Ovarian Tumors of Low Malignant Potential

Controversies surrounding ovarian tumors of low malignant potential (LMP) largely arise from gaps in our understanding of their etiology and biology, as well as the lack of effective adjuvant therapy for the small proportion of patients who experience progressive disease. These gaps in knowledge have also led to confusion in nomenclature.[1] These tumors have been called "ovarian tumors of low malignant potential," based on the observation that most are associated with a benign prognosis, "borderline ovarian tumors," based on the faulty premise that they represent an intermediate stage between benign ovarian tumors and frank epithelial carcinoma, and "atypical proliferating tumors," a more recent title introduced to escape from the cancer mindset associated with the terms "borderline" and "malignant." Classification Recent developments in molecular biology suggest that what we call "epithelial ovarian carcinoma" really represents a potpourri of different histologies, each with a differing etiology, molecular biology, prognosis, and response to therapy.[2,3] Similarly, both clinical and laboratory reports suggest that ovarian tumors of LMP represent a "grab bag" of tumors, again with different etiologies, molecular biologies, and prognoses.[4] LMP tumors do not appear to represent an intermediate stage between benign ovarian epithelium and carcinoma.[5] To date, however, the lack of consensus among gynecologic pathologists as to the appropriate nomenclature of the various subtypes has hampered communication among investigators. In addition, in many reports, the diagnoses have been based on light microscopy, with variable use of molecular biology for more specific characterization. Another pathologic controversy centers on the role of "invasive" vs "noninvasive" peritoneal implants. Several retrospective reports suggest that the presence of "invasive" implants conveys a worse prognosis.[6-10] To date, however, gynecologic pathologists have not been able to reach a consensus as to what histopathologic features denote invasion. We would argue that the presence of invasion signifies a primary peritoneal carcinoma, rather than a tumor of LMP. Yet another controversy swirls around the origin of mucinous ovarian tumors associated with pseudomyxoma peritonei. Recent developments suggest that the majority of these cases are associated with a primary tumor in the appendix or elsewhere in the gastrointestinal tract.[11,12] We concur, and would recommend removing this group of intra-abdominal tumors from the larger classification of ovarian LMP tumors. Several groups have also argued that certain ovarian tumors with "micropapillary"features are associated with a higher rate of disease progression.[ 13-15] The term "micropapillary serous carcinoma" has been proposed for this subset of tumors, again removing them from the category of ovarian LMP tumors. To date, however, this delineation has not been universally accepted. Incidence Data on the incidence of LMP tumors should not be considered particularly reliable. Published incidence data stems from both population-based tumor registries and single-institution series. In most cases, the reports from population-based registries have not included central pathology review. We do not know whether changes in awareness of this entity may have affected reported changes in incidence, nor how reliable pathologists are in making this diagnosis. Moreover, cancer registries have not routinely collected information on benign ovarian tumors, a category in which ovarian tumors of LMP are sometimes included. The Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute (NCI), for example, stopped collecting data on LMP tumors in 1997. Comparison between institutional series has been difficult, as individual pathologists have often used different diagnostic criteria for different subtypes of LMP tumors. Ovarian tumors of LMP have been reported among Caucasian, black African, Chinese, Japanese, and Indian populations.[16] Among populations that are predominantly Caucasian, they appear to comprise from 4% to 14% of all ovarian neoplasms. Shi et al have reported that the ratio of borderline epithelial ovarian tumors to epithelial carcinomas in China is 1:5.9, with a similar incidence of mucinous and serous tumors of LMP.[17] Harlow et al survyeyed ovarian tumors in western Washington State between 1975 and 1983.[18] They found that the incidence of ovarian tumors of LMP was 26.2 per million women per year in the white population and 16.5 per million women per year in the nonwhite population. They also noted an increase in the incidence of serous and mucinous tumors of LMP from 7.1 per million women per year to 39 per million women per year between 1975 and 1983. Prognosis
As with incidence, collective data on prognosis is hardly robust, as it is largely based on population-based cancer registries without central pathology review or predominantly retrospective institutional series with inconsistent pathologic criteria between institutions. In addition, some reports lump all LMP patients together, while others split off the most common subtypes (serous and mucinous). Nonetheless, the vast majority of reports indicate that most of these tumors convey a benign prognosis, particularly if patients with mucinous tumors of gastrointestinal origin or primary peritoneal cancers are excluded. Data from the NCI's SEER program, for example, which tracked 2,818 women, showed a 10-year relative survival of 97% for women with stage I ovarian tumors of LMP, 90% for those with stage II, 88% for those with stage III, and 69% for those with stage IV.[19] Similarly, Bjorge et al found 93% 5-year survival for women diagnosed with LMP tumors in Norway between 1970 and 1993.[20] Seidman and Kurman summarized 97 reports, which included 4,129 serous tumors of LMP.[21] They calculated the overall disease-specfic survival rate, with a mean follow-up of 7 years, to be 99% for women with stage I disease and 95% for women with stage III disease. Zanetta and colleagues found disease-free survival to be 99% for 277 women with stage I disease, 96% for 24 women with stage II disease, and 89% for 28 women with stage III disease, with a mean follow-up of 70 months.[22] Epidemiology We observe different patterns in epidemiology between ovarian tumors of LMP and epithelial ovarian cancer.[23] The most common inherited predilections for epitheial ovarian cancer- namely BRCA1 and BRCA2 mutations, and hereditary nonpolyposis colorectal cancer (HNPCC)-do not appear to increase the risk of LMP tumors.[24] The use of fertility drugs appears to increase the risk of LMP tumors but not ovarian carcinomas.[25,26] LMP tumors are more common among younger women, whereas epithelial ovarian tumors are more common among older women.[27] On the other hand, factors that reduce the number of ovulations and increase exposure to progestins-ie, pregnancy, lactation, and the use of oral contraceptives-do decrease the risk of both LMP tumors and epithelial ovarian carcinomas.[28] Preoperative Diagnosis The absence of an effective screening test or algorithm for epithelial ovarian cancer is well known. Similarly, no pattern of tumor marker expression or imaging study has been shown to reliably predict ovarian tumors of LMP or to distinguish between ovarian cancer and LMP tumors. CA-125 can be elevated in serous tumors of LMP but is generally not elevated with mucinous tumors of LMP. Ovarian tumors of LMP share the same presenting symptoms as those of any ovarian enlargement, benign or malignant. In general, the most commly reported symptoms are abdominal discomfort, abdominal pain, and/or a feeling of abdominal enlargement or sensation of a mass. Intraoperative Diagnosis The diagnosis of an ovarian tumor of LMP is best made on permanent section, after inspection of representative slides prepared from an ovarian tumor as well as other intra-abdominal and retroperitoneal samples. Consultation with a pathologist with expertise in the diagnosis of ovarian malignancies may be helpful.[29] Frozen- section diagnosis of an LMP tumor is often unreliable.[30-33] The operating surgeon, therefore, is often told that a tumor is "at least LMP." He or she must then make intraoperative decisions with the knowledge that the final diagnosis may be LMP or invasive cancer. Surgical Management Preoperative counseling remains critically important. One must explore with a patient and her family her desire for future fertility and to retain normal tissue. As noted above, LMP tumors are more common among younger women. The gynecologic surgeon should, therefore, consider conservative management for LMP tumors. A review of treatment patterns in the US found that half of women less than 40 years of age undergo conservative, fertility-sparing surgery when diagnosed with an ovarian tumor of LMP.[19] It may be appropriate in certain instances to plan a second surgical procedure once final pathology is available.[ 34] If an individual with appropriate training in the operative and clinical management of ovarian cancer is not available when an LMP tumor is diagnosed, then a second operation should be scheduled when a gynecologic surgeon with the relevant expertise is available. expertise is available. Nonetheless, certain principles of gynecologic oncology should be observed. First, these patients need a complete exploration of the pelvis and abdomen. Any suspicious lesions should be biopsied or removed. Surgical staging should be performed, with conservation of the uterus and contralateral tube and ovary as indicated.[ 35] Women who undergo ovarian cystectomy are at greater risk for tumor persistence or recurrence than those who undergo oophorectomy. A laparoscopic approach may be appropriate, as long as the operator has expertise in both laparoscopic techniques and the operative management of gynecologic cancers.[36,37] Postoperative Adjuvant Therapy To date, no adjuvant therapy- whether chemotherapy, hormonal therapy, or radiation therapy-has been shown to prolong survival in women with LMP tumors. This observation has held true both in the few prospective trials completed as well as in the many retrospective reports.[16,38,39] The 1994 National Institutes of Health Consensus Statement on Ovarian Cancer concluded that adjuvant chemotherapy was not of benefit to women with ovarian tumors of low malignant potential.[40] As mentioned above, however, these women should undergo complete surgical exploration and staging to rule out the diagnosis of a primary peritoneal cancer. As noted earlier, several retrospective reports suggest that women with "micropapillary carcinoma"-a subset of patients who have been grouped with LMP tumor patients in the past- are at high risk for recurrence and progression.[13-15] Even these micropapillary tumors, however, are relatively indolent; they have limited response to the standard chemotherapeutic agents we use for epithelial ovarian carcinoma. Prognostic Markers Several groups of investigators have sought to identify progostic markers that might identify LMP tumor patients at high risk for disease progression. Markers identified to date in single-institution, retrospective studies include DNA ploidy, p53 overexpression, MIB1, and histomorphometry.[ 41-44] However, none of these have been validated in prospective, multi-institutional studies. Risk of Malignant Transformation The risk of malignant transformation of LMP tumors is controversial. Kurman and Trimble summarized 22 studies, including 953 women with serous LMP tumors, and found a risk of malignant transforation of 0.7%, similar to the risk of malignant transformation of uterine leiomyomata.[45] They excluded from this analysis women with invasive peritoneal implants at time of diagnosis. Crispens et al, however, found that 36 of 49 patients with recurrent LMP tumors had low-grade serous carcinomas at the time of recurrence.[46] This retrospective series ran from 1956 to 1997; many patients did not have definitive surgical staging at the time of initial diagnosis. We suspect that the vast majority of cases in which the malignant transformation of an LMP tumor has been posited actually reflect either a primary peritoneal cancer missed at initial diagnosis or a new ovarian or primary peritoneal cancer. Therapy for Progressive Disease Surgical debulking remains the mainstay of treatment for the few women who experience disease progression. Ovarian tumors of LMP, even those that progress, are generally so indolent that they do not respond to standard chemotherapeutic agents. That said, experimental cytostatic or biologic regimens should be considered in this setting. Surveillance After Conservative Surgery Women who undergo conservative surgery are at risk for persistent or progressive disease. They should be counseled regarding this risk. Zanetta et al have reported on a small series of 24 women who underwent conservative surgery.[47] They found that transvaginal ultrasound was more sensitive than physical examination or CA-125 in diagnosing recurrences. In their series, 18 of 19 women found to have recurrences had an abnormal adnexal mass on ultrasound. Pregnancy After Diagnosis Numerous retrospective reports suggest that women who undergo pregnancy after the diagnosis of an LMP tumor do not increase the risk of tumor progression.[48-52] We should note, however, that these women are at risk for the presence of persistent LMP disease at the time of cesarean section, laparoscopy, or hysterectomy after completion of childbearing. There have also been case reports documenting the use of in vitro fertilization among women with a history of an ovarian tumor of LMP.[53] Hormone Replacement Therapy After Diagnosis We have scant data regarding the safety of hormonal replacement therapy (HRT) after the diagnosis of an LMP tumor. Prospective clinical trials evaluating the safety of HRT are not feasible, given the low baseline risk of progression, the relatively low incidence of LMP tumors, and recent studies demonstrating the limited health benefits of HRT. A woman who has acute menopausal symptoms as well as a diagnosis of an LMP tumor should be considered for HRT. She should be counseled, however, about the absence of data supporting the safety of HRT in this clinical setting. Future Directions for Research We need to go beyond light microscopy to molecular biology to characterize ovarian tumors of LMP. In this instance, classifications will be more useful with further "splitting" rather than "lumping." Once we have defined the molecular biology of these tumors, then we need to identify both prognostic markers, to help identify women at risk for progression, and targets for therapy. Once we have identified therapeutic targets, then we need to develop and test interventions in the appropriate patient populations. Given the small number of women who may benefit from postoperative adjuvant therapy, as well as the small number of women who experience progression of disease, we will need to set up multiinstitutional phase II trials for such evaluations.