and associates.” Briefly, 7 ml. of blood were drawn into tubes containing sodium edetate (Na-EDTA) and the plasma was withdrawn following centrifugation. Then 1 ml. aliquots of plasma were extracted with 0.6M perchloric acid and the samples were analyzed immediately by radioimmunoassay for CEA by the Hansen Z-gel procedure. Using this method, Chu and Reynoso’ were able to distinguish reliably between plasma levels of 0 and 0.5 ng. per milliliter. CEA reagents were kindly provided by Hoffmann-La Roche Laboratories, Nutley, New Jersey. A plasma CEA value of 2.5 ng. per milliliter was taken as the upper limit of normal.
Results Plasma CEA values in the patients studied are presented in Table I. Absolute plasma levels of CEA varied from 0 to 25 ng. per milliliter (mean 1.7 ng. per milliliter). Twenty-nine per cent of patients with CIN had elevated (>2.5 ng. per milliliter) levels of plasma CEA. This was significantly (p < 0.002) higher than that in normal healthy volunteers. Of those patients with abnormally elevated CEA titers, nine (27.6 per cent) had values above 5.0 ng. per milliliter and one
Fig. 2. Dyskeratosis-keratin-production by a normally nonkeratizing stratified squamous epithelium. (Hernatox\lin and eosin; X 100.)
patient had a plasma CEA value greater than 10.0 ng. per milliliter. Patients with CIN had higher plasma CEA values than those with benign gynecologic disease, but this difference was not significant (p < 0.16) in the numbers of patients studied. The clinical characteristics of patients with elevated plasma CEA values did not differ from those with normal CEA values (Table II). This included the presence of cigarette smoking which was essentially the same in both groups. Histomorphologic findings in conization specimens are illustrated in Table III. Abnormally elevated plasma CEA levels were significantly (p < 0.001) correlated with the extent of CIN as manifested by the number of sections involved and the extension of tumor into the endocervical glands. Likewise, the production of keratin by cervical epithelium was associated with elevated plasma CEA titers. Both lymphoplasmacytic infiltration and benign adenomatoid hyperplasia of the endocervical glands were unrelated to the level of CEA. The incidence of abnormally elevated CEA increased with the progrrs-
CEA in cervicalintraepithelial neoplasia 107
Volume 126 Number 1
Table
I.
Carcinoembryonic
antigen
values in patients
studied
--___ CE.4 dues
No. of patients
Healthy volunteers Benign gynecologic disease Cervical intraepithelial neoplasia
176 95
157 (89%) 78 (82%) 71 (71%)
100
Table II. Clinical characteristics of patients plasma carcinoembryonic antigen values
related
to
Plasma CEA lemds 12.5
Patients Age (v-4
Height (in.) Weight (lb.) Parity Smol&g (packs per day)
sion of the histologic epithelial moderate dysplasia to carcinoma
ng.lml.
>2.5
71 32 62.5
ng,iml.
3.4
29 35 63 143 3.8
0.8
0.8
abnormalities in situ (Table
from IV).
145
<2.5
FOllOW-Up
Twenty-six of the 29 patients with abnormal plasma CE,4 levels returned for repeat CEA determinations 2 days and 8 weeks following conization. Plasma CEA titers returned to normal by the second postoperative day in 23 per cent of the patients and by the eighth week following surgery in 77 per cent of the patients. Hvsterectomy c\‘as performed 8 weeks after conization in 19 patients. Six patients had residual intraepithelial carcinoma in the uterus following conization. Plasma CEA determinations 8 weeks postconization in patients with residual uterine carcinoma in situ ranged from 2.0 to 1 1.7 ng. per milliliter (mean 5.4 ng. per milliliter). In contrast, patients with no evidence of residual disease had plasma CEA values from 0 to 4.8 (mean 2.0 ng. per milliliter). Three patients had plasma CEA values greater than 5.0 ng. per milliliter 8 weeks following conization, and all three had residual intraepithelial carcinoma in the uterine specimen.
Comment Intraepithelial neoplasia had been considered by many investigators to be an intermediate step in the biologic progression toward invasive malignancy, particularly in the uterine cervix. Neoplastic change in cervical epithelium has been recognized colposcopitally, histologically, and even genetically.“* “-” Richart and Wilbanks15 and Granberg,’ for example, noted an increase in abnormal chromosome numbers in cervical
(r&g. /ml.)
2.5-4.9
5.0-10.1)
17 (10%)
2 (l’dj
20 (20%)
8 (8%i
13(14%)
>lO 0 (0%)
4 (4%‘)
0 (0%j --
1 (I%)
epithelium cells progressing from in situ tc, invasive cervical carcinoma. Very little has been reported, however, concerning the expression of antigenit determinants in intraepithelial cancer. Carcinoembryonic antigen was originall\~ thought to be a tumor antigen specificall! assoc.iated with adenocarcinoma of the colon.‘. ’ Howevet-, it has more recently been identified in the plasma of’ normal individuals.’ The difference in CEA levels between benign and malignant tissues appears. therefore. r~ he cluantitative rather than qualitative. In addition. L.oGerfo and colleagues” have demonstrated an antigenic site on CL4 common to both entoclermall~ and nonentodermally derived tissues. This has been suhsl,lntiated h) the identification of CEA in the tumors an(l plasma of patients with invasive cervical and ovarian cancer which has physicochemical and immunologic properties similar to colonic cancer CEX.‘2. “I. ‘I The findings of the present study demonstrate the lack of specificity of CEA as a diagnostic method in patients with CIN. Only 29 per cent of these patients had elevated plasma CEA values. This was significantly higher than that in normal control patients, but lower
than that in patients with iiivkrsi!,c 85ntxologic malignancy. The correlation of plasma CEA to the extent of cervical intraepithelial disease is consistent cvith similar findings in patients with invasive cervical cancer.“. lH Elevation of plasma CE.4 seetns related IO the amount of tumor present whether the tumor is intraepithelial or has invaded the cervical stroma. Histomorphologic criteria of the cervix were of limited value in predicting CEA elevation. The association of increased plasma CEA levels with the production of keratin by the cervical epithelium (dyskeratosis) may simply indicate a positive relationship between CL4 expression and chronic cervical inflammation. Plasma CEA has previously been reported to he increased in certain inflammatory- diseases of the bowel and pelvis.‘. “3 ” The lack of correlation between Iymphoplasmacytic infiltration and CEA elevation is not surprising since this nonspecific finding was present in 78 per cent of’ all sections reviewed. Analysis of follow-rrp data indicate that complete
van Nagell et al.
108
Table
III. Histomorphologic
characteristics
of cone specimens Plasma CEA lnds 1
<2.5 ne.iml.
I
A52/791 140 7%) -l-. .-- ,-.. ,_,
Sections involved with tXN Glandular involvement with ClN L-P infiltrate Adenomatoid hyperplasia Dyskeratosis
25171 53171 5171 18/71
(35.2%) (74.6%) (7.0%) (25.4%)
I
>2.5 np. lml. 21 l/338 21129 2WPQ 3/29 17129 ^-.
--
(62.4%) (72.4%;) (86.2%~ (10.7%) (58.6%) \--.-
,“,
Siffnificancc p < 0.001 p < 0.001 NS* NS p < 0.01
*NS = Nonsignificant Table IV. Carcinoembryonic antigen the severity of cervical intraepithelial
Moderate dysplasia Severe dysplasia Carcinoma in situ Total
26
19 55 100
values related neoplasia
20 (76.9%) 37(67.3%)
6 (23.1%) 5 (26.3%) fi (32.7%)
71(71%)
29 (20%)
14 (73.7%)
to
excision of intraepithelial cancer was associated with a return to normal plasma CEA titers within 8 weeks in over three fourths of the cases. This is similar to previous findings in patients with cervical, ovarian, and colonic carcinomas7, ‘2 “, *’ The use of CEA as an aid in determining residual in situ carcinoma in the cervix following cervical conization is potentially important. Conization is curative in approximately 80 per cent of patients with intraepithedial cervical carcinoma, but it is often difficult to predict
residual cervical or uterine disease by examination of the conization specimen. Although a persistently elevated plasma CEA (>5.0 ng. per milliliter) following conization was uniformly associated with residual in situ carcinoma in the cervix, the numbers of patients with this finding was small. Consequently, this observation will have to be evaluated in larger numbers of patients. Further studies concerning tissue localization and metabolism of CEA in patients with gynecologic maiignancies are needed. However, these results and those of other investigations’* I6 suggest that sequential CEA determinations may be of value in the follow-up of those cervical cancer patients who have high initial plasma or tumor levels of antigen. The authors would like to express their thanks to Dr. Hans Hansen and Hoffmann-La Roche, Inc., for supplying reagents for the radioimmunoassay, and to John Haley, Ph.D., of the Department of Behavioral Science, University of Kentucky Medical Center, for statistical evaluation of the data.
REFERENCES
1. Chu, T. M., and Reynoso, G.: Evaluation of a new radioimmunoassay method for carcinoembryonic antigen in plasma with the use of Zirconyl phosphate gel, Clin. Chem. 18: 918, 1972. 2. Chu, T. M., and Reynoso, G.: Demonstration of carcinoembryonic antigen 238: 152, 1972. 3.
4.
5.
6.
7.
in normal
human
plasma,
8.
9.
Nature
DiSaia, P. J., Haverback, B. J., Dyce, B. J., and Morrow, C. P.: Carcinoembryonic antigen in patients with gynecologic malignancies, AM. J. OBSTET. GYNECOL. 121: 159, 1975. Gold, P., and Freedman, S. 0.: Demonstration of tumor specific antigens in human colonic carcinomata by immunological tolerance and absorption techniques, J. Exp. Med. 121: 439, 1965. Gold, P., and Freedman, S. 0.: Specific carcinoembryonic antigens of the human digestive system, J. Exp. Med. 121: 467, 1965. Granberg, I.: Chromosomes in preinvasive, microinvasive, and invasive cervical carcinoma, Hereditas 68: 165, 1971. Holyoke, D., Reynoso, G., and Chu, T. M.: Carcinoem-
10.
11.
12.
13.
bryonic antigen in patients with carcinoma of the digestive tract, Ann. Surg. 176: 559, 1972. Johnson, L. D., Easterday, C. L., Gore, II., and Hertig, A. T.: The histogenesis of carcinoma in situ of the uterine cervix, Cancer 17: 213, 1964. Khoo, S. K., and MacKay, E. V.: Carcinoembryonic antigen in cancer of the female reproductive systemsequential levels and effects of treatment. Aust. N. Z. J. Obstet. Gynecol. 13: 1, 1973. LoGerfo, P., Krupey, J.. and Hansen, H. J.: Demonstration of an antigen common to several varieties of neoplasia-assay using Zirconyl phosphate gel, N. Engl. J. Med. 285: 138, 1971. Meeker, W. R., Kashmiri, K., Hunter, L., Clapp, W., and Griffen, W. 0.: Clinical evaluation of carcinoembryonic antigen test, Arch. Surg. 107: 266, 1973. Pletsch, Q. A., van Nagell, J. R., Jr., and Goldenberg, D. M.: Characterization of carcinoembryonic antigen in the plasma and tumors of patients with ovarian cancer. Proc. Am. Assoc. Cancer Res. 15: 146, 1974. Reynoso, G., Chu, T. M., Guinan, P., and Murphy, G. P.: Carcinoembryonic antigen in patients with tumors of the urogenital tract, Cancer 30: 1, 1972.
Volume Number
126 L
14. Reynoso, G., Chu, T. M., Holyoke, D., Cohen, E., Nemoto, T., Wang, J. J.. Chuang, J., Cuinan, P., and Murphy, G. P.: Carcinoembryonic antigen in patients with different cancers, J. A. M. A. 220: 361, 1972. 15. Richart, R., and Wilhanks, G.: The chromosomes of human intraepitheliaf neoplasia: Report of cases of cervical intraepithelial neoplasia and review, Cancer Res. 26: 60, 1966. 16. Staff, A., and Mattingly, R. F.: Colposcopic diagnosis of cervical neoplasia, Obstet. Gynecol. 41: 168, 1973. 17. Stanley, M. A., and Kirkland, J. A.: Chromosome and histologic patterns of preinvasive lesions of the cervix, Acta Cytol. 19: 142. 1975. 18. ‘Townsend, D. E.. Ostergard, D. R., Mishell, D. R., and Hirose, F. M.: Abnormal papanicolaou smearsevaluation by colposcopy, biopsies and endocervical curettage, AM. J. ORsxr. GYSECOL. 108: 429, 1970.
CE3
19.
in cervical
van Nagell, j. R., Jr., Meeker,
intraepithelial
neopiasia
109
W. R., Parker, J. (..:.)I . . and Harralson. J. D.: Carcinoembryonic antigen in patients with gynecologic malignancy, Cancer 35: 1.772. 197.5. 20. van Nagell, J. R., Jr., Pletsch, (r. A., and Goldenberg, 1). M.: A study of cyst fluid and plasma (.arcillormhr~r)ni~ antigen in patients with rvstir ovariar~ neoplasn~s. t :itntet Res. 35: 1433. 1975. 21. van Nagell, J. R.. Jr.. Pletsch, v. A., and t;o\dcnhcrg. 1). M.: Carcinoemhryrmic antigen in squamott< t eli carcinoma of. the cervix, Gvnecol. Invest. 6: 40. 1!)7.?. 22. Zamcheck. N., Moore, T. I.., Dhar. P.. and Kupcltik. H.: Immunologic diagnosis and prognosis of’ huntan digestive cancer-carcinoembryonic antigens. S. Enql. 1, hlctl. 286: 83. 1972.
R.
W. J.
VAN
R. C.
NAGELL.
MEEKER, PARKER,
RAFIAH
JR.,
VIOLA
McCOLLUM. IYentwfry
M.D.
M.D. M.D.
JR.,
KASHMIRI.
Lt?xinglon,
neoplasia
M.S. B.A.
Carcinoembryonic antigen (CEA) was elevated (>2.5 ng. per milliliter) in 29 of IO0 with cervical intraepithelial neoplasia (CIN). CEA concentration was related to the amount of‘ intraepithelial neoplasia and to the presence of glandular extension. Lymphoplasmacytic infiltration of tumor cells was unrelated to CEA levels. CEA va1ue.s returned to normal within 8 weeks following surgery in 77 per cent of patients. A jwkstently elevated (>5.0 ng. per milliliter) plasma CEA value following conization u~as associated with residual GIN in the cervix. These results suggest that sequential CEA determinations may be of value in the follow-up of those cervical cancer patients who initially huvr hug-h plasma antigen leuels. patients
Material and methods
SINCE ITS original description by Gold and Freedman”‘ ’ in 1965. carcinoembryonic antigen (CEA) has been identified in the tumors and plasma of patients with a variety of invasive cancers.“* i3* I43 2o In a previous investigation from this institution,” 80 per cent of patients with invasive squamous cell carcinoma of the cervix were found to have elevated (>2.5 ng. per milliliter) plasma CEA levels. The incidence of elevated plasma CEA was directly related to the stage of disease and to the presence of vessel invasion by tumor cells. Very little has been reported, however, concerning the presence of this antigen in patients with noninvasive cervical lesions. The present prospective study was undertaken to determine (1) the incidence of elevated plasma CEA in patients with cervical intraepithelial neoplasia (CIN), (2) the relationship of CEA elevation to the extent and histomorphologic characteristics of cervical intraepithelial tumors, and (3) the effect of surgical removal of these tumors on plasma CEA. From the Depurtments Su~gryy, und P&olo~, Cmtw Rec&vd,for Accvptud
publication
Clinical material. From January, 1954, through April, 1975, 100 patients with CIN underwent cervical conization on the Gynecologic Oncology Service at the University of Kentucky Medical Center. Clinical cvaluation of each patient was based upon physical. radiologic, and laboratory findings prior to surgery. Data were collected concerning age, parity. height, weight, and history of cigarette smoking for all patients. Plasma samples for CEA were dravtn on the day prior to conization, 2 days fc>llo~~ing surgery. and X weeks thereafter. Eight to 20 sections (rut,an 11 .O) of each conization specimen were cut in orrlcr to define the extent of intraepithelial neoplasia. Eac,h histologic section was reviewed and classified accctrding to the following criteria: cell type, extension of moplasia to endocervical glands, benign adenomatoid or endocervital glandular hyperplasia (Fig. 1). d~keratosis or abnormal keratin production (Fig. 2). and lymphoplasmacytic infiltration. In addition, cervical epithelial abnormalities were further subdivided Into moderate dyspiasia, severe dysplasia, or carcinoma in situ according to the morphologic criteria report4 by Johnson and colleagues.x A group of 176 healthy vohmtccrs and 95 patients with benign gynecologic rliwasc studied previously at this institution by van Na;gcll and associates” served as control populations lirt- this investigation.
of Obsietrics and Gynecology, L+ziniwrsity of Kentucky Medical October
1, 1975.
Drcembe1- 12, 1975.
Hepint requests: Dr. John R. uan Nag&, Department of Obstetrics and Gynecology, Kentucky, Lexingioton, Kentucky 40506.
Jr., University
of 105
106
van Nagell et al.
Fig. 1. Benign glandular (adenomatoid) hyperplasia characterized by prominent small mucus-producing, back-to-back endocervical glands with scattered mononuclear inflammatory cells. (Hematoxylin and eosin: X 100.)
Methods Plasma CEA levels were determined as previously described by LoGerfo and associates”’ and Meeker and associates.” Briefly, 7 ml. of blood were drawn into tubes containing sodium edetate (Na-EDTA) and the plasma was withdrawn following centrifugation. Then 1 ml. aliquots of plasma were extracted with 0.6M perchloric acid and the samples were analyzed immediately by radioimmunoassay for CEA by the Hansen Z-gel procedure. Using this method, Chu and Reynoso’ were able to distinguish reliably between plasma levels of 0 and 0.5 ng. per milliliter. CEA reagents were kindly provided by Hoffmann-La Roche Laboratories, Nutley, New Jersey. A plasma CEA value of 2.5 ng. per milliliter was taken as the upper limit of normal.
Results Plasma CEA values in the patients studied are presented in Table I. Absolute plasma levels of CEA varied from 0 to 25 ng. per milliliter (mean 1.7 ng. per milliliter). Twenty-nine per cent of patients with CIN had elevated (>2.5 ng. per milliliter) levels of plasma CEA. This was significantly (p < 0.002) higher than that in normal healthy volunteers. Of those patients with abnormally elevated CEA titers, nine (27.6 per cent) had values above 5.0 ng. per milliliter and one
Fig. 2. Dyskeratosis-keratin-production by a normally nonkeratizing stratified squamous epithelium. (Hernatox\lin and eosin; X 100.)
patient had a plasma CEA value greater than 10.0 ng. per milliliter. Patients with CIN had higher plasma CEA values than those with benign gynecologic disease, but this difference was not significant (p < 0.16) in the numbers of patients studied. The clinical characteristics of patients with elevated plasma CEA values did not differ from those with normal CEA values (Table II). This included the presence of cigarette smoking which was essentially the same in both groups. Histomorphologic findings in conization specimens are illustrated in Table III. Abnormally elevated plasma CEA levels were significantly (p < 0.001) correlated with the extent of CIN as manifested by the number of sections involved and the extension of tumor into the endocervical glands. Likewise, the production of keratin by cervical epithelium was associated with elevated plasma CEA titers. Both lymphoplasmacytic infiltration and benign adenomatoid hyperplasia of the endocervical glands were unrelated to the level of CEA. The incidence of abnormally elevated CEA increased with the progrrs-
CEA in cervicalintraepithelial neoplasia 107
Volume 126 Number 1
Table
I.
Carcinoembryonic
antigen
values in patients
studied
--___ CE.4 dues
No. of patients
Healthy volunteers Benign gynecologic disease Cervical intraepithelial neoplasia
176 95
157 (89%) 78 (82%) 71 (71%)
100
Table II. Clinical characteristics of patients plasma carcinoembryonic antigen values
related
to
Plasma CEA lemds 12.5
Patients Age (v-4
Height (in.) Weight (lb.) Parity Smol&g (packs per day)
sion of the histologic epithelial moderate dysplasia to carcinoma
ng.lml.
>2.5
71 32 62.5
ng,iml.
3.4
29 35 63 143 3.8
0.8
0.8
abnormalities in situ (Table
from IV).
145
<2.5
FOllOW-Up
Twenty-six of the 29 patients with abnormal plasma CE,4 levels returned for repeat CEA determinations 2 days and 8 weeks following conization. Plasma CEA titers returned to normal by the second postoperative day in 23 per cent of the patients and by the eighth week following surgery in 77 per cent of the patients. Hvsterectomy c\‘as performed 8 weeks after conization in 19 patients. Six patients had residual intraepithelial carcinoma in the uterus following conization. Plasma CEA determinations 8 weeks postconization in patients with residual uterine carcinoma in situ ranged from 2.0 to 1 1.7 ng. per milliliter (mean 5.4 ng. per milliliter). In contrast, patients with no evidence of residual disease had plasma CEA values from 0 to 4.8 (mean 2.0 ng. per milliliter). Three patients had plasma CEA values greater than 5.0 ng. per milliliter 8 weeks following conization, and all three had residual intraepithelial carcinoma in the uterine specimen.
Comment Intraepithelial neoplasia had been considered by many investigators to be an intermediate step in the biologic progression toward invasive malignancy, particularly in the uterine cervix. Neoplastic change in cervical epithelium has been recognized colposcopitally, histologically, and even genetically.“* “-” Richart and Wilbanks15 and Granberg,’ for example, noted an increase in abnormal chromosome numbers in cervical
(r&g. /ml.)
2.5-4.9
5.0-10.1)
17 (10%)
2 (l’dj
20 (20%)
8 (8%i
13(14%)
>lO 0 (0%)
4 (4%‘)
0 (0%j --
1 (I%)
epithelium cells progressing from in situ tc, invasive cervical carcinoma. Very little has been reported, however, concerning the expression of antigenit determinants in intraepithelial cancer. Carcinoembryonic antigen was originall\~ thought to be a tumor antigen specificall! assoc.iated with adenocarcinoma of the colon.‘. ’ Howevet-, it has more recently been identified in the plasma of’ normal individuals.’ The difference in CEA levels between benign and malignant tissues appears. therefore. r~ he cluantitative rather than qualitative. In addition. L.oGerfo and colleagues” have demonstrated an antigenic site on CL4 common to both entoclermall~ and nonentodermally derived tissues. This has been suhsl,lntiated h) the identification of CEA in the tumors an(l plasma of patients with invasive cervical and ovarian cancer which has physicochemical and immunologic properties similar to colonic cancer CEX.‘2. “I. ‘I The findings of the present study demonstrate the lack of specificity of CEA as a diagnostic method in patients with CIN. Only 29 per cent of these patients had elevated plasma CEA values. This was significantly higher than that in normal control patients, but lower
than that in patients with iiivkrsi!,c 85ntxologic malignancy. The correlation of plasma CEA to the extent of cervical intraepithelial disease is consistent cvith similar findings in patients with invasive cervical cancer.“. lH Elevation of plasma CE.4 seetns related IO the amount of tumor present whether the tumor is intraepithelial or has invaded the cervical stroma. Histomorphologic criteria of the cervix were of limited value in predicting CEA elevation. The association of increased plasma CEA levels with the production of keratin by the cervical epithelium (dyskeratosis) may simply indicate a positive relationship between CL4 expression and chronic cervical inflammation. Plasma CEA has previously been reported to he increased in certain inflammatory- diseases of the bowel and pelvis.‘. “3 ” The lack of correlation between Iymphoplasmacytic infiltration and CEA elevation is not surprising since this nonspecific finding was present in 78 per cent of’ all sections reviewed. Analysis of follow-rrp data indicate that complete
van Nagell et al.
108
Table
III. Histomorphologic
characteristics
of cone specimens Plasma CEA lnds 1
<2.5 ne.iml.
I
A52/791 140 7%) -l-. .-- ,-.. ,_,
Sections involved with tXN Glandular involvement with ClN L-P infiltrate Adenomatoid hyperplasia Dyskeratosis
25171 53171 5171 18/71
(35.2%) (74.6%) (7.0%) (25.4%)
I
>2.5 np. lml. 21 l/338 21129 2WPQ 3/29 17129 ^-.
--
(62.4%) (72.4%;) (86.2%~ (10.7%) (58.6%) \--.-
,“,
Siffnificancc p < 0.001 p < 0.001 NS* NS p < 0.01
*NS = Nonsignificant Table IV. Carcinoembryonic antigen the severity of cervical intraepithelial
Moderate dysplasia Severe dysplasia Carcinoma in situ Total
26
19 55 100
values related neoplasia
20 (76.9%) 37(67.3%)
6 (23.1%) 5 (26.3%) fi (32.7%)
71(71%)
29 (20%)
14 (73.7%)
to
excision of intraepithelial cancer was associated with a return to normal plasma CEA titers within 8 weeks in over three fourths of the cases. This is similar to previous findings in patients with cervical, ovarian, and colonic carcinomas7, ‘2 “, *’ The use of CEA as an aid in determining residual in situ carcinoma in the cervix following cervical conization is potentially important. Conization is curative in approximately 80 per cent of patients with intraepithedial cervical carcinoma, but it is often difficult to predict
residual cervical or uterine disease by examination of the conization specimen. Although a persistently elevated plasma CEA (>5.0 ng. per milliliter) following conization was uniformly associated with residual in situ carcinoma in the cervix, the numbers of patients with this finding was small. Consequently, this observation will have to be evaluated in larger numbers of patients. Further studies concerning tissue localization and metabolism of CEA in patients with gynecologic maiignancies are needed. However, these results and those of other investigations’* I6 suggest that sequential CEA determinations may be of value in the follow-up of those cervical cancer patients who have high initial plasma or tumor levels of antigen. The authors would like to express their thanks to Dr. Hans Hansen and Hoffmann-La Roche, Inc., for supplying reagents for the radioimmunoassay, and to John Haley, Ph.D., of the Department of Behavioral Science, University of Kentucky Medical Center, for statistical evaluation of the data.
REFERENCES
1. Chu, T. M., and Reynoso, G.: Evaluation of a new radioimmunoassay method for carcinoembryonic antigen in plasma with the use of Zirconyl phosphate gel, Clin. Chem. 18: 918, 1972. 2. Chu, T. M., and Reynoso, G.: Demonstration of carcinoembryonic antigen 238: 152, 1972. 3.
4.
5.
6.
7.
in normal
human
plasma,
8.
9.
Nature
DiSaia, P. J., Haverback, B. J., Dyce, B. J., and Morrow, C. P.: Carcinoembryonic antigen in patients with gynecologic malignancies, AM. J. OBSTET. GYNECOL. 121: 159, 1975. Gold, P., and Freedman, S. 0.: Demonstration of tumor specific antigens in human colonic carcinomata by immunological tolerance and absorption techniques, J. Exp. Med. 121: 439, 1965. Gold, P., and Freedman, S. 0.: Specific carcinoembryonic antigens of the human digestive system, J. Exp. Med. 121: 467, 1965. Granberg, I.: Chromosomes in preinvasive, microinvasive, and invasive cervical carcinoma, Hereditas 68: 165, 1971. Holyoke, D., Reynoso, G., and Chu, T. M.: Carcinoem-
10.
11.
12.
13.
bryonic antigen in patients with carcinoma of the digestive tract, Ann. Surg. 176: 559, 1972. Johnson, L. D., Easterday, C. L., Gore, II., and Hertig, A. T.: The histogenesis of carcinoma in situ of the uterine cervix, Cancer 17: 213, 1964. Khoo, S. K., and MacKay, E. V.: Carcinoembryonic antigen in cancer of the female reproductive systemsequential levels and effects of treatment. Aust. N. Z. J. Obstet. Gynecol. 13: 1, 1973. LoGerfo, P., Krupey, J.. and Hansen, H. J.: Demonstration of an antigen common to several varieties of neoplasia-assay using Zirconyl phosphate gel, N. Engl. J. Med. 285: 138, 1971. Meeker, W. R., Kashmiri, K., Hunter, L., Clapp, W., and Griffen, W. 0.: Clinical evaluation of carcinoembryonic antigen test, Arch. Surg. 107: 266, 1973. Pletsch, Q. A., van Nagell, J. R., Jr., and Goldenberg, D. M.: Characterization of carcinoembryonic antigen in the plasma and tumors of patients with ovarian cancer. Proc. Am. Assoc. Cancer Res. 15: 146, 1974. Reynoso, G., Chu, T. M., Guinan, P., and Murphy, G. P.: Carcinoembryonic antigen in patients with tumors of the urogenital tract, Cancer 30: 1, 1972.
Volume Number
126 L
14. Reynoso, G., Chu, T. M., Holyoke, D., Cohen, E., Nemoto, T., Wang, J. J.. Chuang, J., Cuinan, P., and Murphy, G. P.: Carcinoembryonic antigen in patients with different cancers, J. A. M. A. 220: 361, 1972. 15. Richart, R., and Wilhanks, G.: The chromosomes of human intraepitheliaf neoplasia: Report of cases of cervical intraepithelial neoplasia and review, Cancer Res. 26: 60, 1966. 16. Staff, A., and Mattingly, R. F.: Colposcopic diagnosis of cervical neoplasia, Obstet. Gynecol. 41: 168, 1973. 17. Stanley, M. A., and Kirkland, J. A.: Chromosome and histologic patterns of preinvasive lesions of the cervix, Acta Cytol. 19: 142. 1975. 18. ‘Townsend, D. E.. Ostergard, D. R., Mishell, D. R., and Hirose, F. M.: Abnormal papanicolaou smearsevaluation by colposcopy, biopsies and endocervical curettage, AM. J. ORsxr. GYSECOL. 108: 429, 1970.
CE3
19.
in cervical
van Nagell, j. R., Jr., Meeker,
intraepithelial
neopiasia
109
W. R., Parker, J. (..:.)I . . and Harralson. J. D.: Carcinoembryonic antigen in patients with gynecologic malignancy, Cancer 35: 1.772. 197.5. 20. van Nagell, J. R., Jr., Pletsch, (r. A., and Goldenberg, 1). M.: A study of cyst fluid and plasma (.arcillormhr~r)ni~ antigen in patients with rvstir ovariar~ neoplasn~s. t :itntet Res. 35: 1433. 1975. 21. van Nagell, J. R.. Jr.. Pletsch, v. A., and t;o\dcnhcrg. 1). M.: Carcinoemhryrmic antigen in squamott< t eli carcinoma of. the cervix, Gvnecol. Invest. 6: 40. 1!)7.?. 22. Zamcheck. N., Moore, T. I.., Dhar. P.. and Kupcltik. H.: Immunologic diagnosis and prognosis of’ huntan digestive cancer-carcinoembryonic antigens. S. Enql. 1, hlctl. 286: 83. 1972.