Serum tissue polypeptide antigen (S-TPA) in bladder cancer as a tumor marker. A prospective study

394

Serum Tissue Polypeptide Antigen (STPA) in Bladder Cancer as a Tumor Marker A Prospective Study C. Maulard, M.D.,* M . E. Touberf, M.D.,§ Y. Chrefien, M.D.,t S. Delanian, M.D.,* B. Dufour, M.D.,t and M . Housset, M.D.*

Background. Tissue polypeptide antigen (TPA) is a differentiation and a proliferation tissue marker of nonsquamous epithelia. Increased urinary and serum TPA (S-TPA)levels were found in some patients with invasive bladder cancer. The authors report the results of a prospective study evaluating the role of serum TPA (S-TPA) in bladder carcinoma. Methods. S-TPA concentrations were measured by radioimmunoassay in 53 patients with invasive bladder tumor before treatment, at the end of treatment, and during follow-up. The upper normal limit of the test was set at 80 UI/ml. Results. With a specificity of 100%, the diagnostic sensitivity of the test was 54.7%. S-TPA was increased in 88% of patients with N1 + N2 disease compared with 38.8% of the patients with NO disease (P = 0.01)and in 100% of patients with metastatic disease and 48% of patients with nonmetastatic disease (P = 0.01).S-TPA was increased in 23% of patients with total macroscopic debulking and in 68% of patients with persistent macroscopic disease ( P = 0.004).For patients staged NO MO, no statistical correlation between S-TPA level and debulking by transurethral resection (TUR) was found (P = 0.15).In the subset of patients with normal pretherapeutic S-TPA levels, 75% achieved a complete response, compared with 44.8% of the patients with initial elevated STPA levels ( P = 0.04).However, there was no statistically significant relationship between pretherapeutic S-TPA levels and immediate response to treatment according to the stratification for tumor volume after initial debulking by TUR. For a mean follow-up of 15 months -+ 7 months, median survival time and 1-year survival rates were studied in the subset of patients with limited disease (NO MO) according to the pretherapeutic S-TPA levels. The median survival time was not reached, and the From the *Service d’Oncologie-RadiothCrapie and tService d’lirologie B, Hapital Necker, Paris, France; and the GService de MCdecine NuclCaire, Hdpital Saint-Louis, Pans, France. Address for reprints: C. Maulard, Service dOncologie-Radiothhrapie, H6pital Necker, 149 Rue de Sevres, 75015 Paris, France. Accepted for publication .

1-year survival rate was 80% when the initial S-TPA level was normal; these were 10 months and 44%, respectively, when the S-TPA level was high (P < 0.01).Among the 31 patients who achieved a complete response, 9 experienced a relapse, with an increase of the S-TPA level in 8 patients. Conclusions. The S-TPA level is correlated with initial tumor volume. It appears to be a prognostic factor and a valuable parameter for follow-up. Cancer 1994; 73~394-8.

Key words: tumor marker, serum tissue polypeptide antigen, bladder cancer, radioimmunoassay.

Tissue polypeptide antigen (TPA) is a tumor-related protein, originally isolated from extracts of pooled tumors. Although the epithelial distribution and biochemical characteristics of TPA show some similarity to intermediate filament proteins (i.e., low molecular weight cytokeratins), recent data have demonstrated that active and clinically relevant TPA is distinct from cytokeratins.’ Actually, it can be considered as a differentiation and a proliferation epithelial tissue marker of normal nonsquamous epithelia and derived neoplasms. The serum content of TPA was found to be elevated in various malignant disorders’ and in some inflammatory conditions such as Crohn‘s disease and he pa ti ti^.^ TPA is present in normal, inflammatory, and in malignant ~rothelium.~ No tumor marker has been available for diagnosis and follow-up of bladder cancer, so TPA could be helpful with both aims. Several studies have shown elevated urinary TPA levels in bladder-invasive tumors, probably because of an increased urothelial cell turnover and a u t ~ l y s i s but , ~ , ~only a studies few have evaluated serum TPA (S-TPA) levels in such urothelial We report the results of a prospective study evaluating the role of S-TPA in diagnosis and follow-up of bladder carcinoma.

Bladder Cancer and S-TPA/Maulard et al.

Patients and Methods Patients Between September 1989 and September 1991,53 patients were referred to Necker Hospital (Paris) for biopsy -proven transitional bladder carcinoma. The median age was 68 years +- 9 years (range, 45-85 years). Forty-six (87%) patients were men, and 7 (13%) were women. Among the patients, there were 1 TO, 2 TI, 27 T2 (51%), 17 T3 (32%), 4 T4, and 2 Tx tumors (1987 TNM classification). All patients but one with tumor Stage TO and T l had pelvic relapse, a node, or metastatic involvement of a previously treated bladder cancer. One patient with a large and multirecurrent T1 Grade G3 tumor was included in the study. Histologic grading of the primary was G1 in 2 patients, G2 in 7 patients, G3 in 40 patients (75%), and unknown in 4 patients. The nodal extent was evaluated on pelvic and abdominal computed tomography scan. Thirty-six patients were free of radiologic adenopathies (Stage NO). Seventeen (32%) had nodal involvement: of the pelvic area in 10 patients ( N l ) and of the para-aortic areas in 7 patients (N2). Six (11Yo) patients had metastatic disease (Ml). The location of metastatic disease was as follows: bone, four patients; skin, two patients; adrenal gland, one patient.

Treatment Treatment was based on surgery, chemotherapy, or radiation therapy according to the tumor extent. Twentyeight patients were treated with a 5-fluorouracil/cisplatin combination and concomitant pelvic radiation therapy (total dose to 65 Gy); 20 patients preoperatively received the same combination (total dose to 45 Gy), followed by a total cystectomy; all patients with a metastatic disease received chemotherapy alone (Methotrexate, Vinblastine, Adriamycin, Cyclophosphamide regimen). FOllOW-Up

For patients treated with curative intent, clinical examination, urinary cytology, cystoscopy (with biopsy if necessary), and pelvic computed tomography scan were systemically performed every 6 months and before the reappearance of symptoms. For patients with metastatic disease, the follow-up depended on performance status, symptoms, and location of metastasis.

Laboratory Methods For each patient, the S-TPA level was analyzed by radioimmunoassay" using a commercially available kit

395

(Prolifigen R-TPA IRMA, Pharmacia, Santec Medical, Sweden) before the beginning of treatment, at the end of treatment, and during follow-up. The upper normal limit of the test was set at 80 UI/ml and was based on studies of serum from 89 healthy control subjects (95% of the control subjects had S-TPA values of less than 65 UI/ml; for 100°/~of control subjects, the value was less than 80 UI/ml). For patients without metastatic extent, the mean interval between transurethral resection (TUR) and pretherapeutic S-TPA dosage was 36 days (range, 15-75 days).

Statistical Methods Analysis was performed in June 1992. Data were evaluated in an univariate study using contingency table analysis and the chi-square test (with Yates adjustment or with Fisher's exact test when the number of patients was too small). Survival times were measured from the date of diagnosis. Survival curves were calculated by the Kaplan-Meier method and compared by log-rank testing.

Results Evaluation of Pretherapeutic S-TPA The pretherapeutic S-TPA level was determined in all the 53 patients studied. Twenty-nine had elevated levels of S-TPA before treatment. The sensitivity of S-TPA was 54.7%. Specificity and predictive value of positive test were 100% because of the upper limit chosen (see Laboratory Methods). Predictive value of negative test was 79%. Tumor extent. As shown in Table 1, there was a statistically significant correlation between the pretherapeutic level of S-TPA and the nodal or metastatic staging. S-TPA was increased in 88% of patients with N1 N2 disease compared with 38.8% of patients with NO disease (P = 0.01). In the same way, all of the six patients with metastatic disease had an elevated pretherapeutic level of S-TPA and only 48% of the patients with nonmetastatic disease did (P = 0.01). No correlation between S-TPA level and primary tumor staging was seen. Finally, pretherapeutic S-TPA level was studied according to the residual tumor mass after TUR in all but one patient (1 Mx disease). S-TPA levels were increased in 4 of 17 (23%) patients for whom a total macroscopic debulking had been achieved (TO NO MO) and in 24 of 35 (68%) patients with persistent macroscopic disease (P = 0.004). For patients with disease staged NO MO, no correlation was seen between S-TPA levels and debulking by TUR: S-TPA was elevated in 9 of 18 (50%) patients with incomplete resection and in 23% of

+

396

CANCER January 15,1994, Volume 73,No. 2 Table 1. Correlation of Pretherapeutic S-TPA Level to Tumor Staging and Histological Grading No. of patients

No. of patients with elevated S-TPA

T2* T3 + T4 NOt N 1 + N2

27 21 36 17

14

MOS M1 GI§

46

22 6 0

G2

6 2 7

G3

40

% of patients with elevated S-TPA

51 52

11

14 15

38.8 88 48 100

57 57.5

4 23

P value NS 0.01 0.01

NS

*TO:1 patient; TI: 2 patients; Tx: 2 patients (see text).

t NO: no clinical evidence of nodal involvement, $ Staging unknown for 1 patient. 5 Grading unknown for 4 patients. No = number of patients; S-TPA:serum-tissue polypeptide antigen.

patients after complete resection, as previously described (P = 0.15). Histologic grading. No correlation between S-TPA level and primary tumor histologic grading was found (see Table 1). Immediate response to treatment. All patients were evaluable for response to treatment: 31 (58.5%) had disease in complete remission, 7 had disease in partial remission, and 15 had stable or progressive disease. As shown in Table 2, there was a statistically significant correlation between the pretherapeutic level of S-TPA and the response to treatment. In the subset of patients with normal pretherapeutic S-TPA levels, 75% achieved a complete remission, compared with 44.8% of the patients with initial elevated S-TPA levels (P = 0.04). This correlation was no longer significant when the residual tumor mass after debulking by TUR was considered, but the number of patients is small in each group. In the subset of patients with complete resection, 10 of 13 (77%) with normal pretherapeutic S-TPA levels achieved a complete remission, compared with 2 of 4 with elevated S-TPA levels. In the subset of patients with persistent macroscopic disease, 8 of 11 (72%) with

Evaluation of Posttherapeutic S-TPA Levels for Follow-up

Table 2. Correlation of Pretherapeutic Serum-Tissue Polypeptide Antigen Level to Immediate Response to Treatment Complete response Normal 24 patients Elevated 29 patients

18 75 % 13

44.8%

normal pretherapeutic S-TPA levels achieved complete disease remission, compared with 10 of 24 (42%) with elevated S-TPA (P = 0.15). Survival. The mean follow-up was 15 months f 7 months. Survival was studied for patients with comparable tumor extent, staged NO MO, according to the pretherapeutic S-TPA level. Twenty-two patients had a normal pretherapeutic S-TPA level and 13 an increased S-TPA level. There was no significant difference between the two groups related to the quality of resection by TUR: complete debulking was achieved in 59% and 31% of these patients, respectively (P = 0.15). For the patients with initial normal S-TPA levels, the median survival time was not reached, and the 1-year survival rate was 80%. For the patients with initial elevated STPA levels, the median time and l-year survival rate were 10 months and 44%, respectively (P < 0.01; see Fig. 1).The cause of death was as follows: in the subset of patients with initial normal S-TPA levels, four patients died, with bladder cancer progression in three (14%) patients; in the other group, six patients died, with bladder cancer progression in five (65%; P = 0.16) patients.

Failure 6 25% 16 55.2%

Response to treatment. All but two patients with normal pretherapeutic S-TPA levels had normal posttherapeutic S-TPA levels. Of the patients with increasing S-TPA levels, one had a nodal para-aortic progression developed at the end of treatment; the other was free of disease at 21 months of follow-up. Twentysix of the 29 patients with elevated pretherapeutic lev-

Bladder Cancer and S-TPAIMaulard et al.

397

':i 60

un 20

-

D

C

0.01

0

els were evaluable for subsequent S-TPA sampling during follow-up. Of these, the S-TPA level was normalized in seven patients; all of these patients responded (six had histologic complete responses and one had complete clinical response). Of the 19 remaining patients with persistent increased S-TPA levels, only 5 (26%) had response (4 had clinical complete responses and 1 had histologic response). Four experienced relapse at 5, 6, 6, and 11 months, respectively, and the latter had a lung carcinoma develop 12 months later. Relapse. Among the 31 patients who achieved a complete response, 26 were evaluable for S-TPA levels during follow-up. Nine experienced disease relapse, with an increase of the S-TPA level in eight patients at the time of the relapse. Discussion

Tumor markers are potentially useful in screening for cancer, monitoring the course of the disease, and detecting relapse. Unfortunately, no tumor marker is routinely available for bladder carcinoma. TPA was first isolated in 1957 from membranes of various cancers and has been successfully demonstrated in cancer cells by immunocytochemical assay. S-TPA concentrations were found to be elevated in many types of malignant disease, but there are little data in the literature regarding serum TPA levels in patients with bladder cancer. The interest of S-TPA in bladder cancer has been especially noted since 1984 by Adolphs and Oehr,'-" by Oehr et al.," and by Ruther et al." In these studies, the sensitivity of S-TPA was high, (approximately 70-80% for invasive tumors), but the threshold was low (37-50 UI/ml). In our series, the sensitivity was only 54.7% because of an upper threshold chosen (80 UI/ml); it reached 83% if the upper normal limit was set at 37 UI/ml (data not shown). However, the specificity of the test dropped for this cutoff level, and the test failed to distinguish superficial and invasive turn or^.^'^

We failed to demonstrate any correlation between S-TPA level and primary tumor staging. One can note that this staging could be somewhat doubtful depending on the deepness of the tumor resection. In previous published series, the S-TPA level was classically compared for noninvasive and invasive tumors without any analysis related to the tumor extent through the bladder wall for the latter. In addition, we failed to demonstrate any correlation between S-TPA level and primary tumor grading, possibly because of the disproportion in the number of patients between the three groups (Gl, G2, and G3). On the contrary, we found a statistically significant correlation between the pretherapeutic level of S-TPA and the nodal or metastatic staging. Such a correlation has been reported by Adolphs and Oehr.'," In these studies and in ours, the S-TPA level is correlated to the tumor volume at the beginning of treatment. In our series, considering a subset of patients with disease exclusively located to the bladder, quality of debulking by TUR seems to interfere with pretherapeutic S-TPA level: increased S-TPA level was found in twice as many patients after incomplete TUR than after complete resection. Nevertheless, this correlation between S-TPA and TUR was not significant, probably because there were too few patients studied. One important observation was that patients who responded to treatment classically achieved a normalization of TPA level, whereas patients with a progressive clinical course continued to have high serum TPA levels. Immediate complete response to treatment is more often obtained in the subset of patients with initial normal S-TPA levels. This relationship between S-TPA level and immediate response to treatment was not significant when residual tumor mass after debulking by TUR was considered, but the number of patients in each subset was small. A trend can be noted: in the group of patients with persistent disease, 72% with normal pretherapeutic S-TPA levels achieved a complete response, compared with 42% with initially increased S-TPA levels. For patients with normal pretherapeutic S-TPA levels and limited disease, the median survival time is longer, and overall survival rates may be better, although the follow-up time has been insufficient to reach that conclusion with certainty. In this group, 14% of patients died of bladder cancer; 65% of patients with initially increased S-TPA levels died of bladder cancer. In conclusion, the S-TPA level is correlated with bladder tumor volume. It appears as a prognostic factor for extra-bladder tumor extent, possibly for response to treatment and survival. It probably is not useful in the primary diagnosis because a low specificity is associated with a cutoff of 80 UI/ml. However, the S-TPA level

398

CANCER Ianuary 15, 1994, Volume 7 3 , No. 2

could be useful in the course of follow-up and in patients with an increase in the level for monitoring of relapse or metastatic disease.

6. Carbin BE, Ekman P, Eneroth P, Nilsson B. Urine TPA (tissue

References

I", Manna A, Marafini V, Valli C, Marchei P, et al. Tissue and serum TPA in subjects at risk for bladder cancer. Int ] Biol Markers 1989; 4:13-7. 8. Adolphs HD, Oehr P. Significance of plasma tissue polypeptide antigen determination for diagnosis and follow-up of urothelial bladder cancer. Urol Res 1984; 12125-8. 9. Khanna OP, Wu B. Tissue polypeptide antigen (TPA) as a predictor for genitourinary cancers and their metastases. Urology

1. Bjorklund B, Bjorklund V. The enigma of human tumor marker: TPA revisted. Third International Conference on Human Tumor Markers: Biology and clinical applications, Naples. 1986:24. 2. Luthgens M, Schlegel G. Combined use of carcinoembryonic antigen and tissue polypeptide antigen in Oncologic therapy and surveillance. Cancer Detect Prevent 1983; 6:51-9. 3. Nathrath WBJ, Heidenkummer P, Bjorklund V, Bjorklund B. Distribution of tissue polypeptide antigen (TPA) in normal human tissue: immunohistochemical study on unfixed methanol-, ethanol-, and formalin-fixed tissue. ] Histochem Cytochem 1985; 33:99-109. 4. Carbin BE, Collins VP, Ekman P. Tissue polypeptide antigen

(TPA), some cytokeratins and epithelial membrane antigen (EMA) in normal, inflamed and malignant urothelium. Urol Res 1987; 15:191-4. 5. Senatore S, Zizzi L, Blasi C, Alfieri G, Saccatani Jotti G, Gabrielli M, et al. Tissue polypeptide antigen (TPA) in urinary bladder cancer cytology: a follow-up study. Oticology 1990; 47:37-42.

polypeptide antigen), flow cytometry and cytology as markers for tumor invasiveness in urinary bladder carcinoma. Urol Res 1989; 17:269-72. 7. Giovagnori MR, Di Fraia

1987; 30~106-10. 10. Oehr P, Westrour A, Derigs G , Borman R. Evaluation and charac-

terization of tumor associated antigens by the inverse distribution function. Tumor Diagn 1981; 2195. 11. Oehr P, Vogel J, Maisey R, Jager J, Vahlensieck N. TPA assay in diagnosis and follow-up of urinary bladder carcinomas. In: Wust G, editor. Tumor markers, current aspects and clinical relevance. Proceedings. Munster, West Germany, 1986. 12. Ruther U, Luthgen M, Baeurle K, Rasseweiler J, Bub P, Eissenberger F, et al. Surveillance and M-VEC therapy monitoring of patients with advanced carcinoma of the urinary tract by TPA. Tumor Marker Oncol 1989; 4:303.