[CANCER RESEARCH 42, 2506-2513, June 1982]0008-5472/82/0042-0000$02.00

Serum Levels of Carcinoembryonic Antigen and a Tumor-extractedCarcinoembryonic Antigen-related Antigen in Cancer Patients1

John E. Shively,2 Vickie Spayth, Fong-Fu Chang, Gerald E. Metter, Leonard Klein, Cary A. Present, and

Charles W. Todd

Division of Immunology ¡J.E. S., V. S., C. W. T.], City of Hope Research Institute, and Departments of Medical Oncology [F-F. C., L. K., C. A. P.] and Biostatistics¡G.M.I, City of Hope National Medical Center. Duarte. California 91010

ABSTRACT

Levels of Carcinoembryonic antigen (CEA) and a tumor-extracted CEA-related antigen (TEX) were determined in sera

from patients with carcinomas of the breast, colon, lung, headand neck, and a number of miscellaneous categories. Thepurpose of this study was to compare the levels of each antigenat various stages of malignant disease. Competitive radioim-

munoassays developed in our laboratory were shown to besufficiently specific to detect either of the.se two antigensindependent of each other. The results indicate that: (a) withour specific assays, CEA was not significantly elevated insmoker controls, but TEX was elevated in 29% of smokercontrols; (fa)TEX was equivalent to CEA as a tumor marker forcolonie cancer. TEX was better than CEA as a marker for lungcancer and, based on limited data, there is a possibility thatTEX is a significantly better tumor marker than is CEA in earlylung cancer; (c) TEX was superior to CEA as a tumor markerfor breast and head and neck cancers; (d) there is a strongindication that serial determinations of TEX can be used aseffectively as CEA in the monitoring of disease progress. Thesepreliminary results must be confirmed by increasing the numberof cancer patients and including nonmalignant disease controls.

INTRODUCTION

CEA3 was first described by Gold and Freedman (6, 7) in

1965 as a tumor marker for colonie cancer. Subsequently, anumber of CEA-related antigens have been described (for

review, see Refs. 8 and 14) including NCA (17), found in lungand spleen tissue, and TEX (10), coisolated with CEA from livermétastasesof colonie carcinoma. Structural studies on CEA,NCA, and TEX (5, 13, 14) indicate that NCA and TEX haveidentical NH2-terminal amino acid sequences, similar molecularweights (M, ~ 100,000), and nearly identical amino acid andcarbohydrate compositions. Both differ from CEA in that CEAhas a higher molecular weight (M, ~ 180,000), a higher per

centage of carbohydrate (50 to 60% versus 30 to 35% for NCAand TEX), valine at position 21 in its NH2-terminal amino acid

sequence in place of alanine for NCA and TEX, and no detectable methionine (NCA and TEX contain approximately 4 to 6methionines). The development of a radioimmunoassay for

1Supported by Grants CA 16434 from the National Cancer Institute and CA

19163 from the National Large Bowel Program. The authors are members of theCity of Hope Cancer Research Center.

2 To whom requests for reprints should be addressed.3 The abbreviations used are: CEA, Carcinoembryonic antigen; NCA. nonspe

cific cross-reacting antigen; TEX, tumor extracted CEA-related antigen; CMF,cyclophosphamide-methotrexate-5-fluorouracil.

Received December 22. 1980; accepted March 11,1982.

NCA by Burtin, von Kleist, et al. (16, 18) enabled them toevaluate the potential use of NCA as a tumor marker. Theyconcluded that NCA was a very poor marker for cancer.

Studies in our laboratory indicated that TEX possessed an-tigenic determinants found in CEA but not in NCA, thus suggesting the possibility that TEX may behave differently fromNCA as a tumor marker. In order to evaluate whether TEX wasa better marker than CEA for various cancers, a pilot study wasinitiated at the City of Hope National Medical Center. In 331cancer patients with a wide variety of carcinomas, the serumlevels of CEA and TEX were determined by competitive radioim-munoassays in which a high degree of antigen specificity wasexhibited. The results indicate that TEX may be a superiortumor marker to CEA in all types of cancer studied except forcolonie cancer, in which case it was equivalent to CEA. Itshould be stressed that these results require confirmation byinclusion of larger numbers of patients with cancer and controlswith nonmalignant diseases.

MATERIALS AND METHODS

Serum Samples. Serum samples from 100 healthy volunteers (59nonsmokers, 41 smokers) were used as normal controls. Serial serumsamples were drawn from 331 cancer patients who were seen at theCity of Hope Medical Center over a 1-year period. Patient selection

criteria consisted of histologically proven diagnosis of cancer. Sampleswere drawn serially at 2- to 6-week intervals when possible and werestored at —20°.Voluntary informed consents were obtained prior to

blood drawing.Radioimmunoassays. CEA and TEX used as assay standards were

isolated by the perchloric acid method as previously described (1,10).Radioimmunoassays were performed by a triple-isotope, double-antibody method (2) utilizing 57Co as a supernatant marker (3). The CEA

assay used goat anti-CEA as the primary antibody, and the TEX assayused rabbit anti-TEX. Under the assay conditions established, TEX did

not interfere (was not detected as CEA) in the CEA assay up to 300ng/ml, and CEA did not interfere in the TEX assay up to 600 ng/ml.Examples of the levels of interference are the following: TEX (1.5 ¡ig/ml) is equivalent to CEA (8 ng/ml) in the CEA assay, and CEA (12 /¿g/ml) is equivalent to TEX (50 ng/ml) in the TEX assay. In general, theseassay conditions were suitable for independent measurements of CEAand TEX, since even for high values of one antigen in its assay thecorresponding levels in the other assay were very low and could becorrected for if desired.

Statistical Analysis. Serum levels of CEA and TEX are presented asmean ±S.E. Tests for equality of serum levels were made by 2-sidedf tests using a pooled estimate of the variance. McNemar's test for

matched samples, corrected for continuity, was used to test for equalityof proportions positive for CEA and TEX.

RESULTS

Controls. The serum levels of CEA and TEX found in non-

2506 CANCER RESEARCH VOL. 42

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CEA and TEX Serum Levels in Cancer Patients

smoker, healthy controls are shown in Chart 1. The meanserum levels are 10.4 ±4.3 ng/ml for CEA and 140.1 ±51.9ng/ml for TEX. It should be noted that the reported normalrange for the serum levels of CEA using a triple-isotope, double-

antibody assay is 5.6 to 12.6 ng/ml (4). Cutoffs of 20 ng/mlfor CEA and 200 ng/ml for TEX were chosen as the upperranges for controls in this study. Using these cutoffs, valuesdivided by 10 in the TEX assay are equivalent to values in theCEA assay. The cutoff for CEA is roughly 2 S.D. above themean, and that for TEX is 1 S.D. above the mean. It can beseen from data shown in subsequent charts that the number ofelevations do not change significantly for a cutoff at 1 or 2 S.D.above the mean. By these criteria, 5% of the nonsmokingcontrols had slightly elevated levels of CEA, and 8% hadelevated levels of TEX. One nonsmoking control gave a repeated TEX value of 400 ng/ml. Because of the newness ofthis test, it is not clear whether such a finding is rare. If thisvalue is excluded, 6% of controls show slight elevations forTEX. In nonsmoking controls, 3 had elevated values of CEAbut normal TEX levels, and 5 had elevated TEX levels butnormal CEA levels.

The serum levels of CEA and TEX for smokers are shown inChart 2. The mean serum level is 11.9 ±4.2 ng/ml for CEAand 194.1 ±58.6 ng/ml for TEX. Based on the cutoff valuesfor nonsmokers, 5% of smokers showed elevation in the levelof CEA and 29% in TEX. In this limited survey, these resultssuggest that a smoking history does not necessarily increasethe frequency of CEA elevations but does increase that forTEX. In smoking controls, 2 had elevated levels of CEA, one ofwhich had normal levels of TEX; 12 had elevated TEX, 11 ofwhich had normal levels of CEA.

16 n r-t

12-

{¡»

4-

CEA3/59 =5%

16-1

12-

8

4-

ng/ml

TEX5/59 =8%

OOOOOOOOOOOOOOOQOOOOOra<9Q N v tt oo o c\j <r <—<\ievjcvjc\ieMioi»

ng/ml

Chart 1. Serum levels of CEA and TEX in nonsmoker controls. Shaded columns, levels above 20 ng/ml for CEA and above 200 ng/ml for TEX.

Cancer Patients. The largest number of patients were accumulated for 4 major primary tumor sites: breast; colon; lung;and head and neck. Sample sizes in other groups were toosmall to permit statistically significant interpretations of thedata. The overall breakdown of elevations for CEA and TEX isshown in Table 1. An attempt was made to perform severaldeterminations on all patients. The results in Table 1 reflect thehighest determination of serum levels for CEA or TEX obtainedduring the time the patients were in the study. These datasuggest that as general tumor markers CEA and TEX areroughly equivalent for colon and lung carcinomas but that TEXmay be a better marker than CEA for breast and for head andneck carcinomas.

The observation that the increase in frequency of elevations

16-1

12-

8,

4-

CEA

2/41 = 5%

ng/ml16-1

-12-1

8-34--TEX£—12/41

=29%—H^Ii

^ ^OOOOOOOOOOOOQOOOOOwwOwvl»COOnJVwjQOra*vwQ)O

ng/ml

Chart 2. Serum levels of CEA and TEX \n smoker controls. For details, ¡legend to Chart Ì.

Table 1

Frequency of elevation of serum levels of CEA and TEX in various primarycarcinomas

No.elevated/totalPrimary

siteBreastColonLungHead

+neckProstateMelanomaOvaryGastricBladderOtherTotalCEA27/86

(31)a22/38

(58)38/67(57)8/42(19)4/15(27)4/11(36)5/9(56)5/9(56)3/6

(50)24/48(50)140/331

(42)TEX46/86

(53)22/38(58)50/67(75)28/42

(67)6/15(40)5/11(45)5/9(56)4/9(44)2/6

(33)14/48(29)182/331

(55)'' Numbers in parentheses, percentage of total.

JUNE 1982 2507

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

J. E. Shively et al.

Table2 Comparison of mean serum levels of CEA and

LocalRegionalSmoker

Nonsmoker SmokerNonsmokerCEA

TEX CEA TEXCEABreast28 ±8a 263 ±36 (6)" 15 ±2 179 ±23 (7) 21 ±5Colon

22 ±6 327 ±66 (2)c 20 ±3 187 ± 8 (7) 23 ±6Head-neck

15 ±2 216 ±20 (9) (0) 58 ±34Lung

17 ±3 292 ±48 (5) (0) 25 ±9*Mean ±S.E.

6 Numbers in parentheses, sample size within a group.c Sample sizes which are too small to permit aconclusion.of

TEX compared to CEA in breast cancer was similar to thatobserved in smoking versus nonsmoking controls prompted usto stratify the data according to smoking habit. This breakdown,TEX

CEATEX278

±45 (5) 14 ±2 197 ±21(9)283

±14 (5) 17 ±0 160 ± 0(if290

±27 (8) 19 ±5 238 ± 8(3)c256

±35 (9)(0)^

LOCAL TOTAL•15

O] REGIONAL TOTAL •15

O] METASTATIC TOTAL •56

which also includes extent of disease (local, regional, or met-

astatic), is shown in Table 2. For breast cancer, a significantrise in the mean of TEX values is observed for smokers (f =2.80, 63 d.f., p < .01) but is not comparable to that seen innormal controls (140 ±6.8 for nonsmokers, 194 ±9.2 forsmokers). The mean for nonsmokers (234 ±12) is considerably higher than that for nonsmoker controls (f = 6.49, 98 d.f.,

p < 0.000001). Thus, the contribution of smoking to increasedlevels of TEX can be largely excluded for breast cancer.

For colonie cancer, nonsmokers showed a higher meanelevation of TEX than smokers (t = 3.77, 24 d.f., p < 0.001),but in terms of number of elevations (Table 1) TEX was elevatedjust as often as was CEA. Thus, the contribution of smoking incolonie cancer can be neglected in this context. The reasonfor a very high mean value in the nonsmoking group is that anumber of patients had exceptionally high levels of bothmarkers (see later discussion).

For head and neck cancer and lung cancer, virtually all thepatients were smokers; thus, a statistically significant conclusion cannot be reached at this time.

Carcinoma of the Breast. Of 86 patients entered in thiscategory, 56 had metastatic disease; of these patients, 36%had elevated levels for CEA and 54% had elevated levels forTEX. Although the number of patients with localized diseasewas small (15), 4 (27%) elevations of CEA were observed, and8 (53%) showed elevations for TEX. In the group with regionaldisease, 20% had elevated CEA values, and 53% had elevatedTEX values. Scattergrams giving the values for each markerand the correlation of the 2 markers are shown in Chart 3.Concordance of the 2 markers (both elevated) was observedin about 30% of the case studies. TEX alone was elevated inabout 40% of the cases, and CEA alone was elevated in 3% ofthe cases. The overall incidence of CEA elevations for breastcancer was 31 % for CEA and 53% for TEX (McNemar's test:X2 = 15.43, p < 0.001). These results indicate that TEX is a

good marker for local, regional, and metastatic breast cancer.One drawback to this study is that the majority of patients hadadvanced disease. In the group of 24 smokers, 9 had elevatedTEX only, 9 had both markers elevated, and the remainder hadneither elevated. In the group of 41 nonsmokers, 9 had elevated TEX only, 4 had elevated CEA only, 11 had both markerselevated, and the remainder had neither elevated.

Carcinoma of the Lung. Of 67 patients with lung cancer, 11had regional disease, 51 had metastatic disease, and 5 had

604020CEATEXm.

I25)(I)BOU+-1—1-+J'(19)--\++

1000

too -

«I

010

10,000

1,000

33

100

10LOCAL REGIONAL METASTASIZED

Chart 3. Correlation diagrams and scartergrams for serum levels of CEA andTEX in breast cancer. Top, correlations broken down according to extent ofdisease (local, regional, or metastatic). Histogram gives percentage of totalnumber of patients with each disease stage who had elevated CEA only, elevatedTEX only, elevation of neither CEA nor TEX, and elevated CEA and TEX. Numberof patients in each category (e.g. 19 patients with metastatic disease wereelevated for both CEA and TEX) shown in parentheses. Bottom, scattergrams forserum levels of CEA and TEX.

2508 CANCER RESEARCH VOL. 42

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

*'T*"'.V":'"-'-

TEX in smoking and nonsmoking cancer patients

CEA and TEX Serum Levels in Cancer Patients

Table 2

SmokerCEA18

±3109

±3626

±7743

±577MetastalicTotalNonsmoker

SmokerTEX

CEA2952092852072QnD±±±±42

(13) 45 ±131

7 (4)c 1305 ±84369

(9) 16 ±01439

(38) 7 ±4LOCAL

TOTAL'S

REGIONAL TOTAL 01

METASTATIC TOTAL -51TEX

CEATEX325

± 60 (25) 22 ± 12781556

±865 (7) 51 ±15273251

±0 (1)c 33 ± 42641

58 ± 27 (3)° 262 ±58873±±±±3188144TEX

(McNemar's test:

the correlation of the 2the number of patient!(24)(11)(26)(52)NonsmokerCEA25

±447

±18

±7

±319014X2

= 7.56, p < 0.01).

markers are shown ins with early presentatiTEX234

±634

±245

±158

±12

(41)205

(15)5

(4)°27

(3fScattergrams

andChart 4. Althoughon of lung cancer

80

60

40

20CEATEX+-H3*(3)lì-+(0)1(11)--(1)1!31)*+

1000

KX5

10

K>,000

1,000

!100

LOCAL REGIONAL METASTASIZED10

Chart 4. Correlation diagram and scattergrams tor serum levels of CEA andTEX in lung cancer. For details, see legend to Chart 3.

localized disease. In the group with regional disease, 36% hadelevations in CEA, and 64% had elevations in TEX. In the groupwith distant métastases,65% had elevated CEA, and 74% hadelevated TEX. Although the group with localized disease wassmall (5), 20% had elevated CEA, and 100% had elevatedTEX. Overall, 57% had elevated CEA, and 75% had elevated

in this study was small, there is a possibility that TEX is asignificantly better marker than is CEA in this group.

Carcinoma of the Colon. Out of 38 patients entered in thestudy with colonie cancer, 19 had metastatic disease, 7 hadregional disease, and 12 had localized disease. In the groupwith metastatic disease, 74% had elevations for CEA, and 63%had elevations for TEX; in the group with regional disease,43% had elevations for CEA, and 70% had elevations for TEX.Of patients with localized disease, 42% had elevations for CEA,and 42% had elevations for TEX. Scattergrams and the correlation of the 2 markers are shown in Chart 5. TEX appears tobe equivalent to CEA as a marker in all 3 disease stages. Wecould not demonstrate any advantage to the use of TEX overCEA as a tumor marker for colonie cancer. In the group of 11smokers, 3 had elevated TEX only, 7 had both markers elevated, and 1 had neither elevated. In the group of 15 non-

smokers, 1 had elevated TEX only, 2 had elevated CEA only,7 had both markers elevated, and the remainder had neitherelevated.

Carcinoma of the Head and Neck. Of 42 patients enteredfor head and neck cancer, 14 had localized disease, 14 hadregional disease, and 14 had distant metastatic disease. CEAlevels were elevated in 3% of patients \d$h local disease, in29% with regional disease, and in 21 % with metastatic disease.TEX levels were elevated in 43% of patients with local disease,in 86% with regional disease, and in 71% with metastaticdisease. Scattergrams and the correlation of the 2 markers areshown in Chart 6. There is a clear trend indicating that TEX isa better marker than CEA for head and neck cancer. Theoverall breakdown shows CEA elevated in 19% of the casesand TEX elevated in 67% of the cases (McNemar's test: x2 =

18.05, p < 0.001).Serial Levels of CEA and TEX in Colonie Cancer. An

attempt was made to obtain serial blood samples on all patientsin this study. Since the majority of patients were outpatientsand were entered in the study over a 1-year period only, the

average number of samples obtained per patient was only 2.The lack of extensive data from serial sampling limits our abilityto judge the clinical utility of serial TEX levels for the majordisease categories studied. In the few individual cases in which5 to 10 serial samples were obtained, the correlation betweendisease progress (as indicated by radiological or physicalexamination) and serum levels of CEA and TEX was excellent.

Of 7 patients with elevated CEA levels, 6 returned to normalwithin 1 to 6 weeks after surgical removal of their primarytumors. Of 3 patients with normal levels of CEA prior to surgery,

JUNE 1982 2509

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

J. E. Shively et al.

80604020^

LOCAL TOTAL=120

REGIONAL TOTAL=7Q

METASTATIC TOTAL=19(12)(5)

Õ3(2,•CEATEX2)+-2)1D)-+f^

«'1--i++

1000

100

§10

1000

S

loo

ioLOCAL REGIONAL METASTASIZED

Chart 5. Correlation diagrams and scattergrams for serum levels of CEA andTEX in colon cancer. For details, see legend to Chart 3.

Hin80n4020DCEATEXLOCAL TOTAL •14REGIONAL

TOTALO4METASTATIC

TOTAL-1418)

(B)(51+-^-+13--"1(30•f;

1000p

100-

§10

10000

1000

5<§^

100

10

LOCAL REGIONAL METASTASIZED

Chart 6. Correlation diagrams and scattergrams for serum levels of CEA andTEX in head and neck cancer. For details, see legend to Chart 3.

2 remained normal and one became significantly elevatedwithin 2 weeks. In this same group, TEX levels closely followedCEA levels. In the group with metastatic disease, one patientshowed decreasing levels of CEA (but not a return to normal),increasing levels of TEX, and evidence of increasing disease.A second patient had CEA and TEX levels return to normal andremained disease free. A third patient exhibited decreasingCEA and TEX levels (but not a return to normal) after surgery,then increasing levels followed by decreasing levels after theonset of radiation therapy, followed by increasing levels againuntil further radiation therapy was initiated (at which point thelevels dropped, and this study was terminated). In a fourthpatient with very high levels of CEA and TEX (1013 and 1315ng/ml, respectively), the levels of CEA but not TEX droppedafter surgery, and both increased within an 8-week period.

Eventually, the levels of both markers exceeded 4000 ng/ml

prior to the death of the patient. Several other terminal patientsexhibited CEA and TEX levels as high as 4000 ng/ml. In agroup of 15 patients judged disease free following surgery,none showed increasing levels of CEA, and only one showedincreasing levels of TEX. This patient had a colonie tumorresected in 1966 and returned in 1978 with elevated CEA andTEX levels. A short section of colon was resected to remove aconstriction; subsequently, TEX levels fell to normal, and CEAlevels fell but continued to remain elevated over a 1-year

period.Serial Levels of CEA and TEX in Breast Cancer. Plots of

the serum levels of CEA and TEX and their correlation toclinical status (no evidence of disease, improved, stable, orworse) for 8 breast cancer patients are shown in Charts 7 and8. An additional 8 patients (charts not shown) are also discussed. Clinical evaluations simultaneous to serum acquisition

2510 CANCER RESEARCH VOL. 42

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CEA and TEX Serum Levels in Cancer Patients

were made by the physician without prior knowledge of thetumor marker levels.

Patient 1 (Chart 7) had a right radical mastectomy in September 1964 with pectoral and lymph node involvement. The tumortested positive for estrogen receptors, and stilbestrol therapywas begun in 1975. A total abdominal hysterectomy includingoophorectomy was performed in June 1976, and a bilateraladrenalectomy in October 1976. In April 1977, CMF chemotherapy was initiated. Radiation therapy (5000 rads) was performed on the right posterior neck in July 1978 and continuedthrough October 1978. During this period, métastases werefound in the pleura and bone, and chemotherapy with doxorub-

icin and vincristine was begun. In May 1979, mitomycin andmedroxyprogesterone were administered. Objective clinicalcriteria indicated a continued worsening of the disease for thispatient, with the levels of CEA paralleling the disease state.TEX shows a brief period of improvement (return to normal)followed by steadily increasing values above normal.

Patient 2 (Chart 7) had a right radical mastectomy in January1979 with 4 of 26 axillary nodes involved. Chemotherapy(CMF) was begun in February 1979. The patient has shown nosubsequent evidence of disease, and both tumor markers havefallen and stayed with the normal range.

Patient 3 (Chart 7) had a left modified radical mastectomy inAugust 1978 with 7 of 30 axillary nodes and one pectoral nodeinvolved. Radiation therapy and chemotherapy (CMF) was begun the same month. Chemotherapy was continued throughAugust 1979 with no evidence of disease. CEA levels are withinthe normal range and show some fluctuations with an overalltendency to predict improvement. TEX levels exhibit a cyclicpattern with the majority of values above normal.

Patient 4 (Chart 7) had a left modified radical mastectomy inNovember 1978 with 2 of 13 axillary nodes involved. Chemotherapy (CMF) was administered from December 1978 to June1979, and 5-fluorouracil and methotrexate were given from

June 1979 to the present. The patient showed no evidence of

disease during this period, with TEX levels in the normal rangeand stable and CEA levels rising above normal from November1978 to January 1979 but falling to normal by March 1979.

Patient 5 (Chart 8) had a left modified radical mastectomy inOctober 1977 and was given chemotherapy (CMF) from February 1978 to September 1978, and tamoxifen from March1978 to May 1978. In December 1978, métastasesto the boneand lung were detected. When pleural métastaseswere foundin January 1979, radiât'on therapy was started. Clinically, thepatient's condition worsened through the period February

1979 to April 1979, and the patient expired in May 1979. CEAand TEX levels were above normal throughout the periodmonitored and showed dramatic increases from February 1979to April 1979.

Patient 6 (Chart 8) had a right modified radical mastectomyin August 1973 with no nodal involvement. In September 1978,the patient received 200 rads of radiation daily to rib métastases and was given chemotherapy (CMF). Treatment wascontinued through June 1979, when métastasesto the cranialvault were detected. CEA and TEX levels were above normalduring the entire period of monitoring, with CEA showing adramatic increase from November 1978 to February 1979, aperiod during which metastatic disease was progressing. TEXlevels remained stable during this interval.

Patient 7 (Chart 8) had a right radical mastectomy in August1975 with a localized tumor. In October 1978, metastatic bonedisease was found, and chemotherapy (5-fluorouracil and cy-

clophosphamide) and steroid therapy (prednisone) were begunand continued through June 1979. During this period, thepatient was clinically stable with improvement noted by June1979. Both CEA and TEX levels are within the normal rangeand fall to lower levels by June 1979.

Patient 8 (Chart 8) had a right radical mastectomy in May1976 with recurrence in June 1977. In August 1977, a rightoophorectomy was performed. In March 1978, métastasestothe chest wall were observed, and chemotherapy (CMF) plus

Chart 7. Serial serum levels of CEA and TEX in breastcancer patients. CEA and TEX levels in 4 breast cancerpatients 11 to 4). Sample dates are shown on the abscissawith a clinical evaluation shown above. W, worse; S,stable; /, improvement; M, mastectomy; N, no evidence ofdisease. , upper limit of normal range. See text forfurther details and examples.

s ?

JUNE 1982 2511

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

J. E. Shively et al.

Chart 8. Correlation diagram and scattergrams for serum levels of CEA and TEX in breast cancer (Patients 5 to8). Sample dates are shown on the abscissa with a clinicalevaluation shown above. W, worse; S, stable; C, expired;/, improvement; W. no evidence of disease.

steroid therapy (prednisone) were begun and continuedthrough January 1979. From October 1978 to January 1979,no evidence of disease was found, and CEA and TEX levelsfluctuated somewhat around the normal range.

Patient 9 (data not shown) had a left modified radical mastectomy in June 1979, received L-phenylalanine mustard from

June 1975 to June 1977, and had a bilateral oophorectomy inMay 1978. In August 1978, metastatic bone disease wasfound, and tamoxifen therapy was administered from December1978 to April 1979, when our study terminated. Clinically, thepatient was stable from November 1978 to December 1978and showed improvement thereafter. Both CEA and TEX levelsbegan to climb above normal during this period, suggestingthat the patient's condition was worsening. Follow-up in August

1979 showed new metastatic disease.Patient 10 (data not shown) had a right radical mastectomy

in November 1968 and received 2000 rads of radiation to thehip during September 1971. In December 1971, an adrenal-

ectomy and oophorectomy were performed. Tamoxifen therapywas begun in November 1978. Radiation therapy (3000 rads)was administered from December 1978 to January 1979 at theend of which metastatic bone disease had progressed. Tamoxifen therapy continued through July 1979, at which time moremétastasesto the ribs were found. During the period of study,TEX levels fell from above normal to normal, and CEA levelswere above normal and then fell to normal. On the contrary,the clinical picture suggested initially a worsening condition.Four months later, disease was controlled.

Patient 11 (data not shown) had a right modified radicalmastectomy in October 1978. In December 1978, a recurrencewas detected at the site of surgery, but otherwise the patientshowed no further evidence of disease. TEX levels startedhigh, fell to normal, and began to rise again by January 1979.CEA levels started low and continuously climbed through January 1979. Taken together, the tumor markers suggest anincrease in disease. Follow-up through April 1979 shows no

evidence of disease.

¿ s

Patient 12 (data not shown) had a left mastectomy in January1971. A right radical mastectomy was performed in June 1973with 28 of 28 axillary nodes involved, and a bilateral oophorectomy was performed as well. In February 1978, 5-fluorouracil-methotrexate-prednisone therapy was begun. In March 1978,

métastases to the pleura were found. Chemotherapy (CMF)was administered from October 1978 to January 1979, duringwhich time disease was present but relatively stable. CEAlevels remained below normal, TEX levels rose above normal,and both showed a fall followed by a rise during the middle ofthe chemotherapy period.

Patient 13 (data not shown) had a right modified radicalmastectomy in August 1978 with 2 of 18 axillary nodes involved. Chemotherapy (CMF) was administered from October1978 to February 1979, during which no evidence of diseasewas found. Except for a brief increase in TEX levels, bothmarkers remained constant and within the normal range.

Patient 14 (data not shown) had a right radical mastectomyin September 1978 with 3 of 12 axillary nodes involved.Chemotherapy (CMF) was administered from November 1978to July 1979. In March 1979, several recurrences in the rightaxillary area were removed; and by July 1979, lymphaticmétastases were detected. Although the clinical picture wasworsening, both CEA and TEX levels were stable and withinthe normal range.

Patient 15 (data not shown) had a left modified radicalmastectomy in October 1977 with 4 of 16 axillary nodesinvolved. Radiation therapy (4500 rads) and chemotherapy (5-

fluorouracil and cyclophosphamide) was administered fromOctober 1977 to August 1978. From August 1978 to November1978, radiation therapy was continued (3000 rads) with newchemotherapy (doxorubicin, 5-fluorouracil, and tamoxifen). In

November 1978, métastasesto the bone and lung were found.Chemotherapy was continued through January, 1979, at whichtime possible cranial métastases were observed. CEA levelsfell from above normal to normal during the clinically stableperiod and began to rise again with increasing disease. TEX

2512 CANCER RESEARCH VOL. 42

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CEA and TEX Serum Levels in Cancer Patients

levels paralleled the CEA levels but remained within the normalrange.

Patient 16 (data not shown) had a right radical mastectomyin October 1978 with 20 of 30 axillary nodes involved. Radiation therapy was administered from December 1978 to February 1979, during which no evidence of disease was observed.CEA and TEX levels remained within the normal range butclimbed steadily during this period. A 4-month follow-up on this

patient shows a definite increase in disease.

DISCUSSION

The purpose of this study was to evaluate the use of serialTEX assays for the detection and monitoring of a number ofmalignant diseases and to compare the results to those obtained with CEA assays. The data obtained in this work suggestthat the TEX assay is equivalent to the CEA assay in colonieand metastatic lung cancer but that it may be superior to theCEA assay for breast and head and neck cancers. In addition,it was found that, under the assay conditions used in our work,CEA was not significantly elevated in smoker controls but TEXwas elevated. Since both CEA and TEX can be purified fromthe same tumor tissue, it is possible that in some cases immu-noassays utilizing radiolabeled CEA and anti-CEA actually contain a mixture of CEA and TEX and anti-CEA and anti-TEX. Ifsuch were the case, these assays would measure CEA andTEX simultaneously. Since our results suggest that TEX, butnot CEA, is elevated in some smokers, it is possible that theidea that CEA is often elevated in smokers is incorrect. Ifimmunoassay kits were shown to be specific for CEA with noreactivity towards TEX, it is likely that smokers would show nohigher incidences of elevated CEA than nonsmokers. Thisresult has indeed been reported at a recent CEA consensusmeeting for a CEA assay kit under development.* It is unlikely

that TEX is identical to colon carcinoma antigen III reported inthe papers of Newman ef al. (11 ) and Primus et al. (12) whichis also elevated in smokers' serum, since colon carcinoma

antigen III is reported not to cross-react in CEA immunoassaysand is present in very high concentrations (700 to 1500 ng/ml) in normal plasma (12). The results shown in Chart 2 suggesta bimodal distribution for TEX in smokers. If larger numbersconfirm that smokers can be grouped into 2 distinct categories,those with normal and those with elevated TEX values, then itwould be of interest to perform long-term follow-up studies on

both groups to determine whether or not one group is at ahigher risk for lung cancer. One point of interest is that all ofthe long-term smokers who had quit smoking for over 1 year

had normal TEX values. Monitoring of patients with colonie orbreast cancer with serial determinations of CEA and TEX serum

* C. Miller, Beckman Diagnostics, Fullerton, Calif., personal communication.

Data from 596 serum samples assayed by 4 participating institutions werepresented at a CEA consensus meeting (sponsored by the NIH) held at the endof October 1980 in Washington. D. C.

levels was encouraging. It is likely that either one or bothmarkers will be useful in predicting improvement or progressionof the disease state.

These results and the well-documented correlation of CEA

levels with disease state in the literature (9, 15) suggest thatan enlarged study is merited for the evaluation of monitoringthe progress of disease and serial TEX assays.

ACKNOWLEDGMENTS

The authors wish to acknowledge the expert technical assistance of LeeGanteaume, Fran Crawford, Karen Rickard, Westley Dentón, and Diana Esparza.We are particularly indebted for the help of Drs. Arthur Sarauw and DinaRappaport in the initial stages of this study.

REFERENCES

1. Coligan, J. E., Lautenschleger, J. T., Egan, M. L.. and Todd, C. W. Isolationand characterization of carcinoembryonic antigen. Immunochemistry, 9.377-386, 1972.

2. Egan, M. L., Lautenschleger, J. T., Coligan, J. E., and Todd, C. W. Radioimmune assay of carcinoembryonic antigen. Immunochemistry, 9. 289-299,1972.

3. Egan, M. L., Todd. C. W., and Knight. W. S. Co": a volume marker for the

triple-isotope, double-antibody radioimmune assay. Immunochemistry, 14:611-613, 1977.

4. Pritsche, H. A., Tashima, C. K., Romsdahl, M. M., Holoye, P. Y.. and Geitner,A. A clinical evaluation of the triple-isotope double-antibody radioimmunoas-say for carcinoembryonic antigen. Am. J. Clin. Pathol.. 69. 140-146, 1978.

5. Fuks. A., Banjo, C., Shuster, J., Freedman, S. O., and Gold, P. Carcinoembryonic antigen (CEA): molecular biology and clinical significance. Biochim.Biophys. Acta. 417: 123-152, 1974.

6. Gold. P., and Freedman, S. O. Demonstration of tumor-specific antigens inhuman colonie carcinomata by immunological tolerance and absorptiontechniques. J. Exp. Med., 121: 439-462, 1965.

7. Gold, P., and Freedman, S. O. Specific carcinoembryonic antigens of thehuman digestive system. J. Exp. Med., )22. 467-481, 1965.

8. Hammarstrom, S., Svenberg. T., Hedin, A., and Sunblad, G. Antigens relatedto carcinoembryonic antigen. Scand. J. Immunol.. 7 (Suppl. 6). 33-46,

1978.9. Jubert, A. V.. Talbott, T. M., and Maycroft, T. M. Characteristics of adeno-

carcinomas of the colorectum with low levels of preoperative plasma carcinoembryonic antigen (CEA). Cancer (Phila.), 42: 635-639, 1978.

10. Kessler, M. J., Shively, J. E., Pritchard, D. G., and Todd, C. W. Isolation,immunological characterizations and structural studies of a tumor antigenrelated to carcinoembryonic antigen. Cancer Res., 38: 1041-1048, 1978.

11. Newman, E. S., Petras, S. E., Georgiadis, A., and Hansen, H. S. Interrelationship of carcinoembryonic antigen and colon carcinoma antigen-Ill. Cancer Res., 34: 2125-2130, 1974.

12. Primus, F. J., Newman, E. S., and Hansen, H. J. Affinity in radioimmunoassayof antibody crossreactive with carcinoembryonic antigen (CEA) and coloncarcinoma antigen-Ill (CC-III). J. Immunol.. 118: 55-61, 1977.

13. Shively, J. E., and Todd, C. W. Carcinoembryonic antigen. Scand. J.Immunol., 7(suppl. 6): 19-31, 1978.

14. Shively, J. E., and Todd, C. W. Carcinoembryonic antigen A: chemistry andbiology. In: S. Sells (ed.). Cancer Markers, pp. 295-314. Clifton, N. J.:Humana Press. 1980.

15. Sugarbaker, P. H., Zamcheck, N., and Moore, F. D. Assessment of serialcarcinoembryonic antigen (CEA) assays in postoperative detection of recurrent colorectal cancer. Cancer (Phila.), 38. 2310-2315, 1976.

16. Troupel, S., von Kleist, S., and Burtin, P. Dosage radioimmunologique duCEA et du NCA dans le serum de nouveau-nés a terme et prematures. Bull.Cancer, 6. 655-660, 1976.

17. von Kleist, S., Chavanel, G., and Burtin, P. Identification of an antigen fromnormal human tissue that crossroads with the carcinoembryonic antigen.Proc. Nati. Acad. Sei. U. S. A., 69: 2492-2494. 1972.

18. von Kleist, S., Troupel, S., King, M., and Burtin, P. Etude comparative dutaux serique du NCA et du CEA. Bull. Cancer, 63. 627-632, 1976.

JUNE 1982 2513

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1982;42:2506-2513. Cancer Res   John E. Shively, Vickie Spayth, Fong-Fu Chang, et al.   Cancer PatientsTumor-extracted Carcinoembryonic Antigen-related Antigen in Serum Levels of Carcinoembryonic Antigen and a

  Updated version

  http://cancerres.aacrjournals.org/content/42/6/2506

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/42/6/2506To request permission to re-use all or part of this article, use this link

on April 11, 2020. © 1982 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from