t .Nl
0360-3016/92 $5.00 Copyright 0 1992 Pergamon Pres Ltd.
Im. J Radialron Oncology Biol Phys.. Vol. 22, pp. 867-874 Printed in the U.S.A All rights reserved.
??Special Feature
BORDERLINE EPITHELIAL OVARIAN TUMORS: A REVIEW OF 81 CASES WITH AN ASSESSMENT OF THE IMPACT OF TREATMENT L. A. MANCHUL, B.Sc., M.D., F.R.C.P.C.,’ J. SIMM, M.SC.,~ W. LEVIN, M.B., F.R.C.P.C.,’ A. W. FYLES, M.D.,
F.R.C.P.C.,’
A. J. DEMBO, M.B.,
F.R.C.P.C.,4
J. F. PRINGLE, M.B., F.R.C.P.C.,’ G. A. RAWLINGS, M.D., F.R.C.P.C.,’ J. F. G. STURGEON, M.B., F.R.C.P.C.3 AND G. M. THOMAS, M.D., F.R.C.P.C.4 ‘Departmentsof Radiation Oncology, ‘Biostatistics,and ‘Medicine, PrincessMargaretHospital, and 4Departmentof Radiation Oncology, Toronto-BayviewRegional CancerCentre,Universityof Toronto, Toronto, Ontario, Canada Optimal management of borderline epithelial ovarian tumors remains controversial because of the lack of clear, universally accepted pathologic criteria for diagnosis, the lack of complete understanding of the significance of intraperitoneal implants, and the desire to employ more limited surgery in young women. We reviewed the experience with borderline epithelial ovarian tumors at Princess Margaret Hospital in order to assess the natural history of the disease, to determine prognostic factors that would aid in management decisions, and to determine if adjuvant therapy influenced outcome. Eighty-one patients were analyzed. The mean age was 48 years. Seventy-two percent of tumors were of the serous histologic sub-type and 28% were mutinous. Seventy-eight percent were Stage I, 11% Stage II, and 11% Stage III. Peritoneal washings contained malignant cells in 14 of 32 patients (not recorded or obtained in 49), cyst rupture occurred in 25%, surface excrescences in 402, and adhesions in 46%. None of these factors had a significant effect on recurrence rate or survival. Eleven patients received adjuvant radiation therapy (10 abdomino-pelvic and 1 pelvic alone), four adjuvant chemotherapy, and one both radiation therapy and chemotherapy. The rest (65) received no adjuvant therapy. Due to the small numbers and infrequent events, it was not possible to analyze and thus draw valid conclusions regarding the effect of adjuvant therapy on survival or recurrence. The overall survival (OS) and cause specific survival (CSS) were 85% and 96% at 10 years, respectively. No Stage I patient died of tumor. OS for Stage I patients was 90% at 10 years, the majority of whom (61 of 63) received no adjuvant therapy, and is thus unnecessary in Stage I disease. The adequacy of unilateral oophorectomy or ovarian cystectomy could not be confirmed because of small numbers. The 10 year OS and disease-free survival in Stage II and III were 75% and 50%, respectively, despite the use of adjuvant radiation therapy, chemotherapy, or both. It is necessary to create a multi-center tumor registry in order to acquire a prospective data base from which to develop sound therapeutic decisions. Borderline epithelial ovarian neoplasms, Ovarian tumors of low malignant potential, Therapeutic management.
malignant” having some but not all of the morphologic features of malignancy: stratification of epithelial cells, epithelial budding and detachment, with some degree of nuclear atypia and increased mitotic activity but without stromal invasion (28). Borderline epithelial ovarian tumors account for approximately 10 to 20% of all ovarian tumors ( 1,20). They affect a younger group of women than do malignant epithelial ovarian tumors, and frequently women who would wish to maintain their child-bearing potential. They generally present at an earlier stage than do malignant tumors: anywhere from 50 to 85% present as Stage I ( 1, 2, 5, 18, 23,25). Patients who present with Stage I disease have an
INTRODUCTION
In 1929 Taylor described a group of papillary ovarian tumors that appeared clinically malignant with overt peritoneal metastases, demonstrating some of the histologic criteria of malignancy, but that behaved in a “benign” fashion (3 1). It was not until 197 1 that FIG0 officially recognized epithelial ovarian tumors of “low potential malignancy” with epithelial proliferation and nuclear abnormalities but without infiltrative destructive growth (15). In 1973 the WHO Histologic Classification of Tumours included a form of common epithelial tumor “intermediate between one clearly benign and obviously
Reprint requests to: Dr. L. A. Manchul, Department of Radiation Oncology, Princess Margaret Hospital, 500 Sherboume St., Toronto, Ontario, M4X lK9 Canada. Acknowledgements-We appreciatethe expertiseof DebbieTsuji
and June Saunders for data collection and management and Colette Plasway for secretarial assistance. Accepted for publication 6 September 199 I. 867
868
1. J. Radiation Oncology 0 Biology 0 Physics
excellent prognosis; however, late recurrences and distant metastases have been reported (1, 2, 5, 10, 18). Patients who present with disseminated intra-abdominal disease fare worse than those without, but their survival exceeds that of patients with equivalent stage invasive epithelial ovarian cancer (1, 3, 6, 11, 14, 16, 18, 19, 23, 29, 34). Optimal management of patients with borderline epithelial ovarian tumors is controversial and is difficult to define clearly for a number of reasons. First, the pathologic criteria for diagnosis are not universally accepted, and “microinvasive” tumors have been recently described (4? 30), leaving in doubt the uniformity of the patient population from which current recommendations are derived. Second, there is controversy regarding the nature and significance of the peritoneal implants found in patients with presumed advanced disease with evidence emerging to suggest that invasive implants carry a worse prognosis than non-invasive implants (4, 22). There is some doubt that non-invasive implants represent true metastases but rather autochthonous extra-ovarian neoplasia and should be managed differently from the invasive implants (12, 26, 27). Third, there is an increasing tendency to recommend more limited surgery in patients presenting with unilateral Stage I disease based on the results of small series (21, 32). Although more limited surgery respects the patient’s desire to maintain fertility, it is not yet known in which women this approach is most appropriate. The value of adjuvant therapy is unclear and cannot be accurately extrapolated from the literature on the subject, which consists largely of retrospective reviews of small numbers of patients subjected to a variety of therapeutic modalities over a long period of time. Pathologic diagnostic criteria are not always consistent from series to series, staging investigations are not uniform, and different end points are often assessed. We reviewed the experience of borderline epithelial ovarian tumors at the Princess Margaret Hospital to gain an insight into the natural history of the disease and to determine prognostic factors that would aid in providing a rational approach to therapeutic management. METHODS
AND MATERIALS
The cohort of patients who formed the basis for this analysis consisted of those patients who were referred to the Princess Margaret Hospital between 1967 and 1988 with a diagnosis of borderline ovarian cancer or with a diagnosis of ovarian cancer where subsequent review at this Institution confirmed borderline histology. The analyses were confined to those patients referred with a newly diagnosed borderline ovarian tumor. Patients who were initially referred with a relapse or who presented with a concurrent invasive cancer were excluded from the analysis. Information on age, clinical features and investigations at presentation, surgical findings, adjuvant postoperative
Volume 22, Number 5, 1992
treatment, and pathologic criteria for the diagnosis was obtained from the patients’ charts or from the referring institution or physician as necessary. Information regarding such possible prognostic factors as the presence of ascites, surface excrescences, tumor rupture, adherence, tumor size, the amount of residual disease, and the presence of malignant cells in peritoneal washings was obtained where available. All patients were retrospectively staged according to UICC 1987 criteria. All pathology slides were reviewed by one of three expert gynecologic pathologists at the Princess Margaret Hospital but were not re-reviewed specifically in conjunction with this study. Those tumors which demonstrated areas of microinvasion without destructive stromal infiltration were identified. Peritoneal implants were assessed as representing either atypical endosalpingiosis, borderline tumor implants, or frankly invasive metastases with a borderline primary. Initial therapy consisted of either a total abdominal hysterectomy with bilateral salpingoophorectomy (BSOH), a unilateral oophorectomy or salpingoophorectomy, or rarely, a cystectomy alone. Adjuvant radiation therapy consisted usually of whole abdomino-pelvic radiation (AP-RT) with a pelvic boost or, rarely, pelvic radiation alone. Pelvic doses ranged from 40 Gy to 47 Gy (mid-plane) in 20-35 fractions over 28-98 days. Abdominal doses ranged from 22.5 Gy to 26.4 Gy in 22-25 fractions over 31 to 84 days. Rarely, the abdomino-pelvic strip technique (previously described) was employed, utilizing 22.5 Gy in 2.25 Gy fractions to the whole abdomen and a boost to the pelvis of 22.5 Gy in 2.25 Gy fractions (9). Chemotherapy consisted of several regimens including melphalan alone or cyclophosphamide and cis-platinum (or carboplatinum) with or without adriamycin. Patient status was recorded at the time of analysis as alive or dead, noting the presence or absence of ovarian tumor. The site of relapse and the cause of death were recorded. Overall, cause-specific and disease-free survival were determined using Kaplan-Meier estimates (17). Patients who died from another cause without the presence of ovarian tumor were censored at the time of death in the cause-specific survival calculation. The Chi-square test was used for assessing the significance of the associations between various prognostic factors and survival or recurrence. RESULTS Ninety charts were obtained for analysis. Nine patients were deemed ineligible: 3 were excluded because subsequent pathological review showed the tumors to be benign adenomas, 4 patients had been referred with recurrent disease, one patient had a borderline tumor found incidentally at the time of BSOH for cancer of the endometrium, and 1 patient had a probable primary tumor of the large bowel, leaving 8 1 patients available for analysis. The mean age was 48 years with a range of 16 to 90
etal.
Borderline ovarian tumors 0 L. A. MANCHUL
Table 3. Type of adjuvant
Table 1. Frequency by stage Stage
I
Percent
No.
43 12 8
11
81
100
III IIIA IIIB IIIC Total
11
2 4 3 9 3 5 1
IIA IIB IIC
Proportion who died
Adhesions
Peritoneal
Tumour
Cytology
rupture
Absent Present Unknown None Sharp Blunt Both Unknown Negative Positive Unknown No Yes Unknown
XRT
Chemotherapy** +APRT Total
Table 2. Proportion of relapses or deaths in relation to the presence or absence of tumor excrescences, adhesions, rupture, or malignant cells in peritoneal washings
II
III
4 5% 5’ 6% 0
0
Total
0
0
5% 1 1%
63 78%
9 11%
9 11%
81 100%
4 5% 4”
* 3 died of other causes without tumor present. + 1 died of other causes without tumor present. t 3 cyclophosphphamide and cis-platinum, 1 melphalan. B2 died of ovarian tumor. ** cyclophosphamide and carboplatinum. AP RT = Abdomino-pelvic radiation therapy.
tients. Table 2 depicts the relationship between these factors and recurrence or survival; none of these factors were found to have a significant effect. The type of adjuvant therapy given by stage at presentation is outlined in Table 3. Eleven patients or 14% received radiation therapy, four patients received chemotherapy, and one patient received both radiation and chemotherapy. The rest, or 80%, received no adjuvant therapy. Six patients died, four of other causes without evidence of ovarian tumor present at the time of death. Two patients with Stage III disease at presentation died with ovarian tumor present despite aggressive chemotherapy with cyclophosphamide and cis-platinum, 2 and 4 years after initial diagnosis. Details of these patients are summarized in Table 4. Four patients were alive with persistent or relapsed ovarian tumor. One of these patients developed an adenocarcinoma of the cervix 20 years after undergoing pelvic radiation therapy for bilateral borderline ovarian tumors. The outcome (survival or relapse)
Proportion who relapsed*
2143 2133 215 O/31 2110 3126 O/2 l/6 l/18 l/14 4149 l/34 3120 2121
* Relapse is defined as relapse or persistent
by stage
65 80% 11 14% 4 5% 1 1%
61* 75% 2 2% 0
Chemotherapy*
years. Pain was the presenting symptom in 46% of patients, abnormal bleeding in 7%, a mass in 17%, and ascites in two patients. A pelvic mass was palpable in 43%. Fifty-eight or 72% of tumors were of the serous histological sub-type; 23 or 28% were mutinous. There were no borderline endometrioid or clear cell tumors in the cohort. Thirty-three percent of tumors were bilateral. Table 1 demonstrates the breakdown by stage: 78% were Stage 1, 11% Stage II, and 11% Stage III. Only two patients or 2.4% had ascites at laparotomy. Peritoneal washings were positive for malignant cells in 14 patients of 32 where peritoneal washings were assessed, but peritoneal fluid was either not obtained or not recorded in the charts in 49 patients. Rupture of the ovarian tumors occurred in 25%, and surface excrescences were noted in 40%. Adhesions were present in 46%, adhesions requiring blunt dissection were present in 32%, adhesions requiring sharp dissection in 12%, and both sharp and blunt dissection in two pa-
Excrescences
I Observation
9
II
therapy Stage
18
63 IA IB IC
869
o/43 4133 O/5 l/37 o/10 O/26 212 l/6 O/18 o/14 4149 2134 2120 o/27 disease.
Table 4. Details of patients
Pt. no.
Stage
73
IA
78 54 74 22 37
IB IC IIB IIIC IIIB
Adjuvant treatment None None None AP-RT CP CT* CP CT*
who died
Cause of death Congenital heart defect Colon cancer Lung cancer Lung cancer Ovarian cancer Ovarian cancer
* Never in remission. CP CT = Cyclophosphamide, cis-platinum AP RT = Abdomino-pelvic radiation.
Tumor present No No No No Yes Yes
chemotherapy;
870
I. J.
RadiationOncology0 Biology0 Physics
Volume 22, Number 5, 1992
Table 5. Details of patients alive with disease relapse or Dersistence Adjuvant Stage treatment
Pt.
no.
Relapse/persistence
55
IB
Pelvic RT
56
IB
AP RT
Relapse at 20 yrs. ? 2nd primary cancer, cervix Relapse at 8 yrs
17
IIB
None
lung*, liver metastases Pelvic relapse
72
IIIB
CP CT
Persistent
disease
Status Alive, with disease 20.
Alive, with disease Alive, with disease Alive, with disease
* Lung lesion biopsy-proven borderline histology. AP RT = Abdominopelvic radiation therapy; CP CT = Cyclophosphamide, cis-platinum chemotherapy. according to stage and type of treatment received is further detailed in Table 5. Table 6 summarizes the frequency of relapse, persistent disease and death from disease by stage and type of therapy. It is not possible to analyze the effectiveness of the various therapies employed because of the small numbers in each cell. The overall actuarial survival, cause-specific survival, and disease-free rate are depicted in Figure 1. Cause-specific survival and disease-free rate, by stage, are depicted in Figure 2. The median follow-up was 3.9 years with a range of 2 months to 23 years. Follow-up was complete to within 15 months of analysis for 80% of patients. Fortynine patients were followed for 5 or fewer years, 27 or 35% for 5 to 10 years, and 5 or 7% for 10 years or more. Table 7 outlines the recurrence rate by type of surgery, either complete BSOH or less than complete BSOH. Of the three patients who recurred, all had had a BSOH. None of the patients who received less than a BSOH recurred, but one of these patients died of other causes. Two patients treated by cystectomy only have not recurred. Eight patients were thought to have focal areas of microinvasion within the primary tumor on pathologic review at our institution. Seven of these patients are alive without evidence of tumor. Six received no adjuvant
0
0
10
20 Years after diagnosis
N-al-at”* 28
5
Fig. 1. Survival of borderline ovarian tumor patients. CSS = Cause specific survival; OS = Overall survival; DES = Diseasefree survival.
therapy following initial surgery (five Stage I and one Stage II) and one (Stage IIIB) received abdomino-pelvic radiation therapy. One patient with Stage IIIB disease is alive and well with tumor, having never had a response (as assessed by laparoscopy), following platinum-based chemotherapy. One patient was noted to have omental deposits considered to be benign endosalpingiosis consisting of glandlike structures lined with a single layer of columnar epithelium without atypia. Another patient had multiple implants throughout the pelvis and abdomen considered to be atypical endosalpingiosis without the criteria to allow classification as borderline tumors and is alive and well 1 year following the diagnosis. Seven patients had histologically proven borderline peritoneal implants. One of these patients died of disease despite chemotherapy. Four underwent abdomino-pelvic RT (3 Stage II, 1 Stage III), two of whom have moderate side effects from therapy (one with intermittent small bowel obstruction and one with pelvic fibrosis, pelvic pain and dyspareunia) with no clinical evidence of disease. Two remain well with no evidence of disease 1 year following initial surgical resection without adjuvant therapy. Five patients who had evidence
Table 6. Death from disease and relapse or persistence by stage and type of therapy Stage Adjuvant Observation XRT Chemotherapy Chemotherapy Total
therapy
+ XRT
I
II
III
Total
Of6 1 l/2 -
l/4 l/5 -
O/4 314 O/l
65 11 4
63
9
9
81
1
0
0
10
20 Years after diagnosis
Fig. 2. Survival by stage. CSS = Cause specific survival; DFS = Disease-free survival. The curve for Stage II/III DFS originates at less than 100% due to the patients who did not achieve remission.
Borderline ovarian tumors ??L. A. MANCHUL
Table 7. Type of initial surgery and relapse status No relapse
Relapse
Progressive disease
61 1 11 2
3 1* 0 0
2 0 0 0
Complete BSOH BSO us0 Cystectomy
* Adenocarcinoma of the cervix 20 years after pelvic radiation. BSOH = Bilateral salpingo-oophorectomy and hysterectomy; BSO = Bilateral salpingo-oophorectomy; US0 = Unilateral salpingo-oophorectomy.
of invasive pelvic or abdominal peritoneal implants with a borderline primary were optimally surgically debulked at initial laparotomy with no or minimal residual disease. These patients received a variety of treatments: two underwent abdomino-pelvic RT, two chemotherapy (one melphalan for 1 year and one cyclophosphamide and cisplatinum), and one had both AP-RT and cyclophosphamide, carboplatinum chemotherapy. All are alive but one developed liver, lung, and mediastinal nodal metastases 7 years following radiation therapy for Stage IIB disease. The status of patients with peritoneal implants is summarized in Table 8. DISCUSSION
This analysis was initially undertaken with specific goals in mind: 1. To assess the natural history of a cohort of patients with borderline ovarian malignancy. 2. To determine if any factors present at the time of initial diagnosis are predictive for survival or future recurrence which may provide a basis for sound management decisions. 3. To determine if adjuvant therapy influences outcome. 4. To determine if more limited surgery can be considered equivalent to the traditional BSOH in younger patients wishing to maintain fertility who present with Stage I disease. 5. To determine if patients with invasive extra-ovarian implants fare more poorly than those with non-invasive extra-ovarian implants and thus should be treated more aggressively. 6. To assess the impact of microinvasion on patient outcome. Table 8. Patient status according to invasive status of peritoneal implants Implant status
Alive, no disease
Alive, with disease
Dead with disease
Invasive Non-invasive
415 6/7
l/S 0
0 l/7
etal.
871
This analysis of 81 patients with borderline ovarian tumors confirms many observations that have been previously reported. These patients present at an earlier age and stage than invasive ovarian cancer (1, 2, 5, 18, 23, 25). The mean age of this cohort was 48, which is somewhat older than expected, but is likely a result of some degree of referral bias. Princess Margaret Hospital is a tertiary referral center providing an oncologic consultative service with facilities for administration of radiation therapy and chemotherapy but does not provide primary surgical management. Thus, there are probably many young patients with early stage, borderline tumors that are managed outside of our institution and not referred for an opinion regarding further management. Approximately 80% of this cohort presented with Stage I disease, with approximately 10% with Stage II and 10% with Stage III. No patient had Stage IV disease at presentation. This is consistent with previously reported series where the majority of patients presented with early stage disease. Six patients died, resulting in an overall actuarial survival of 85% at 10 and 20 years. Two patients died of advanced borderline ovarian tumor and four of other, unrelated causes without ovarian tumor present at the time of death resulting in a 1O-year cause-specific survival of 96%. Our Stage I patients, as expected, have an excellent prognosis, with an overall survival of almost 90% at 10 years and a cause-specific survival of 100%. No Stage I patient died of or with disease. One Stage I patient who underwent pelvic radiation therapy in 1967 and whose pathology was subsequently reviewed and found to be borderline tumor developed an adenocarcinoma of the cervix 20 years following treatment. It is not certain whether this represented a relapse or a new primary tumor. The majority (61 of 63) patients with Stage I disease underwent no further therapy following initial surgery, which in 77% consisted of BSOH. It is apparent that following a careful staging laparotomy with removal of the ovaries, no further therapy is necessary for Stage I patients. The extent of primary surgery necessary at the time of initial presentation in Stage I disease is not clearly defined at present. Lim Tan et al. reported an ipsilateral persistence or recurrence rate of 8% per ovary and 100% survival in a group of 35 patients with serous borderline tumours subjected to cystectomy alone (2 1). Thirty-three of these patients had Stage I tumors. Julian and Woodruff reported a 100% survival rate in 15 patients with Stage I borderline serous ovarian tumors who underwent unilateral adnexal removal but did not report the local recurrence rate ( 16). Tazelaar et al. reported a 15% recurrence rate in 20 patients who underwent limited surgery (cystectomy in four and oophorectomy in 16) for Stage IA borderline ovarian turnours, all of whom were alive and well following salvage therapy (32). In our cohort of patients, 15 underwent limited surgery consisting of bilateral salphingoophorectomy in two, unilateral salpingoophorectomy without a contralateral wedge resection in 11, and cystectomy alone in
872
1. J. Radiation Oncology 0 Biology 0 Physics
two. One patient who underwent a bilateral salpingoophorectomy followed by pelvic radiation developed a second primary adenocarcinoma of the cervix 20 years following treatment, which may or may not represent recurrence. None of the remaining 14 have developed recurrence, but follow-up is quite short in 1 of the patients who underwent cystectomy alone. Whether cystectomy alone is adequate therapy for Stage I borderline ovarian tumours is unresolved. Although there appears to be a higher local recurrence rate following cystectomy alone, salvage therapy appears to be effective, but is critically dependent upon close, intense follow-up to allow early discovery of a recurrence. Lim Tan et al. suggest that cystectomy alone is insufficient therapy in the presence of multiple cysts or a positive margin, as the majority of their recurrences occurred in this sub-group (21). The value of a contralateral ovarian wedge resection is doubtful, as only 5- 10% of normal appearing ovaries will contain histologic evidence of tumor and can result in impairment of fertility (2 I, 32, 33). Based on the results of a small number of patients in our cohort and previous reports, oophorectomy alone with a careful staging laparotomy for unilateral borderline tumor or bilateral salpingoophorectomy for bilateral disease with careful follow-up appears to be adequate treatment for Stage I disease. It is not possible to draw valid conclusions on the effectiveness of cystectomy alone, as only two patients in our cohort underwent this limited surgery. Further follow-up on a larger number of patients than those currently reported will be necessary before a definitive answer can be provided. The value of such prognostic factors as peritoneal cytology, ascites, cyst rupture, surface excrescences, tumor size, and cyst adherence requiring sharp or blunt dissection from other structures has not been previously reported for borderline ovarian tumors. Although certain of these factors have been found to carry prognostic significance in invasive epithelial ovarian cancer (8), none of these factors had a significant impact on survival or tumor recurrence in our study. Unfortunately, this information could not in all cases be obtained in this retrospective analysis so that the numbers for analysis were small. It is also difficult to assess the significance of prognostic factors when the number of events (death, relapse) was so small. Tavassoli and others have recently suggested that borderline ovarian tumors that contain focal area(s) of microinvasion without a destructive stromal reaction can be managed similarly to borderline ovarian tumors without invasion because of their similar natural history (30). lndeed, none of our patients with Stage I microinvasive borderline tumor relapsed, and it appears a moot point whether or not the term “microinvasive” has clinical significance in Stage I patients. The outcome for properly staged Stage I well-differentiated invasive epithelial ovarian cancer is extremely good, with greater than 90% 5 year survival (7), which is equivalent to similarly managed
Volume 22. Number 5. 1992
Stage I patients with borderline ovarian tumors. The presence of microinvasion may, however, be significant in Stage II and Ill patients where tumor deposits, whether non-invasive autochthonous “metastases” or true metastases from the primary borderline tumor, are demonstrated. The presence of microinvasion in any case should be regarded with caution and should prompt a more thorough pathologic assessment of the primary tumor. The value of adjuvant therapy in Stage II and Ill disease is unclear. Consistent with other reported series, our group of Stage II and Ill patients had a less favourable prognosis than those with Stage I disease, with a 75% IO-year survival and 50% IO-year disease-free survival, despite the use of adjuvant radiation therapy, chemotherapy, or both. It has been our usual policy to recommend radiation therapy in Stage II or Ill disease when microscopic or small macroscopic disease remained following staging laparotomy. There was a tendency to use chemotherapy in Stage Ill disease when residual tumor was more bulky. One of five Stage II patients who underwent abdominopelvic RT relapsed in the liver and lung, and one of four Stage II patients who underwent no further therapy relapsed in the pelvis. Of our nine Stage Ill patients, four underwent abdominopelvic radiation with no recurrences. Two of four patients who underwent chemotherapy with cyclophosphamide and cis-platinum died with progressive disease and one patient has persistent disease. One patient received both chemotherapy and radiation therapy and is still alive without disease at 2.5 years. It is not possible to draw valid conclusions regarding the impact of adjuvant therapy on outcome from this analysis, as the numbers available for analysis were small and events occurring few. It is also not possible, in retrospect, to ensure the similarity of the different treatment groups to allow valid comparisons. Other centers have experienced similar difficulties in advancing sound therapeutic recommendations for Stage II and Ill disease with a wide disparity in opinion concerning the value of adjuvant therapy. Hopkins et al. reported on 68 patients with borderline ovarian tumors, suggesting that radiation therapy prolonged survival in Stage Ill disease, as compared to observation or chemotherapy ( 14). The number of patients available for assessment was small (17 for the three groups), the length of follow-up was not equivalent for the chemotherapy and radiation therapy groups, and the equivalence between the groups of patients subjected to each therapeutic modality is uncertain. Both Nation et al. and O’Quinn et al. reported the ineffectiveness of chemotherapy in management of patients with advanced disease (13,29). Nation et al, reported that all Stage II and Ill patients with macroscopic disease following initial laparotomy had a positive second look laparotomy following chemotherapy with melphalan or cyclophosphamide and cis-platinum (23). However, two cases with minimal disease had a negative second look laparotomy. O’Quinn reported that mel-
Borderline ovarian tumors 0 L. A. MANCHUL
phalan did not eradicate residual tumour in their series of patients who underwent second-look laparotomy (24). Kliman et al. reported that only one patient of 12 who had residual disease following initial laparotomy responded to chemotherapy (19). Fort et al., in a careful analysis of their series of patients with borderline tumors, suggest that adjuvant therapy can eradicate residual disease (11). Of 15 patients who had residual disease at initial laparotomy, 12 were free of disease at second-look laparotomy following therapy consisting of chemotherapy, radiation therapy, or both. There are inherent difficulties in attempting to define treatment recommendations from these retrospective reviews. The patient numbers are too small to allow valid conclusions, the similarity of the patient populations undergoing each form of therapy (or no therapy) is doubtful, the rigor of the initial debulking procedure is not uniform, the pathologic criteria for diagnosis are not uniform, and end points assessed vary from clinical to surgico-pathologic. As well, there are biologic reasons that may explain the apparent lack of effectiveness of adjuvant therapy. Borderline ovarian tumors have a slow doubling time and may remain indolent for several years, rendering them less sensitive to cycle-specific chemotherapeutic agents and radiation. The true nature of the peritoneal implants in presumed advanced stage borderline ovarian tumors is controversial. At the benign end of the spectrum are the gland-like inclusions of endosalpingiosis or atypical endosalpingiosis, which should not be considered metastases from the primary borderline ovarian tumor. In the past, if these were mistakenly identified as metastases, che-
ela/.
873
motherapy or radiation therapy would not be expected to be effective. There have been suggestions that the true borderline peritoneal lesions may represent autochthonous (or simultaneous) extra-ovarian neoplasia and not true metastases and may carry a better prognosis than clearly invasive peritoneal lesions. Bell has suggested that the presence of invasion in peritoneal implants carries a significantly poorer prognosis (3). In our series, however, only 1 of 5 patients with invasive implants has recurred. These patients received adjuvant radiation therapy, chemotherapy, or both. It may be that these invasive lesions are more poorly differentiated with a higher growth fraction and shorter doubling time and might respond better to aggressive therapy than the borderline implants. Maximum debulking of the non-invasive implants should be carried out at initial laparotomy, and should be the mainstay of initial therapy. It is necessary to initiate a national borderline ovarian tumor registry in order to prospectively follow all patients with this diagnosis so that rational therapeutic management recommendations may be made in the future. Preferably a multicenter prospective randomized trial should be undertaken to try to resolve the current controversies. In the interim, optimum therapeutic recommendations must take into account the age of the patient and her desire for fertility, patient willingness to accept a somewhat higher recurrence rate with limited surgery, the ability to carry out frequent follow-up, the ability to resect all apparent disease completely, the extent of disease, and the invasive or non-invasive nature of peritoneal implants.
REFERENCES 1. Aure, J. C.; Hoeg, K.; Kolstad, P. Clinical and histologic studies of ovarian carcinoma: Long term follow-up of 990 cases. Obstet. Gynecol. 37: 1-9; 197 1. 2. Bamhill, D.; Heller, P.; Brzozowski, P.; Advani, H.; Gallup, D.; Park, R. Epithelial ovarian carcinoma of low malignant potential. Obstet. Gynecol. 65: 53-59; 1985. 3. Bell, D. A.; Weinstock, M. A.; Scully, R. E. Peritoneal implants of ovarian serous borderline tumors. Cancer 62: 2212-2222; 1988. 4. Bell, D. A.; Scully, R. E. Ovarian serous borderline tumours with stromal microinvasion: a report of 2 1 cases. Hum. Pathol. 2 1: 397-403; 1990. 5. Bostwick, D. G.; Tazelaar, H. D.; Ballon, S. C.; Hendrickson, M. R.; Kempson, R. L. Ovarian epithelial tumors of borderline malignancy. A clinical and pathologic study of 109 cases. Cancer 58: 2052-2065; 1986. 6. Chambers, J. T.; Merino, M. J.; Kohom, E. I.; Schwartz, P. T. Borderline ovarian tumors. Am. J. Obstet. Gynecol. 159: 1088-1094; 1988. 7. Dembo, A. J. Abdominopelvic radiotherapy in ovarian cancer: a 10 year experience. Cancer 55: 2285-2290; 1985. 8. Dembo, A. J.; Stenwig, A. E.; Berle, E. J.; Bush, R. S.; Kjorstad, K. Prognostic factors in patients with Stage I epithelial ovarian cancer. Obstet. Gynecol. 75: 263-273; 1990. 9. Dembo, A. J.; VanDyk, J.; Japp B.; Bean, H. A.; Beale, F.; Pringle, J. F.; Bush, R. S. Whole abdominal radiation by a
10.
1 I.
12. 13.
14.
15.
16.
17.
moving strip technique for patients with ovarian cancer. Int. J. Radiat. Oncol. Biol. Phys. 5: 1933-1942; 1979. Dgani, R.; Blickstein, I.; Stroham, Z.; Czemobilsky, B.; BenHur, H.; Shani, A.; Katz, Z.; Borenstein, R. Clinical aspects of ovarian tumors of low malignant potential. Eur. J. Obstet. Gynecol. Reprod. Biol. 35: 25 l-258; 1990. Fort, M. G.; Pierce, V. K.; Saigo, P. E.; Hoskins, W. J.; Lewis, J. L. Evidence for the efficacy of adjuvant therapy in epithelial ovarian tumors of low malignant potential. Gynecol. Oncol. 32: 269-272; 1989. Fox, H. The concept of borderline malignancy in ovarian tumors: a reappraisal. Curr. Top. Path. 78: 111-134; 1989. Gershenson, D. M.; Silva, E. G. Serous ovarian tumors of low malignant potential with peritoneal implants. Cancer 65: 578-585; 1990. Hopkins, M. P.; Kumar, N. B.; Morley, G. W. An assessment of pathologic features and treatment modalities in ovarian tumors of low malignant potential. Obstet. Gynecol. 70: 923-929; 1987. International Federation of Gynecology and Obstetrics. Classification and staging of malignant tumors of the female pelvis. Acta. Obstet. Gynecol. Scandinav. 50: l-7; 197 1. Julian, C. G.; Woodruff, J. D. The biologic behaviour of low grade papillary serous carcinoma of the ovary. Obstet. Gynecol. 40: 860-867; 1972. Kaplan, E. L.; Meier, P. Nonparametric estimation from
874
18.
19.
20.
21.
22.
23.
24.
25.
I. J. Radiation Oncology 0 Biology 0 Physics
incomplete observations. J. Am. Stat. Assoc. 53: 457-48 1; 1953. Katzenstein, A. A.; Mazur, M. T.; Morgan, T. E.; Kao, M. S. Proliferative serous tumors of the ovary: histologic features and prognosis. Am. J. Surg. Pathol. 2: 339-355; 1978. Kliman, L.; Rome, R. M.; Fortune, D. W. Low malignant potential tumors of the ovary: a study of 76 cases. Obstet. Gynecol. 68: 338-344; 1986. Kottmeier, H. L. Annual report of the results of treatment in gynecological cancer. Int. Fed. Gynecol. Obstet. Stockholm: Pogo Print; 1979. Lim-Tan, S. K.; Cajigas, H. E.; Scully, R. E. Ovarian cystectomy for serous borderline tumors: a follow-up study of 35 cases. Obstet. Gynecol. 72: 775-781; 1988. McCaughey, W. T. E.; Kirk, M. E.; Lester, W.; Dardick, I. Peritoneal epithelial lesions associated with proliferative serous tumors of ovary. Histopathol. 8: 195-208; 1984. Nation, J. G.; Krepart, G. V. Ovarian carcinoma of low malignant potential: staging and treatment. Am. J. Obstet. Gynecol. 154: 290-293; 1986. O’Quinn, A. G.; Hannigan, E. V. Epithelial ovarian neoplasms of low malignant potential. Gynecol. Oncol. 2 1: 177185; 1985. Russell, P.; Merkur, H. Proliferating ovarian “epithelial” tumors: a clinico-pathologic analysis of 144 cases. Aust. N. Z. J. Obstet. Gynaec. 19: 45-51; 1979.
Volume 22, Number 5, 1992
26. Russel, P.; Bannatyne, P. M.; Solomon, H. J.; Stoddard, L. D.; Tattersall, M. H. N. Multifocal tumorigenesis in the upper female genital tract-implications for staging and management. Int. J. Gynecol. Pathol. 4: 192-210; 1985. 27. Scully, R. E. Common epithelial tumors of borderline malignancy (carcinomas of low malignant potential). Bull. Cancer (Paris) 69: 228-238; 1982. 28. Serov, S. F.; Scully, R. E.; Sobin, L. H. International histological classification of tumors, No. 9. Histological typing of ovarian tumors. Geneva: World Health Organization; 1973: 37-38. 29. Tasker, M.; Langley, F. A. The outlook for women with borderline epithelial tumors of the ovary. Brit. J. Obstet. Gynecol. 92: 969-973; 1985. 30. Tavassoli, F. A. Serous tumor of low malignant potential with early stromal invasion (serous LMP with microinvasion). Mod. Pathol. 1: 407-414; 1988. 31. Taylor, H. C. Malignant and semimalignant tumors of the ovary. Surg. Gynecol. Obstet. 48: 204-230; 1929. 32. Tazelaar, H. D.; Bostwick, D. G.; Ballon, S. C.; Hendrickson, M. R.; Kempson, R. L. Conservative treatment of borderline ovarian tumors. Obstet. Gynecol. 66: 417-422; 1985. 33. Williams, T. J.; Dockerty, M. B. Status of the contralateral ovary in encapsulated low grade malignant tumors of the ovary. Surg. Gynecol. Obstet. 143: 763-766; 1976. 34. Yoonesi, M.; Crickard, K.; Celik, C.; Yoonessi, S. Borderline epithelial tumors of the ovary: ovarian intraepithelial neoplasia. Obstet. Gynecol. Survey. 43: 435-444; 1988.
0360-3016/92 $5.00 Copyright 0 1992 Pergamon Pres Ltd.
Im. J Radialron Oncology Biol Phys.. Vol. 22, pp. 867-874 Printed in the U.S.A All rights reserved.
??Special Feature
BORDERLINE EPITHELIAL OVARIAN TUMORS: A REVIEW OF 81 CASES WITH AN ASSESSMENT OF THE IMPACT OF TREATMENT L. A. MANCHUL, B.Sc., M.D., F.R.C.P.C.,’ J. SIMM, M.SC.,~ W. LEVIN, M.B., F.R.C.P.C.,’ A. W. FYLES, M.D.,
F.R.C.P.C.,’
A. J. DEMBO, M.B.,
F.R.C.P.C.,4
J. F. PRINGLE, M.B., F.R.C.P.C.,’ G. A. RAWLINGS, M.D., F.R.C.P.C.,’ J. F. G. STURGEON, M.B., F.R.C.P.C.3 AND G. M. THOMAS, M.D., F.R.C.P.C.4 ‘Departmentsof Radiation Oncology, ‘Biostatistics,and ‘Medicine, PrincessMargaretHospital, and 4Departmentof Radiation Oncology, Toronto-BayviewRegional CancerCentre,Universityof Toronto, Toronto, Ontario, Canada Optimal management of borderline epithelial ovarian tumors remains controversial because of the lack of clear, universally accepted pathologic criteria for diagnosis, the lack of complete understanding of the significance of intraperitoneal implants, and the desire to employ more limited surgery in young women. We reviewed the experience with borderline epithelial ovarian tumors at Princess Margaret Hospital in order to assess the natural history of the disease, to determine prognostic factors that would aid in management decisions, and to determine if adjuvant therapy influenced outcome. Eighty-one patients were analyzed. The mean age was 48 years. Seventy-two percent of tumors were of the serous histologic sub-type and 28% were mutinous. Seventy-eight percent were Stage I, 11% Stage II, and 11% Stage III. Peritoneal washings contained malignant cells in 14 of 32 patients (not recorded or obtained in 49), cyst rupture occurred in 25%, surface excrescences in 402, and adhesions in 46%. None of these factors had a significant effect on recurrence rate or survival. Eleven patients received adjuvant radiation therapy (10 abdomino-pelvic and 1 pelvic alone), four adjuvant chemotherapy, and one both radiation therapy and chemotherapy. The rest (65) received no adjuvant therapy. Due to the small numbers and infrequent events, it was not possible to analyze and thus draw valid conclusions regarding the effect of adjuvant therapy on survival or recurrence. The overall survival (OS) and cause specific survival (CSS) were 85% and 96% at 10 years, respectively. No Stage I patient died of tumor. OS for Stage I patients was 90% at 10 years, the majority of whom (61 of 63) received no adjuvant therapy, and is thus unnecessary in Stage I disease. The adequacy of unilateral oophorectomy or ovarian cystectomy could not be confirmed because of small numbers. The 10 year OS and disease-free survival in Stage II and III were 75% and 50%, respectively, despite the use of adjuvant radiation therapy, chemotherapy, or both. It is necessary to create a multi-center tumor registry in order to acquire a prospective data base from which to develop sound therapeutic decisions. Borderline epithelial ovarian neoplasms, Ovarian tumors of low malignant potential, Therapeutic management.
malignant” having some but not all of the morphologic features of malignancy: stratification of epithelial cells, epithelial budding and detachment, with some degree of nuclear atypia and increased mitotic activity but without stromal invasion (28). Borderline epithelial ovarian tumors account for approximately 10 to 20% of all ovarian tumors ( 1,20). They affect a younger group of women than do malignant epithelial ovarian tumors, and frequently women who would wish to maintain their child-bearing potential. They generally present at an earlier stage than do malignant tumors: anywhere from 50 to 85% present as Stage I ( 1, 2, 5, 18, 23,25). Patients who present with Stage I disease have an
INTRODUCTION
In 1929 Taylor described a group of papillary ovarian tumors that appeared clinically malignant with overt peritoneal metastases, demonstrating some of the histologic criteria of malignancy, but that behaved in a “benign” fashion (3 1). It was not until 197 1 that FIG0 officially recognized epithelial ovarian tumors of “low potential malignancy” with epithelial proliferation and nuclear abnormalities but without infiltrative destructive growth (15). In 1973 the WHO Histologic Classification of Tumours included a form of common epithelial tumor “intermediate between one clearly benign and obviously
Reprint requests to: Dr. L. A. Manchul, Department of Radiation Oncology, Princess Margaret Hospital, 500 Sherboume St., Toronto, Ontario, M4X lK9 Canada. Acknowledgements-We appreciatethe expertiseof DebbieTsuji
and June Saunders for data collection and management and Colette Plasway for secretarial assistance. Accepted for publication 6 September 199 I. 867
868
1. J. Radiation Oncology 0 Biology 0 Physics
excellent prognosis; however, late recurrences and distant metastases have been reported (1, 2, 5, 10, 18). Patients who present with disseminated intra-abdominal disease fare worse than those without, but their survival exceeds that of patients with equivalent stage invasive epithelial ovarian cancer (1, 3, 6, 11, 14, 16, 18, 19, 23, 29, 34). Optimal management of patients with borderline epithelial ovarian tumors is controversial and is difficult to define clearly for a number of reasons. First, the pathologic criteria for diagnosis are not universally accepted, and “microinvasive” tumors have been recently described (4? 30), leaving in doubt the uniformity of the patient population from which current recommendations are derived. Second, there is controversy regarding the nature and significance of the peritoneal implants found in patients with presumed advanced disease with evidence emerging to suggest that invasive implants carry a worse prognosis than non-invasive implants (4, 22). There is some doubt that non-invasive implants represent true metastases but rather autochthonous extra-ovarian neoplasia and should be managed differently from the invasive implants (12, 26, 27). Third, there is an increasing tendency to recommend more limited surgery in patients presenting with unilateral Stage I disease based on the results of small series (21, 32). Although more limited surgery respects the patient’s desire to maintain fertility, it is not yet known in which women this approach is most appropriate. The value of adjuvant therapy is unclear and cannot be accurately extrapolated from the literature on the subject, which consists largely of retrospective reviews of small numbers of patients subjected to a variety of therapeutic modalities over a long period of time. Pathologic diagnostic criteria are not always consistent from series to series, staging investigations are not uniform, and different end points are often assessed. We reviewed the experience of borderline epithelial ovarian tumors at the Princess Margaret Hospital to gain an insight into the natural history of the disease and to determine prognostic factors that would aid in providing a rational approach to therapeutic management. METHODS
AND MATERIALS
The cohort of patients who formed the basis for this analysis consisted of those patients who were referred to the Princess Margaret Hospital between 1967 and 1988 with a diagnosis of borderline ovarian cancer or with a diagnosis of ovarian cancer where subsequent review at this Institution confirmed borderline histology. The analyses were confined to those patients referred with a newly diagnosed borderline ovarian tumor. Patients who were initially referred with a relapse or who presented with a concurrent invasive cancer were excluded from the analysis. Information on age, clinical features and investigations at presentation, surgical findings, adjuvant postoperative
Volume 22, Number 5, 1992
treatment, and pathologic criteria for the diagnosis was obtained from the patients’ charts or from the referring institution or physician as necessary. Information regarding such possible prognostic factors as the presence of ascites, surface excrescences, tumor rupture, adherence, tumor size, the amount of residual disease, and the presence of malignant cells in peritoneal washings was obtained where available. All patients were retrospectively staged according to UICC 1987 criteria. All pathology slides were reviewed by one of three expert gynecologic pathologists at the Princess Margaret Hospital but were not re-reviewed specifically in conjunction with this study. Those tumors which demonstrated areas of microinvasion without destructive stromal infiltration were identified. Peritoneal implants were assessed as representing either atypical endosalpingiosis, borderline tumor implants, or frankly invasive metastases with a borderline primary. Initial therapy consisted of either a total abdominal hysterectomy with bilateral salpingoophorectomy (BSOH), a unilateral oophorectomy or salpingoophorectomy, or rarely, a cystectomy alone. Adjuvant radiation therapy consisted usually of whole abdomino-pelvic radiation (AP-RT) with a pelvic boost or, rarely, pelvic radiation alone. Pelvic doses ranged from 40 Gy to 47 Gy (mid-plane) in 20-35 fractions over 28-98 days. Abdominal doses ranged from 22.5 Gy to 26.4 Gy in 22-25 fractions over 31 to 84 days. Rarely, the abdomino-pelvic strip technique (previously described) was employed, utilizing 22.5 Gy in 2.25 Gy fractions to the whole abdomen and a boost to the pelvis of 22.5 Gy in 2.25 Gy fractions (9). Chemotherapy consisted of several regimens including melphalan alone or cyclophosphamide and cis-platinum (or carboplatinum) with or without adriamycin. Patient status was recorded at the time of analysis as alive or dead, noting the presence or absence of ovarian tumor. The site of relapse and the cause of death were recorded. Overall, cause-specific and disease-free survival were determined using Kaplan-Meier estimates (17). Patients who died from another cause without the presence of ovarian tumor were censored at the time of death in the cause-specific survival calculation. The Chi-square test was used for assessing the significance of the associations between various prognostic factors and survival or recurrence. RESULTS Ninety charts were obtained for analysis. Nine patients were deemed ineligible: 3 were excluded because subsequent pathological review showed the tumors to be benign adenomas, 4 patients had been referred with recurrent disease, one patient had a borderline tumor found incidentally at the time of BSOH for cancer of the endometrium, and 1 patient had a probable primary tumor of the large bowel, leaving 8 1 patients available for analysis. The mean age was 48 years with a range of 16 to 90
etal.
Borderline ovarian tumors 0 L. A. MANCHUL
Table 3. Type of adjuvant
Table 1. Frequency by stage Stage
I
Percent
No.
43 12 8
11
81
100
III IIIA IIIB IIIC Total
11
2 4 3 9 3 5 1
IIA IIB IIC
Proportion who died
Adhesions
Peritoneal
Tumour
Cytology
rupture
Absent Present Unknown None Sharp Blunt Both Unknown Negative Positive Unknown No Yes Unknown
XRT
Chemotherapy** +APRT Total
Table 2. Proportion of relapses or deaths in relation to the presence or absence of tumor excrescences, adhesions, rupture, or malignant cells in peritoneal washings
II
III
4 5% 5’ 6% 0
0
Total
0
0
5% 1 1%
63 78%
9 11%
9 11%
81 100%
4 5% 4”
* 3 died of other causes without tumor present. + 1 died of other causes without tumor present. t 3 cyclophosphphamide and cis-platinum, 1 melphalan. B2 died of ovarian tumor. ** cyclophosphamide and carboplatinum. AP RT = Abdomino-pelvic radiation therapy.
tients. Table 2 depicts the relationship between these factors and recurrence or survival; none of these factors were found to have a significant effect. The type of adjuvant therapy given by stage at presentation is outlined in Table 3. Eleven patients or 14% received radiation therapy, four patients received chemotherapy, and one patient received both radiation and chemotherapy. The rest, or 80%, received no adjuvant therapy. Six patients died, four of other causes without evidence of ovarian tumor present at the time of death. Two patients with Stage III disease at presentation died with ovarian tumor present despite aggressive chemotherapy with cyclophosphamide and cis-platinum, 2 and 4 years after initial diagnosis. Details of these patients are summarized in Table 4. Four patients were alive with persistent or relapsed ovarian tumor. One of these patients developed an adenocarcinoma of the cervix 20 years after undergoing pelvic radiation therapy for bilateral borderline ovarian tumors. The outcome (survival or relapse)
Proportion who relapsed*
2143 2133 215 O/31 2110 3126 O/2 l/6 l/18 l/14 4149 l/34 3120 2121
* Relapse is defined as relapse or persistent
by stage
65 80% 11 14% 4 5% 1 1%
61* 75% 2 2% 0
Chemotherapy*
years. Pain was the presenting symptom in 46% of patients, abnormal bleeding in 7%, a mass in 17%, and ascites in two patients. A pelvic mass was palpable in 43%. Fifty-eight or 72% of tumors were of the serous histological sub-type; 23 or 28% were mutinous. There were no borderline endometrioid or clear cell tumors in the cohort. Thirty-three percent of tumors were bilateral. Table 1 demonstrates the breakdown by stage: 78% were Stage 1, 11% Stage II, and 11% Stage III. Only two patients or 2.4% had ascites at laparotomy. Peritoneal washings were positive for malignant cells in 14 patients of 32 where peritoneal washings were assessed, but peritoneal fluid was either not obtained or not recorded in the charts in 49 patients. Rupture of the ovarian tumors occurred in 25%, and surface excrescences were noted in 40%. Adhesions were present in 46%, adhesions requiring blunt dissection were present in 32%, adhesions requiring sharp dissection in 12%, and both sharp and blunt dissection in two pa-
Excrescences
I Observation
9
II
therapy Stage
18
63 IA IB IC
869
o/43 4133 O/5 l/37 o/10 O/26 212 l/6 O/18 o/14 4149 2134 2120 o/27 disease.
Table 4. Details of patients
Pt. no.
Stage
73
IA
78 54 74 22 37
IB IC IIB IIIC IIIB
Adjuvant treatment None None None AP-RT CP CT* CP CT*
who died
Cause of death Congenital heart defect Colon cancer Lung cancer Lung cancer Ovarian cancer Ovarian cancer
* Never in remission. CP CT = Cyclophosphamide, cis-platinum AP RT = Abdomino-pelvic radiation.
Tumor present No No No No Yes Yes
chemotherapy;
870
I. J.
RadiationOncology0 Biology0 Physics
Volume 22, Number 5, 1992
Table 5. Details of patients alive with disease relapse or Dersistence Adjuvant Stage treatment
Pt.
no.
Relapse/persistence
55
IB
Pelvic RT
56
IB
AP RT
Relapse at 20 yrs. ? 2nd primary cancer, cervix Relapse at 8 yrs
17
IIB
None
lung*, liver metastases Pelvic relapse
72
IIIB
CP CT
Persistent
disease
Status Alive, with disease 20.
Alive, with disease Alive, with disease Alive, with disease
* Lung lesion biopsy-proven borderline histology. AP RT = Abdominopelvic radiation therapy; CP CT = Cyclophosphamide, cis-platinum chemotherapy. according to stage and type of treatment received is further detailed in Table 5. Table 6 summarizes the frequency of relapse, persistent disease and death from disease by stage and type of therapy. It is not possible to analyze the effectiveness of the various therapies employed because of the small numbers in each cell. The overall actuarial survival, cause-specific survival, and disease-free rate are depicted in Figure 1. Cause-specific survival and disease-free rate, by stage, are depicted in Figure 2. The median follow-up was 3.9 years with a range of 2 months to 23 years. Follow-up was complete to within 15 months of analysis for 80% of patients. Fortynine patients were followed for 5 or fewer years, 27 or 35% for 5 to 10 years, and 5 or 7% for 10 years or more. Table 7 outlines the recurrence rate by type of surgery, either complete BSOH or less than complete BSOH. Of the three patients who recurred, all had had a BSOH. None of the patients who received less than a BSOH recurred, but one of these patients died of other causes. Two patients treated by cystectomy only have not recurred. Eight patients were thought to have focal areas of microinvasion within the primary tumor on pathologic review at our institution. Seven of these patients are alive without evidence of tumor. Six received no adjuvant
0
0
10
20 Years after diagnosis
N-al-at”* 28
5
Fig. 1. Survival of borderline ovarian tumor patients. CSS = Cause specific survival; OS = Overall survival; DES = Diseasefree survival.
therapy following initial surgery (five Stage I and one Stage II) and one (Stage IIIB) received abdomino-pelvic radiation therapy. One patient with Stage IIIB disease is alive and well with tumor, having never had a response (as assessed by laparoscopy), following platinum-based chemotherapy. One patient was noted to have omental deposits considered to be benign endosalpingiosis consisting of glandlike structures lined with a single layer of columnar epithelium without atypia. Another patient had multiple implants throughout the pelvis and abdomen considered to be atypical endosalpingiosis without the criteria to allow classification as borderline tumors and is alive and well 1 year following the diagnosis. Seven patients had histologically proven borderline peritoneal implants. One of these patients died of disease despite chemotherapy. Four underwent abdomino-pelvic RT (3 Stage II, 1 Stage III), two of whom have moderate side effects from therapy (one with intermittent small bowel obstruction and one with pelvic fibrosis, pelvic pain and dyspareunia) with no clinical evidence of disease. Two remain well with no evidence of disease 1 year following initial surgical resection without adjuvant therapy. Five patients who had evidence
Table 6. Death from disease and relapse or persistence by stage and type of therapy Stage Adjuvant Observation XRT Chemotherapy Chemotherapy Total
therapy
+ XRT
I
II
III
Total
Of6 1 l/2 -
l/4 l/5 -
O/4 314 O/l
65 11 4
63
9
9
81
1
0
0
10
20 Years after diagnosis
Fig. 2. Survival by stage. CSS = Cause specific survival; DFS = Disease-free survival. The curve for Stage II/III DFS originates at less than 100% due to the patients who did not achieve remission.
Borderline ovarian tumors ??L. A. MANCHUL
Table 7. Type of initial surgery and relapse status No relapse
Relapse
Progressive disease
61 1 11 2
3 1* 0 0
2 0 0 0
Complete BSOH BSO us0 Cystectomy
* Adenocarcinoma of the cervix 20 years after pelvic radiation. BSOH = Bilateral salpingo-oophorectomy and hysterectomy; BSO = Bilateral salpingo-oophorectomy; US0 = Unilateral salpingo-oophorectomy.
of invasive pelvic or abdominal peritoneal implants with a borderline primary were optimally surgically debulked at initial laparotomy with no or minimal residual disease. These patients received a variety of treatments: two underwent abdomino-pelvic RT, two chemotherapy (one melphalan for 1 year and one cyclophosphamide and cisplatinum), and one had both AP-RT and cyclophosphamide, carboplatinum chemotherapy. All are alive but one developed liver, lung, and mediastinal nodal metastases 7 years following radiation therapy for Stage IIB disease. The status of patients with peritoneal implants is summarized in Table 8. DISCUSSION
This analysis was initially undertaken with specific goals in mind: 1. To assess the natural history of a cohort of patients with borderline ovarian malignancy. 2. To determine if any factors present at the time of initial diagnosis are predictive for survival or future recurrence which may provide a basis for sound management decisions. 3. To determine if adjuvant therapy influences outcome. 4. To determine if more limited surgery can be considered equivalent to the traditional BSOH in younger patients wishing to maintain fertility who present with Stage I disease. 5. To determine if patients with invasive extra-ovarian implants fare more poorly than those with non-invasive extra-ovarian implants and thus should be treated more aggressively. 6. To assess the impact of microinvasion on patient outcome. Table 8. Patient status according to invasive status of peritoneal implants Implant status
Alive, no disease
Alive, with disease
Dead with disease
Invasive Non-invasive
415 6/7
l/S 0
0 l/7
etal.
871
This analysis of 81 patients with borderline ovarian tumors confirms many observations that have been previously reported. These patients present at an earlier age and stage than invasive ovarian cancer (1, 2, 5, 18, 23, 25). The mean age of this cohort was 48, which is somewhat older than expected, but is likely a result of some degree of referral bias. Princess Margaret Hospital is a tertiary referral center providing an oncologic consultative service with facilities for administration of radiation therapy and chemotherapy but does not provide primary surgical management. Thus, there are probably many young patients with early stage, borderline tumors that are managed outside of our institution and not referred for an opinion regarding further management. Approximately 80% of this cohort presented with Stage I disease, with approximately 10% with Stage II and 10% with Stage III. No patient had Stage IV disease at presentation. This is consistent with previously reported series where the majority of patients presented with early stage disease. Six patients died, resulting in an overall actuarial survival of 85% at 10 and 20 years. Two patients died of advanced borderline ovarian tumor and four of other, unrelated causes without ovarian tumor present at the time of death resulting in a 1O-year cause-specific survival of 96%. Our Stage I patients, as expected, have an excellent prognosis, with an overall survival of almost 90% at 10 years and a cause-specific survival of 100%. No Stage I patient died of or with disease. One Stage I patient who underwent pelvic radiation therapy in 1967 and whose pathology was subsequently reviewed and found to be borderline tumor developed an adenocarcinoma of the cervix 20 years following treatment. It is not certain whether this represented a relapse or a new primary tumor. The majority (61 of 63) patients with Stage I disease underwent no further therapy following initial surgery, which in 77% consisted of BSOH. It is apparent that following a careful staging laparotomy with removal of the ovaries, no further therapy is necessary for Stage I patients. The extent of primary surgery necessary at the time of initial presentation in Stage I disease is not clearly defined at present. Lim Tan et al. reported an ipsilateral persistence or recurrence rate of 8% per ovary and 100% survival in a group of 35 patients with serous borderline tumours subjected to cystectomy alone (2 1). Thirty-three of these patients had Stage I tumors. Julian and Woodruff reported a 100% survival rate in 15 patients with Stage I borderline serous ovarian tumors who underwent unilateral adnexal removal but did not report the local recurrence rate ( 16). Tazelaar et al. reported a 15% recurrence rate in 20 patients who underwent limited surgery (cystectomy in four and oophorectomy in 16) for Stage IA borderline ovarian turnours, all of whom were alive and well following salvage therapy (32). In our cohort of patients, 15 underwent limited surgery consisting of bilateral salphingoophorectomy in two, unilateral salpingoophorectomy without a contralateral wedge resection in 11, and cystectomy alone in
872
1. J. Radiation Oncology 0 Biology 0 Physics
two. One patient who underwent a bilateral salpingoophorectomy followed by pelvic radiation developed a second primary adenocarcinoma of the cervix 20 years following treatment, which may or may not represent recurrence. None of the remaining 14 have developed recurrence, but follow-up is quite short in 1 of the patients who underwent cystectomy alone. Whether cystectomy alone is adequate therapy for Stage I borderline ovarian tumours is unresolved. Although there appears to be a higher local recurrence rate following cystectomy alone, salvage therapy appears to be effective, but is critically dependent upon close, intense follow-up to allow early discovery of a recurrence. Lim Tan et al. suggest that cystectomy alone is insufficient therapy in the presence of multiple cysts or a positive margin, as the majority of their recurrences occurred in this sub-group (21). The value of a contralateral ovarian wedge resection is doubtful, as only 5- 10% of normal appearing ovaries will contain histologic evidence of tumor and can result in impairment of fertility (2 I, 32, 33). Based on the results of a small number of patients in our cohort and previous reports, oophorectomy alone with a careful staging laparotomy for unilateral borderline tumor or bilateral salpingoophorectomy for bilateral disease with careful follow-up appears to be adequate treatment for Stage I disease. It is not possible to draw valid conclusions on the effectiveness of cystectomy alone, as only two patients in our cohort underwent this limited surgery. Further follow-up on a larger number of patients than those currently reported will be necessary before a definitive answer can be provided. The value of such prognostic factors as peritoneal cytology, ascites, cyst rupture, surface excrescences, tumor size, and cyst adherence requiring sharp or blunt dissection from other structures has not been previously reported for borderline ovarian tumors. Although certain of these factors have been found to carry prognostic significance in invasive epithelial ovarian cancer (8), none of these factors had a significant impact on survival or tumor recurrence in our study. Unfortunately, this information could not in all cases be obtained in this retrospective analysis so that the numbers for analysis were small. It is also difficult to assess the significance of prognostic factors when the number of events (death, relapse) was so small. Tavassoli and others have recently suggested that borderline ovarian tumors that contain focal area(s) of microinvasion without a destructive stromal reaction can be managed similarly to borderline ovarian tumors without invasion because of their similar natural history (30). lndeed, none of our patients with Stage I microinvasive borderline tumor relapsed, and it appears a moot point whether or not the term “microinvasive” has clinical significance in Stage I patients. The outcome for properly staged Stage I well-differentiated invasive epithelial ovarian cancer is extremely good, with greater than 90% 5 year survival (7), which is equivalent to similarly managed
Volume 22. Number 5. 1992
Stage I patients with borderline ovarian tumors. The presence of microinvasion may, however, be significant in Stage II and Ill patients where tumor deposits, whether non-invasive autochthonous “metastases” or true metastases from the primary borderline tumor, are demonstrated. The presence of microinvasion in any case should be regarded with caution and should prompt a more thorough pathologic assessment of the primary tumor. The value of adjuvant therapy in Stage II and Ill disease is unclear. Consistent with other reported series, our group of Stage II and Ill patients had a less favourable prognosis than those with Stage I disease, with a 75% IO-year survival and 50% IO-year disease-free survival, despite the use of adjuvant radiation therapy, chemotherapy, or both. It has been our usual policy to recommend radiation therapy in Stage II or Ill disease when microscopic or small macroscopic disease remained following staging laparotomy. There was a tendency to use chemotherapy in Stage Ill disease when residual tumor was more bulky. One of five Stage II patients who underwent abdominopelvic RT relapsed in the liver and lung, and one of four Stage II patients who underwent no further therapy relapsed in the pelvis. Of our nine Stage Ill patients, four underwent abdominopelvic radiation with no recurrences. Two of four patients who underwent chemotherapy with cyclophosphamide and cis-platinum died with progressive disease and one patient has persistent disease. One patient received both chemotherapy and radiation therapy and is still alive without disease at 2.5 years. It is not possible to draw valid conclusions regarding the impact of adjuvant therapy on outcome from this analysis, as the numbers available for analysis were small and events occurring few. It is also not possible, in retrospect, to ensure the similarity of the different treatment groups to allow valid comparisons. Other centers have experienced similar difficulties in advancing sound therapeutic recommendations for Stage II and Ill disease with a wide disparity in opinion concerning the value of adjuvant therapy. Hopkins et al. reported on 68 patients with borderline ovarian tumors, suggesting that radiation therapy prolonged survival in Stage Ill disease, as compared to observation or chemotherapy ( 14). The number of patients available for assessment was small (17 for the three groups), the length of follow-up was not equivalent for the chemotherapy and radiation therapy groups, and the equivalence between the groups of patients subjected to each therapeutic modality is uncertain. Both Nation et al. and O’Quinn et al. reported the ineffectiveness of chemotherapy in management of patients with advanced disease (13,29). Nation et al, reported that all Stage II and Ill patients with macroscopic disease following initial laparotomy had a positive second look laparotomy following chemotherapy with melphalan or cyclophosphamide and cis-platinum (23). However, two cases with minimal disease had a negative second look laparotomy. O’Quinn reported that mel-
Borderline ovarian tumors 0 L. A. MANCHUL
phalan did not eradicate residual tumour in their series of patients who underwent second-look laparotomy (24). Kliman et al. reported that only one patient of 12 who had residual disease following initial laparotomy responded to chemotherapy (19). Fort et al., in a careful analysis of their series of patients with borderline tumors, suggest that adjuvant therapy can eradicate residual disease (11). Of 15 patients who had residual disease at initial laparotomy, 12 were free of disease at second-look laparotomy following therapy consisting of chemotherapy, radiation therapy, or both. There are inherent difficulties in attempting to define treatment recommendations from these retrospective reviews. The patient numbers are too small to allow valid conclusions, the similarity of the patient populations undergoing each form of therapy (or no therapy) is doubtful, the rigor of the initial debulking procedure is not uniform, the pathologic criteria for diagnosis are not uniform, and end points assessed vary from clinical to surgico-pathologic. As well, there are biologic reasons that may explain the apparent lack of effectiveness of adjuvant therapy. Borderline ovarian tumors have a slow doubling time and may remain indolent for several years, rendering them less sensitive to cycle-specific chemotherapeutic agents and radiation. The true nature of the peritoneal implants in presumed advanced stage borderline ovarian tumors is controversial. At the benign end of the spectrum are the gland-like inclusions of endosalpingiosis or atypical endosalpingiosis, which should not be considered metastases from the primary borderline ovarian tumor. In the past, if these were mistakenly identified as metastases, che-
ela/.
873
motherapy or radiation therapy would not be expected to be effective. There have been suggestions that the true borderline peritoneal lesions may represent autochthonous (or simultaneous) extra-ovarian neoplasia and not true metastases and may carry a better prognosis than clearly invasive peritoneal lesions. Bell has suggested that the presence of invasion in peritoneal implants carries a significantly poorer prognosis (3). In our series, however, only 1 of 5 patients with invasive implants has recurred. These patients received adjuvant radiation therapy, chemotherapy, or both. It may be that these invasive lesions are more poorly differentiated with a higher growth fraction and shorter doubling time and might respond better to aggressive therapy than the borderline implants. Maximum debulking of the non-invasive implants should be carried out at initial laparotomy, and should be the mainstay of initial therapy. It is necessary to initiate a national borderline ovarian tumor registry in order to prospectively follow all patients with this diagnosis so that rational therapeutic management recommendations may be made in the future. Preferably a multicenter prospective randomized trial should be undertaken to try to resolve the current controversies. In the interim, optimum therapeutic recommendations must take into account the age of the patient and her desire for fertility, patient willingness to accept a somewhat higher recurrence rate with limited surgery, the ability to carry out frequent follow-up, the ability to resect all apparent disease completely, the extent of disease, and the invasive or non-invasive nature of peritoneal implants.
REFERENCES 1. Aure, J. C.; Hoeg, K.; Kolstad, P. Clinical and histologic studies of ovarian carcinoma: Long term follow-up of 990 cases. Obstet. Gynecol. 37: 1-9; 197 1. 2. Bamhill, D.; Heller, P.; Brzozowski, P.; Advani, H.; Gallup, D.; Park, R. Epithelial ovarian carcinoma of low malignant potential. Obstet. Gynecol. 65: 53-59; 1985. 3. Bell, D. A.; Weinstock, M. A.; Scully, R. E. Peritoneal implants of ovarian serous borderline tumors. Cancer 62: 2212-2222; 1988. 4. Bell, D. A.; Scully, R. E. Ovarian serous borderline tumours with stromal microinvasion: a report of 2 1 cases. Hum. Pathol. 2 1: 397-403; 1990. 5. Bostwick, D. G.; Tazelaar, H. D.; Ballon, S. C.; Hendrickson, M. R.; Kempson, R. L. Ovarian epithelial tumors of borderline malignancy. A clinical and pathologic study of 109 cases. Cancer 58: 2052-2065; 1986. 6. Chambers, J. T.; Merino, M. J.; Kohom, E. I.; Schwartz, P. T. Borderline ovarian tumors. Am. J. Obstet. Gynecol. 159: 1088-1094; 1988. 7. Dembo, A. J. Abdominopelvic radiotherapy in ovarian cancer: a 10 year experience. Cancer 55: 2285-2290; 1985. 8. Dembo, A. J.; Stenwig, A. E.; Berle, E. J.; Bush, R. S.; Kjorstad, K. Prognostic factors in patients with Stage I epithelial ovarian cancer. Obstet. Gynecol. 75: 263-273; 1990. 9. Dembo, A. J.; VanDyk, J.; Japp B.; Bean, H. A.; Beale, F.; Pringle, J. F.; Bush, R. S. Whole abdominal radiation by a
10.
1 I.
12. 13.
14.
15.
16.
17.
moving strip technique for patients with ovarian cancer. Int. J. Radiat. Oncol. Biol. Phys. 5: 1933-1942; 1979. Dgani, R.; Blickstein, I.; Stroham, Z.; Czemobilsky, B.; BenHur, H.; Shani, A.; Katz, Z.; Borenstein, R. Clinical aspects of ovarian tumors of low malignant potential. Eur. J. Obstet. Gynecol. Reprod. Biol. 35: 25 l-258; 1990. Fort, M. G.; Pierce, V. K.; Saigo, P. E.; Hoskins, W. J.; Lewis, J. L. Evidence for the efficacy of adjuvant therapy in epithelial ovarian tumors of low malignant potential. Gynecol. Oncol. 32: 269-272; 1989. Fox, H. The concept of borderline malignancy in ovarian tumors: a reappraisal. Curr. Top. Path. 78: 111-134; 1989. Gershenson, D. M.; Silva, E. G. Serous ovarian tumors of low malignant potential with peritoneal implants. Cancer 65: 578-585; 1990. Hopkins, M. P.; Kumar, N. B.; Morley, G. W. An assessment of pathologic features and treatment modalities in ovarian tumors of low malignant potential. Obstet. Gynecol. 70: 923-929; 1987. International Federation of Gynecology and Obstetrics. Classification and staging of malignant tumors of the female pelvis. Acta. Obstet. Gynecol. Scandinav. 50: l-7; 197 1. Julian, C. G.; Woodruff, J. D. The biologic behaviour of low grade papillary serous carcinoma of the ovary. Obstet. Gynecol. 40: 860-867; 1972. Kaplan, E. L.; Meier, P. Nonparametric estimation from
874
18.
19.
20.
21.
22.
23.
24.
25.
I. J. Radiation Oncology 0 Biology 0 Physics
incomplete observations. J. Am. Stat. Assoc. 53: 457-48 1; 1953. Katzenstein, A. A.; Mazur, M. T.; Morgan, T. E.; Kao, M. S. Proliferative serous tumors of the ovary: histologic features and prognosis. Am. J. Surg. Pathol. 2: 339-355; 1978. Kliman, L.; Rome, R. M.; Fortune, D. W. Low malignant potential tumors of the ovary: a study of 76 cases. Obstet. Gynecol. 68: 338-344; 1986. Kottmeier, H. L. Annual report of the results of treatment in gynecological cancer. Int. Fed. Gynecol. Obstet. Stockholm: Pogo Print; 1979. Lim-Tan, S. K.; Cajigas, H. E.; Scully, R. E. Ovarian cystectomy for serous borderline tumors: a follow-up study of 35 cases. Obstet. Gynecol. 72: 775-781; 1988. McCaughey, W. T. E.; Kirk, M. E.; Lester, W.; Dardick, I. Peritoneal epithelial lesions associated with proliferative serous tumors of ovary. Histopathol. 8: 195-208; 1984. Nation, J. G.; Krepart, G. V. Ovarian carcinoma of low malignant potential: staging and treatment. Am. J. Obstet. Gynecol. 154: 290-293; 1986. O’Quinn, A. G.; Hannigan, E. V. Epithelial ovarian neoplasms of low malignant potential. Gynecol. Oncol. 2 1: 177185; 1985. Russell, P.; Merkur, H. Proliferating ovarian “epithelial” tumors: a clinico-pathologic analysis of 144 cases. Aust. N. Z. J. Obstet. Gynaec. 19: 45-51; 1979.
Volume 22, Number 5, 1992
26. Russel, P.; Bannatyne, P. M.; Solomon, H. J.; Stoddard, L. D.; Tattersall, M. H. N. Multifocal tumorigenesis in the upper female genital tract-implications for staging and management. Int. J. Gynecol. Pathol. 4: 192-210; 1985. 27. Scully, R. E. Common epithelial tumors of borderline malignancy (carcinomas of low malignant potential). Bull. Cancer (Paris) 69: 228-238; 1982. 28. Serov, S. F.; Scully, R. E.; Sobin, L. H. International histological classification of tumors, No. 9. Histological typing of ovarian tumors. Geneva: World Health Organization; 1973: 37-38. 29. Tasker, M.; Langley, F. A. The outlook for women with borderline epithelial tumors of the ovary. Brit. J. Obstet. Gynecol. 92: 969-973; 1985. 30. Tavassoli, F. A. Serous tumor of low malignant potential with early stromal invasion (serous LMP with microinvasion). Mod. Pathol. 1: 407-414; 1988. 31. Taylor, H. C. Malignant and semimalignant tumors of the ovary. Surg. Gynecol. Obstet. 48: 204-230; 1929. 32. Tazelaar, H. D.; Bostwick, D. G.; Ballon, S. C.; Hendrickson, M. R.; Kempson, R. L. Conservative treatment of borderline ovarian tumors. Obstet. Gynecol. 66: 417-422; 1985. 33. Williams, T. J.; Dockerty, M. B. Status of the contralateral ovary in encapsulated low grade malignant tumors of the ovary. Surg. Gynecol. Obstet. 143: 763-766; 1976. 34. Yoonesi, M.; Crickard, K.; Celik, C.; Yoonessi, S. Borderline epithelial tumors of the ovary: ovarian intraepithelial neoplasia. Obstet. Gynecol. Survey. 43: 435-444; 1988.
