Serum Sialyl Tn as an Independent Predictor of Poor Prognosis in Patients With Epithelial Ovarian Cancer By Hiroshi Kobayashi, Toshihiko Terao, and Yoshiro Kawashima Purpose: Monoclonal antibody (moAB) TKH-2 directed to the tumor-associated O-linked sialyl 2-6-at-Nacetylgalactosaminyl (sialyl Tn; STN) epitope was generated by immunization with ovine submaxillary mucin (Kleldsen et al, Cancer Res 48:2214-2220, 1988). We investigated whether circulating serum levels of STN antigen might influence the prognosis of patients with ovarian cancer. Patients and Methods: Serum samples were obtained from 126 healthy nonpregnant women, 157 patients with benign gynecologic disease, and 89 patients with histologically proven epithelial ovarian cancer. Circulating serum STN-antigen concentrations (U/mL) were determined by a competitive radiolmmunoassay kit (Otsuka Assay Laboratories, Tokushima, Japan) In a one-step procedure. Results: Serum antigen levels were elevated in 48.3% of the patients. The levels of STN antigen were signifi-
T
cantly higher in the sera of patients with cancer when compared with levels in benign and healthy controls (P < .05). The 5-year survival rate for patients with STNnegative (serum STN levels < 50.0 U/mL) versus STNpositive (> 50 U/mL) tumors was 76.9% versus 10.8%, respectively (P < .05). The progression-free interval (PFI) at 5 years was 51.9% versus 5.4%, respectively (P < .05). The overall survival probability and PFI were worse in patients with STN-positive sera. Multivariate regression analysis revealed that stage, residual tumor size, positive STN, performance status, and histologic grade were the five important variables for predicting overall survival. Conclusion: We conclude that a positive STN-antigen level in sera is an independent predictor of poor prognosis in ovarian cancer. J Clin Oncol 10:95-101. G 1992 by American Society of Clinical Oncology.
patients with ovarian cancer."5 However, the serum test is not sensitive enough to be useful for screening. Higher
esis) participate in a variety of specific biologic functions. 2 Prognostically important differences in tumor biology may still be due to qualitative changes in tumor mucin." Qualitative rather than quantitative mucin alterations might be important in the biology of cancer. To investigate this possibility, we have examined the correlation between the prognosis of patients with epithelial ovarian cancer (overall survival and progres-
mean CA125 levels are also observed in endometriosis,
sion-free interval [PFI]) and circulating serum levels of a
pelvic inflammatory disease, unexplained infertility, and
mucin carbohydrate antigen (STN) associated with the earliest steps of mucin glycosylation.
HE ELEVATION
of various circulating serum
tumor markers have been reported to correlate with the occurrence of malignancy in cancer.' 2 CA125, a glycoprotein antigen recognized by the monoclonal antibody (moAB) OC125,' 4 has been the most useful antigen for detection of ovarian cancer and follow-up of
during menstruation.'-"
In addition to its high false-
positive rate, the insufficient sensitivity of CA125 in mucinous-type ovarian cancer is another shortcoming. New moABs recognizing a core structure of a mucintype carbohydrate chain have been prepared by Kjeldsen et al. 1 MoABs TKH-1 and TKH-2 directed to the tumor-associated
O-linked
sialyl 2-6-a-N-acctylgalac-
tosaminyl (sialyl Tn; STN) epitope were generated by immunization with ovine submaxillary mucin. The moABs that recognize the carbohydrate structures of cell-
surface glycoconjugates of cancer cells have been regarded as useful tools for characterizing cell types and for cancer diagnosis via assays of antigenic glycoproteins secreted into the bloodstream. The potential for a serum immunoassay of the antigenic determinant STN as a
blood test to screen for ovarian cancer and, subsequently, to observe its clinical course, has not yet been realized. Cell-surface glycolipids or glycoproteins (the composition has been shown to change during tumorigen-
PATIENTS AND METHODS Some patients were recruited from the well-characterized population reported in a previous study on STN." We obtained serum samples from September 1985 through March 1989 at the Department of Obstetrics and Gynecology, Hamamatsu University Hospital, and its related hospitals. Samples were obtained from 157 patients with benign gynecologic disease including 37 uterine myomas, 35 cases adenomyosis, 85 benign ovarian tumors, and 89 cases of histologically proven epithelial ovarian cancer. Table 1 summarizes the clinicopathologic data of the 89 patients with
From the Department of Obstetrics and Gynecology, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka, Japan. Submitted May 20, 1991; acceptedAugust 7, 1991. Address reprintsrequest to Hiroshi Kobayashi, MD, Department of Obstetrics and Gynecology, Hamamatsu University School of Medicine, Handa-cho 3600, Hamamatsu, Shizuoka, 431-31, Japan. C 1992 by American Society of Clinical Oncology. 0732-183X/92 /1001-0016$3.00/0
Journalof Clinical Oncology, Vol 10, No 1 (January), 1992: pp 95-101
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
95
96
KOBAYASHI, TERAO, AND KAWASHIMA Table 1. Distribution of Serum STN-Antigen Levels No. of
Patient Population
Patients
Healthy controls Benign disease Ovarian cancer Stage I Stage II Stage III1 Stage IV
126 157 23 18 38 10
Mean -- SD
(U/mL) 23.3 ± 8.9 29.8 - 9.6 35.5 + 41.7 48.0 ± 62.6 148.0 + 173.1 283.4 ± 305.6
No. of Positive
Cases (%) 5(4.0) 15 (9.6) 6(26.1) 6 (33.3) 24 (63.2) 7 (70.0)
NOTE. Cut off value of serum STN antigen is 41.1 U/mL (mean ± 2 SD).
ovarian cancer included in the current study. Diagnoses of all patients were confirmed by review of operative reports and pathology reports. Staging of ovarian cancer according to the International Federation of Gynecology and Obstetrics (FIGO) classification showed 23 patients with stage I disease, 18 with stage II, 38 with stage III, and 10 with stage IV. All the assays were obtained before initial surgery for those patients with ovarian cancer and benign disease, and stored at -80 0 C until use. All patients were initially treated with optimal debulking surgery followed by combination chemotherapy including cisplatin 50 mg/m2, Adriamycin (doxorubicin; Adria Laboratories, Columbus, OH) 50 mg/m 2, and cyclophosphamide 500 mg/m2. Judgment of disease progression was based on chart review, image diagnosis, and evaluation of objective intraoperative observation. Serum samples from 126 healthy female volunteers, ranging in age from 25 to 71 years, served as normal controls. Circulating serum STN-antigen concentrations (U/mL) were determined by a competitive radioimmunoassay kit (Otsuka Assay Laboratories, Tokushima, Japan) that uses moAB TKH-2 in a one-step procedure. Technical characteristics of the assay are 14 described elsewhere. Briefly, beads coated with STN-antigen purified from meconium were incubated with 50 .LL of serum and 200 CIL of radioactive iodine-labeled TKH-2 antibody solution for 90 minutes. Radioactivity bound to beads was measured by a gammacounter. All samples were assayed in duplicate. Levels more than 41.1 U/mL (normal mean + 2 SD) were considered elevated determinations. Survival curves were calculated from the first day of treatment by the Kaplan-Meier method," and tests of differences in survival and 2 PFI distributions were based on a generalized Wilcoxon test.'" A x 7 test was used to test for differences between groups of subjects.' ,• Cox's proportional hazards model was used to determine the effect of the prognostic factors on survival using the likelihood ratio test."
RESULTS
Sera from healthy controls and patients with benign disease and ovarian cancer at various stages were assayed for STN-antigen levels (Table 1). The cutoff value of STN was arbitrarily set at the mean - 2 SD of healthy controls (41.1 U/mL) to define the positive rate. The mean value and its standard deviation was 23.3 -+ 8.9 U/mL. STN-antigen levels were elevated in five control patients, all over the age of 45 years, with a false-positive rate of 4.0%. The serum level of the antigen was not affected by menstrual cycle (data not
shown). Of the 157 patients with nonmalignant gynecologic disease, 15 (9.6%) had STN-antigen levels of more than 41.1 U/mL: two of 37 patients with uterine myomas had levels of 58 and 62 U/mL, four were cases of adenomyosis, five were cases of serous cystadenoma of the ovary, one case was a dermoid cyst, and the other three cases were mucinous cystadenomas. Elevated levels of serum STN antigen (mean, 113.9 U/mL) were seen in patients with ovarian cancer when compared with those of patients with benign disease and normal controls. The difference in STN-antigen levels between cancer patients and controls or patients with benign disease was significant (P < .05). Of the 89 patients with ovarian cancer, 43 (48.3%) showed STN levels over the cut-off value: six patients with stage I disease (26.1%), six with stage II (33.3%), 24 with stage III (63.2%), and seven with stage IV (70.0%) had elevated antigen levels. Stage I and II cases showed a significantly low frequency of elevation and low values compared with more advanced cases. Significant correlation was seen between elevation of serum STN-antigen levels and the size of the primary tumor (data not shown). These results indicate that serum STN-antigen levels correlate to the stage of disease. The diameter of the largest remaining tumor nodule after laparotomy was greater than 10 cm in eight patients, 2 to 10 cm in 26, and less than 2 cm in the remaining 55 patients. Of the 89 patients, 43 had an elevated serum STN level. The residual tumor mass was greater than 10 cm in seven patients, 2 to 10 cm in 16, and less than 2 cm in 20 (Table 2). The frequency of elevated serum STN levels varied with residual tumor size, and this phenomenon was statistically significant (> 10/< 10 cm and > 2/< 2 cm, P < .05). The relationship between the STN-antigen levels and histology for the patients with cancer is shown in Table 3. The high STN-antigen levels and the high sensitivity in patients with mucinous adenocarcinoma (59.3%) are particularly interesting. Table 4 indicates the relationship between serum STN-antigen status and clinicopathologic features of the tumor. In the present study, patients with serum STN levels more than 50 U/mL (mean - 3 SD) were chosen as STN-positive, because this cutoff value (50 U/mL) could most significantly reflect patient outcome as to Table 2. Serum STN Level and Residual Tumor Size Serum STN Level
Residual
Tumor Size (cm)
Elevated
Normal
>10 2-10 41.1 U/mL(%)
prognosis. There was no statistically significant difference between STN status and age, histologic category, and histologic grade. Of the 89 patients' sera tested, 40 of 52 patients with STN-negative sera (serum STNantigen levels < 50 U/mL) survived for 5 years, whereas those with STN-positive sera (serum STN-antigen levels > 50 U/mL) had a significantly worse outcome (10.8% overall 5-year survival v 76.9%, respectively). The overall survival probability was better in patients with STN-negative sera (P < .05). The median duration of survival of patients with STN-positive sera was 17 months, whereas that of patients with STN-negative sera has not been reached and will exceed 60 months (Fig 1). The estimated 5-year PFI was also better in patients with STN-negative sera compared with those with STNpositive sera (51.9% v 5.4%; P < .05; Fig 2). When 5-year survival was analyzed by both stage and STN status, patients with STN-negative sera had a better prognosis in all stages (Fig 3). The overall 5-year survival of patients as a function of histologic category plus STN status is shown in Fig 4. Even in mucinous carcinoma, patients with STN-positive sera had a significantly poorer prognosis than those with STN-negative sera (P < .05).
Variable
---- ---
U) u L
Months On Study
Fig 1. Overall 5-year survival rate for patients with STN-positive (serum STN levels o50 U/mL, n = 37; --) v STN-negotive (serum STN levels < 50 U/mL, n = 52;- - -) tumors.
The influences of both histologic grade and STN status are demonstrated in Fig 5. For each histologic grade, the overall 5-year survival was worse in patients with STNpositive sera (P < .05). A similar tendency was noted in PFI for the effects of STN status plus stages, histologic category, and histologic grade (Figs 6, 7, and 8, respectively). Subsequently, the overall 5-year PFI was also worse in patients with STN-positive sera as each function of STN status plus stage, histologic category, or histologic grade (P < .05). STN has been shown to be associated with early relapse of this malignancy. Multivariate regression analysis was performed to evaluate further potential prognostic factors (Table 5). Cox regression analysis revealed that stage (stages III and IV), residual tumor size (> 2 cm), STN-positivity (serum STN levels > 50 U/mL), performance status (3
STN-Positive* (n = 37)
STN-Negative (n = 52)
54.9 (35-74)
57.0 (34-70)
4 5
19 13
C 0
22
16
0
6
4
14 16 5 2
24 11 9 8
10 16 11
17 22 13
*Patients with serum STN levels > 50 U/mL were labeled as STNpositive.
5,
o= 5, C
a al
Months On Study Fig 2. 5-year progression-free survival for patients with STNpositive (serum STN levels > 50 U/mL, n = 37; --) v STNnegative (serum STN levels < 50 U/mL, n = 52; - - -) tumors.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
98
KOBAYASHI, TERAO, AND KAWASHIMA .. A
>
0
(nC:
41 U
"
0) 0o
S60 40
-
20 C
C
12
24
36
48
60
D
12
24
36
48
12
24
36
60
48
60
Months On Study
Months On Study
Fig 3. Overall survival of patients by stage of disease and STN (A) Stage I, (B) ), STN-negative (---). status: STN-positive (stage II, (C) stage III, and (D) stage IV.
and 4), and histologic grade (grade 3) were significant negative predictors of survival. STN has been shown to be independently associated with prognosis. DISCUSSION
Oncogenic transformation is often associated with changes in glycosylation in glycoproteins or glycolipids in cancer cells. Carbohydrate antigens with changes in glycosylation have been recently used as tumor markers. Tumor-associated antigens identified by moABs have been used for serologic detection of malignancies. Many tumor markers have the structure of carbohydrate antigens. '"2' STN is considered to be the antigen of the core region of mucin oligosaccharides and is defined by its moAB, TKH-2." In the present study, we monitored the levels of circulating serum STN antigen in patients with ovarian cancer. We further investigated whether
Fig 5. Overall survival of patients as a function of the histologic grade and STN status: STN-positive ( ), STN-negative (- - -). (A) Grade 1, (B) grade 2, and (C) grade 3.
serum STN status plays a role in the biologic behavior of this malignancy. First, we examined the clinical usefulness of the management of STN-antigen levels for the diagnosis of patients with ovarian cancer. The results showed that 48.3% of patients with surgically demonstrated ovarian cancer had elevated STN-antigen levels (>_ mean + 2 SD). The demonstration that elevated concentrations of STN circulate in the blood of patients with ovarian cancer suggests that cell-surface STN glycoconjugate is shed into the systemic circulation. In contrast, STN expression in normal and nonmalignant sera is highly restricted. The specificity of the STN-antigen assay was much greater than that of CA125. The extremely low concentrations of STN antigen in normal or nonmalignant sera can be interpreted to mean that (1) little STN antigen is 10
---------
60 40 20
4)
SA 100o
]oo-
-
12
.100 o0
24
36
48
1-DO-
----
-,
20 12
24
36
4
so
D
12
24
36
48
0.
Fig 4.
201
12
24
-
2
4
36 648 44
60
Months On Study
Months On Study Overall survival of patients by histologic category and STN
), STN-negative (---). status: STN-positive (-mucinous, (C) endometrioid, and (D) clear.
C
- -
40
Q- 40
C
60
(A) Serous, (B)
Fig 6. Progression-free survival of patients by disease stage and STN status: STN-positive (-), STN-negative (-- -). (A) Stage I, (B) stage II, (C) stage III, and (D) stage IV.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
SIALYL TN IN OVARIAN CANCER 100 - -
C
a)
-
- .
12
24
99
20
A
610
4
36
o
43 0
100
0,C
a,
L_____.
0 20
D
12
24
36
48
60
Months On Study
Fig 7. Progression-free survival of patients by histologic category and STN status: STN-positive (), STN-negative (---). (A) Serous, (B) mucinous, (C) endometrioid, and (D) clear.
synthesized or (2) the linkage N-acetylgalactosamine (GalNAc) of normal mucin preferentially receives an N-acetylglucosamine (GIcNAc) residue first, because adding a sialic acid first would prevent further oligosaccharide elongation. 21 '2 2 However, ovarian cancer patients with stage I and II disease showed a significantly low frequency of elevation and low values compared with more advanced cases. The correlation between the positivity of the serum STN test and the disease stage seems to be entirely based on the extent of disease. The frequent expression of the antigen in sera of cancer patients suggests that STN may be a tumor-associated antigen secreted into the circulation. The appearance of core-region antigens provides evidence for incomplete glycosylation in cancer cells. The considerable expression of STN antigen represents another example of precursor accumulation in cancer cells. 22
Q
0o Ln
U) ft. 0
4100
b6o-•-----_
_
40 20C
? ... 12
24
36
4
6b
Months On Study Fig 8. Progression-free survival of patients as a function of histologic grade and STN status: STN-positive (), STN-negative (- - -). (A) Grade 1, (B)grade 2, and (C)grade 3.
Regarding histologic type, it is interesting to note the high positive rate of STN in patients with mucinous carcinoma of the ovary. The qualitative mucin alterations that occur in ovarian cancer, particularly in mucinous-type ovarian cancer, may be important in the biologic behavior of the cancer relating to poor prognosis. This supports the observations by Itzkowitz et al21 concerning the histologic type of colorectal cancer and the strong staining-intensity and high proportion of cells expressing Tn (GalNAc-R) and/or STN. These antigens appear to be good markers of colon cancer mucin that is elaborated by both signet ring-cell carcinoma and colloid carcinoma. As patients with mucinous carcinoma and poorly differentiated adenocarcinoma of the colon often have a poor prognosis, the Tn and STN antigens may be useful markers for these subsets of patients with 2 1 22 colon cancer. ' Our results demonstrate that those ovarian cancer patients with serum STN-antigen levels of 2 50 U/mL had a shorter survival rate than patients with negative STN-antigen levels (< 50 U/mL). In addition, the patients with positive STN-antigen levels had a shorter PFI than those with nonelevated values. The poor survival rate of patients with elevated STN-antigen levels may reflect their higher tumor burden. The apparently better outcome of patients with STNnegative sera cannot be ascribed to confounding factors such as younger age or a well-differentiated type of cancer. By multivariate regression analysis, stage, residual tumor size, and STN status were variables of importance in predicting both PFI and overall survival. STN has been shown to be independently associated with prognosis and a strong predictor of survival. This antigen may be of considerable importance in deciding which postresection patients require further therapy. It is not known whether the levels of circulating serum STN antigen reflect the changes in expression of this antigen on cancer-cell membranes. We are currently investigating the relationship between the expression of STN antigen in cancer tissues by immunohistochemical staining and circulating serum levels of this antigen. We have demonstrated an apparent correlation between the expression of STN by immunohistochemical staining and the concentration of serum STN antigen (unpublished data, October 1991). In addition, the expression of STN antigen in cancer tissues is also correlated with the expression of urokinase-type plasminogen activator (uPA; data not shown). Malignant cells of various types typically synthesize high levels of such proteases as uPA, collagenase, cathepsin B, or cathepsin D. These proteases may facilitate the translocation of tumor cells
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
100
KOBAYASHI, TERAO, AND KAWASHIMA Table 5. Proportional Hazards General Linear Model of Survival Time Variables
Age Stage Histology Grade STN status Residual tumor Performance status
3 -0.07352467 3.13585068 -0.02179025 1.57468468 -2.09452792 -2.36481201 1.90546703
SE
X2
P
0.02974634 0.89436015 0.20427873 0.53127939 1.00834570 0.83468001 0.51252389
3.08 18.05 0.01 11.20 16.12 14.56 21.34
.0228 .0001 .8460 .0001 .0001 .0001 .0001
contributes to survival time. Namely, stage, residual tumor size (< 2 cm), STN status (STN-negativity), performance NOTE. The absolute value of 13 status, and grade of differentiation (grade 1) were positive predictors of survival or favorable prognostic factors, in this order.
through the extracellular matrix.23 Consequently, qualitative alterations of cell-surface glycoproteins or glycolipids and augmented expression of these proteases are considered to be a common feature of the malignant phenotype. Springer 24 has shown that the relative proportions of T (galactose 1,3 GlcNAc-R) and Tn antigens on human carcinomas frequently correlate with the carcinoma's aggressiveness or invasiveness. Recently, we have simultaneously measured serum CA72-4- and CA54/61-antigen levels by a sandwich radioimmunoassay using moABs B72.3 and CC49, and moABs MA54 and MA61, respectively, in patients with ovarian cancer. 25-28 There were many points of similarity in data among these three antigens (STN, CA72-4, and CA54/61). CA72-4 and STN antigens may have an identical antigenic determinant epitope, NeuAc 2-6 Gal NAc 1-Ser/Thr..12 . This finding seems to explain why the results of the present study of STN are similar to those of our previously reported studies of CA72-4 or
CA54/61 (data not shown). It will be necessary to confirm the differences among these antigens by measuring their serum levels in the same samples from a greater number of subjects and by biochemical studies to characterize the oligosaccharide structures of these antigens. We feel that the immunoassay for CA72-4 or CA54/61 in serum also may have reasonable clinical potential, not only as a cancer diagnostic tool or monitoring test for cancer, but also as a prognostic factor. A large series of additional prospective studies should evaluate this issue.
ACKNOWLEDGMENT We are greatly indebted to the personnel at Otsuka Assay Laboratories (Manager, Dr Sadahito Shin, Tokushima, Japan) for the extensive assistance provided to us in the measurement of serum STN antigen levels. We also thank Dr Kazuhiro Sumimoto (Hamamatsu University School of Medicine, Hamamatsu, Japan) and Dr Tetsuya Tachikawa (Otsuka Assay Laboratories, Tokushima, Japan) for the statistical analysis and their valuable advice.
REFERENCES 1. Del Villano BC, Brennan S, Brock P, et al: Radioimmunometric assay for a monoclonal antibody-defined tumor marker, CA19-9. Clin Chem 29:549-552, 1983 2. Herlyn M, Sears HF, Steplewski Z, et al: Monoclonal antibody detection of a circulating tumor-associated antigen. 1. Presence of antigen in sera of patients with colorectal, gastric, and pancreatic carcinoma. J Clin Immunol 2:135-140, 1982 3. Kabawat SE, Bast RC, Bahn AK, et al: Tissue distribution of a coelomic epithelium-related antigen recognized by the monoclonal antibody OC125. Int J Gynecol Pathol 2:275-285, 1983 4. Davis HM, Zurawsky VR, Bast RC Jr, et al: Characterization of the CA125 antigen associated with human epithelial ovarian carcinoma. Cancer Res 46:6143-6151, 1986 5. Bast RC Jr, Feeney M, Lazarus H, et al: Reactivity of a monoclonal antibody with human ovarian carcinoma. J Clin Invest 68:1331-1337, 1981 6. Bast RC, Klug TL, St. John E, et al: A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N Engl J Med 309:883-887, 1983 7. Klug TL, Bast RC Jr, Niloff JM, et al: A monoclonal antibody immunoradiometric assay for an antigenic determinant (CA125)
associated with human epithelial ovarian carcinomas. Cancer Res 44:1048-1053, 1984 8. Barbieri RL, Niloff JM, Bast RC Jr, et al: Elevated serum concentrations of CA125 in patients with advanced endometriosis. Fertil Steril 45:630-634, 1986 9. Patton PE, Field CS, Harms RW, et al: CA125 levels in endometriosis. Fertil Steril 45:770-773, 1986 10. Pittaway DE, Fayez JA: The use of CA125 in the diagnosis and management of endometriosis. Fertil Steril 46:790-795, 1986 11. Kjeldsen T, Clausen H, Hirihashi S, et al: Preparation and characterization of monoclonal antibodies directed to the tumorassociated O-linked sialosyl 2-6-a-N-acetylgalactosaminyl (sialosyl Tn) epitope. Cancer Res 48:2214-2220, 1988 12. Hakomori S: Aberrant glycosylation in cancer cell membranes as focused on glycolipids: Overview and perspective. Cancer Res 45:2405-2414, 1985 13. Fukushi Y, Kannagi R, Hakomori S, et al: Location and distribution of difucoganglioside (VI' NeuAc V3111 Fuc2 nLc6) in normal and tumor tissues defined by its monoclonal antibody FH-6. Cancer Res 45:3711-3717, 1985 14. Kobayashi H, Terao T, Kawashima Y: Clinical evaluation of
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
SIALYL TN IN OVARIAN CANCER circulating serum sialyl Tn antigen levels in patients with epithelial ovarian cancer. J Clin Oncol: 983-987, 1991 15. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 16. Gehan EA: A generalized Wilcoxon test for comparing arbitrarily singly-censored samples. Biometrika 52:203-223, 1965 17. Peto R, Pike MC, Armittage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. I. Introduction and design. Br J Cancer 34:585-612, 1976 18. Peto R, Pike MC, Armittage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35:1-39, 1977 19. Hamajima N, Sasaki R, Shibata A, et al: Explanatory assay for researchers who plan to analyze their clinical data using Cox's Proportional Hazard Model. J Jpn Soc Cancer Ther 24:579-584, 1989 20. Springer GF, Desai PR, Robinson MK, et al: The fundamental and diagnostic role of T and Tn antigens in breast carcinoma at earliest histologic stage and throughout, in Dao T, Brodie A, Ip C (eds): Tumor Markers and Their Significance in the Management of Breast Cancer. New York, NY, Liss, 1986, pp 47-70 21. Itzkowitz SH, Bloom EJ, Kokal WA, et al: Sialosyl Tn. A
101 novel mucin antigen associated with prognosis in colorectal cancer patients. Cancer 66:1960-1966, 1990 22. Itzkowitz SH, Yuan M, Montgomery CK, et al: Expression of Tn, sialosyl-Tn, and T antigens in human colon cancer. Cancer Res 49:197-204, 1989 23. Liotta LA, Rao CN, Barsky SH: Tumor invasion and the extracellular matrix. Lab Invest 49:636-649, 1983 24. Springer GF: T and Tn, general carcinoma autoantigens. Science 224:1198-1206, 1984 25. Thor A, Ohuchi W, Szpak CA, et al: Distribution of oncofetal antigen tumor-associated glycoprotein-72 defined by monoclonal antibody B72.3. Cancer Res 46:3118-3124, 1986 26. Kobayashi H: Clinical usefulness of serum CA72-4 analysis in patients with ovarian cancer. Acta Obstet Gynecol Jpn 41:85-87, 1989 27. Katari RS, Fernsten PD, Schlom J: Characterization of the shed form of the human tumor-associated glycoprotein (TAG-72) from serous effusions of patients with different types of carcinomas. Cancer Res 50:4885-4890, 1990 28. Nozawa S, Yajima M, Kojima K, et al: Tumor-associated mucin-type glycoprotein (CA54/61) defined by two monoclonal antibodies (MA54 and MA61) in ovarian cancers. Cancer Res 49:493-498, 1989
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
T
cantly higher in the sera of patients with cancer when compared with levels in benign and healthy controls (P < .05). The 5-year survival rate for patients with STNnegative (serum STN levels < 50.0 U/mL) versus STNpositive (> 50 U/mL) tumors was 76.9% versus 10.8%, respectively (P < .05). The progression-free interval (PFI) at 5 years was 51.9% versus 5.4%, respectively (P < .05). The overall survival probability and PFI were worse in patients with STN-positive sera. Multivariate regression analysis revealed that stage, residual tumor size, positive STN, performance status, and histologic grade were the five important variables for predicting overall survival. Conclusion: We conclude that a positive STN-antigen level in sera is an independent predictor of poor prognosis in ovarian cancer. J Clin Oncol 10:95-101. G 1992 by American Society of Clinical Oncology.
patients with ovarian cancer."5 However, the serum test is not sensitive enough to be useful for screening. Higher
esis) participate in a variety of specific biologic functions. 2 Prognostically important differences in tumor biology may still be due to qualitative changes in tumor mucin." Qualitative rather than quantitative mucin alterations might be important in the biology of cancer. To investigate this possibility, we have examined the correlation between the prognosis of patients with epithelial ovarian cancer (overall survival and progres-
mean CA125 levels are also observed in endometriosis,
sion-free interval [PFI]) and circulating serum levels of a
pelvic inflammatory disease, unexplained infertility, and
mucin carbohydrate antigen (STN) associated with the earliest steps of mucin glycosylation.
HE ELEVATION
of various circulating serum
tumor markers have been reported to correlate with the occurrence of malignancy in cancer.' 2 CA125, a glycoprotein antigen recognized by the monoclonal antibody (moAB) OC125,' 4 has been the most useful antigen for detection of ovarian cancer and follow-up of
during menstruation.'-"
In addition to its high false-
positive rate, the insufficient sensitivity of CA125 in mucinous-type ovarian cancer is another shortcoming. New moABs recognizing a core structure of a mucintype carbohydrate chain have been prepared by Kjeldsen et al. 1 MoABs TKH-1 and TKH-2 directed to the tumor-associated
O-linked
sialyl 2-6-a-N-acctylgalac-
tosaminyl (sialyl Tn; STN) epitope were generated by immunization with ovine submaxillary mucin. The moABs that recognize the carbohydrate structures of cell-
surface glycoconjugates of cancer cells have been regarded as useful tools for characterizing cell types and for cancer diagnosis via assays of antigenic glycoproteins secreted into the bloodstream. The potential for a serum immunoassay of the antigenic determinant STN as a
blood test to screen for ovarian cancer and, subsequently, to observe its clinical course, has not yet been realized. Cell-surface glycolipids or glycoproteins (the composition has been shown to change during tumorigen-
PATIENTS AND METHODS Some patients were recruited from the well-characterized population reported in a previous study on STN." We obtained serum samples from September 1985 through March 1989 at the Department of Obstetrics and Gynecology, Hamamatsu University Hospital, and its related hospitals. Samples were obtained from 157 patients with benign gynecologic disease including 37 uterine myomas, 35 cases adenomyosis, 85 benign ovarian tumors, and 89 cases of histologically proven epithelial ovarian cancer. Table 1 summarizes the clinicopathologic data of the 89 patients with
From the Department of Obstetrics and Gynecology, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka, Japan. Submitted May 20, 1991; acceptedAugust 7, 1991. Address reprintsrequest to Hiroshi Kobayashi, MD, Department of Obstetrics and Gynecology, Hamamatsu University School of Medicine, Handa-cho 3600, Hamamatsu, Shizuoka, 431-31, Japan. C 1992 by American Society of Clinical Oncology. 0732-183X/92 /1001-0016$3.00/0
Journalof Clinical Oncology, Vol 10, No 1 (January), 1992: pp 95-101
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
95
96
KOBAYASHI, TERAO, AND KAWASHIMA Table 1. Distribution of Serum STN-Antigen Levels No. of
Patient Population
Patients
Healthy controls Benign disease Ovarian cancer Stage I Stage II Stage III1 Stage IV
126 157 23 18 38 10
Mean -- SD
(U/mL) 23.3 ± 8.9 29.8 - 9.6 35.5 + 41.7 48.0 ± 62.6 148.0 + 173.1 283.4 ± 305.6
No. of Positive
Cases (%) 5(4.0) 15 (9.6) 6(26.1) 6 (33.3) 24 (63.2) 7 (70.0)
NOTE. Cut off value of serum STN antigen is 41.1 U/mL (mean ± 2 SD).
ovarian cancer included in the current study. Diagnoses of all patients were confirmed by review of operative reports and pathology reports. Staging of ovarian cancer according to the International Federation of Gynecology and Obstetrics (FIGO) classification showed 23 patients with stage I disease, 18 with stage II, 38 with stage III, and 10 with stage IV. All the assays were obtained before initial surgery for those patients with ovarian cancer and benign disease, and stored at -80 0 C until use. All patients were initially treated with optimal debulking surgery followed by combination chemotherapy including cisplatin 50 mg/m2, Adriamycin (doxorubicin; Adria Laboratories, Columbus, OH) 50 mg/m 2, and cyclophosphamide 500 mg/m2. Judgment of disease progression was based on chart review, image diagnosis, and evaluation of objective intraoperative observation. Serum samples from 126 healthy female volunteers, ranging in age from 25 to 71 years, served as normal controls. Circulating serum STN-antigen concentrations (U/mL) were determined by a competitive radioimmunoassay kit (Otsuka Assay Laboratories, Tokushima, Japan) that uses moAB TKH-2 in a one-step procedure. Technical characteristics of the assay are 14 described elsewhere. Briefly, beads coated with STN-antigen purified from meconium were incubated with 50 .LL of serum and 200 CIL of radioactive iodine-labeled TKH-2 antibody solution for 90 minutes. Radioactivity bound to beads was measured by a gammacounter. All samples were assayed in duplicate. Levels more than 41.1 U/mL (normal mean + 2 SD) were considered elevated determinations. Survival curves were calculated from the first day of treatment by the Kaplan-Meier method," and tests of differences in survival and 2 PFI distributions were based on a generalized Wilcoxon test.'" A x 7 test was used to test for differences between groups of subjects.' ,• Cox's proportional hazards model was used to determine the effect of the prognostic factors on survival using the likelihood ratio test."
RESULTS
Sera from healthy controls and patients with benign disease and ovarian cancer at various stages were assayed for STN-antigen levels (Table 1). The cutoff value of STN was arbitrarily set at the mean - 2 SD of healthy controls (41.1 U/mL) to define the positive rate. The mean value and its standard deviation was 23.3 -+ 8.9 U/mL. STN-antigen levels were elevated in five control patients, all over the age of 45 years, with a false-positive rate of 4.0%. The serum level of the antigen was not affected by menstrual cycle (data not
shown). Of the 157 patients with nonmalignant gynecologic disease, 15 (9.6%) had STN-antigen levels of more than 41.1 U/mL: two of 37 patients with uterine myomas had levels of 58 and 62 U/mL, four were cases of adenomyosis, five were cases of serous cystadenoma of the ovary, one case was a dermoid cyst, and the other three cases were mucinous cystadenomas. Elevated levels of serum STN antigen (mean, 113.9 U/mL) were seen in patients with ovarian cancer when compared with those of patients with benign disease and normal controls. The difference in STN-antigen levels between cancer patients and controls or patients with benign disease was significant (P < .05). Of the 89 patients with ovarian cancer, 43 (48.3%) showed STN levels over the cut-off value: six patients with stage I disease (26.1%), six with stage II (33.3%), 24 with stage III (63.2%), and seven with stage IV (70.0%) had elevated antigen levels. Stage I and II cases showed a significantly low frequency of elevation and low values compared with more advanced cases. Significant correlation was seen between elevation of serum STN-antigen levels and the size of the primary tumor (data not shown). These results indicate that serum STN-antigen levels correlate to the stage of disease. The diameter of the largest remaining tumor nodule after laparotomy was greater than 10 cm in eight patients, 2 to 10 cm in 26, and less than 2 cm in the remaining 55 patients. Of the 89 patients, 43 had an elevated serum STN level. The residual tumor mass was greater than 10 cm in seven patients, 2 to 10 cm in 16, and less than 2 cm in 20 (Table 2). The frequency of elevated serum STN levels varied with residual tumor size, and this phenomenon was statistically significant (> 10/< 10 cm and > 2/< 2 cm, P < .05). The relationship between the STN-antigen levels and histology for the patients with cancer is shown in Table 3. The high STN-antigen levels and the high sensitivity in patients with mucinous adenocarcinoma (59.3%) are particularly interesting. Table 4 indicates the relationship between serum STN-antigen status and clinicopathologic features of the tumor. In the present study, patients with serum STN levels more than 50 U/mL (mean - 3 SD) were chosen as STN-positive, because this cutoff value (50 U/mL) could most significantly reflect patient outcome as to Table 2. Serum STN Level and Residual Tumor Size Serum STN Level
Residual
Tumor Size (cm)
Elevated
Normal
>10 2-10 41.1 U/mL(%)
prognosis. There was no statistically significant difference between STN status and age, histologic category, and histologic grade. Of the 89 patients' sera tested, 40 of 52 patients with STN-negative sera (serum STNantigen levels < 50 U/mL) survived for 5 years, whereas those with STN-positive sera (serum STN-antigen levels > 50 U/mL) had a significantly worse outcome (10.8% overall 5-year survival v 76.9%, respectively). The overall survival probability was better in patients with STN-negative sera (P < .05). The median duration of survival of patients with STN-positive sera was 17 months, whereas that of patients with STN-negative sera has not been reached and will exceed 60 months (Fig 1). The estimated 5-year PFI was also better in patients with STN-negative sera compared with those with STNpositive sera (51.9% v 5.4%; P < .05; Fig 2). When 5-year survival was analyzed by both stage and STN status, patients with STN-negative sera had a better prognosis in all stages (Fig 3). The overall 5-year survival of patients as a function of histologic category plus STN status is shown in Fig 4. Even in mucinous carcinoma, patients with STN-positive sera had a significantly poorer prognosis than those with STN-negative sera (P < .05).
Variable
---- ---
U) u L
Months On Study
Fig 1. Overall 5-year survival rate for patients with STN-positive (serum STN levels o50 U/mL, n = 37; --) v STN-negotive (serum STN levels < 50 U/mL, n = 52;- - -) tumors.
The influences of both histologic grade and STN status are demonstrated in Fig 5. For each histologic grade, the overall 5-year survival was worse in patients with STNpositive sera (P < .05). A similar tendency was noted in PFI for the effects of STN status plus stages, histologic category, and histologic grade (Figs 6, 7, and 8, respectively). Subsequently, the overall 5-year PFI was also worse in patients with STN-positive sera as each function of STN status plus stage, histologic category, or histologic grade (P < .05). STN has been shown to be associated with early relapse of this malignancy. Multivariate regression analysis was performed to evaluate further potential prognostic factors (Table 5). Cox regression analysis revealed that stage (stages III and IV), residual tumor size (> 2 cm), STN-positivity (serum STN levels > 50 U/mL), performance status (3
STN-Positive* (n = 37)
STN-Negative (n = 52)
54.9 (35-74)
57.0 (34-70)
4 5
19 13
C 0
22
16
0
6
4
14 16 5 2
24 11 9 8
10 16 11
17 22 13
*Patients with serum STN levels > 50 U/mL were labeled as STNpositive.
5,
o= 5, C
a al
Months On Study Fig 2. 5-year progression-free survival for patients with STNpositive (serum STN levels > 50 U/mL, n = 37; --) v STNnegative (serum STN levels < 50 U/mL, n = 52; - - -) tumors.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
98
KOBAYASHI, TERAO, AND KAWASHIMA .. A
>
0
(nC:
41 U
"
0) 0o
S60 40
-
20 C
C
12
24
36
48
60
D
12
24
36
48
12
24
36
60
48
60
Months On Study
Months On Study
Fig 3. Overall survival of patients by stage of disease and STN (A) Stage I, (B) ), STN-negative (---). status: STN-positive (stage II, (C) stage III, and (D) stage IV.
and 4), and histologic grade (grade 3) were significant negative predictors of survival. STN has been shown to be independently associated with prognosis. DISCUSSION
Oncogenic transformation is often associated with changes in glycosylation in glycoproteins or glycolipids in cancer cells. Carbohydrate antigens with changes in glycosylation have been recently used as tumor markers. Tumor-associated antigens identified by moABs have been used for serologic detection of malignancies. Many tumor markers have the structure of carbohydrate antigens. '"2' STN is considered to be the antigen of the core region of mucin oligosaccharides and is defined by its moAB, TKH-2." In the present study, we monitored the levels of circulating serum STN antigen in patients with ovarian cancer. We further investigated whether
Fig 5. Overall survival of patients as a function of the histologic grade and STN status: STN-positive ( ), STN-negative (- - -). (A) Grade 1, (B) grade 2, and (C) grade 3.
serum STN status plays a role in the biologic behavior of this malignancy. First, we examined the clinical usefulness of the management of STN-antigen levels for the diagnosis of patients with ovarian cancer. The results showed that 48.3% of patients with surgically demonstrated ovarian cancer had elevated STN-antigen levels (>_ mean + 2 SD). The demonstration that elevated concentrations of STN circulate in the blood of patients with ovarian cancer suggests that cell-surface STN glycoconjugate is shed into the systemic circulation. In contrast, STN expression in normal and nonmalignant sera is highly restricted. The specificity of the STN-antigen assay was much greater than that of CA125. The extremely low concentrations of STN antigen in normal or nonmalignant sera can be interpreted to mean that (1) little STN antigen is 10
---------
60 40 20
4)
SA 100o
]oo-
-
12
.100 o0
24
36
48
1-DO-
----
-,
20 12
24
36
4
so
D
12
24
36
48
0.
Fig 4.
201
12
24
-
2
4
36 648 44
60
Months On Study
Months On Study Overall survival of patients by histologic category and STN
), STN-negative (---). status: STN-positive (-mucinous, (C) endometrioid, and (D) clear.
C
- -
40
Q- 40
C
60
(A) Serous, (B)
Fig 6. Progression-free survival of patients by disease stage and STN status: STN-positive (-), STN-negative (-- -). (A) Stage I, (B) stage II, (C) stage III, and (D) stage IV.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
SIALYL TN IN OVARIAN CANCER 100 - -
C
a)
-
- .
12
24
99
20
A
610
4
36
o
43 0
100
0,C
a,
L_____.
0 20
D
12
24
36
48
60
Months On Study
Fig 7. Progression-free survival of patients by histologic category and STN status: STN-positive (), STN-negative (---). (A) Serous, (B) mucinous, (C) endometrioid, and (D) clear.
synthesized or (2) the linkage N-acetylgalactosamine (GalNAc) of normal mucin preferentially receives an N-acetylglucosamine (GIcNAc) residue first, because adding a sialic acid first would prevent further oligosaccharide elongation. 21 '2 2 However, ovarian cancer patients with stage I and II disease showed a significantly low frequency of elevation and low values compared with more advanced cases. The correlation between the positivity of the serum STN test and the disease stage seems to be entirely based on the extent of disease. The frequent expression of the antigen in sera of cancer patients suggests that STN may be a tumor-associated antigen secreted into the circulation. The appearance of core-region antigens provides evidence for incomplete glycosylation in cancer cells. The considerable expression of STN antigen represents another example of precursor accumulation in cancer cells. 22
Q
0o Ln
U) ft. 0
4100
b6o-•-----_
_
40 20C
? ... 12
24
36
4
6b
Months On Study Fig 8. Progression-free survival of patients as a function of histologic grade and STN status: STN-positive (), STN-negative (- - -). (A) Grade 1, (B)grade 2, and (C)grade 3.
Regarding histologic type, it is interesting to note the high positive rate of STN in patients with mucinous carcinoma of the ovary. The qualitative mucin alterations that occur in ovarian cancer, particularly in mucinous-type ovarian cancer, may be important in the biologic behavior of the cancer relating to poor prognosis. This supports the observations by Itzkowitz et al21 concerning the histologic type of colorectal cancer and the strong staining-intensity and high proportion of cells expressing Tn (GalNAc-R) and/or STN. These antigens appear to be good markers of colon cancer mucin that is elaborated by both signet ring-cell carcinoma and colloid carcinoma. As patients with mucinous carcinoma and poorly differentiated adenocarcinoma of the colon often have a poor prognosis, the Tn and STN antigens may be useful markers for these subsets of patients with 2 1 22 colon cancer. ' Our results demonstrate that those ovarian cancer patients with serum STN-antigen levels of 2 50 U/mL had a shorter survival rate than patients with negative STN-antigen levels (< 50 U/mL). In addition, the patients with positive STN-antigen levels had a shorter PFI than those with nonelevated values. The poor survival rate of patients with elevated STN-antigen levels may reflect their higher tumor burden. The apparently better outcome of patients with STNnegative sera cannot be ascribed to confounding factors such as younger age or a well-differentiated type of cancer. By multivariate regression analysis, stage, residual tumor size, and STN status were variables of importance in predicting both PFI and overall survival. STN has been shown to be independently associated with prognosis and a strong predictor of survival. This antigen may be of considerable importance in deciding which postresection patients require further therapy. It is not known whether the levels of circulating serum STN antigen reflect the changes in expression of this antigen on cancer-cell membranes. We are currently investigating the relationship between the expression of STN antigen in cancer tissues by immunohistochemical staining and circulating serum levels of this antigen. We have demonstrated an apparent correlation between the expression of STN by immunohistochemical staining and the concentration of serum STN antigen (unpublished data, October 1991). In addition, the expression of STN antigen in cancer tissues is also correlated with the expression of urokinase-type plasminogen activator (uPA; data not shown). Malignant cells of various types typically synthesize high levels of such proteases as uPA, collagenase, cathepsin B, or cathepsin D. These proteases may facilitate the translocation of tumor cells
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
100
KOBAYASHI, TERAO, AND KAWASHIMA Table 5. Proportional Hazards General Linear Model of Survival Time Variables
Age Stage Histology Grade STN status Residual tumor Performance status
3 -0.07352467 3.13585068 -0.02179025 1.57468468 -2.09452792 -2.36481201 1.90546703
SE
X2
P
0.02974634 0.89436015 0.20427873 0.53127939 1.00834570 0.83468001 0.51252389
3.08 18.05 0.01 11.20 16.12 14.56 21.34
.0228 .0001 .8460 .0001 .0001 .0001 .0001
contributes to survival time. Namely, stage, residual tumor size (< 2 cm), STN status (STN-negativity), performance NOTE. The absolute value of 13 status, and grade of differentiation (grade 1) were positive predictors of survival or favorable prognostic factors, in this order.
through the extracellular matrix.23 Consequently, qualitative alterations of cell-surface glycoproteins or glycolipids and augmented expression of these proteases are considered to be a common feature of the malignant phenotype. Springer 24 has shown that the relative proportions of T (galactose 1,3 GlcNAc-R) and Tn antigens on human carcinomas frequently correlate with the carcinoma's aggressiveness or invasiveness. Recently, we have simultaneously measured serum CA72-4- and CA54/61-antigen levels by a sandwich radioimmunoassay using moABs B72.3 and CC49, and moABs MA54 and MA61, respectively, in patients with ovarian cancer. 25-28 There were many points of similarity in data among these three antigens (STN, CA72-4, and CA54/61). CA72-4 and STN antigens may have an identical antigenic determinant epitope, NeuAc 2-6 Gal NAc 1-Ser/Thr..12 . This finding seems to explain why the results of the present study of STN are similar to those of our previously reported studies of CA72-4 or
CA54/61 (data not shown). It will be necessary to confirm the differences among these antigens by measuring their serum levels in the same samples from a greater number of subjects and by biochemical studies to characterize the oligosaccharide structures of these antigens. We feel that the immunoassay for CA72-4 or CA54/61 in serum also may have reasonable clinical potential, not only as a cancer diagnostic tool or monitoring test for cancer, but also as a prognostic factor. A large series of additional prospective studies should evaluate this issue.
ACKNOWLEDGMENT We are greatly indebted to the personnel at Otsuka Assay Laboratories (Manager, Dr Sadahito Shin, Tokushima, Japan) for the extensive assistance provided to us in the measurement of serum STN antigen levels. We also thank Dr Kazuhiro Sumimoto (Hamamatsu University School of Medicine, Hamamatsu, Japan) and Dr Tetsuya Tachikawa (Otsuka Assay Laboratories, Tokushima, Japan) for the statistical analysis and their valuable advice.
REFERENCES 1. Del Villano BC, Brennan S, Brock P, et al: Radioimmunometric assay for a monoclonal antibody-defined tumor marker, CA19-9. Clin Chem 29:549-552, 1983 2. Herlyn M, Sears HF, Steplewski Z, et al: Monoclonal antibody detection of a circulating tumor-associated antigen. 1. Presence of antigen in sera of patients with colorectal, gastric, and pancreatic carcinoma. J Clin Immunol 2:135-140, 1982 3. Kabawat SE, Bast RC, Bahn AK, et al: Tissue distribution of a coelomic epithelium-related antigen recognized by the monoclonal antibody OC125. Int J Gynecol Pathol 2:275-285, 1983 4. Davis HM, Zurawsky VR, Bast RC Jr, et al: Characterization of the CA125 antigen associated with human epithelial ovarian carcinoma. Cancer Res 46:6143-6151, 1986 5. Bast RC Jr, Feeney M, Lazarus H, et al: Reactivity of a monoclonal antibody with human ovarian carcinoma. J Clin Invest 68:1331-1337, 1981 6. Bast RC, Klug TL, St. John E, et al: A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N Engl J Med 309:883-887, 1983 7. Klug TL, Bast RC Jr, Niloff JM, et al: A monoclonal antibody immunoradiometric assay for an antigenic determinant (CA125)
associated with human epithelial ovarian carcinomas. Cancer Res 44:1048-1053, 1984 8. Barbieri RL, Niloff JM, Bast RC Jr, et al: Elevated serum concentrations of CA125 in patients with advanced endometriosis. Fertil Steril 45:630-634, 1986 9. Patton PE, Field CS, Harms RW, et al: CA125 levels in endometriosis. Fertil Steril 45:770-773, 1986 10. Pittaway DE, Fayez JA: The use of CA125 in the diagnosis and management of endometriosis. Fertil Steril 46:790-795, 1986 11. Kjeldsen T, Clausen H, Hirihashi S, et al: Preparation and characterization of monoclonal antibodies directed to the tumorassociated O-linked sialosyl 2-6-a-N-acetylgalactosaminyl (sialosyl Tn) epitope. Cancer Res 48:2214-2220, 1988 12. Hakomori S: Aberrant glycosylation in cancer cell membranes as focused on glycolipids: Overview and perspective. Cancer Res 45:2405-2414, 1985 13. Fukushi Y, Kannagi R, Hakomori S, et al: Location and distribution of difucoganglioside (VI' NeuAc V3111 Fuc2 nLc6) in normal and tumor tissues defined by its monoclonal antibody FH-6. Cancer Res 45:3711-3717, 1985 14. Kobayashi H, Terao T, Kawashima Y: Clinical evaluation of
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
SIALYL TN IN OVARIAN CANCER circulating serum sialyl Tn antigen levels in patients with epithelial ovarian cancer. J Clin Oncol: 983-987, 1991 15. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 16. Gehan EA: A generalized Wilcoxon test for comparing arbitrarily singly-censored samples. Biometrika 52:203-223, 1965 17. Peto R, Pike MC, Armittage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. I. Introduction and design. Br J Cancer 34:585-612, 1976 18. Peto R, Pike MC, Armittage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35:1-39, 1977 19. Hamajima N, Sasaki R, Shibata A, et al: Explanatory assay for researchers who plan to analyze their clinical data using Cox's Proportional Hazard Model. J Jpn Soc Cancer Ther 24:579-584, 1989 20. Springer GF, Desai PR, Robinson MK, et al: The fundamental and diagnostic role of T and Tn antigens in breast carcinoma at earliest histologic stage and throughout, in Dao T, Brodie A, Ip C (eds): Tumor Markers and Their Significance in the Management of Breast Cancer. New York, NY, Liss, 1986, pp 47-70 21. Itzkowitz SH, Bloom EJ, Kokal WA, et al: Sialosyl Tn. A
101 novel mucin antigen associated with prognosis in colorectal cancer patients. Cancer 66:1960-1966, 1990 22. Itzkowitz SH, Yuan M, Montgomery CK, et al: Expression of Tn, sialosyl-Tn, and T antigens in human colon cancer. Cancer Res 49:197-204, 1989 23. Liotta LA, Rao CN, Barsky SH: Tumor invasion and the extracellular matrix. Lab Invest 49:636-649, 1983 24. Springer GF: T and Tn, general carcinoma autoantigens. Science 224:1198-1206, 1984 25. Thor A, Ohuchi W, Szpak CA, et al: Distribution of oncofetal antigen tumor-associated glycoprotein-72 defined by monoclonal antibody B72.3. Cancer Res 46:3118-3124, 1986 26. Kobayashi H: Clinical usefulness of serum CA72-4 analysis in patients with ovarian cancer. Acta Obstet Gynecol Jpn 41:85-87, 1989 27. Katari RS, Fernsten PD, Schlom J: Characterization of the shed form of the human tumor-associated glycoprotein (TAG-72) from serous effusions of patients with different types of carcinomas. Cancer Res 50:4885-4890, 1990 28. Nozawa S, Yajima M, Kojima K, et al: Tumor-associated mucin-type glycoprotein (CA54/61) defined by two monoclonal antibodies (MA54 and MA61) in ovarian cancers. Cancer Res 49:493-498, 1989
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
