Int. J. Cancer (Pred. Oncol.): 84, 525–528 (1999) r 1999 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer CONCENTRATION OF FREE hCG␤ SUBUNIT IN SERUM AS A PROGNOSTIC MARKER FOR SQUAMOUS-CELL CARCINOMA OF THE ORAL CAVITY AND OROPHARYNX Johan HEDSTRÖM1*, Reidar GRENMAN2, Hans RAMSAY1, Patrik FINNE3, Johan LUNDIN4, Caj HAGLUND5, Henrik ALFTHAN3 and Ulf-Håkan STENMAN3 1Department of Otorhinolaryngology, University of Helsinki, Helsinki, Finland 2Department of Otorhinolaryngology-Head and Neck Surgery and Medical Biochemistry, University of Turku, Turku, Finland 3Department of Clinical Chemistry, University of Helsinki, Helsinki, Finland 4Department of Oncology, University of Helsinki, Helsinki, Finland 5Department of Surgery, University of Helsinki, Helsinki, Finland This study was conducted to evaluate the clinical usefulness of serum hCG␤ in the diagnosis and prognosis of patients (n ⴝ 59) with cancers of the oral cavity and oropharynx. As a reference marker we used squamous-cell carcinoma antigen (SCCAg). A blood sample was obtained from all patients before primary surgery. Serum hCG␤ was determined by a time-resolved immunofluorometric assay (IFMA) and SCCAg by a solid phase immunoenzymometric assay. Elevated preoperative hCG␤ levels were observed in 8 (14%) and elevated SCCAg in 12 (20%) out of 59 patients. Patients with preoperatively elevated hCG␤ had a shorter recurrence-free survival when compared with those with normal hCG␤ levels (logrank Chi-squared ⴝ 6.83, p ⴝ.009), and the risk-ratio for recurrence during follow-up for those was 3.6 (95% CI ⴝ 1.29–9.94). In a Cox multivariate model hCG␤ (p ⴝ 0.039) and stage (p ⴝ 0.044) were independent prognostic factors. SCCAg showed no correlation with recurrence-free survival. We conclude that determination of hCG␤ in serum is a potential marker in the prognostic evaluation of patients with SCC of the oral cavity and oropharynx. Int. J. Cancer (Pred. Oncol.) 84:525–528, 1999. r 1999 Wiley-Liss, Inc. Head and neck cancer is the sixth most common type of cancer in the world (Åkervall, 1998). Histopathologically, squamous-cell carcinoma (SCC) is by far the most frequent histological type of this disease (Rassekh et al., 1994). After curative treatment, about 50% of the patients develop a recurrence, and 80% of the relapses occur within the first 2 years of follow-up (Terhaard et al., 1992). About 50% of the patients with SCC of the head and neck (SCCHN) die from cancer with no significant improvement during recent decades (Rassekh et al., 1994). In the management of patients with SCCHN, treatment decisions are still based on the TNM classification system, although T and N status do not reliably predict the clinical outcome in individual cases (Bailey, 1991; Hanna et al., 1990). Thus, effective treatment planning would be enhanced by identification of prognostic indicators that would reflect the biological behavior of a tumor. The identification of a reliable circulating tumor marker therefore has great potential. Although new potential prognostic markers are continuously being introduced and evaluated in SCCHN, so far none has influenced standard treatment (Rassekh et al., 1994; Hanna et al., 1990; Snyderman et al., 1995). Human choriogonadotropin (hCG) is a glycoprotein hormone produced by the placenta. It is widely used for diagnosis of pregnancy, pregnancy-associated disorders and trophoblastic disease. hCG consists of 2 dissimilar subunits designated ␣ and ␤. During pregnancy, the proportion of hCG␤ is 1 to 5% of hCG immunoreactivity, and in healthy nonpregnant subjects, the proportion of hCG␤ to hCG is 10 to 30% (Alfthan et al., 1992a,b). In gestational trophoblastic disease, the serum and urine levels of hCG␤ are almost always elevated (Fan et al., 1987), and elevated in 35% in testicular cancer (Saller et al., 1990). In nontrophoblastic cancers, elevated levels have most often been observed in patients with tumors of the pancreas (30–72%) (Marcillac et al., 1992; Alfthan et al., 1992b), stomach (9%) (Marcillac et al., 1992), and liver (11%) (Marcillac et al., 1992), in gynecological cancers (26%) (Marcillac et al., 1992) and in bladder cancer (35–47%) (Marcillac et al., 1992; Iles et al., 1996). In urothelial and ovarian cancer determination of hCG␤ is of prognostic value (Iles et al., 1996; Ind et al., 1997). hCG␤ expression has been identified in 29 of 45 (64%) of SCC from the oral cavity (Bhalang et al., 1999), indicating that hCG␤ is produced by the majority of these tumors. One study was conducted to determine serum concentration of hCG␤ in head and neck cancer. In this study, all 7 patients showed hCG␤ levels below the detection limit of the assay (Marcillac et al., 1992). We have now studied whether hCG␤ has any diagnostic and prognostic value in the evaluation of patients with cancer of the oral cavity and oropharynx. As a reference marker, we used squamous-cell-carcinoma antigen (SCCAg), one of the few useful markers for patients with SCC of the head and neck (Eibling et al., 1989; Snyderman et al., 1995). Among different forms of SCC of the head-and-neck region, tumors of the oral cavity and oropharynx are particularly associated with poor prognosis. We therefore included only patients with tumors of this type. MATERIAL AND METHODS Patient samples The study included 59 consecutive patients presenting with SCC of the oral cavity or oropharynx admitted to the Department of Otorhinolaryngology—Head-and-Neck Surgery of Turku University Central Hospital between 1989 and 1996 (Table I). All these patients underwent surgery, and 50 patients (85%) received radiation therapy. A blood sample was obtained from each patient before surgery and before radiation therapy. Of the patients, 27 were female and 32 male. Ages ranged from 37 to 92 years, with a mean of 67 years. Tumor histology, site, stage, grade, and recurrences were analyzed; 38 patients (64%) showed no recurrence within the follow-up period that ranged from 14 to 93 months (median 33 months). The follow-up program included clinical evaluation and, when indicated, endoscopic examination every third month for 2 years, and thereafter during 1 to 3 annual follow-up visits. Recurrences were detected in 21 patients (36%), and 7 patients Grant sponsors: Research Foundation Finska Läkaresällskapet, Research Foundation Liv och Hälsa and Research Institute, Helsinki University Central Hospital. *Correspondence to: Department of Otorhinolaryngology, Helsinki University Central Hospital, FIN-00290 Helsinki, Finland. Fax: ⫹358–04714804. E-mail: firstname.lastname@example.org Received 29 March 1999; Revised 5 June 1999 HEDSTRÖM ET AL. 526 TABLE I – SITE, STAGE AND GRADUS OF 59 PATIENTS WITH SQUAMOUS-CELL CARCINOMA OF ORAL CAVITY AND OROPHARYNX INCLUDED IN THIS STUDY Site of SCC Oral cavity Tongue Floor of mouth Alveolus and gingiva Hard palate Buccal mucosa Lip Oropharynx Tonsil n Stage Grade I II III IV I II III 22 6 12 3 5 5 6 1 1 1 1 3 5 2 3 0 3 1 7 0 3 0 0 3 4 3 5 2 1 0 11 3 6 2 2 4 9 2 3 1 2 1 2 1 3 0 1 0 6 0 1 3 2 2 3 1 (12%) died of the disease. Clinical data on recurrence and overall survival were retrieved from clinical records. Serum samples were stored at ⫺20°C until assayed. Assays Serum SCCAg was determined by the IMX method (Abbott, Abbott Park, IL). The upper reference limit used for SCCAg in serum was 2.5 µg/l. Serum hCG␤ was determined by a time-resolved immunofluorometric assay (IFMA) based on 2 monoclonal antibodies to hCG␤ (Alfthan et al., 1988). The detection limit of the assay is 0.2 pmol/l and the upper reference limit based on the 97.5 percentile of apparently healthy controls is 2.0 pmol/l. The cross-reaction with hCG, LH and LH␤ is ⱕ0.1% (Alfthan et al., 1992a). Statistical analysis Life-tables were calculated according to Kaplan-Meier. Patients were divided into groups having a pre-operative tumor marker value above or below the upper reference limit, and their recurrencefree survival was compared. The statistical significance of the difference in recurrence-free survival between the groups was calculated using the log-rank test. In addition, a Cox multivariate survival analysis was performed, including the following covariates: age, gender, stage and grade of the tumor. The correlation between the values of hCG␤ and SCCAg was calculated by the Mann-Whitney U test. Recurrence-free survival and overall survival were defined by the time intervals between primary surgical treatment and detection of recurrent disease, and death from SCC, respectively. TABLE II – UNIVARIATE ANALYSIS OF THE RELATIONSHIP BETWEEN CLINICAL CHARACTERISTICS AND RECURRENCE-FREE SURVIVAL IN 59 PATIENTS WITH SQUAMOUS-CELL CARCINOMA OF ORAL CAVITY AND OROPHARYNX* Clinicopathological variable Pre-operative hCG␤ ⬎2.0 pmol/l ⱕ2.0 pmol/l Pre-operative hCG␤ Continuous variable Pre-operative SCCAg ⬎2.5 µg/l ⱕ2.5 µg/l Age, years ⬎64 ⱕ64 Gender Female Male Grade I II III Stage I II III IV Nodal state N⫺ N⫹ Number of Median recurrencepatients free survival in months 8 51 4.4 15.1 59 2 p 6.83 0.009 6.71 0.01 12 47 7.5 14.2 0.27 NS 30 29 13.8 10.7 0.04 NS 27 32 15.2 11.6 0.63 NS 30 21 8 9.1 15.2 10.7 0.52 NS 13 15 14 17 15.3 14.2 6.7 5.9 7.7 0.006 37 22 14.2 6.7 5.05 0.025 *2 RESULTS The median values for hCG␤ was 0.73 pmol/l and the range 0.10–12.2 pmol/l. Elevated pre-operative hCG␤ levels were observed in 8 out of 59 patients (14 %) and elevated SCCAg (median 1.6 µg/l, range 0.30–10.0 µg/l) in 12 (20%) of the patients (Table II) (Figs. 1 and 2). There was no correlation between preoperative hCG␤ and SCCAg levels (r ⫽ 0.09, p ⫽ 0.498). Patients with pre-operatively elevated hCG␤ (⬎2.0 pmol/l) had a shorter recurrence-free survival when compared with those with normal hCG␤ levels (p ⫽ 0.009) (Fig.1). The risk ratio for recurrence during follow-up for those with elevated hCG␤ was 3.6 (95% CI ⫽1.29–9.94). Of the patients with pre-operatively elevated serum hCG␤ 70% had recurrences within 12 months of follow-up, while recurrences were detected in only 20% of those with normal levels during the same period. Entering hCG␤ as a continuous variable in a Cox regression analysis also showed a significant association (p ⫽ 0.01) with recurrence-free survival. There was no association between recurrence-free survival and gender (p ⫽ 0.43), age (p ⫽ 0.84) or tumor grade (p ⫽ 0.77). However, there was a significant association with stage of the disease (p ⫽ 0.023, RR 3.03, 95% CI ⫽ 1.17–7.84). In a multivariate survival analysis, hCG␤ and stage were independent and corresponding p values were calculated by the log-rank test. When the variable examined had 3 or more ordered categories, the log-rank test for trend was used. NS, not significant. prognostic factors (Table III). SCCAg showed no correlation with recurrence-free survival (p ⫽ 0.605) (Table II) (Fig. 2). Two of the patients with elevated hCG␤ (25%) and 5 of the patients with normal hCG␤ (10%) died during follow-up (median 33 months). The corresponding figures for SCCAg were 2 patients (17%), and 5 patients (11%), respectively. Because of the small number of events, the difference in overall survival was not determined. DISCUSSION Our results showed that patients with SCC of the oral cavity and oropharynx with a pre-operatively elevated hCG␤ levels in serum (⬎2.0 pmol/l) (14%) had shorter recurrence-free survival than those with normal hCG␤ (p ⫽ 0.009). The risk for recurrent disease in patients with pre-operatively elevated hCG␤ was 3.6-fold that of patients with normal hCG␤. HCG␤ IN SERUM AND ORAL CANCER PROGNOSIS 527 FIGURE 1 – Kaplan-Meier life-tables for patients with squamous-cell carcinoma of the oral cavity and oropharynx (n ⫽ 59) grouped by pre-operative serum hCG␤ value. The median recurrence-free survival for 51 patients (dashed line) with an hCG␤ value ⬍2.0 pmol/l was 15 months, that for 8 patients (solid line) with an hCG␤ values ⱖ2.0 pmol/l, 4.4 months (p ⫽ 0.009). FIGURE 2 – Kaplan-Meier life-tables for patients with squamous-cell carcinoma of the oral cavity and oropharynx (n ⫽59) grouped by pre-operative serum SCCAg value. The median recurrence-free survival for 47 patients (dashed line) with an SCCAg value ⬍2.5 pmol/l was 14 months, that for 12 patients (solid line) with an SCCAg value ⱖ2.5 pmol/l, 7.5 months (NS). Several markers have been studied with a view to improving evaluation of patients with SCCHN; they are: carcino-embryonic antigen, CA 19–9, CA 125, alkaline phosphatase, CYFRA 21–1, immune complexes, lipid-associated sialic acid, and SCCAg (Doweck et al., 1995; Hanna et al., 1990; Rassekh et al., 1994). However, although some of these are useful for monitoring some patients, they do not provide prognostic information independent of stage. SCCAg, which was used as a reference marker in the present study, was originally isolated and purified from a SCC of the uterine cervix (Kato and Torigoe, 1977). Elevated serum levels of SCCAg have been reported in patients with carcinoma of the uterine cervix, esophagus and lung and are used for monitoring these diseases (Kato and Torigoe, 1977). Earlier studies have shown significant differences in postoperative SCCAg levels between patients who remain disease-free and those who suffer relapse (Eibling et al., 1989; Snyderman et al., 1995). However, in accordance with the results of the present study, pre-operative SCCAg-levels do not differentiate between the 2 groups (Eibling et al., 1989; Snyderman et al., 1995). HCG␤ has been shown to be expressed by a variety of non-trophoblastic tumors (Iles et al., 1990). HCG␤ lacks gonadotropic activity (Marcillac et al., 1992), and little is known about its biological function. However, has been shown to stimulate phospholipid methylation in Leydig cells (Ronco et al., 1993) and also the growth of bladder cancer cell lines (Gillott et al., 1996). Thus, hCG␤ may act as an autocrine growth factor for some tumor cells. TABLE III – MULTIVARIATE ANALYSIS (COX PROPORTIONAL-HAZARDS MODEL, BACKWARD STEPWISE, p TO REMOVE 0.05) OF RECURRENCE-FREE SURVIVAL IN 59 PATIENTS WITH SQUAMOUS-CELL CARCINOMA OF ORAL CAVITY AND OROPHARYNX* Co-variate p 95% CI hCG␤ Stage Grade Gender Age 0.039 0.044 NS NS NS 1.01–1.33 1.03–7.20 *hCG␤ was entered as ⬍2.0 pmol/l ⫽ 0, ⱖ2.0 pmol/l ⫽ 1, stage as an ordinal variable (stage I–II ⫽ 0, stage III–IV ⫽ 1), grade as an ordinal variable, high ⫽ 0, moderate ⫽ 1, low ⫽ 2, gender as male ⫽ 0, female ⫽ 1, and age as ⬍64 years ⫽ 0, ⱖ64 years ⫽ 1. CI, confidence interval; NS, not significant. Such a function could explain why expression of hCG␤ is a sign of poor prognosis. Our results suggest that hCG␤ is an independent prognostic marker in cancers of the oral cavity and oropharynx. Elevated hCG␤ in serum identifies a high-risk group of patients. This finding could be used to improve planning of surgical treatment and subsequent follow-up in order to detect recurrent disease at an early stage. We conclude that hCG␤ is a new potential marker in the prognostic evaluation of patients with SCC of oral cavity and oropharynx. Further studies to confirm these results should be performed in larger patient groups and with longer follow-up. REFERENCES ÅKERVALL, J., Prognostic factors in squamous-cell carcinoma of the head and neck, p. 103, University of Lund, Lund (1998). ALFTHAN, H., HAGLUND, C., DABEK, J. and STENMAN, U.H., Concentrations of human choriogonadotropin, its beta-subunit, and the core fragment of the beta-subunit in serum and urine of men and nonpregnant women. Clin. Chem., 38, 1981–1987 (1992). ALFTHAN, H., HAGLUND, C., ROBERTS, P. and STENMAN, U.H., Elevation of free beta subunit of human choriogonadotropin and core beta fragment of human choriogonadotropin in the serum and urine of patients with malignant pancreatic and biliary disease. Cancer Res., 52, 4628–4633 (1992). ALFTHAN, H., SCHRODER, J., FRASER, R., KOSKIMIES, A., HALILA, H. and STENMAN, U.H., Choriogonadotropin and its beta subunit separated by hydrophobic-interaction chromatography and quantified in serum during pregnancy by time-resolved immunofluorometric assays. Clin. Chem., 34, 1758–1762 (1988). BAILEY, B.J., beyond the ‘‘new’’ TNM classification. Arch. Otolaryngol. Head Neck Surg., 117, 369–370 (1991). BHALANG, K., KAFRAWY, A. and MILES, D., Immunohistochemical study of the expression of human chorionic gonadotropin-beta in oral SCC. Cancer, 85, 757–762 (1999). DOWECK, I., BARAK, M., GREENBERG, E., URI, N., KELLNER, J., LURIE, M. and GRUENER, N., Cyfra 21–1. A new potential tumor marker for squamouscell carcinoma of head and neck. Arch. Otolaryngol. Head Neck Surg., 121, 177–181 (1995). 528 HEDSTRÖM ET AL. EIBLING, D.E., JOHNSON, J.T., WAGNER, R.L. and SU, S., SCC-RIA in the diagnosis of squamous-cell carcinoma of the head and neck. Laryngoscope, 99, 117–124 (1989). FAN, C., GOTO, S., FURUHASHI, Y. and TOMODA, Y., Radioimmunoassay of the serum free beta-subunit of human chorionic gonadotropin in trophoblastic disease. J. clin. Endocrinol. Metab., 64, 313–318 (1987). GILLOTT, D.J., ILES, R.K. and CHARD, T., The effects of beta-human chorionic gonadotrophin on the in vitro growth of bladder cancer cell lines. Brit. J. Cancer, 73, 323–326 (1996). HANNA, E.Y., PAPAY, F.A., GUPTA, M.K., LAVERTU, P. and TUCKER, H.M., Serum tumor markers of head and neck cancer: current status. Head Neck, 12, 50–59 (1990). ILES, R.K., PERSAD, R., TRIVEDI, M., SHARMA, K.B., DICKINSON, A., SMITH, P. and CHARD, T., Urinary concentration of human chorionic gonadotrophin and its fragments as a prognostic marker in bladder cancer. Brit. J. Urol., 77, 61–69 (1996). ILES, R.K., PURKIS, P.E., WHITEHEAD, P.C., OLIVER, R.T., LEIGH, I. and CHARD, T., Expression of beta human chorionic gonadotrophin by nontrophoblastic non-endocrine ‘‘normal’’ and malignant epithelial cells. Brit. J . Cancer, 61, 663–666 (1990). IND, T., ILES, R., SHEPHERD, J. and CHARD, T., Serum concentrations of cancer antigen 125, placental alkaline phosphatase, cancer-associated serum antigen and free beta human chorionic gonadotrophin as prognostic markers for epithelial ovarian cancer. Brit. J. Obstet. Gynaecol., 104, 1024–1029 (1997). KATO, H. and TORIGOE, T., Radioimmunoassay for tumor antigen of human cervical squamous-cell carcinoma. Cancer, 40, 1621–1628 (1977). MARCILLAC, I., AND 17 OTHERS. Free human chorionic gonadotropin beta subunit in gonadal and nongonadal neoplasms. Cancer Res., 52, 3901–3907 (1992). RASSEKH, C.H., JOHNSON, J.T. and EIBLING, D.E., Circulating markers in squamous-cell carcinoma of the head and neck: a review. Europ. J. Cancer B. Oral Oncol., 30B, 23–28 (1994). RONCO, A.M., LLANOS, M.N. and VALLADARES, L.E., Human chorionic gonadotropin and free beta sub-units stimulate phospholipid methylation in intact rat Leydig cells. Steroids, 58, 314–319 (1993). SALLER, B., CLARA, R., SPOTTL, G., SIDDLE, K. and MANN, K., Testicular cancer secretes intact human choriogonadotropin (hCG) and its free beta-subunit: evidence that hCG (⫹hCG-beta) assays are the most reliable in diagnosis and follow-up. Clin. Chem., 36, 234–239 (1990). SNYDERMAN, C.H., D’AMICO, F., WAGNER, R. and EIBLING, D.E., A reappraisal of the squamous-cell carcinoma antigen as a tumor marker in head and neck cancer. Arch. Otolaryngol. Head Neck Surg., 121, 1294– 1297 (1995). TERHAARD, C.H., HORDIJK, G.J., VAN DEN BROEK, P., DE JONG, P.C., SNOW, G.B., HILGERS, F.J., ANNYAS, B.A., TJHO-HESLINGA, R.E. and DE JONG, J.M., T3 laryngeal cancer: a retrospective study of the Dutch Head and Neck Oncology Cooperative Group: study design and general results. Clin. Otolaryngol., 17, 393–402 (1992).