Brain Metastases from Epithelial Ovarian Cancer: Overview and Optimal Management

Incidence

Brain metastases from ovarian cancer have traditionally been considered to be a very rare form of relapse since its first description in 1978 by Mayer et al. (5). In their autopsy report of 567 cases between 1973 and 1979, they found an incidence of 0.9%. After this first report, subsequent investigators reported a higher incidence of 1-2.5% (6-8). This rise in incidence has been contributed to several factors. The most plausible one reflects the advances and increased availability of imaging techniques, leading to an earlier and more sensitive diagnosis of brain lesions. Another factor is stated to be the prolonged survival achieved by better therapy regimens including more aggressive cytoreductive surgery and systemic chemotherapy. This results in brain metastases being a variation of the basic disease, a finding that was not possible 20 years ago because of the short time interval between first diagnosis and death. Other theories claim a connection to modern chemotherapy and the blood-brain barrier (BBB), stating either that cancer cells can implant more easily through an impaired BBB after chemotherapy, or that the BBB acts as a safe haven from chemotherapeutic agents.

There are several authors who state that the incidence of CNS relapse in ovarian cancer is rising even more rapidly during the last years and is approaching 12%. However, after a careful review of the current literature, reports of an incidence over 2.5% are the exception and limited only to a few articles which are now mostly outdated. Even though these reports are often cited in current articles as proof of an ongoing sharp rise in incidence, most of the authors who studied larger collectives in the last years reported a stable incidence between 1-2.5% (Table I). Despite the fact that the incidence reported has not changed rapidly in the last decades, 1-2.5% is most certainly still a strong underestimate of the true figure. This is due to the fact that nearly all articles published on this subject are clinical studies rather than autopsy studies. Because brain imaging is not part of the routine follow-up care of ovarian cancer patients, the reported incidence of 1-2.5% is more likely a figure just below the rate of symptomatic brain lesions in ovarian cancer, given that the often unspecific symptoms will not always result in further diagnostic steps.

Site and Appearance of Metastatic Manifestation

Metastases from EOC can present themselves in a variety of phenotypic manifestations (Figure 1). The tumour lesions can be solid, cystic and mixed, whereas mixed solid-cystic seems to be the most common form. At the time of CNS relapse from ovarian cancer, the number of patients with a single brain lesion and multiple lesions is balanced in most reports (2, 8). However, there seems to be a trend towards multiple lesions being the more common form of presentation, as currently reported by several authors (Table II) (7, 9). This observation is in concordance with experience from other malignancies such as lung cancer and melanoma. In these carcinomas, which regularly cause brain metastases, multiple brain lesions occur more frequently than they do in ovarian cancer (10). In terms of regional distribution, the frontal lobe was reported to be the most frequent localisation of metastatic manifestation, followed by the parietal lobe, the temporal lobe and the cerebellum (11). At time of diagnosis of brain metastases, proof of extra cranial disease is found in over half of the patients.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Cranial images of patients with brain metastases from ovarian cancer diagnosed at Medical University of Berlin. A, Axial magnetic resonance (MR) imaging of the brain (T2-weighted) showing a prominent tumor in the left parietal lobe with low density signal surounding the tumor mass representing an edema. B, Axial enhanced CT scan of the brain showing 33 mm right inhomogeneous cerebellar lesion with perifocal edema. C, Frontal MR imaging of the brain (T1-weighted) demontrates a right inhomogeneous cerebellar lesion with surrounding focal edema. D, Axial MR imaging of the brain (T1-weighted) demonstrates different inhomogeneous signal density compatible to multiple metastatic cerebellar lesions.

Prognostic Factors

In the last few decades many investigations have been performed in order to gain a better understanding of ovarian cancer metastatic to the brain. One of the main perspectives of these efforts has been to identify prognostic factors for the survival of these patients. Many different factors have been investigated but partly with conflicting results. Only two factors were able to achieve acceptance by most authors of the current literature. One factor is the presence of extracranial disease at the time of brain relapse, which was shown to have a significantly negative impact on survival. Anupol et al. reported a median survival of 8 months for the occurrence of brain lesions for patients without any evidence of extracranial disease, in comparison to 3 months for cases which did yield that evidence (p=0.005) (2). Cohen et al. affirmed these findings and reported a median survival of 12.2 versus 3.5 months (p=0.001) (7). The other independent factor is the performance status of the patients at the time of recurrent disease. It was shown that a high performance status is likewise related to a higher median survival from the time of central metastatic relapse (31).

For the factors “single” versus “multiple” metastases and the time interval between primary diagnosis and metastatic manifestation, no consistent conclusion can be drawn from the recent literature because the results are not significant in multivariate analysis. A solitary lesion most likely promises a slight advantage over presentation with multiple lesions, as reported by Anupol et al. with a median survival of 7 versus 5 months (p=0.07), but this factor does not withstand multivariate analysis in most reports (2). A short time interval between primary diagnosis and brain relapse was reported to have a negative impact on survival. Cormio et al. for instance found a median survival of 9 months for patients with an interval shorter than 40 months versus a median survival of 16 months for patients with a longer interval (p=0.03) (12). But again, this factor did not pass multivariate analysis. On the other hand, for most of the other factors which have been investigated in the past, like tumour stage, grade, histotype, age at diagnosis of CNS relapse, and site of lesion, it is widely agreed upon in the current literature that these are not related to survival (12).

Therapeutic Options

Therapeutic options for patients with brain metastases from EOC include best supportive care with or without corticosteroids, surgery, radiation (including SRS) and chemotherapy. Death in these patients often results from cranial herniation in the wake of highly increased intracranial pressure. This can be due to large, mostly cystic lesions causing a mass effect, or due to a cerebral oedema surrounding the lesions. A therapeutic approach to this problem is the treatment with corticosteroids, which often results in a satisfactory symptom reduction. However, it was shown by McMeekin et al. that this treatment is of strictly palliative nature and is equivalent to no treatment in terms of mortality, whereas any other form of therapy (radiation, surgery, gamma-knife, chemotherapy), not focusing on the treatment modality, does translate to a prolongation of median survival from 2 to 7 months (4). When compared with each other individually, the conventional neurosurgical approach promises the best results. Even though success of surgical therapy options is often boosted by selection bias, there is evidence that surgery is the most promising individual technique when applicable in the individual patient. Despite that fact, the most commonly used individual therapy strategy is WBRT. This is mostly due to contraindications to surgery, inaccessibility of lesions and the presence of multiple metastases. WBRT applied individually has the potential to decrease neurological symptoms and prolong median survival up to 3-6 months (7, 21). The technique is easily applicable to most patients, even with multiple and otherwise inaccessible lesions, but can cause severe dementia and brain atrophy (in case the patient lives long enough to experience late sequelae).

Even though chemotherapy is an inherent part of standard treatment for ovarian cancer and even extracranial recurrence, its role in the treatment of brain metastases still remains somewhat controversial with regards to the management of EOC. Case reports have shown good results and even complete remission with platinum-based agents such as carboplatin and cisplatin (32, 33). There are very few reports on any experience with chemotherapy as an individual therapeutic strategy. Melichar et al. reported a median survival of 16 months after chemotherapy alone, while other studies observed a median survival of 11-15 months only, even when combined with surgery or radiation (3, 6, 31). Up to this point there is not enough evidence on individually applied chemotherapy, and further trials are needed to support promising results. Also, the focus in future studies should shift towards agents with the potential of reaching a high concentration in the cerebrospinal fluid, because of their ability to cross the BBB. Most agents of platinum-based chemotherapy - the standard choice for ovarian cancer - do not possess this potential (34). Cisplatin for example reaches only a mere concentration of 0-2% compared to the blood level, while agents like topotecan can achieve a concentration of up to 40% (34, 35). Even although it is known that the BBB is already impaired in the metastatic lesion itself, there is experience, from other malignancies frequently causing brain metastases, which favours agents with that potential (36). Although an individual therapy strategy has benefits because of simple application and limited side-effects, many investigators have proven that the therapeutic potential can be greatly increased by the combination of different modalities into a multimodal approach. Cohen and colleagues report a median survival of 5.6 months for surgery or WBRT when these are employed individually, but they also observed a multiplication of the median survival to 23.1 months after a combination of both strategies (7). Great efforts have been undertaken to identify the optimal combination of modalities, but the results remain to a certain extent a matter of dispute. Two strategies have evolved as most promising from studies with large collectives and meta-analysis. One is the combination of surgery and radiation, the other a threefold approach with additional chemotherapy. Pectasides et al. reported a median survival of 21 months after treatment with surgery plus radiotherapy versus 20 months with additional chemotherapy (16). Melichar et al. on the other hand observed a median survival of 17 months after surgery and radiotherapy compared to 20 months after a triple therapy with added chemotherapy (31). Anupol et al. found a median survival of 16 months after radiation plus surgery versus 22 months after added chemotherapy (2). However, this benefit achieved through additional chemotherapy has to be weighed against the potential side effects of this option. Even though chemotherapeutical agents are usually well tolerated, a number of frequent side effects are known to exist. These effects might not pose a great threat, but most chemotherapy-related side-effects result in a decrease of quality of life for the patient. Considering that brain relapse in ovarian cancer is generally a palliative situation, quality of life should be one of the main objectives that has to be taken into account in the decision on the optimal therapy strategy. In the last few years SRS (for instance linear accelerator- or gamma-knife-based) has come into focus as another promising therapy option in brain metastases from ovarian cancer. SRS is mostly applied in patients with no more than 3 cerebral lesions, which are treated with one single high-dose radiation fraction. Lee et al. even observed a remarkable median survival of 29 months after treatment with SRS compared to 6 months after WBRT (37). Few investigators have combined this option with other strategies, despite its promising properties. Even though the findings of Lee et al. might be subject to selection bias, SRS is a very promising option that should be further investigated in future trials and integrated into a multimodal approach.