A new model of ovarian carcinogenesis may influence early detection strategies

Ovarian cancers, more than 95% epithelial in origin, represent the most lethal site of gynecologic malignancy in the United States, with an estimated 22,430 new cases and 15,280 deaths per year according to 2007 American Cancer Society Statistics.1 Although an estimated 10-25% improvement in median survival for patients with advanced (stage III and stage IV) disease has been observed since the mid-1980s and attributed to advances in surgical management, the incorporation of taxanes into standard platinum based primary chemotherapy,2, 3 and the discovery of several other active nonplatinum cytotoxic agents4; nevertheless, long-term survival rates have been disappointing. Ultimately traditional therapy has been limited by drug resistance and lack of specificity to mechanisms of disease progression. Despite the established benefit of oral contraceptives and prophylactic surgery in primary prevention, the US incidence has remained quite flat since the mid-1990s.1

For patients with a diagnosis of epithelial ovarian cancer, extent of disease, defined by surgical-pathologic stage, and extent of residual disease at the conclusion of primary surgery have been considered the 2 most important prognostic variables. Advances in early detection strategies to identify preclinical neoplasia should theoretically have a major impact on mortality; however, as Kurman and colleagues point out in this issue’s article, “Early detection and treatment of ovarian cancer: Shifting from early stage to minimal volume of disease based on a new model of carcinogenesis,” the assumption that this could be accomplished through identification of early-stage tumors that would otherwise progress in stage is erroneous, given evolving molecular, histopathologic, and anatomic evidence. On this basis the authors propose a new model for ovarian carcinogenesis, which stratifies epithelial ovarian cancers into type I and II tumors.

Type I tumors are described as relatively uncommon and are generally confined to the ovary at the time of diagnosis. They are characterized by indolent growth, identifiable precursors, mutations in distinct oncogenes and tumor suppressor genes, and overall genetic stability. These tumors include borderline and low grade, carcinomas with serous, mucinous, and endometrioid patterns. Type II tumors, on the other hand, represent the vast majority of epithelial ovarian cancers and are usually metastatic early in their development. They are characterized by an explosive onset and rapid growth rate, mutations in the tumor suppressor gene p53 and genetic instability. This group includes high-grade serous and poorly differentiated carcinomas of ovarian and extraovarian origin.

The data presented by Kurman and colleagues reaffirm the heterogeneity of epithelial ovarian cancers on the molecular level, most likely resulting from genetic instability in the more common, type II tumors. This presents a tremendous challenge with respect to molecular screening approaches. In light of such observations, promising multimarker serologic panels have been developed through genomic discovery,5 which will likely be tested in prospective population based screening trials.

The proposed model of Kurman and colleagues provides us with a useful foundation on which to base early detection approaches. From the public health standpoint, type II tumors should be the focus of a population based effort because they represent the majority of tumors, and the consequences of disease would likely justify the costs of screening. The authors point out, however, that based on clinical behavior and biologic characteristics, these tumors are unlikely to be identified in a preclinical state or, once invasive, confined to 1 organ of origin. The authors conclude that the aim of screening for ovarian cancer should be to identify “low-volume” rather than “early-stage” disease. Disease detection prior to clinical presentation would likely result in a relatively healthier population women affected by the disease, enhance the safety and completeness of surgical resection, and optimize the tolerability and response to systemic or regional therapeutics, thus improving prognosis. Such screening methods should address not only the ovaries as primary sites but also should take into account the zones of spread, including the pelvis and upper abdomen.

There is already evidence to support this approach. Perhaps the best example is the prospective early detection study by Jacobs et al6 in which more than 20,000 asymptomatic women aged 45 years or older were randomized to a screening program vs observation. The screening program consisted of 3 annual serum CA 125 assessments, pelvic ultrasound in response to a CA 125 elevation, and referral to a gynecologic surgeon for a suspicious pelvic ultrasound. Both the screening and observation cohorts were followed for a median of 7 years. Although the total number of epithelial ovarian cancers diagnosed within the entire period of the study was not significantly different between the cohorts (16 in the screening group, 20 in the control group), the median survival of women with ovarian cancer in the screening group was significantly prolonged at 73 months, compared with 42 months for women with ovarian cancer in the observation group.

Although lead time bias may account for some of this difference, based on the biologic behavior of these tumors, it is just as likely that identification of stage III or IV tumors before the onset of symptoms may have in fact improved the outcomes for patients who would have otherwise been identified with more burdensome disease and deteriorating health. Larger-scale randomized trials are ongoing to address this issue.

Along the same line, it is also conceivable that earlier recognition of advanced disease by history and physical examination alone might have a beneficial effect on survival and cure rate. In 2007 the Gynecologic Cancer Foundation, Society of Gynecologic Oncologists, and American Cancer Society published an Ovarian Cancer Symptoms Consensus Statement regarding symptoms commonly preceding the diagnosis of ovarian cancer, including bloating, pelvic or abdominal pain, dysphagia, early satiety, and urinary urgency or frequency.7 This statement has been endorsed by more than 30 ovarian cancer alliances and is intended to raise awareness among the general population and primary health care professionals, such that these symptoms would initiate more immediate formal diagnostic evaluations.