The incidence of ovarian cancer has remained fairly steady over the past decade, with more than 22,000 newly diagnosed cases predicted in the United States in 2005. Unfortunately, 60% to 70% of women will be diagnosed with advancedstage disease due to lack of specific symptoms as well as effective screening. For these reasons, ovarian cancer is the fifth leading cause of cancerrelated mortality in women, associated with more than 16,000 deaths per year. The current standard of care for women with advanced ovarian cancer is cytoreductive surgery followed by adjuvant platinum-based chemotherapy. The most commonly employed regimen in the United States is carboplatin(Drug information on carboplatin) and paclitaxel(Drug information on paclitaxel), which produces an expected response rate of 70% to 80%.[2,3] Unfortunately, up to 75% of patients with advanced ovarian cancer will develop a recurrence following initial treatment and require subsequent therapy.
A paradigm shift has occurred over the past decade, with many physicians now viewing the treatment of recurrent epithelial ovarian cancer as management of a chronic disease. Certainly there are patients who do well from the start and experience long-term cures or long disease-free intervals. Additionally, there is a small group who progress during first-line chemotherapy and have a relatively short overall survival. However, the majority of women will develop recurrent ovarian cancer at varying intervals, with symptoms that wax and wane in response to repetitive cycles of various chemotherapeutic agents. During the course of their disease, these women will display a variety of symptoms related to disease in their abdomen, as well as more distant sites.
In this article, we will review the literature regarding the management of recurrent disease. Additionally, we will share our own approach to therapeutic decisions based on initial tu- mor characteristics and treatment, symptoms of disease, and toxicities.
Diagnosis of Recurrent Disease
Symptoms of recurrent ovarian cancer are typically vague abdominal complaints, much like those found prior to initial diagnosis. However, most women will be aware of recurrence prior to symptoms due to clinical surveillance schemas that include physical exam, serial CA-125 measurement, and computed tomography (CT) scans. One area of controversy has been the treatment of women with asymptomatic recurrences, either biochemical or based on clinically identifiable tumor. Although most physicians will treat the latter even in asymptomatic women, many physicians struggle with the best time to institute therapy in the setting of an isolated increase in CA-125 levels. Certainly as patients become more informed, they may wish to proceed with aggressive treatment of their disease in the hope that chemotherapy will be able to completely eradicate recurrent small-volume disease.
The Gynecologic Oncology Group (GOG) is currently conducting a randomized trial evaluating tamoxifen vs thalidomide(Drug information on thalidomide) (Thalomid) in women with biochemical recurrences of ovarian cancer. Although no data are available to date, this is an attractive approach for many clinicians as these drugs are generally well tolerated with limited toxicities. This strategy has the potential to extend the platinumfree interval, which may be beneficial in improving response to subsequent platinum-based regimens.
Management Based on Time to Recurrence
Over the past 2 decades, we have explored many new therapeutic options for the treatment of patients with ovarian cancer. However, two major conceptual changes have greatly influenced our approach to recurrent disease. First, we now tend to treat lower-stage or less residual disease more aggressively in the adjuvant setting, believing that the patient with more favorable disease will benefit more from improved survival rather than less aggressive therapy. Recent data have confirmed that there is a survival advantage in women with advanced disease who are cytoreduced to no gross visible disease compared to those with residual disease of greater than 1 cm, when all are treated with combination chemotherapy.
The other important change has been the concept of recurrent ovarian cancer as a chronic disease. Historically, when a patient developed a recurrence, it was believed that she must be resistant to the last chemotherapy regimen received. Therefore, a different drug with a different mechanism of action was chosen for treatment of the recurrence. The result was often rapid movement through a variety of therapies with their collective toxicities resulting in poor long-term care for the patient. In some patients this may not have made a difference, but for many, opportunities were missed for long-term stabilization of disease with good tolerance to drugs that have limited cumulative toxicity.
Current thinking divides patients into groups based on time from completion of initial therapy. Patients developing recurrent disease less than 6 months after completion of primary platinumbased therapy are typically referred to as platinum-resistant or platinum-refractory.[ 6,7] It is safe to assume that patients in this group have disease that is highly resistant to the therapy already received. In this population, an assessment should be made as to the overall extent of disease, the identification of clear measurable parameters, and then a choice made generally from systemic therapeutic options that are non-platinum-cross-resistant. It is very important to enter many of these patients on clinical trials to help identify new therapeutic strategies.
Patients diagnosed with recurrence after a 1-year disease-free interval are typically referred to as platinum-sensitive and are treated with one or both of the two initial drugs. Data evaluating treatment of primary recurrence with platinum-based regimens have shown that a longer disease-free interval correlates with improved response when rechallenging with platinum. A recent study by Markman et al demonstrated response rates of 33% for those less than 12 months from prior therapy, 55% for those 12 to 18 months from prior therapy, and 75% for those more than 18 months from prior therapy. Objective response rates in women with platinumsensitive ovarian cancer are expected to be 30% to 50% when treated with single-agent platinum, compared to 50% to 80% when treated with combination platinum/paclitaxel.[9-13]
The International Collaborative Ovarian Neoplasm (ICON4) trial compared retreatment with platinum and paclitaxel vs conventional platinumbased regimens in women with a treatment-free interval of more than 6 months. Hazard ratios of 0.82 for 2-year survival (P = .02) and 0.76 for progression-free survival (P = .0004) were noted in women treated with the platinum/paclitaxel regimen. There has been criticism of this study regarding prior exposure to paclitaxel, since the subgroup of patients that appeared to derive the most benefit from combination therapy were those patients who did not receive paclitaxel as part of their primary therapy. However, a combination regimen appears to be a reasonable choice in patients with a treatment-free interval longer than 12 months, but may be of little added benefit in patients recurring 6 to 11 months after platinum-based therapy.
In our practice, patients who experience treatment-free intervals of greater than 6 months are treated in much the same way as noted above. Our choice for initial therapy of recurrent disease is generally one of the drugs used for initial therapy, ie, platinum or paclitaxel. Although patients recurring in the 6-month to 1-year interval probably have relatively resistant disease, we still tend to re-treat with one of the first-line drugs for second-line therapy. The vast majority of patients are given carboplatin because it tends to be a better tolerated drug than paclitaxel and the patient does not lose her hair for a second time. In this setting, the drugs are generally started and continued until there is clear evidence of tumor progression or until an unacceptable toxicity limits the therapy. Often, patients can then be switched to the other drug used for primary therapy, they again respond, and they can be maintained for several additional months.
'Highly' Platinum-Sensitive Disease
Patients who recur 3 or 4 years out from primary therapy are generally assessed with an eye to retreatment as if in primary treatment mode. Workup of these patients should include assessment of the extent of disease, followed by debulking in some cases and then primary multiagent chemotherapy.
Various groups have evaluated the role of secondary cytoreduction. Recent studies have shown that a survival benefit can be achieved when patients have more than a 24-month disease-free interval and are cytoreduced to microscopic disease.[12,13] Eisenkop and colleagues also showed a 25-month survival advantage in patients achieving optimal secondary cytoreduction to microscopic disease prior to salvage chemotherapy (P = .005) and a 2-month survival advantage for women whose largest tumor measured < 10 cm (P = .04). We tend to use secondary cytoreduction in women with a disease-free interval longer than 24 months with few masses present on exam and/or CT scan. We also rechallenge these pa- tients with front-line chemotherapy, for which expected response rates are greater than 70%.
Beyond Platinum and Paclitaxel
Today, many options exist for the treatment of recurrent ovarian cancer. Although platinum and paclitaxel remain the standards for primary therapy for ovarian cancer in the United States, many other compounds have been shown to possess activity against the disease. A selected list of these compounds can be found in Table 1.[10,15-34] When considering which of these agents to use when platinum and/or paclitaxel has been used as first-, second-, or even thirdline chemotherapy, we tend to balance the chances of a good response with the possibility of new or additive toxicities. Additionally, many of the trials referred to in Table 1 had a significant number of patients who exhibited stable disease as their best response to therapy. The goal of disease stabilization with limited toxicity may be a reasonable approach in certain patients who have minimal symptoms from their disease.
An excellent example of this approach is a heavily pretreated patient who we have begun treating with liposomal doxorubicin(Drug information on doxorubicin) (Doxil), with good results. Previous studies have shown response rates ranging from 10% to 26%,[22,23,35] but an additional 30% to 40% of patients may have stable disease while on therapy. Combining this drug's excellent tolerability and safety profile with its high response and disease stabilization rates, patients can continue on therapy for prolonged periods with monthly infusions of 20 to 40 mg/m2.[36,37] Using this approach, we have had several patients who have been on liposomal doxorubicin for more than 24 months. Controversy exists regarding an appropriate time to stop such therapy, but we believe that choosing agents with few or tolerable toxicities that can be given on a convenient schedule over a prolonged period of time is an acceptable way to manage recurrent disease. Additionally, we have found that most patients are unwilling to stop taking the agent once clinical tumor progression has ceased, particularly if they remain asymptomatic.
Several other options exist for treatment in recurrent ovarian cancer including the topoisomerase inhibitors etoposide(Drug information on etoposide) and topotecan (Hycamtin), the microtubule stabilizer docetaxel(Drug information on docetaxel) (Taxotere), and the nucleoside analog gemcitabine(Drug information on gemcitabine) (Gemzar). Although response rates to these and other drugs range from 10% to 47%, it is difficult to tell which drug(s) will offer which patient the best response.[16-20,29,30,38] Importantly, none of these drugs are cross-resistant to platinum, and therefore, they have the potential to induce responses or disease stabilization even in patients that have been refractory to platinum-based therapy.
As discussed previously, the ICON4 trial showed progression-free and overall survival benefits for patients treated with combination platinumbased regimens for platinum-sensitive disease. While this is the first time that this finding has been identified in a prospective randomized fashion, other authors have shown similar benefits with other combinations.
One of the combinations receiving significant attention is that of gemcitabine and platinum. Gemcitabine is a nucleoside analog that inhibits DNA synthesis by causing premature nucleoside substitution-based chain termination as well as other mechanisms. This combination has been shown to possess synergistic activity in vitro that may be at least partly due to reversal of ERCC1-induced platinum resistance. Several trials have shown objective response rates ranging from 55% to 64% when carboplatin and gemcitabine are used both in platinum-sensitive and platinum-resistant populations, and this combination has been associated with quicker palliation of symptoms and longer time to progression.[33,34,40-42]
However, this improvement in activity comes at the cost of toxicity, with as many as 78% of patients exhibiting grade 3/4 bone marrow toxicity. Heavily pretreated patients have more frequent grade 3/4 toxicities and often require growth factor support and/or blood transfusions. Nevertheless, due to its high response rate, we are using this palliative regimen, particularly in highly symptomatic patients with gastrointestinal complaints.
Although most clinicians tend to consider mostly cytotoxic therapy for recurrent ovarian cancer, several hormonal options have shown benefit in clinical trials. Tamoxifen(Drug information on tamoxifen), megestrol(Drug information on megestrol) acetate, and some of the more recently applied aromatase inhibitors are less likely to produce complete responses but have the potential for prolonged disease stabilization. Response rates to tamoxifen have ranged from 13% to 18%, with an additional 30% to 40% achieving stable disease in a platinum-resistant setting.[27,28]
It is probably worthwhile to check the primary tumor for estrogen/progesterone receptors, although one can with great reliability assume that mucinous tumors and most clear cell tumors will not show significant receptor expression and therefore tend to respond poorly to hormonal therapy. We feel that the patient who has progressive disease after two or three prior courses of chemotherapy and who is fortunate enough to be feeling well is an excellent candidate for hormonal therapy. Again, in a patient with no symptoms, it may be appropriate to simply stabilize disease. Under these circumstances, most patients would be happy to begin a nontoxic therapy in hopes of preventing progressive disease.
Management Based on Location of Recurrence
As a rule, ovarian cancer tends to recur in the abdomen. Patients with recurrence in the abdomen may present similar to patients with primary disease, with abdominal aching and symptoms of increased gas, bloating, or nausea. When these signs initially appear, we usually first place the patient on a low-residue diet (Table 2) to try to forestall any obstructive symptoms. This approach often allows us to initiate the next course of chemotherapy, frequently producing enough of a response to avoid complete obstruction. The initiation of a low-residue diet also puts a degree of control back in the patient's hands, as she learns what she can do to prevent a more complete obstruction.
A GI series or CT scan is often done to first identify and then to try to localize the site of obstruction. CT has the added advantage of being able to assess the volume and site(s) of disease. When deciding on various modes of therapy, it is extremely important to integrate the acute event with overall disease status. The earlier an obstruction occurs in a patient's disease course or the more distal the obstruction is, the more likely a surgical intervention will be considered.
Malignant Bowel Obstruction
Options for management of malignant bowel obstruction include medical therapy, surgical bypass or excision of the obstruction, intraluminal stenting, and, occasionally, localized radiation therapy. All of these options must be considered in light of the patient's age and overall medical status. For example, if the patient has received a number of therapies and the recurrence is very proximal, it may be that a decompressing gastrostomy is the only reasonable choice. In properly chosen patients, this may be just as successful at palliating symptoms with fewer complications and may offer the patient equal quality of life and survival when compared to open surgical procedures such as intestinal bypass or ostomy formation. Placement of a gastrostomy tube can occasionally be accompanied by a liquid diet. In the obstructed patient who has been documented to have a very slowly progressive tumor that is not causing other symptoms, the use of home parenteral nutrition may be appropriate.
- • Medical Therapy—Malignant bowel obstructions are typically managed initially by nasogastric tube decompression and intravenous fluids to treat nausea, vomiting, and dehydration. We often add intravenous steroids to this regimen, both for their theoretical anti-inflammatory effect on the bowel wall and their antiemetic properties. Although the beneficial effects of steroids have not been proven in randomized controlled trials, a recent meta-analysis appears to show a benefit of steroids in this setting. We have been successful in managing patients initially with intravenous steroids and then converting them to lower doses, which they continue along with a liquid or low-residue diet as outpatients.
In addition to nausea and vomiting, many patients experience generalized abdominal pain as well as colic. When medical therapy fails to relieve an obstruction, the options for surgical intervention vs palliative care should be discussed with the patient. Again, counseling must be undertaken keeping in mind the overall health and disease status of the patient. When the decision is made to proceed with palliative care, a variety of medications can be used to abate the patient's symptoms (Table 3).
- •Invasive Procedures—Attempts at exploration may be appropriate even in patients with an obstruction that is very proximal, but these occasions will be limited. The risks are high and attempts at surgical palliation can result in little more than hospitalizing the patient for a significant percentage of her remaining days. The problem comes not necessarily from the proximal obstruction, but basically from the protection it provides for obstructions that are more distal. At surgery, therefore, one is faced with the task of determining what is obstructed and what is not, among multiple loops of adherent tumor-coated bowel. This becomes a rather impossible task, especially in very advanced recurrent disease. Although bypass of a proximal obstruction may indeed be successful, the back pressure created by the next obstruction can be sufficient to rupture the newly formed anastomosis. If the first obstruction is more than 180 to 200 cm from the ligament of Treitz, an ileostomy may be formed which will bypass the obstruction and still allow enough absorptive surface for the maintenance of nutrition. The more distal the first obstruction, the more options that become available, such as an ileotransverse anastomosis or ostomy formation.
When distal small bowel obstructions are identified on imaging, we use a limited gastrograffin enema to first document that there is no obstruction between the rectum and the transverse colon. If there is a satisfactory amount of small intestine above the obstruction, a small incision can often be made so that the most distal dilated loop can be quickly identified, and a stapled anastomosis to the transverse colon is created followed by abdominal closure. This approach works well because of careful preoperative planning.
Under most circumstances, extensive abdominal exploration is avoided and a secure anastomosis is almost always assured because of the absence of distal obstruction. This is one of the special circumstances nicely managed by a long intestinal tube. If caught early in the course of a patient's symptoms, a long intestinal tube could be manipulated into place and passed to the level of the obstruction. Then the upper intestinal series can be done by pushing the contrast through the tube. This is a much easier process for the patient, requiring less contrast material. On entering the abdomen, the surgeon can locate the end of the tube as the point of the small intestine to use for the anastomosis.
- •Stenting—Intraluminal stents have been effective in a limited number of small series. To date, expandable coated metal stents have been used successfully in palliating up to 75% of patients presenting with malignant bowel obstructions. Both colonic and gastroduodenal stents have been used in the management of malignant bowel obstructions in ovarian cancer patients.[46,47] Problems with the placement of intraluminal stents include reocclusion and stent migration, but effective palliation of symptoms can usually be achieved with this less invasive procedure.
- •Radiotherapy—Localized areas of colonic obstruction can also be palliated with radiation therapy. However, symptom control is usually of very brief duration and is typically accompanied by morbidity due to the adherent bowel around the obstruction. This modality is generally most successful in treating malignant obstructions of the rectum or rectosigmoid.