Int J Gynaecol Obstet 16: 388-393, 1979
Carcinoembryonic Antigen and /^-Microglobulin as Serum Tumor Markers in Women with Genital Cancer Soo Keat Khoo, Brian Daunter and Eric Mackay Department of Obstetrics and Gynaecology, University of Queensland, Royal Brisbane Hospital, Brisbane, Australia
ABSTRACT Khoo SK, Daunter B, Mac/cay EV (Dept of Obstetrics and Gynaecology, University of Queensland, Royal Brisbane Hospital, Brisbane, Australia). Carcinoembryonic antigen and ^-microglobulin as serum tumor markers in women with genital cancer. Int J Gynaecol Obstet 16: 388-393, 1979 A comparative study of carcinoembryonic antigen (CEA) and B2-microglobulin (B2-MG) in serum was made by radioimmunoassay in 77 women with genital cancer. With a positive level defined as 5 ng of CEA/ml and 3.0 fig of B2-MG/ml, CEA was positive in 31% of the women with cancer of the corpus, 36% of those with cancer of the cervix and 36% of those with cancer of the ovary; the corresponding figures for B2-MG were 6%, 27% and 56%, respectively. The additional use of B2-MG provided an increase in positive results, especially in cases of cancer of the ovary. A direct relationship between the extent of tumor and serum marker level was more evident for B2-MG than CEA. There was no correlation between serial levels of CEA and B2-MG in most patients. CEA levels appeared to predict subsequent tumor behavior more accurately in patients with good prognoses (ie, complete or partial tumor response), whereas B2-MG levels gave the same prediction in those with bad prognoses (ie, nonresponsive or progressive tumor).
INTRODUCTION It is generally accepted that the diagnosis and follow-up of patients with cancer can be improved with the use of appropriate serum tumor markers. However, there is at present no known marker which appears to meet all of the following required criteria: its detection in all tumors, specificity restricted to patients with cancer, reasonably rapid clearance from the circulation and correlation of levels in serum with tumor load and activity. In the past decade, the search for tumor markers has been
InlJ Gynaecol Obstet 16
successful in identifying two useful oncofetal antigens, a-fetoprotein and carcinoembryonic antigen (CEA). T h e former marker has been strongly associated with hepatocellular and germ cell carcinomas, and the latter with a larger spectrum of tumors, in particular, those of the gastrointestinal tract (19). T h e clinical significance of CEA, a large glycoprotein with a molecular weight of 200 000, a n d its limitations are becoming better known (4); its predictive value in tumor monitoring has been reported in women who h a d minimal or no postoperative residual ovarian cancer (6). T h e recent demonstration of ^-microglobulin (/?2-MG) in serum of some patients with cancer highlights its possible use as another tumor marker (7, 14, 16). H u m a n B2-MG is a small protein of molecular weight 11 800 found in high concentrations in fetal serum (7). Originally isolated from the urine of patients with renal tubular disorders (1), it has been shown to be synthesized and secreted by lymphocytes in culture (13); furthermore, malignant cells produce higher concentrations than nonmalignant cells in culture (12). Although its biological function is unknown, it has been found to be structurally related to immunoglobulin G (15) a n d to the common portion of the major histocompatibility complex of man (11). In this preliminary study, a comparative evaluation of CEA and B2-MG as serum tumor markers has been made in women with genital cancer at the time of initial diagnosis and during follow-up.
MATERIALS A N D M E T H O D S Blood samples were obtained by venipuncture from 77 women with histologically confirmed genital cancer. Because these patients were unselected and their admission to the study was made when they presented to the gynecologic and oncologic
Serum tumor markers in genital cancer
Table 1 . Distribution of patients by tumor type and stage. Tumor Stage Tumor Site
No. of Patients
I
II
III
IV
Corpus Cervix Ovary
16 36 25
13 23 7
2 9 3
1 3 9
0 1 6
wards, their distribution according to the stage of the tumor reflected the usual extent of disease for the three major types of genital cancer (Table I). As expected, the majority of patients with cancer of the corpus or cervix h a d stage I tumor, while more than half of those with cancer of the ovary had stage III or IV tumors. All of the tumors were carcinomas, either squamous cell carcinoma or adenocarcinoma. Blood samples were also obtained from 29 apparrently healthy subjects, aged 22-45 years, who served as controls in this study. Serial blood samples, taken at monthly intervals, were obtained from 19 of the patients with cancer of the ovary who were followed-up for at least 12 months. All but one had postoperative residual disease. Radiotherapy to the pelvis and intermittent chemotherapy (melphelan) were administered to all. The status of the disease was assessed independently at a consultative clinic at the end of 12 months. T h e patients were grouped as follows: group A, no evidence of disease; group B, existing residual tumor which had shown at least 50% reduction in size and, therefore, seemed to show a favorable response; group C, existing residual disease which had shown less than 50% reduction in tumor size; and group D, progressive disease, as evidenced by continued tumor growth. Serum urea a n d creatinine values of each patient were checked by multichannel analysis. All results were within the normal range of 3.50-7.50 m m o l e / liter for urea and 0.04-0.11 mmole/liter for creatinine, indicating no biochemical evidence of renal impairment. Radioimmunoassay for CEA and /82-MG in serum The assay for CEA was established using a double antibody technique, based on inhibition of binding between anti-CEA antiserum and radiolabeled 125I CEA by C E A in the serum with precipitation of the complexes by a second antibody, a rabbit antiserum to goat y-globulin (5, 10). T h e sensitivity of the assay was between 3 a n d 50 ng C E A / m l , and a level of 5 n g / m l or more was considered a positive result (6).
389
T h e assay for /?2-MG was performed by a solid phase technique, as described in the Phadebas /?2micro test (Pharmacia Diagnostics, Sydney, Australia). A volume of 100 jul of serum or /S2-MG standard was added to 100 jul 125I /? 2 -MG and 100 fil of Sephadex-anti-/?2-MG before incubation on a shaker for three hours at room temperature. After the addition of 2 ml of normal saline, the tubes were washed three times by centrifugation at 2000 X g for two minutes. T h e radioactivity in the precipitate was counted in duplicate. T h e sensitivity of the test ranged between 1.5 and 96.0 ¡ig of j S ^ M G / m l and a level of 3.0 /ig/ml or more was considered a positive result. Interpretation of serial levels From our previous experience with serial CEA levels and tumor monitoring (6), the following patterns, observed for both CEA and /?2-MG, were categorized: 1. There was no detectable marker at the time of initial presentation a n d thereafter during the period of follow-up. T h e level of detectability was taken as 3.0 n g / m l for CEA and 1.5 jug/ml for j3 2 -MG. 2. After an initial elevated level, there was a rapid disappearance of the marker from detection in serum. 3. There was a fall in serum levels but not to nondetectable levels. 4. There were fluctuating low levels during the period of follow-up, whereby the C E A level did not exceed 10 n g / m l , and the /62-MG level did not exceed 5 jug/ml. 5. There were persistently high levels, exceeding 10 ng of C E A / m l or 5 fig of /? 2 -MG/ml. 6. There was a rise in serum levels, evidenced by an at least 25% increase in concentration from the previous level. O n this basis, it was considered that, when pattern 1 was obtained, monitoring was not possible. T o assess the predictive value of the serial levels, pattern 2 was taken to represent a complete tumor remission, whereby no clinical evidence of tumor was present. Pattern 3 was taken to be an incomplete remission, whereby a significant therapeutic response was noted. Pattern 4 was seen as no significant change in the residual tumor. Patterns 5 a n d 6 were classified as progressive disease.
RESULTS Positive results according to tumor origin Table II shows the results with CEA and /82-MG,
IntJ Gynaecol Obstei 16
390
Khoo et al
Table II. Pretreatment levels of carcinoembryonic antigen (CEA) and ^-microglobulin (/?2-MG) in cancer of the corpus, cervix and ovary. Positive Tumor Marker Both
Tumor Site Corpus Cervix Ovary Total cancer Controls
Only f}2 -MG
Only CEA
At Least One
No. of Subjects
No.
%
No.
%
No.
%
No.
%
16 36 25 77 29
0 6 5 11 0
0 16 20 14 0
5 13 9 27 0
31 36 36 35 0
1 10 14 25 3
6 27 56 32 10
6 17 18 41 3
37 47 72 53 10
•a a
m
• •
E
"or\ 3.
10
'-da-
•
o
m
>
nn
10
10 •
•
•
•
a
Z
o
-J 3 ID
O
• ••
S
•
•î"^focï'°
nacao
S WU.-BÎ.ÇÇP.j
Control
•:•
îji-
!•••• 0
•
O
¥
K:
O
!• •
•:•
• ••
CM C
un
• n
o
>
• • • o
•
5
o oc
m
OD
S.MM.
Corpus
is**. •jQúDcn
JiOJiSH.
Cervix
•
m •"••
3 00 -I
o
rn m m m
CC U
" " o
JI..0E.1ÎS—. -0
ODD ODD • CD QUO ana oaa ODD ODD
Ovary
DDO
Fig. 1. Individual values of CEA and /?2-MG in control subjects and patients with cancer of the corpus, cervix and ovary.
alone and concurrently, in controls and women with cancer of the corpus, cervix and ovary before commencement of treatment. Although none of the controls showed a positive CEA result, a positive /?2-MG result was obtained in three (10%), at levels of 3.0, 3.0 a n d 3.4 /ig/ml, respectively. T h e incidence of CEA positivity in the three groups ranged from 31% to 36%. Positive /?2-MG levels were much more frequent in patients with cancer of the ovary (56%) than in those with cancer of the cervix (27%) or cancer of the corpus (6%). It is interesting that concurrently positive CEA and /?2-MG results were obtained infrequently—only in six patients (16%) of those with cancer of the cervix and in five (20%) of those with cancer of the ovary. As would be expected from this, use of the two markers was associated with an increase in overall marker positivity (53%) in the cancer group. There was a wider scatter of positive CEA values
IntJ Gynaecol Obstei 16
TL
l Stage
M
m
of D i s e a s e
Fig. 2. The distribution of CEA and /82-MG values according to stage of disease in women with genital cancer.
than j8 2 -MG values in patients with cancer (Fig. 1). There were observable differences in the proportion of patients and levels of /?2-MG according to tumor origin, with the highest values obtained in patients with ovarian cancer. Serum levels according to the stage of disease Fig. 2 shows the distribution of CEA and /?2-MG values according to stage of disease. In stage I disease, /?2-MG was positive in only seven patients (16%) and C E A was positive in only ten (23%) of 43 patients. However, it was noted that highly elevated CEA levels ( > 50 ng/ml) were present in three patients, even at this early stage of disease. With more advanced stages, there was a progressive increase in /?2-MG values from 42% positive in stage
Serum tumor markers in genital cancer
391
Table III. Long-term patterns of CEA and B2-MG in relation to tumor behavior. Pattern of Serial Levels CEA Group A: No evidence of disease Case 1 Case 2 Case 3 Case 4
(82-MG
Rapid fall to negative Rapid fall to negative Undetectable 3 Undetectable 8
Rapid fall to negative Fluctuating, low Undetectable" Undetectable"
Group B: Residual disease/significant response Case 5 Falling Case 6 Falling—eventually negative Case 7 Fluctuating, low Case 8 Rising
Persistently high Undetectable" Falling Undetectable"
Group C: Residual disease/no change Case 9 Case 10 Case 11 Case 12 Case 13
Undetectable Rapid fall to negative Undetectable 3 Undetectable" Rising
Fluctuating, low Fluctuating, low Fluctuating, low Undetectable" Undetectable"
Group D: Progressive disease Case 14 Case 15 Case 16 Case 17 Case 18 Case 19
Rising Rising Rising Persistently high Undetectable 8 Persistently low
Rising Rising Undetectable" Fluctuating, low Persistently low Rising
'All levels
Carcinoembryonic Antigen and /^-Microglobulin as Serum Tumor Markers in Women with Genital Cancer Soo Keat Khoo, Brian Daunter and Eric Mackay Department of Obstetrics and Gynaecology, University of Queensland, Royal Brisbane Hospital, Brisbane, Australia
ABSTRACT Khoo SK, Daunter B, Mac/cay EV (Dept of Obstetrics and Gynaecology, University of Queensland, Royal Brisbane Hospital, Brisbane, Australia). Carcinoembryonic antigen and ^-microglobulin as serum tumor markers in women with genital cancer. Int J Gynaecol Obstet 16: 388-393, 1979 A comparative study of carcinoembryonic antigen (CEA) and B2-microglobulin (B2-MG) in serum was made by radioimmunoassay in 77 women with genital cancer. With a positive level defined as 5 ng of CEA/ml and 3.0 fig of B2-MG/ml, CEA was positive in 31% of the women with cancer of the corpus, 36% of those with cancer of the cervix and 36% of those with cancer of the ovary; the corresponding figures for B2-MG were 6%, 27% and 56%, respectively. The additional use of B2-MG provided an increase in positive results, especially in cases of cancer of the ovary. A direct relationship between the extent of tumor and serum marker level was more evident for B2-MG than CEA. There was no correlation between serial levels of CEA and B2-MG in most patients. CEA levels appeared to predict subsequent tumor behavior more accurately in patients with good prognoses (ie, complete or partial tumor response), whereas B2-MG levels gave the same prediction in those with bad prognoses (ie, nonresponsive or progressive tumor).
INTRODUCTION It is generally accepted that the diagnosis and follow-up of patients with cancer can be improved with the use of appropriate serum tumor markers. However, there is at present no known marker which appears to meet all of the following required criteria: its detection in all tumors, specificity restricted to patients with cancer, reasonably rapid clearance from the circulation and correlation of levels in serum with tumor load and activity. In the past decade, the search for tumor markers has been
InlJ Gynaecol Obstet 16
successful in identifying two useful oncofetal antigens, a-fetoprotein and carcinoembryonic antigen (CEA). T h e former marker has been strongly associated with hepatocellular and germ cell carcinomas, and the latter with a larger spectrum of tumors, in particular, those of the gastrointestinal tract (19). T h e clinical significance of CEA, a large glycoprotein with a molecular weight of 200 000, a n d its limitations are becoming better known (4); its predictive value in tumor monitoring has been reported in women who h a d minimal or no postoperative residual ovarian cancer (6). T h e recent demonstration of ^-microglobulin (/?2-MG) in serum of some patients with cancer highlights its possible use as another tumor marker (7, 14, 16). H u m a n B2-MG is a small protein of molecular weight 11 800 found in high concentrations in fetal serum (7). Originally isolated from the urine of patients with renal tubular disorders (1), it has been shown to be synthesized and secreted by lymphocytes in culture (13); furthermore, malignant cells produce higher concentrations than nonmalignant cells in culture (12). Although its biological function is unknown, it has been found to be structurally related to immunoglobulin G (15) a n d to the common portion of the major histocompatibility complex of man (11). In this preliminary study, a comparative evaluation of CEA and B2-MG as serum tumor markers has been made in women with genital cancer at the time of initial diagnosis and during follow-up.
MATERIALS A N D M E T H O D S Blood samples were obtained by venipuncture from 77 women with histologically confirmed genital cancer. Because these patients were unselected and their admission to the study was made when they presented to the gynecologic and oncologic
Serum tumor markers in genital cancer
Table 1 . Distribution of patients by tumor type and stage. Tumor Stage Tumor Site
No. of Patients
I
II
III
IV
Corpus Cervix Ovary
16 36 25
13 23 7
2 9 3
1 3 9
0 1 6
wards, their distribution according to the stage of the tumor reflected the usual extent of disease for the three major types of genital cancer (Table I). As expected, the majority of patients with cancer of the corpus or cervix h a d stage I tumor, while more than half of those with cancer of the ovary had stage III or IV tumors. All of the tumors were carcinomas, either squamous cell carcinoma or adenocarcinoma. Blood samples were also obtained from 29 apparrently healthy subjects, aged 22-45 years, who served as controls in this study. Serial blood samples, taken at monthly intervals, were obtained from 19 of the patients with cancer of the ovary who were followed-up for at least 12 months. All but one had postoperative residual disease. Radiotherapy to the pelvis and intermittent chemotherapy (melphelan) were administered to all. The status of the disease was assessed independently at a consultative clinic at the end of 12 months. T h e patients were grouped as follows: group A, no evidence of disease; group B, existing residual tumor which had shown at least 50% reduction in size and, therefore, seemed to show a favorable response; group C, existing residual disease which had shown less than 50% reduction in tumor size; and group D, progressive disease, as evidenced by continued tumor growth. Serum urea a n d creatinine values of each patient were checked by multichannel analysis. All results were within the normal range of 3.50-7.50 m m o l e / liter for urea and 0.04-0.11 mmole/liter for creatinine, indicating no biochemical evidence of renal impairment. Radioimmunoassay for CEA and /82-MG in serum The assay for CEA was established using a double antibody technique, based on inhibition of binding between anti-CEA antiserum and radiolabeled 125I CEA by C E A in the serum with precipitation of the complexes by a second antibody, a rabbit antiserum to goat y-globulin (5, 10). T h e sensitivity of the assay was between 3 a n d 50 ng C E A / m l , and a level of 5 n g / m l or more was considered a positive result (6).
389
T h e assay for /?2-MG was performed by a solid phase technique, as described in the Phadebas /?2micro test (Pharmacia Diagnostics, Sydney, Australia). A volume of 100 jul of serum or /S2-MG standard was added to 100 jul 125I /? 2 -MG and 100 fil of Sephadex-anti-/?2-MG before incubation on a shaker for three hours at room temperature. After the addition of 2 ml of normal saline, the tubes were washed three times by centrifugation at 2000 X g for two minutes. T h e radioactivity in the precipitate was counted in duplicate. T h e sensitivity of the test ranged between 1.5 and 96.0 ¡ig of j S ^ M G / m l and a level of 3.0 /ig/ml or more was considered a positive result. Interpretation of serial levels From our previous experience with serial CEA levels and tumor monitoring (6), the following patterns, observed for both CEA and /?2-MG, were categorized: 1. There was no detectable marker at the time of initial presentation a n d thereafter during the period of follow-up. T h e level of detectability was taken as 3.0 n g / m l for CEA and 1.5 jug/ml for j3 2 -MG. 2. After an initial elevated level, there was a rapid disappearance of the marker from detection in serum. 3. There was a fall in serum levels but not to nondetectable levels. 4. There were fluctuating low levels during the period of follow-up, whereby the C E A level did not exceed 10 n g / m l , and the /62-MG level did not exceed 5 jug/ml. 5. There were persistently high levels, exceeding 10 ng of C E A / m l or 5 fig of /? 2 -MG/ml. 6. There was a rise in serum levels, evidenced by an at least 25% increase in concentration from the previous level. O n this basis, it was considered that, when pattern 1 was obtained, monitoring was not possible. T o assess the predictive value of the serial levels, pattern 2 was taken to represent a complete tumor remission, whereby no clinical evidence of tumor was present. Pattern 3 was taken to be an incomplete remission, whereby a significant therapeutic response was noted. Pattern 4 was seen as no significant change in the residual tumor. Patterns 5 a n d 6 were classified as progressive disease.
RESULTS Positive results according to tumor origin Table II shows the results with CEA and /82-MG,
IntJ Gynaecol Obstei 16
390
Khoo et al
Table II. Pretreatment levels of carcinoembryonic antigen (CEA) and ^-microglobulin (/?2-MG) in cancer of the corpus, cervix and ovary. Positive Tumor Marker Both
Tumor Site Corpus Cervix Ovary Total cancer Controls
Only f}2 -MG
Only CEA
At Least One
No. of Subjects
No.
%
No.
%
No.
%
No.
%
16 36 25 77 29
0 6 5 11 0
0 16 20 14 0
5 13 9 27 0
31 36 36 35 0
1 10 14 25 3
6 27 56 32 10
6 17 18 41 3
37 47 72 53 10
•a a
m
• •
E
"or\ 3.
10
'-da-
•
o
m
>
nn
10
10 •
•
•
•
a
Z
o
-J 3 ID
O
• ••
S
•
•î"^focï'°
nacao
S WU.-BÎ.ÇÇP.j
Control
•:•
îji-
!•••• 0
•
O
¥
K:
O
!• •
•:•
• ••
CM C
un
• n
o
>
• • • o
•
5
o oc
m
OD
S.MM.
Corpus
is**. •jQúDcn
JiOJiSH.
Cervix
•
m •"••
3 00 -I
o
rn m m m
CC U
" " o
JI..0E.1ÎS—. -0
ODD ODD • CD QUO ana oaa ODD ODD
Ovary
DDO
Fig. 1. Individual values of CEA and /?2-MG in control subjects and patients with cancer of the corpus, cervix and ovary.
alone and concurrently, in controls and women with cancer of the corpus, cervix and ovary before commencement of treatment. Although none of the controls showed a positive CEA result, a positive /?2-MG result was obtained in three (10%), at levels of 3.0, 3.0 a n d 3.4 /ig/ml, respectively. T h e incidence of CEA positivity in the three groups ranged from 31% to 36%. Positive /?2-MG levels were much more frequent in patients with cancer of the ovary (56%) than in those with cancer of the cervix (27%) or cancer of the corpus (6%). It is interesting that concurrently positive CEA and /?2-MG results were obtained infrequently—only in six patients (16%) of those with cancer of the cervix and in five (20%) of those with cancer of the ovary. As would be expected from this, use of the two markers was associated with an increase in overall marker positivity (53%) in the cancer group. There was a wider scatter of positive CEA values
IntJ Gynaecol Obstei 16
TL
l Stage
M
m
of D i s e a s e
Fig. 2. The distribution of CEA and /82-MG values according to stage of disease in women with genital cancer.
than j8 2 -MG values in patients with cancer (Fig. 1). There were observable differences in the proportion of patients and levels of /?2-MG according to tumor origin, with the highest values obtained in patients with ovarian cancer. Serum levels according to the stage of disease Fig. 2 shows the distribution of CEA and /?2-MG values according to stage of disease. In stage I disease, /?2-MG was positive in only seven patients (16%) and C E A was positive in only ten (23%) of 43 patients. However, it was noted that highly elevated CEA levels ( > 50 ng/ml) were present in three patients, even at this early stage of disease. With more advanced stages, there was a progressive increase in /?2-MG values from 42% positive in stage
Serum tumor markers in genital cancer
391
Table III. Long-term patterns of CEA and B2-MG in relation to tumor behavior. Pattern of Serial Levels CEA Group A: No evidence of disease Case 1 Case 2 Case 3 Case 4
(82-MG
Rapid fall to negative Rapid fall to negative Undetectable 3 Undetectable 8
Rapid fall to negative Fluctuating, low Undetectable" Undetectable"
Group B: Residual disease/significant response Case 5 Falling Case 6 Falling—eventually negative Case 7 Fluctuating, low Case 8 Rising
Persistently high Undetectable" Falling Undetectable"
Group C: Residual disease/no change Case 9 Case 10 Case 11 Case 12 Case 13
Undetectable Rapid fall to negative Undetectable 3 Undetectable" Rising
Fluctuating, low Fluctuating, low Fluctuating, low Undetectable" Undetectable"
Group D: Progressive disease Case 14 Case 15 Case 16 Case 17 Case 18 Case 19
Rising Rising Rising Persistently high Undetectable 8 Persistently low
Rising Rising Undetectable" Fluctuating, low Persistently low Rising
'All levels
