Vol. 114, August
THE JOURNAL OF UROLOGY .
Copyright © 1975 by The Williams & Wilkins Co.
Printed in U.S.A.
CLINICAL EVALUATION OF URINARY AND SERUM CARCINOEMBRYONIC ANTIGEN IN BLADDER CANCER RICHARD A. FRASER, MARIO J. RAVRY, JOSEPH W. SEGURA
AND
VAY L. W. GO*
From the Gastroenterology Unit and the Department of Urology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
ABSTRACT
Urinary carcinoembryonic antigen-like activity was increased (more than 1.5 ng. per ml.) in 61 per cent of patients with transitional cell carcinoma of the bladder. The frequency of abnormal carcinoembryonic antigen values correlated with the extent and the grade of the tumor. However, urinary tract infection can produce abnormal carcinoembryonic antigen values in non-cancer patients. The use of urinary carcinoembryonic antigen measurement and urine cytologic examination for diagnosis of transitional cell carcinoma gave a better diagnostic yield than did use of either test alone. In transitional cell carcinoma patients with abnormal preoperative urinary carcinoembryonic antigen values, postoperative urinary carcinoembryonic antigen values were useful for determining completeness of tumor resection. However, with total cystectomy and creation of an ileal conduit or ureterosigmoidostomy, urinary carcinoembryonic antigen values increased markedly postoperatively. Serum carcinoembryonic antigen values were of little value in the diagnosis of transitional cell carcinoma. Advances in the field of cancer have raised hopes of earlier diagnosis of new or recurrent cases through the use of immunologic methods. 1 Recent reports suggest that the carcinoembryonic antigen (CEA) may be useful as a marker protein in the diagnosis of transitional cell carcinoma of the bladder. CEA is a polysaccharide-protein complex that is present in increased concentrations in the serum and tumor tissue of patients with various gastrointestinal and non-gastrointestinal cancers. 2 Reynoso and associates found the CEA concentration to be increased in the serum of 10 of their 20 patients with active transitional cell carcinoma of the bladder; in contrast, only 2 of 24 patients with inactive disease had an abnormal CEA value. 3 Furthermore, Hall and associates found abnormal urinary CEA values in two-thirds of their patients with active bladder cancer. 4 Herein we report the results of a study undertaken to determine the role of CEA in the evaluation of patients with vesical neoplasia and its usefulness as a marker protein or therapeutic indicator in the surgical treatment of these patients. MATERIALS AND METHODS
We studied 236 consecutive patients who were evaluated by the urology service because of a Accepted for publication November 8, 1974. Supported in part by Research Grant CB-23854 and Training Grant Tl-AM-5259 from the National Institutes of Health, Public Health Service, Bethesda, Maryland. * Requests for reprints: Mayo Clinic, Rochester, Minnesota 55901.
definite history or clinical susp1c10n of vesical neoplasia. All patients underwent diagnostic cystoscopic examination of the bladder. In all cases urine samples were collected by bladder catheterization under sterile conditions at the time of cystoscopy. Separate samples were obtained for CEA determination, urinalysis and cytologic examination. The patients were classified as having tumor, cystitis (infection) or no vesical abnormalities on the basis of clinical, cystoscopic, histologic and cytologic findings. Specifically, a patient was judged to have infection if urinalysis showed pyuria and the urine culture was positive. There were 110 untreated patients who had histologically confirmed transitional cell carcinoma of the bladder. In these patients the clinical stage of the disease, as defined by Jewett's criteria, 5 was determined on the basis of objective surgical and clinical evidence. The degree of histologic differentiation of the tumor was described by Broders' grading criteria. 6 Of the remaining 126 patients, 35 had cystitis and 91 had no vesical abnormalities at cystoscopy and no further clinical evidence suggestive of urinary tract cancer. These latter 2 groups constituted the control groups. Urinary CEA was measured by radioimmunoassay, using the method of Hansen after an initial extraction with an equal volume of 1.2 M perchloric acid. 7 There appeared to be no diurnal variation in urinary CEA values in 5 healthy subjects who underwent this evaluation (unpublished observations). Venous blood samples were obtained at the time of diagnosis in 90 of the 110 patients with confirmed vesical malignancy. 226
CLINICAL EVALUATION OF CARC!NOEMBRYONIC ANTIGEN IN BLADDER CANCER
Serum CEA was measured by direct radioimmunoassay, using the method of Hansen as modified by Ravry and associates. 8 Of 531 healthy controls studied by this method in our laboratory, 504 (95 per cent) have had serum CEA values no higher than 3.5 ng. per ml. Therefore, we accept 3.5 ng. per ml. as the upper limit of normal for serum CEA. Vesical tumor tissue was obtained from 20 at the time of operation. Normal bladder tissue also ,;vas obtained from non-invaded areas in 7 with small localized tumors. Tumor concentration was measured radioimthe method of Hansen after and extraction with L2 M perchloric acid. 7 The usefulness of urinary CEA as E marker for determining the completeness of surgical resection for vesical cancer was studied measuring urinary CEA preoperatively and 3 days postoperatively in 70 patients; 56 of these patients underwent transurethral resection or cystectorny, 8 had totai cystectomy with creation of an ilea! conduit and 6 had total cystectomy and ureterositmoidostomy. RESULTS
Serum CEA. The serum CEA concentration was greater than 3 .5 ng. per ml. in only 7 of the 90 patients (8 per cent) with active vesical cancer. Of these 7 patients 5 had Jewett's stage O lesions and only 2 had more extensive tumors. CEA. In control without urinary tract infection, urinary CEA values were higher than L5 ng. per ml. in 3 of the 91 subjects (3 per cent) (fig. 1). In contrast 23 of 35
>,;
v1
~
l;j 11..,
/(TCE!'IOO)
mo
~----0----------.., ~ , c o , ' .. o,r·-------·
I .· .,,el'
80
0
i:::
so
ti
'-!. ~
40
'!.I
~ ·-...
' ~
J /\
infl'lction 135)
' l :
t/~TCE c inf®ction
r~
('!OJ
1
:
20 :4.~ 1:
::i
§ \,.)
l
I .!------Controls c
-!'.!
:~
.. · · , , ,,.·
: 0
5
10
15
20
25
Urin@ CEA,ng/ml FIG. 1. Cumulative percentage distribution of actual urinary CEA values in controls, controls with urinary infection and transitional cell carcinoma patients with and without urinary infection. Presence of infection produces increase in frequency of abnormal urinary CEA values similar to that in transitional cell carcinoma patients v.rith no infection and actual concentrations are greater than in patients with transitional cell cafcino1na.
227
control patients (66 per cent) with infection had urinary CEA values greater than 1.5 ng. per ml., and their urinary CEA concentrations were greater than those in patients without infection. The urinary CEA concentration was more than 1.5 ng. per mL in 61 of the 100 patients with transitional cell carcinoma without concurrent urinary infection. the values were lower than those in controls with urinary infection. In 10 with transitional cell carcinoma with concurrent urinary there was a further increase in CEA concentration to values similar to those seen in control with urinary infection, In with transitional cell carcinoma but without concurrent infection, the of increased urinary CEA values to be related to the stage of the disease: the urinary CEA value was more than 1.5 ng. per mL in 29 of 41 patients (71 per cent) with invasive (Jewett's stage B 2 , C or D) and metastatic cancer but only in 32 of 59 patients (54 per cent) with superficially invasive (Jewett's stage 0, A or B 1) cancer (p equals 0.09). The degree of histologic differentiation of the tumor, available in 74 of the 100 transitional cell carcinoma patients with no concurrent infection, also appeared to correlate with urinary CEA values: CEA values were more than L5 ng. per ml. in 23 of 29 patients (79 per cent) with less differentiated (grade 3 or 4) tumors but in only 22 of 45 patients ( 49 per cent) with well differentiated (grade 1 or 2) tumors (p less than 0.05). We found no significant difference in the frequency of increased urinary CEA values between female and male patients with transitional cell carcinoma. This probably reflects the precautions and sterile method used in collecting urine samples. Four diagnostic procedures were performed on each of 79 patients and the results were positive as follows: urinary CEA value in 58 per cent, urine cytology (malignant cells) in 66 per cent, urinalysis (hematuria) in 80 per cent and biopsy at cystoscopy in 100 per cent. Urinary CEA measurement and cytologic examination, both non-invasive procedures, yielded similar frequencies of positivity. However, simultaneous performance of these 2 tests considerably increased the yield of positive results to 86 per cent. Biopsy at the time of cystoscopy remains the most accurate method for definitive diagnosis.
Comparison of preoperative and postoperative urinary CEA values. Partial cystectomy or transurethral resection: The 55 patients who underwent partial cystectomy or transurethral resection for transitional cell carcinoma were divided into 2 groups depending on their preoperative urinary CEA values: 1) abnormal preoperative CEA value (more than 1.5 ng. per ml.) and 2) normal preoperative CEA value (less than or equal to 1.5 ng. per ml.). Of the 31 patients whose CEA value 11,1as abnormal n~""""'r" (7 4
228
FRASER AND ASSOCIATES
tistically significantly higher than the concentrations found in normal bladder tissue, none of which was higher than 0.2 µg. per gm. wet weight. The concentrations found in these tumors were similar to those found in tumor tissue from 55 patients with colorectal cancer. 9
12
10
DISCUSSION
2
t
Preop
(12)
Postop
FIG. 2. Preoperative and postoperative (day 3) urinary CEA values in patients undergoing partial cystectomy or transurethral resection for transitional cell carcinoma. In 8 other patients urinary CEA concentration increased postoperatively.
per cent) the value had decreased by 3 days postoperatively; in 14 it decreased to normal (fig. 2). In the remaining 8 patients a further increase in CEA values was observed postoperatively; 4 of these patients had a urinary tract infection in the postoperative period. Urinary CEA values were normal preoperatively in 24 patients and in 18 of these they did not change postoperatively. In the remaining 6 patients urinary CEA values increased to abnormal levels postoperatively; 4 of these 6 patients had a urinary infection postoperatively and the other 2 had residual carcinoma. Radical cystectomy: Fourteen patients underwent total cystectomy. Urinary CEA values increased postoperatively in 6 of the 8 patients who had an ilea! conduit and in 5 of the 6 patients who had a ureterosigmoidostomy (fig. 3). In none of the patients could residual tumor be detected postoperatively. Urinary infection was not found postoperatively in those with ilea! conduits. The postoperative increase in urinary CEA values was considerably higher in patients with a ureterosigmoidostomy than in patients with an ileal conduit. Tumor CEA concentration. Vesical tumor tissue from 20 patients and normal vesical tissue from 7 patients were available. Mean (plus or minus standard error) CEA concentration in tumor was 3.65 plus or minus 0.54 µg. per gm. wet weight (range 0.6 to 20 µg. per gm.). This value was sta-
The measurement of CEA-like activity in the urine appears to be useful as an adjunct to the diagnosis of transitional cell carcinoma of the bladder in the absence of urinary tract infections. However, control patients with urinary tract infection had a frequency of abnormal values as high as that seen in transitional cell carcinoma patients. In addition, the presence of concurrent urinary tract infection may cause a further, and frequently substantial, increase in an already abnormal urinary CEA value in patients with transitional cell carcinoma. Hall and associates found a similar effect of urinary tract infection and also found high urinary CEA values in otherwise healthy persons (particularly women) when no ;;pecial p,ecautions were taken during urine collection. 4 By collecting the urine samples under sterile conditions at cystoscopy we apparently eliminated the problem of contamination. The frequency of abnormal urinary CEA values, as well as the magnitude of the values, increased with increasing extent of the cancer. Similar patterns have been found by Hall and associates• in ILEAL CONDUIT
URETEROSIGMOI DOSTOMY
>5,GOO
1,000
'
~i::,, c::
~-
100
~ lb
-!::
:S 10
Preop
Postop
Preop
Postop
FIG. 3. Preoperative and postoperative (day 3) urinary CEA values in patients undergoing radical cystectomy with creation of ilea! conduit or ureterosigmoidostomy. Passage of urine through different parts of intestine produces increase in urinary CEA values. This increase correlates with known secretory rates for CEA-like activity in human ileum and colon.
CLINICAL EVALUATION OF CARCINOEMBRYONIC ANTIGEN IN BLADDER CANCER
transitional cell carcinoma patients and by other investigators 2 in the serum of patients with nonurogenital cancers. Contrary to the results of Hall and associates, 4 we found that the degree of histologic differentiation of the tumor was correlated with the frequency of abnormal urinary CEA values: high grade malignancies were associated with abnormal values significantly more frequently than were low grade malignancies. The measurement of CEA-like activity in urine appears to be useful for diagnosis and prognosis in transitional cell carcinoma patients. Although the urinary CEA value had the lowest rate of the 4 diagnostic procedures used in same when this value was considered with the urine cytologic results, the diagnostic rate was considerably higher than that achieved with either test alone. This has added significance in view of the fact that 59 per cent of our 100 transitional cell carcinoma patients with no concurrent infection had non-invasive or superficially invasive resectable cancers. This diagnostic combination could be useful in screening persons at risk for bladder cancer, such as industrial workers and patients with premalignant lesions. 5 However, histologic confirmation by biopsy remains the definitive and most accurate method for diagnosis of transitional cell carcinoma. Urinary CEA values may be useful for determining the completeness of surgical resection in patients with transitional cell carcinoma. The change in CEA value at 3 days postoperatively correlated well with the extent of resection in the majority of transitional cell carcinoma patients who had abnormal urinary CEA values preoperatively and who underwent partial cystectomy or transurethral resection. Only in the presence of postoperative urinary infection was the correlation equivocal in these patients. When the preoperative urinary CEA value is normal in patients with transitional cell carcinoma, this value postoperatively does not appear to be a useful marker of extent of resection. Of the 24 patients in this group 18 had no postoperative change in the urinary CEA values, which suggests that the tumors probably were not producing CEA. We have observed a similar phenomenon in patients with advanced gastrointestinal cancer with normal serum CEA values. 10 The urinary CEA concentration does not appear to be useful as a marker in patients undergoing radical cystectomy with urinary diversion (ilea! conduit or ureterosigmoidostomy), regardless of the preoperative value. Postoperatively the values increased sharply, more so in patients with ureterosigmoidostomies than in those with ilea] conduits. Hall and associates observed that the CEA concentration was much higher in urine collected from the ilea! stoma than in urine collected at the mouth of the ureters in patients with ilea! conduits. 4 These increases in CEA values seem to be caused the passage of urine through different sections of the intestinal tract and contamination with its intraluminal secretions, Go and asso-
229
ciates (unpublished data) measured CEA-like activity in human intestinal secretions, by using an intestinal perfusion technique, and demonstrated that secretory rates for CEA were highest in the colon (100 times greater than in the ileum), which explains the difference between patients with ureterosigmoidostomies and those with ilea! conduits. Contrary to the favorable results obtained Reynoso and Hall and their associates the serum CEA value appeared to contribute little to the diagnosis of transitional cell carcinoma in our study. 3 , 4 This could be attributed to the fact that most of our patients had localized and resectable lesions or to the use of a higher (3.5 ng. per rnL) upper limit of normal in this study. It is well recognized that the frequency of abnormal serum CEA values increases with increasing extent of cancer. 2 There was no correlation between serum and urinary CEA values. This finding indirectly supports the theory that urinary CEA-like activity is of local origin ( urinary epithelium) and is not a product of renal filtration. There is no evidence thus far to indicate whether urinary CEA and serum CEA are the same. Mrs. Darlene L. Lucas and Mr. William M. Reilly provided technical assistance. REFERENCES
L Laurence, D. J. R
2.
3. 4.
5. 6. 7.
8.
9.
and Neville, A. M.: Foetal antigens and their role in the diagnosis and clinical management of human neoplasms: a review. Brit. J. Cancer, 26: 335, 1972. Dykes, P. W. and King, J.: Carcinoembryonic antigen (CEA). Gut, 13: 1000, 1972. 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. Hall, R R, Laurence, D. J. R, Neville, A. M. and Wallace, D. M.: Carcinoembryonic antigen and urothelial carcinoma. Brit. J. UroL, 45: 88, 1973. Prout, G. R, Jr.: The bladder. In: Cancer Medicine. Edited by J. F. Holland and K Frei, IIL Philadelphia: Lea & Febiger, pp. 1670-1680, 1973. Broders, A. C.: Carcinoma: grading and practical application. Arch. Path. Lab. Med., 2: 376, 1926. Lo Gerfo, P., Krupey, J. and Hansen, ff J.: Demonstration of an antigen common to several varieties of neoplasia: assay using zirconyl phosphate geL New EngL J. Med., 285: 138, 197L Ravry, M., McIntire, K R, Moertel, C. G., Waldmann, T. A, Schutt, A J. and Go, V. L W.: Brief communication: carcinoembryonic antigen and alpha-fetoprotein in the diagnosis of gastric and colonic cancer: a comparative clinical evaluation. J. Nat. Cancer Inst., 52: 1019, 1974. Go, V. L W., Spencer, R J., Ravry, M. J., Shorter, R G. and Huizenga, K. A.: Carcinoembryonic antigen (CEA) in malignant and inflammatory colonic tissue (abstract). Gastroenterology, 64:
734, 1973. 10. Ravry, M., Moertel, C. G., Schutt, A. J. and Go, V.
L. W.: Usefulness of serial serum carcinoembryonic antigen (CEA) determinations during anticancer therapy or long-term followup of gastrointestinal carcinoma. Cancer, ;34: 1230, 1974.
THE JOURNAL OF UROLOGY .
Copyright © 1975 by The Williams & Wilkins Co.
Printed in U.S.A.
CLINICAL EVALUATION OF URINARY AND SERUM CARCINOEMBRYONIC ANTIGEN IN BLADDER CANCER RICHARD A. FRASER, MARIO J. RAVRY, JOSEPH W. SEGURA
AND
VAY L. W. GO*
From the Gastroenterology Unit and the Department of Urology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
ABSTRACT
Urinary carcinoembryonic antigen-like activity was increased (more than 1.5 ng. per ml.) in 61 per cent of patients with transitional cell carcinoma of the bladder. The frequency of abnormal carcinoembryonic antigen values correlated with the extent and the grade of the tumor. However, urinary tract infection can produce abnormal carcinoembryonic antigen values in non-cancer patients. The use of urinary carcinoembryonic antigen measurement and urine cytologic examination for diagnosis of transitional cell carcinoma gave a better diagnostic yield than did use of either test alone. In transitional cell carcinoma patients with abnormal preoperative urinary carcinoembryonic antigen values, postoperative urinary carcinoembryonic antigen values were useful for determining completeness of tumor resection. However, with total cystectomy and creation of an ileal conduit or ureterosigmoidostomy, urinary carcinoembryonic antigen values increased markedly postoperatively. Serum carcinoembryonic antigen values were of little value in the diagnosis of transitional cell carcinoma. Advances in the field of cancer have raised hopes of earlier diagnosis of new or recurrent cases through the use of immunologic methods. 1 Recent reports suggest that the carcinoembryonic antigen (CEA) may be useful as a marker protein in the diagnosis of transitional cell carcinoma of the bladder. CEA is a polysaccharide-protein complex that is present in increased concentrations in the serum and tumor tissue of patients with various gastrointestinal and non-gastrointestinal cancers. 2 Reynoso and associates found the CEA concentration to be increased in the serum of 10 of their 20 patients with active transitional cell carcinoma of the bladder; in contrast, only 2 of 24 patients with inactive disease had an abnormal CEA value. 3 Furthermore, Hall and associates found abnormal urinary CEA values in two-thirds of their patients with active bladder cancer. 4 Herein we report the results of a study undertaken to determine the role of CEA in the evaluation of patients with vesical neoplasia and its usefulness as a marker protein or therapeutic indicator in the surgical treatment of these patients. MATERIALS AND METHODS
We studied 236 consecutive patients who were evaluated by the urology service because of a Accepted for publication November 8, 1974. Supported in part by Research Grant CB-23854 and Training Grant Tl-AM-5259 from the National Institutes of Health, Public Health Service, Bethesda, Maryland. * Requests for reprints: Mayo Clinic, Rochester, Minnesota 55901.
definite history or clinical susp1c10n of vesical neoplasia. All patients underwent diagnostic cystoscopic examination of the bladder. In all cases urine samples were collected by bladder catheterization under sterile conditions at the time of cystoscopy. Separate samples were obtained for CEA determination, urinalysis and cytologic examination. The patients were classified as having tumor, cystitis (infection) or no vesical abnormalities on the basis of clinical, cystoscopic, histologic and cytologic findings. Specifically, a patient was judged to have infection if urinalysis showed pyuria and the urine culture was positive. There were 110 untreated patients who had histologically confirmed transitional cell carcinoma of the bladder. In these patients the clinical stage of the disease, as defined by Jewett's criteria, 5 was determined on the basis of objective surgical and clinical evidence. The degree of histologic differentiation of the tumor was described by Broders' grading criteria. 6 Of the remaining 126 patients, 35 had cystitis and 91 had no vesical abnormalities at cystoscopy and no further clinical evidence suggestive of urinary tract cancer. These latter 2 groups constituted the control groups. Urinary CEA was measured by radioimmunoassay, using the method of Hansen after an initial extraction with an equal volume of 1.2 M perchloric acid. 7 There appeared to be no diurnal variation in urinary CEA values in 5 healthy subjects who underwent this evaluation (unpublished observations). Venous blood samples were obtained at the time of diagnosis in 90 of the 110 patients with confirmed vesical malignancy. 226
CLINICAL EVALUATION OF CARC!NOEMBRYONIC ANTIGEN IN BLADDER CANCER
Serum CEA was measured by direct radioimmunoassay, using the method of Hansen as modified by Ravry and associates. 8 Of 531 healthy controls studied by this method in our laboratory, 504 (95 per cent) have had serum CEA values no higher than 3.5 ng. per ml. Therefore, we accept 3.5 ng. per ml. as the upper limit of normal for serum CEA. Vesical tumor tissue was obtained from 20 at the time of operation. Normal bladder tissue also ,;vas obtained from non-invaded areas in 7 with small localized tumors. Tumor concentration was measured radioimthe method of Hansen after and extraction with L2 M perchloric acid. 7 The usefulness of urinary CEA as E marker for determining the completeness of surgical resection for vesical cancer was studied measuring urinary CEA preoperatively and 3 days postoperatively in 70 patients; 56 of these patients underwent transurethral resection or cystectorny, 8 had totai cystectomy with creation of an ilea! conduit and 6 had total cystectomy and ureterositmoidostomy. RESULTS
Serum CEA. The serum CEA concentration was greater than 3 .5 ng. per ml. in only 7 of the 90 patients (8 per cent) with active vesical cancer. Of these 7 patients 5 had Jewett's stage O lesions and only 2 had more extensive tumors. CEA. In control without urinary tract infection, urinary CEA values were higher than L5 ng. per ml. in 3 of the 91 subjects (3 per cent) (fig. 1). In contrast 23 of 35
>,;
v1
~
l;j 11..,
/(TCE!'IOO)
mo
~----0----------.., ~ , c o , ' .. o,r·-------·
I .· .,,el'
80
0
i:::
so
ti
'-!. ~
40
'!.I
~ ·-...
' ~
J /\
infl'lction 135)
' l :
t/~TCE c inf®ction
r~
('!OJ
1
:
20 :4.~ 1:
::i
§ \,.)
l
I .!------Controls c
-!'.!
:~
.. · · , , ,,.·
: 0
5
10
15
20
25
Urin@ CEA,ng/ml FIG. 1. Cumulative percentage distribution of actual urinary CEA values in controls, controls with urinary infection and transitional cell carcinoma patients with and without urinary infection. Presence of infection produces increase in frequency of abnormal urinary CEA values similar to that in transitional cell carcinoma patients v.rith no infection and actual concentrations are greater than in patients with transitional cell cafcino1na.
227
control patients (66 per cent) with infection had urinary CEA values greater than 1.5 ng. per ml., and their urinary CEA concentrations were greater than those in patients without infection. The urinary CEA concentration was more than 1.5 ng. per mL in 61 of the 100 patients with transitional cell carcinoma without concurrent urinary infection. the values were lower than those in controls with urinary infection. In 10 with transitional cell carcinoma with concurrent urinary there was a further increase in CEA concentration to values similar to those seen in control with urinary infection, In with transitional cell carcinoma but without concurrent infection, the of increased urinary CEA values to be related to the stage of the disease: the urinary CEA value was more than 1.5 ng. per mL in 29 of 41 patients (71 per cent) with invasive (Jewett's stage B 2 , C or D) and metastatic cancer but only in 32 of 59 patients (54 per cent) with superficially invasive (Jewett's stage 0, A or B 1) cancer (p equals 0.09). The degree of histologic differentiation of the tumor, available in 74 of the 100 transitional cell carcinoma patients with no concurrent infection, also appeared to correlate with urinary CEA values: CEA values were more than L5 ng. per ml. in 23 of 29 patients (79 per cent) with less differentiated (grade 3 or 4) tumors but in only 22 of 45 patients ( 49 per cent) with well differentiated (grade 1 or 2) tumors (p less than 0.05). We found no significant difference in the frequency of increased urinary CEA values between female and male patients with transitional cell carcinoma. This probably reflects the precautions and sterile method used in collecting urine samples. Four diagnostic procedures were performed on each of 79 patients and the results were positive as follows: urinary CEA value in 58 per cent, urine cytology (malignant cells) in 66 per cent, urinalysis (hematuria) in 80 per cent and biopsy at cystoscopy in 100 per cent. Urinary CEA measurement and cytologic examination, both non-invasive procedures, yielded similar frequencies of positivity. However, simultaneous performance of these 2 tests considerably increased the yield of positive results to 86 per cent. Biopsy at the time of cystoscopy remains the most accurate method for definitive diagnosis.
Comparison of preoperative and postoperative urinary CEA values. Partial cystectomy or transurethral resection: The 55 patients who underwent partial cystectomy or transurethral resection for transitional cell carcinoma were divided into 2 groups depending on their preoperative urinary CEA values: 1) abnormal preoperative CEA value (more than 1.5 ng. per ml.) and 2) normal preoperative CEA value (less than or equal to 1.5 ng. per ml.). Of the 31 patients whose CEA value 11,1as abnormal n~""""'r" (7 4
228
FRASER AND ASSOCIATES
tistically significantly higher than the concentrations found in normal bladder tissue, none of which was higher than 0.2 µg. per gm. wet weight. The concentrations found in these tumors were similar to those found in tumor tissue from 55 patients with colorectal cancer. 9
12
10
DISCUSSION
2
t
Preop
(12)
Postop
FIG. 2. Preoperative and postoperative (day 3) urinary CEA values in patients undergoing partial cystectomy or transurethral resection for transitional cell carcinoma. In 8 other patients urinary CEA concentration increased postoperatively.
per cent) the value had decreased by 3 days postoperatively; in 14 it decreased to normal (fig. 2). In the remaining 8 patients a further increase in CEA values was observed postoperatively; 4 of these patients had a urinary tract infection in the postoperative period. Urinary CEA values were normal preoperatively in 24 patients and in 18 of these they did not change postoperatively. In the remaining 6 patients urinary CEA values increased to abnormal levels postoperatively; 4 of these 6 patients had a urinary infection postoperatively and the other 2 had residual carcinoma. Radical cystectomy: Fourteen patients underwent total cystectomy. Urinary CEA values increased postoperatively in 6 of the 8 patients who had an ilea! conduit and in 5 of the 6 patients who had a ureterosigmoidostomy (fig. 3). In none of the patients could residual tumor be detected postoperatively. Urinary infection was not found postoperatively in those with ilea! conduits. The postoperative increase in urinary CEA values was considerably higher in patients with a ureterosigmoidostomy than in patients with an ileal conduit. Tumor CEA concentration. Vesical tumor tissue from 20 patients and normal vesical tissue from 7 patients were available. Mean (plus or minus standard error) CEA concentration in tumor was 3.65 plus or minus 0.54 µg. per gm. wet weight (range 0.6 to 20 µg. per gm.). This value was sta-
The measurement of CEA-like activity in the urine appears to be useful as an adjunct to the diagnosis of transitional cell carcinoma of the bladder in the absence of urinary tract infections. However, control patients with urinary tract infection had a frequency of abnormal values as high as that seen in transitional cell carcinoma patients. In addition, the presence of concurrent urinary tract infection may cause a further, and frequently substantial, increase in an already abnormal urinary CEA value in patients with transitional cell carcinoma. Hall and associates found a similar effect of urinary tract infection and also found high urinary CEA values in otherwise healthy persons (particularly women) when no ;;pecial p,ecautions were taken during urine collection. 4 By collecting the urine samples under sterile conditions at cystoscopy we apparently eliminated the problem of contamination. The frequency of abnormal urinary CEA values, as well as the magnitude of the values, increased with increasing extent of the cancer. Similar patterns have been found by Hall and associates• in ILEAL CONDUIT
URETEROSIGMOI DOSTOMY
>5,GOO
1,000
'
~i::,, c::
~-
100
~ lb
-!::
:S 10
Preop
Postop
Preop
Postop
FIG. 3. Preoperative and postoperative (day 3) urinary CEA values in patients undergoing radical cystectomy with creation of ilea! conduit or ureterosigmoidostomy. Passage of urine through different parts of intestine produces increase in urinary CEA values. This increase correlates with known secretory rates for CEA-like activity in human ileum and colon.
CLINICAL EVALUATION OF CARCINOEMBRYONIC ANTIGEN IN BLADDER CANCER
transitional cell carcinoma patients and by other investigators 2 in the serum of patients with nonurogenital cancers. Contrary to the results of Hall and associates, 4 we found that the degree of histologic differentiation of the tumor was correlated with the frequency of abnormal urinary CEA values: high grade malignancies were associated with abnormal values significantly more frequently than were low grade malignancies. The measurement of CEA-like activity in urine appears to be useful for diagnosis and prognosis in transitional cell carcinoma patients. Although the urinary CEA value had the lowest rate of the 4 diagnostic procedures used in same when this value was considered with the urine cytologic results, the diagnostic rate was considerably higher than that achieved with either test alone. This has added significance in view of the fact that 59 per cent of our 100 transitional cell carcinoma patients with no concurrent infection had non-invasive or superficially invasive resectable cancers. This diagnostic combination could be useful in screening persons at risk for bladder cancer, such as industrial workers and patients with premalignant lesions. 5 However, histologic confirmation by biopsy remains the definitive and most accurate method for diagnosis of transitional cell carcinoma. Urinary CEA values may be useful for determining the completeness of surgical resection in patients with transitional cell carcinoma. The change in CEA value at 3 days postoperatively correlated well with the extent of resection in the majority of transitional cell carcinoma patients who had abnormal urinary CEA values preoperatively and who underwent partial cystectomy or transurethral resection. Only in the presence of postoperative urinary infection was the correlation equivocal in these patients. When the preoperative urinary CEA value is normal in patients with transitional cell carcinoma, this value postoperatively does not appear to be a useful marker of extent of resection. Of the 24 patients in this group 18 had no postoperative change in the urinary CEA values, which suggests that the tumors probably were not producing CEA. We have observed a similar phenomenon in patients with advanced gastrointestinal cancer with normal serum CEA values. 10 The urinary CEA concentration does not appear to be useful as a marker in patients undergoing radical cystectomy with urinary diversion (ilea! conduit or ureterosigmoidostomy), regardless of the preoperative value. Postoperatively the values increased sharply, more so in patients with ureterosigmoidostomies than in those with ilea] conduits. Hall and associates observed that the CEA concentration was much higher in urine collected from the ilea! stoma than in urine collected at the mouth of the ureters in patients with ilea! conduits. 4 These increases in CEA values seem to be caused the passage of urine through different sections of the intestinal tract and contamination with its intraluminal secretions, Go and asso-
229
ciates (unpublished data) measured CEA-like activity in human intestinal secretions, by using an intestinal perfusion technique, and demonstrated that secretory rates for CEA were highest in the colon (100 times greater than in the ileum), which explains the difference between patients with ureterosigmoidostomies and those with ilea! conduits. Contrary to the favorable results obtained Reynoso and Hall and their associates the serum CEA value appeared to contribute little to the diagnosis of transitional cell carcinoma in our study. 3 , 4 This could be attributed to the fact that most of our patients had localized and resectable lesions or to the use of a higher (3.5 ng. per rnL) upper limit of normal in this study. It is well recognized that the frequency of abnormal serum CEA values increases with increasing extent of cancer. 2 There was no correlation between serum and urinary CEA values. This finding indirectly supports the theory that urinary CEA-like activity is of local origin ( urinary epithelium) and is not a product of renal filtration. There is no evidence thus far to indicate whether urinary CEA and serum CEA are the same. Mrs. Darlene L. Lucas and Mr. William M. Reilly provided technical assistance. REFERENCES
L Laurence, D. J. R
2.
3. 4.
5. 6. 7.
8.
9.
and Neville, A. M.: Foetal antigens and their role in the diagnosis and clinical management of human neoplasms: a review. Brit. J. Cancer, 26: 335, 1972. Dykes, P. W. and King, J.: Carcinoembryonic antigen (CEA). Gut, 13: 1000, 1972. 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. Hall, R R, Laurence, D. J. R, Neville, A. M. and Wallace, D. M.: Carcinoembryonic antigen and urothelial carcinoma. Brit. J. UroL, 45: 88, 1973. Prout, G. R, Jr.: The bladder. In: Cancer Medicine. Edited by J. F. Holland and K Frei, IIL Philadelphia: Lea & Febiger, pp. 1670-1680, 1973. Broders, A. C.: Carcinoma: grading and practical application. Arch. Path. Lab. Med., 2: 376, 1926. Lo Gerfo, P., Krupey, J. and Hansen, ff J.: Demonstration of an antigen common to several varieties of neoplasia: assay using zirconyl phosphate geL New EngL J. Med., 285: 138, 197L Ravry, M., McIntire, K R, Moertel, C. G., Waldmann, T. A, Schutt, A J. and Go, V. L W.: Brief communication: carcinoembryonic antigen and alpha-fetoprotein in the diagnosis of gastric and colonic cancer: a comparative clinical evaluation. J. Nat. Cancer Inst., 52: 1019, 1974. Go, V. L W., Spencer, R J., Ravry, M. J., Shorter, R G. and Huizenga, K. A.: Carcinoembryonic antigen (CEA) in malignant and inflammatory colonic tissue (abstract). Gastroenterology, 64:
734, 1973. 10. Ravry, M., Moertel, C. G., Schutt, A. J. and Go, V.
L. W.: Usefulness of serial serum carcinoembryonic antigen (CEA) determinations during anticancer therapy or long-term followup of gastrointestinal carcinoma. Cancer, ;34: 1230, 1974.
