CYNECOLOGIC
ONCOLOGY
47, 53-57 (1992)
Transition from Benign to Malignant Epithelium in Mutinous and Serous Ovarian Cystadenocarcinoma LARRY
Division
E. PULS, M.D.,” DEBORAH E. JAMES E. HUNTER, M.D.,* of Gynecologic
Oncology,
*Department
POWELL, RICHARD
of Obstetrics Medical
M.D.,? PAUL D. DEPRIEST, M.D. ,*.I HOLLY H. GALLION, M.D.,* J. KRYSCIO, PH.D.,$ AND J. R. VAN NAGELL, JR., M.D.’ and Gynecology, Center, Lexington,
Departments of tl’athology Kentucky 40536
and $Statistics.
University
of Kentucky
Received January 22, lYY2
The slides of all patients with ovarian cystadenocarcinoma treated at the University of Kentucky Medical Centerfrom 19661990were reviewed. Fifty-four seroustumors and 42 mutinous neoplasmswere identified for further study. Benign epithelium adjacent to an area of borderline or malignant epithelium was observedin 74 tumors (79%) and a site of epithelial transition was noted in 38 cases(40%). The presenceof associatedbenign epitheliumwasmore commonin borderlineor well-differentiated lesionsand in patients with early-stagedisease.These findings are consistentwith epidemiologicand moleculargeneticdata and suggestthat certain benign serousor mutinous ovarian tumors have the potential for malignanttransformation. Removalof these tumors, particularly in postmenopausal women, shouldresult in a subsequent reduction in the frequencyof ovarian cancer. ‘c,1992 Academic
Press, Inc.
INTRODUCTION Since early-stage ovarian cancer produces no symptoms, the majority of patients continue to present with advanced disease where the prognosis is poor. As a result, there has been little change in the overall survival rate of ovarian cancer patients during the past 15 years [l]. These statistics are particularly disturbing since the cure rate of patients with Stage I ovarian cancer may be as high as 90% [2,3]. Recently, there has been increased interest in the development and testing of screening methods which could detect ovarian cancer at a stage when it is potentially curable. In order for a disease to be amenable to screening, it should (1) be reasonably prevalent in the screened population, (2) have a preclinical phase during which detec-
tion is possible by the screening test, and (3) be amenable
to therapy such that the survival rate of patients with early-stage disease is significantly higher than that for patients with advanced-stage disease [4]. Although ovarian cancer fulfills two of these criteria, the duration of the preclinical phase in this disease is unknown. Epithelial ovarian cancer has been postulated to arise by neoplastic transformation of the germinal (surface) epithelium, and benign serous or mutinous ovarian tumors have been thought to have little if any malignant potential. Recent epidemiologic data, however, suggest that certain benign ovarian tumors may have the potential for malignant change [5]. Likewise, electron microscopic observations have confirmed that there is a transition from benign to malignant epithelium in certain ovarian serous tumors [6]. If these tumors could be detected by screening and removed prior to malignant transformation, then the subsequent occurrence of ovarian cancer could be reduced. The following investigation was performed in order to determine the frequency of histologic transition from benign to malignant epithelium in serous and mutinous cystadenocarcinomas of the ovary. MATERIALS
AND METHODS
All patients treated at the University of Kentucky Medical Center from 1966-1990 with the diagnosis of serous or mutinous ovarian cystadenocarcinoma were identified from the institutional Tumor Registry. Histologic material from each of these patients was reviewed by one of the authors (D.E.P.). Tumors were classified as benign, low malignant potential, or malignant according to the World Health Organization (WHO) criteria. Benign epithelium was characterized by a single or minimally stratified layer of cells. These cells were columnar and often ciliated in serous tumors and contained abundant apical cytoplasmic
Presented at the 23rd Annual Meeting of the Society of Gynecologic Oncologists, San Antonio. TX, March 15-18, 1992. ’ American Cancer Society clinical career development awardee. ’ American Cancer Society professor of clinical oncology. 53
(H)90-8258192 $4.00
Copyright 0 1992 by Academic Press.Inc. All rights of reproduction in any form reserved.
54
PULS ET AL.
mucin in mutinous tumors. Atypical epithelium lining tumors of low malignant potential was characterized by cellular proliferation and pleomorphism without stromal invasion. Malignant epithelium demonstrated marked atypia, increased mitotic activity, and stromal invasion. Malignant tumors were further classified as well differentiated (gl), moderately differentiated (g2), or poorly differentiated (g3). This classification was based on both nuclear and architectural features as described by Russell and Bannatyne [7]. The slides of each tumor were examined to determine the presence of benign epithelium adjacent to the area of malignancy. The specific area of histologic change from benign to borderline or malignant epithelium was defined as the site of epithelial transition. Tumors were staged at exploratory laparotomy according to the International Federation of Gynecology and Obstetrics (FIGO) staging system. Following surgery, patients with ovarian tumors of borderline malignancy were treated generally by surgery alone, whereas patients with Stage I serous or mutinous cystadenocarcinomas usually received a 6-month course of adjuvant chemotherapy with a single alkylating agent. Patients with more advanced-stage ovarian malignancies were treated by 6-12 months of combination platinumbased chemotherapy. Patients with normal clinical examinations and normal (~35 u/ml) serum Ca-125 levels after chemotherapy were advised to have second-look laparotomy. After second-look surgery, patients were seen at 3-month intervals for 2 years and every 6 months thereafter.
TABLE
1
Clinical and Histologic Characteristics of Patients Studied 0’ = 94)
Age (years) Height (inches) Weight (pounds) Gravidity
Mean
Range
54 65 149 2.3
18-86 60-17 99-218 O-10
Cell type Serous cystadenocarcinoma Mutinous cystadenocarcinoma
52 42
Histologic differentiation Borderline Grade 1 Grade 2 Grade 3
24 31 17 22
Stage of disease I II III IV
52 5 32 5
except that they were younger (mean age, 51 years vs 55 years). Benign epithelium was noted adjacent to borderline epithelium in all 24 patients with LMP tumors (Table 2). A site of epithelial transition (Fig. 1) was observed in 19 of these patients (79%). Benign epithelium was present adjacent to malignant epithelium in 22 of 39 serous cystadenocarcinomas (56%) and in 28 of 31 mutinous cystadenocarcinomas (90%). The area of transition from be-
RESULTS TABLE
From a total of 173 cases reviewed, 111 were considered eligible for further study. The remainder of the cases were excluded for the following reasons: no primary tumor identified (28), no tumor in the slides reviewed (18), benign tumor only (7), other carcinoma present (5), and undifferentiated carcinoma unable to be classified (4). The clinical and histopathologic characteristics of the patients studied are illustrated in Table 1. Endometrioid ovarian carcinoma was present in 17 patients and these cases were excluded from further investigation. Fifty-two patients had serous ovarian tumors and 42 patients had mutinous ovarian tumors. Twenty-four patients had borderline tumors or tumors of low malignant potential (LMP) and 70 patients had serous or mutinous cystadenocarcinomas. Sixty-two percent of serous ovarian malignancies were moderately or poorly differentiated as compared to only 17% of mutinous cystadenocarcinomas. The clinical characteristics of patients with LMP tumors were similar to those of patients with invasive lesions
2
Frequency of Adjacent Benign Epithelium According to Cell Type, Histologic Differentiation, and Stage in Ovarian Cystadenocarcinomas Cell type Mutinous (N = 42) Borderline Grade 1 Grade 2 Grade 3 Serous (N = 52) Borderline Grade 1 Grade 2 Grade 3 Stage I II II1 IV
No.
ABE
T
11 24 7 0
11 24 4 0
6 8 0 0
13 I 10 22
13 7 6 9
13 4 4 3
52 5 32 5
51 3 18 2
29 1 8 0
Note. ABE, adjacent benign epithelium; T, epithelial transition.
BENIGN TO MALIGNANT
EPITHELIUM
IN OVARIAN
55
CYSTADENOCARCINOMA
FIG. 1. Papillary serous tumor of low malignant potential showing transition from low cuboidal epithelium to papillary configuration exhil 3iting epithelial proliferation. H&E stain; final magnification, 40 x
nign to markedly atypical epithelium was observed in 19 of these tumors (Fig. 2). There was an inverse relationship between the presence of associated benign epithelium and histologic tumor differentiation (Table 2). Benign epithelium was present in 100% of well-differentiated cystadenocarcinomas, 59% of moderately differentiated tumors, and 41% of poorly differentiated tumors. Similarly, the presence of benign epithelium was related directly to stage of disease. Over 95% of patients with Stage I tumors had documented benign epithelium adjacent to neoplastic epithelium as opposed to only 54% in Stage III or IV lesions. Following surgery, 26 patients were treated with 6 months of melphalan chemotherapy (1 mg/kg/5-day course) and 31 patients were treated with 6-12 monthly courses of cisplatin (50 mg/m*) and cyclophosphamide (1000 mg/m*). Thirty-four patients underwent secondlook laparotomy and 8 patients had persistent ovarian cancer. The 3-year survival varied from 94% in patients with Stage I disease to 28% in patients with Stage III or IV disease. The presence of epithelial transition in a tu-
mor was not significant prognosis.
as an independent
indicator
of
DISCUSSION The findings of this investigation indicate that benign epithelium can be documented in a high percentage of serous and mutinous ovarian cystadenocarcinomas. This observation was made in over three-fourths of all the tumors reviewed. Although the exact side of neoplastic epithelial transition was not demonstrated in all cases, it is possible that more complete sectioning of each ovarian tumor would have revealed this area. In a prior related investigation, Plaxe and co-workers [9] evaluated the ovarian epithelium in three low-power fields adjacent to 50 Stage I invasive cancers and 50 normal control ovaries. Nuclear and cellular atypia were present in the epithelium adjacent to invasive cancer in 49 of 50 cases and in none of the control ovaries. Our observations also confirm previous findings demonstrated by scanning electron microscopy. Stenbeck [6], using this technology, reported
56
PULS ET AL.
FIG. 2. Grade 2 serous cystadenocarcinoma. There is an area of transition from a histologically bland single-layered epithelium (top inset) through epithelial atypia to markedly atypical invasive glands (bottom inset). H&E stain; final magnification, 25 x
that there was a continuum in the epithelial lining of certain serous ovarian tumors from benign to borderline to malignant. The age of patients with borderline or LMP tumors in this and other investigations was significantly lower than that of patients with invasive [lO,ll] lesions, indicating that the degree of epithelial abnormality may progress with time in certain cases. These data are also consistent with the major theories of pathogenesis of ovarian cystadenocarcinoma. According to Cramer and Welch [12], cystic ovarian malignancies arise by a series of predictable events. With continuous ovulation, there is entrapment of the germinal epithelium within the ovarian stroma. An inclusion cyst is then formed in the stroma which is lined initially by benign germinal epithelium. Under the influence of hormonal or other oncogenic stimuli, this benign epithelium undergoes transformation to form a malignant ovarian tumor. Recent advances in the molecular genetics and molecular biology of carcinogenesis suggest a complex and sequential process involving oncogenes, suppressor genes, growth factors, and cytokines in tumor development [ 13,141. Model systems based on molecular genetic and cytogenetic data are being developed for human solid tumors, most prominently reti-
noblastoma and the adenoma/colon carcinoma sequence. Proposed models postulate a sequence of genetic alterations, beginning in benign cells which eventuate in an invasive cancer [15]. The strong possibility exists that similar sequences of genetic and cellular events are important in the biology of ovarian cancer. Observations from the present investigation are supported by recent epidemiologic data which suggest that certain benign ovarian tumors may undergo malignant change. Bourne and colleagues [5] reported a significant increase in the frequency of benign epithelial ovarian tumors in primary or secondary relatives of patients with documented ovarian carcinoma. The prevalence of ovarian carcinoma was also increased in this group, leading the authors to postulate that malignant transformation of benign epithelial tumors was one mechanism to explain the origin of genetically related ovarian cancer. Amplification of the HER-2/NEU proto-oncogene has been reported in epithelial ovarian tumors, and the C-erb B2 protein has been demonstrated by immunohistochemical methods in ovarian epithelium [16]. Additionally, tumorspecific losses of chromosomes llp, 13q, and 17 have been observed in ovarian tumors, indicating the presence of
BENIGN
TO MALIGNANT
EPITHELIUM
closely linked tumor suppressor genes [17,18]. Molecular genetic studies are in progress which should help identify those benign or borderline tumors at highest risk to undergo malignant transformation [ 191. With effective ovarian cancer screening, many of these tumors could be detected and removed. Their removal should cause a decrease in the observed number of ovarian cancers in a screened population and a subsequent reduction in ovarian cancer mortality. REFERENCES I.
2.
3.
4. 5.
Boring, C., Squires, T., and Tong, T. Cancer Statistics 1992. Cu Bull. Cancer Prog. 42, 19-38 (1992). Gallion, H., van Nagell, J. R., Donaldson, E. S., Higgins, R. V.. Powell, D. E., and Kryscio, R. J. Adjuvant oral alkylating agent chemotherapy in patients with Stage I epithelial ovarian cancer, Cancer 63, 1070-1073 (1989). Young, R. C., Walton, L. A., Ellenberg, S. S., et al. Adjuvant therapy in Stage I and Stage II epithelial ovarian cancer, N. Engl. J. Med. 322, 1021-1027 (1990). Hulka, B. S. Cancer screening: Degrees of proof and practical application, Cancer 62, 1776-1789 (1989). Bourne, T. H., Whitehead, M. I., Campbell, S., Royston, P., Bhan, V.. and Collins, W. P. Ultrasound screening for familial ovarian cancer, Gynecol. Oncol. 43, 92-97 (1991).
6. Stenback, F. Benign, borderline and malignant serous cystadenomas of the ovary, P&ho/. Rex Pracf. 172, 58-72 (1981). I. Russell. P., and Bannatyne, P. Surgical pathology of the ovaries, Churchill Livingstone, New York, Ch. 23, pp. 190-216. 8. Russell, P., and Bannatyne, P. Surgical pathology of the ovaries, Churchill Livingstone, New York, Ch. 24, pp. 217-235.
IN OVARIAN
CYSTADENOCARCINOMA
57
9. Plaxe, S. C., Deligdisch, L., Dottino, P. R., and Cohen, C. J. Ovarian intraepithelial neoplasia demonstrated in patients with Stage I ovarian carcinoma, Gynecol. Oncol. 38, 367-372 (1990). 10. Snider, D. D., Stuart, G. C., Nation, J. G., and Robertson, D. I. Evaluation of surgical staging in Stage I low malignant potential ovarian tumor, Gynecol. Oncol. 40, 129-132 (1991). 11. Creasman, W. F., Park, R., Norris, H., DiSaia, P. J., Morrow, C. P., and Hreshchyshyn, M. Stage I borderline ovarian tumors, Obstet. Gynecol. 59, 93-96 (1982). 12. Cramer, D. W., and Welch, W. R. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis, JNCI 1971, 717-721 (1983). 13. Weinberg, R. A. Tumor suppressor genes, Science 254, 1138-1146 (1991). 14. Aaronson, S. A. Growth factors and cancer, Science 254, 11461153 (1991). 15. Nowell, P. J. The clonal evolution of tumor cell populations, Science 194, 23-28
(1976).
16 Tyson, F. L., Boyer, C. M., Kaufman, R., O’Briant, K., Cram, G., Crews, J. R., Soper, J. T., Daly, L., Fowler, W. C., Haskill, J. S., and Bast, R. C. Expression and amplification of the HER2/NEU (C-erb B-2) proto-oncogene in epithelial ovarian tumors and cell lines, Am. J. Obstet. Gynecol. 165, 640-646 (1991). 17. Lee, J. H., Kavanagh, J. J., Wildrick, D. M., Wharton, J. T., and Blick, M. Frequent loss of heterozygosity on chromosomes 6q, 11, and 17 in human ovarian carcinomas, Cancer Res. 50, 2724-2728 (1990). 18. Li, S., Schwartz, P. E., Lee, W. H., and Yang-Feng, T. L. Allele loss at the retinoblastoma locus in human ovarian cancer, J. Natl. Cancer Inst. 83, 637-640 (1991). 19. /Gallion, H. H., Powell, D. E., Morrow, J. K., Pieretti, M., Case, E., Turker, M. S., DePriest, M. D., and van Nagell, J. R. Molecular genetic changes in human epithelial ovarian malignancies, Gy.. m press. necof. Uncol.,
ONCOLOGY
47, 53-57 (1992)
Transition from Benign to Malignant Epithelium in Mutinous and Serous Ovarian Cystadenocarcinoma LARRY
Division
E. PULS, M.D.,” DEBORAH E. JAMES E. HUNTER, M.D.,* of Gynecologic
Oncology,
*Department
POWELL, RICHARD
of Obstetrics Medical
M.D.,? PAUL D. DEPRIEST, M.D. ,*.I HOLLY H. GALLION, M.D.,* J. KRYSCIO, PH.D.,$ AND J. R. VAN NAGELL, JR., M.D.’ and Gynecology, Center, Lexington,
Departments of tl’athology Kentucky 40536
and $Statistics.
University
of Kentucky
Received January 22, lYY2
The slides of all patients with ovarian cystadenocarcinoma treated at the University of Kentucky Medical Centerfrom 19661990were reviewed. Fifty-four seroustumors and 42 mutinous neoplasmswere identified for further study. Benign epithelium adjacent to an area of borderline or malignant epithelium was observedin 74 tumors (79%) and a site of epithelial transition was noted in 38 cases(40%). The presenceof associatedbenign epitheliumwasmore commonin borderlineor well-differentiated lesionsand in patients with early-stagedisease.These findings are consistentwith epidemiologicand moleculargeneticdata and suggestthat certain benign serousor mutinous ovarian tumors have the potential for malignanttransformation. Removalof these tumors, particularly in postmenopausal women, shouldresult in a subsequent reduction in the frequencyof ovarian cancer. ‘c,1992 Academic
Press, Inc.
INTRODUCTION Since early-stage ovarian cancer produces no symptoms, the majority of patients continue to present with advanced disease where the prognosis is poor. As a result, there has been little change in the overall survival rate of ovarian cancer patients during the past 15 years [l]. These statistics are particularly disturbing since the cure rate of patients with Stage I ovarian cancer may be as high as 90% [2,3]. Recently, there has been increased interest in the development and testing of screening methods which could detect ovarian cancer at a stage when it is potentially curable. In order for a disease to be amenable to screening, it should (1) be reasonably prevalent in the screened population, (2) have a preclinical phase during which detec-
tion is possible by the screening test, and (3) be amenable
to therapy such that the survival rate of patients with early-stage disease is significantly higher than that for patients with advanced-stage disease [4]. Although ovarian cancer fulfills two of these criteria, the duration of the preclinical phase in this disease is unknown. Epithelial ovarian cancer has been postulated to arise by neoplastic transformation of the germinal (surface) epithelium, and benign serous or mutinous ovarian tumors have been thought to have little if any malignant potential. Recent epidemiologic data, however, suggest that certain benign ovarian tumors may have the potential for malignant change [5]. Likewise, electron microscopic observations have confirmed that there is a transition from benign to malignant epithelium in certain ovarian serous tumors [6]. If these tumors could be detected by screening and removed prior to malignant transformation, then the subsequent occurrence of ovarian cancer could be reduced. The following investigation was performed in order to determine the frequency of histologic transition from benign to malignant epithelium in serous and mutinous cystadenocarcinomas of the ovary. MATERIALS
AND METHODS
All patients treated at the University of Kentucky Medical Center from 1966-1990 with the diagnosis of serous or mutinous ovarian cystadenocarcinoma were identified from the institutional Tumor Registry. Histologic material from each of these patients was reviewed by one of the authors (D.E.P.). Tumors were classified as benign, low malignant potential, or malignant according to the World Health Organization (WHO) criteria. Benign epithelium was characterized by a single or minimally stratified layer of cells. These cells were columnar and often ciliated in serous tumors and contained abundant apical cytoplasmic
Presented at the 23rd Annual Meeting of the Society of Gynecologic Oncologists, San Antonio. TX, March 15-18, 1992. ’ American Cancer Society clinical career development awardee. ’ American Cancer Society professor of clinical oncology. 53
(H)90-8258192 $4.00
Copyright 0 1992 by Academic Press.Inc. All rights of reproduction in any form reserved.
54
PULS ET AL.
mucin in mutinous tumors. Atypical epithelium lining tumors of low malignant potential was characterized by cellular proliferation and pleomorphism without stromal invasion. Malignant epithelium demonstrated marked atypia, increased mitotic activity, and stromal invasion. Malignant tumors were further classified as well differentiated (gl), moderately differentiated (g2), or poorly differentiated (g3). This classification was based on both nuclear and architectural features as described by Russell and Bannatyne [7]. The slides of each tumor were examined to determine the presence of benign epithelium adjacent to the area of malignancy. The specific area of histologic change from benign to borderline or malignant epithelium was defined as the site of epithelial transition. Tumors were staged at exploratory laparotomy according to the International Federation of Gynecology and Obstetrics (FIGO) staging system. Following surgery, patients with ovarian tumors of borderline malignancy were treated generally by surgery alone, whereas patients with Stage I serous or mutinous cystadenocarcinomas usually received a 6-month course of adjuvant chemotherapy with a single alkylating agent. Patients with more advanced-stage ovarian malignancies were treated by 6-12 months of combination platinumbased chemotherapy. Patients with normal clinical examinations and normal (~35 u/ml) serum Ca-125 levels after chemotherapy were advised to have second-look laparotomy. After second-look surgery, patients were seen at 3-month intervals for 2 years and every 6 months thereafter.
TABLE
1
Clinical and Histologic Characteristics of Patients Studied 0’ = 94)
Age (years) Height (inches) Weight (pounds) Gravidity
Mean
Range
54 65 149 2.3
18-86 60-17 99-218 O-10
Cell type Serous cystadenocarcinoma Mutinous cystadenocarcinoma
52 42
Histologic differentiation Borderline Grade 1 Grade 2 Grade 3
24 31 17 22
Stage of disease I II III IV
52 5 32 5
except that they were younger (mean age, 51 years vs 55 years). Benign epithelium was noted adjacent to borderline epithelium in all 24 patients with LMP tumors (Table 2). A site of epithelial transition (Fig. 1) was observed in 19 of these patients (79%). Benign epithelium was present adjacent to malignant epithelium in 22 of 39 serous cystadenocarcinomas (56%) and in 28 of 31 mutinous cystadenocarcinomas (90%). The area of transition from be-
RESULTS TABLE
From a total of 173 cases reviewed, 111 were considered eligible for further study. The remainder of the cases were excluded for the following reasons: no primary tumor identified (28), no tumor in the slides reviewed (18), benign tumor only (7), other carcinoma present (5), and undifferentiated carcinoma unable to be classified (4). The clinical and histopathologic characteristics of the patients studied are illustrated in Table 1. Endometrioid ovarian carcinoma was present in 17 patients and these cases were excluded from further investigation. Fifty-two patients had serous ovarian tumors and 42 patients had mutinous ovarian tumors. Twenty-four patients had borderline tumors or tumors of low malignant potential (LMP) and 70 patients had serous or mutinous cystadenocarcinomas. Sixty-two percent of serous ovarian malignancies were moderately or poorly differentiated as compared to only 17% of mutinous cystadenocarcinomas. The clinical characteristics of patients with LMP tumors were similar to those of patients with invasive lesions
2
Frequency of Adjacent Benign Epithelium According to Cell Type, Histologic Differentiation, and Stage in Ovarian Cystadenocarcinomas Cell type Mutinous (N = 42) Borderline Grade 1 Grade 2 Grade 3 Serous (N = 52) Borderline Grade 1 Grade 2 Grade 3 Stage I II II1 IV
No.
ABE
T
11 24 7 0
11 24 4 0
6 8 0 0
13 I 10 22
13 7 6 9
13 4 4 3
52 5 32 5
51 3 18 2
29 1 8 0
Note. ABE, adjacent benign epithelium; T, epithelial transition.
BENIGN TO MALIGNANT
EPITHELIUM
IN OVARIAN
55
CYSTADENOCARCINOMA
FIG. 1. Papillary serous tumor of low malignant potential showing transition from low cuboidal epithelium to papillary configuration exhil 3iting epithelial proliferation. H&E stain; final magnification, 40 x
nign to markedly atypical epithelium was observed in 19 of these tumors (Fig. 2). There was an inverse relationship between the presence of associated benign epithelium and histologic tumor differentiation (Table 2). Benign epithelium was present in 100% of well-differentiated cystadenocarcinomas, 59% of moderately differentiated tumors, and 41% of poorly differentiated tumors. Similarly, the presence of benign epithelium was related directly to stage of disease. Over 95% of patients with Stage I tumors had documented benign epithelium adjacent to neoplastic epithelium as opposed to only 54% in Stage III or IV lesions. Following surgery, 26 patients were treated with 6 months of melphalan chemotherapy (1 mg/kg/5-day course) and 31 patients were treated with 6-12 monthly courses of cisplatin (50 mg/m*) and cyclophosphamide (1000 mg/m*). Thirty-four patients underwent secondlook laparotomy and 8 patients had persistent ovarian cancer. The 3-year survival varied from 94% in patients with Stage I disease to 28% in patients with Stage III or IV disease. The presence of epithelial transition in a tu-
mor was not significant prognosis.
as an independent
indicator
of
DISCUSSION The findings of this investigation indicate that benign epithelium can be documented in a high percentage of serous and mutinous ovarian cystadenocarcinomas. This observation was made in over three-fourths of all the tumors reviewed. Although the exact side of neoplastic epithelial transition was not demonstrated in all cases, it is possible that more complete sectioning of each ovarian tumor would have revealed this area. In a prior related investigation, Plaxe and co-workers [9] evaluated the ovarian epithelium in three low-power fields adjacent to 50 Stage I invasive cancers and 50 normal control ovaries. Nuclear and cellular atypia were present in the epithelium adjacent to invasive cancer in 49 of 50 cases and in none of the control ovaries. Our observations also confirm previous findings demonstrated by scanning electron microscopy. Stenbeck [6], using this technology, reported
56
PULS ET AL.
FIG. 2. Grade 2 serous cystadenocarcinoma. There is an area of transition from a histologically bland single-layered epithelium (top inset) through epithelial atypia to markedly atypical invasive glands (bottom inset). H&E stain; final magnification, 25 x
that there was a continuum in the epithelial lining of certain serous ovarian tumors from benign to borderline to malignant. The age of patients with borderline or LMP tumors in this and other investigations was significantly lower than that of patients with invasive [lO,ll] lesions, indicating that the degree of epithelial abnormality may progress with time in certain cases. These data are also consistent with the major theories of pathogenesis of ovarian cystadenocarcinoma. According to Cramer and Welch [12], cystic ovarian malignancies arise by a series of predictable events. With continuous ovulation, there is entrapment of the germinal epithelium within the ovarian stroma. An inclusion cyst is then formed in the stroma which is lined initially by benign germinal epithelium. Under the influence of hormonal or other oncogenic stimuli, this benign epithelium undergoes transformation to form a malignant ovarian tumor. Recent advances in the molecular genetics and molecular biology of carcinogenesis suggest a complex and sequential process involving oncogenes, suppressor genes, growth factors, and cytokines in tumor development [ 13,141. Model systems based on molecular genetic and cytogenetic data are being developed for human solid tumors, most prominently reti-
noblastoma and the adenoma/colon carcinoma sequence. Proposed models postulate a sequence of genetic alterations, beginning in benign cells which eventuate in an invasive cancer [15]. The strong possibility exists that similar sequences of genetic and cellular events are important in the biology of ovarian cancer. Observations from the present investigation are supported by recent epidemiologic data which suggest that certain benign ovarian tumors may undergo malignant change. Bourne and colleagues [5] reported a significant increase in the frequency of benign epithelial ovarian tumors in primary or secondary relatives of patients with documented ovarian carcinoma. The prevalence of ovarian carcinoma was also increased in this group, leading the authors to postulate that malignant transformation of benign epithelial tumors was one mechanism to explain the origin of genetically related ovarian cancer. Amplification of the HER-2/NEU proto-oncogene has been reported in epithelial ovarian tumors, and the C-erb B2 protein has been demonstrated by immunohistochemical methods in ovarian epithelium [16]. Additionally, tumorspecific losses of chromosomes llp, 13q, and 17 have been observed in ovarian tumors, indicating the presence of
BENIGN
TO MALIGNANT
EPITHELIUM
closely linked tumor suppressor genes [17,18]. Molecular genetic studies are in progress which should help identify those benign or borderline tumors at highest risk to undergo malignant transformation [ 191. With effective ovarian cancer screening, many of these tumors could be detected and removed. Their removal should cause a decrease in the observed number of ovarian cancers in a screened population and a subsequent reduction in ovarian cancer mortality. REFERENCES I.
2.
3.
4. 5.
Boring, C., Squires, T., and Tong, T. Cancer Statistics 1992. Cu Bull. Cancer Prog. 42, 19-38 (1992). Gallion, H., van Nagell, J. R., Donaldson, E. S., Higgins, R. V.. Powell, D. E., and Kryscio, R. J. Adjuvant oral alkylating agent chemotherapy in patients with Stage I epithelial ovarian cancer, Cancer 63, 1070-1073 (1989). Young, R. C., Walton, L. A., Ellenberg, S. S., et al. Adjuvant therapy in Stage I and Stage II epithelial ovarian cancer, N. Engl. J. Med. 322, 1021-1027 (1990). Hulka, B. S. Cancer screening: Degrees of proof and practical application, Cancer 62, 1776-1789 (1989). Bourne, T. H., Whitehead, M. I., Campbell, S., Royston, P., Bhan, V.. and Collins, W. P. Ultrasound screening for familial ovarian cancer, Gynecol. Oncol. 43, 92-97 (1991).
6. Stenback, F. Benign, borderline and malignant serous cystadenomas of the ovary, P&ho/. Rex Pracf. 172, 58-72 (1981). I. Russell. P., and Bannatyne, P. Surgical pathology of the ovaries, Churchill Livingstone, New York, Ch. 23, pp. 190-216. 8. Russell, P., and Bannatyne, P. Surgical pathology of the ovaries, Churchill Livingstone, New York, Ch. 24, pp. 217-235.
IN OVARIAN
CYSTADENOCARCINOMA
57
9. Plaxe, S. C., Deligdisch, L., Dottino, P. R., and Cohen, C. J. Ovarian intraepithelial neoplasia demonstrated in patients with Stage I ovarian carcinoma, Gynecol. Oncol. 38, 367-372 (1990). 10. Snider, D. D., Stuart, G. C., Nation, J. G., and Robertson, D. I. Evaluation of surgical staging in Stage I low malignant potential ovarian tumor, Gynecol. Oncol. 40, 129-132 (1991). 11. Creasman, W. F., Park, R., Norris, H., DiSaia, P. J., Morrow, C. P., and Hreshchyshyn, M. Stage I borderline ovarian tumors, Obstet. Gynecol. 59, 93-96 (1982). 12. Cramer, D. W., and Welch, W. R. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis, JNCI 1971, 717-721 (1983). 13. Weinberg, R. A. Tumor suppressor genes, Science 254, 1138-1146 (1991). 14. Aaronson, S. A. Growth factors and cancer, Science 254, 11461153 (1991). 15. Nowell, P. J. The clonal evolution of tumor cell populations, Science 194, 23-28
(1976).
16 Tyson, F. L., Boyer, C. M., Kaufman, R., O’Briant, K., Cram, G., Crews, J. R., Soper, J. T., Daly, L., Fowler, W. C., Haskill, J. S., and Bast, R. C. Expression and amplification of the HER2/NEU (C-erb B-2) proto-oncogene in epithelial ovarian tumors and cell lines, Am. J. Obstet. Gynecol. 165, 640-646 (1991). 17. Lee, J. H., Kavanagh, J. J., Wildrick, D. M., Wharton, J. T., and Blick, M. Frequent loss of heterozygosity on chromosomes 6q, 11, and 17 in human ovarian carcinomas, Cancer Res. 50, 2724-2728 (1990). 18. Li, S., Schwartz, P. E., Lee, W. H., and Yang-Feng, T. L. Allele loss at the retinoblastoma locus in human ovarian cancer, J. Natl. Cancer Inst. 83, 637-640 (1991). 19. /Gallion, H. H., Powell, D. E., Morrow, J. K., Pieretti, M., Case, E., Turker, M. S., DePriest, M. D., and van Nagell, J. R. Molecular genetic changes in human epithelial ovarian malignancies, Gy.. m press. necof. Uncol.,
