ABSTRACT: Historically, liver-related metastases associated with melanoma or breast cancer have portended a poor prognosis. Many affected patients are not considered for surgical resection based on the extent and multifocal nature of their disease. For this patient population, treatment includes systemic and/or regional therapy, local destruction (ablation/radiation), and embolization. Despite the best therapeutic regimens, prognosis remains poor. Advances in surgical technique and postoperative care have led to a resurgence in the use of metastasectomy, most notably seen in patients with colorectal-related liver metastases. With the potential for therapeutic durability and a small chance of cure, surgical resection may offer improved survival compared to other therapeutic modalities. This review summarizes the existing literature that addresses the topic of metastasectomy in patients with melanoma and breast cancer.
Advances in perioperative planning, surgical technique, and postoperative care over the past two decades have allowed the indications for hepatic resection to be potentially broadened to include histologies and/or disease burdens previously felt to be contraindications for this treatment modality. This trend has been seen most notably with respect to resection of colorectal-related liver metastases.[1,2] Historically, colorectal liver metastases have been treated with palliative chemotherapy. The improvements seen with surgical resection have made this the gold standard for treatment; expected 5-year survival rates are 50% to 65%, and resection offers the only potential for cure. Hepatic resection for neuroendocrine-related liver metastases is likewise now accepted as appropriate. However, while retrospective reports have demonstrated long-term survival in some highly selected patients with other histologies, including melanoma and breast cancer, the application of hepatectomy in these patients remains controversial.
There are a number of arguments against the use of hepatectomy in the management of metastatic disease. First and foremost is the lack of randomized data to support its utilization, even in patients with neuroendocrine and colorectal histologies. Hepatectomy is associated with severe morbidity rates—20% to 30%—and with 90-day mortality rates of 1% to 5%. Disease recurs in the overwhelming majority of patients following hepatectomy, and the prospect of recurrence exceeds 80% even in the most favorable of histologies. However, similar statements can be made to argue against surgical resection of pancreatic cancer, gastric cancer, esophageal cancer, and primary liver cancer. Ultimately, patients and their oncologists are drawn to the small to modest possibilities of cure that are associated with hepatectomy and other complicated oncologic resections in the face of what would be a near-zero possibility of cure without resection. For young, healthy patients, it is a relatively easy decision, although for older and/or infirm patients the decisions are more complex.
For the vast majority of patients with liver metastases who are not candidates for surgical resection, treatment options include systemic and/or regional therapy, local destruction (ablation/radiation), and embolization. Systemic therapy options have evolved, and there are now a multitude of agents available for metastatic disease. However, response rates have been low and the effect on overall survival is still unclear. Regional therapy, including the use of isolated hepatic artery treatment, has been shown to achieve partial remission, although it has no significant effect on long-term survival. Lastly, embolization techniques have been promising, with reports showing better survival rates than those associated with the best chemotherapy, with minimal toxicity. Still, even this therapy is palliative at best; surgical resection offers the only real opportunity for cure.
The aim of this review is to present an analysis of the existing literature that addresses the topic of metastasectomy in patients with melanoma and breast cancer. Despite improvements in systemic chemotherapy for these diseases, the prognosis for metastasized disease in both cases remains poor.[8,9] With proper patient selection, the added modality of surgical resection may allow for improved survival and, as with colorectal and neuroendocrine cancers, may be the only possibility of cure.
With an estimated 68,130 cases diagnosed in 2010, cutaneous melanoma is approximately 25 times more common than ocular melanoma. While each form has the potential to metastasize to the liver, the rates at which they do so are distinctly different. The difference in the metastatic pattern is likely a direct result of the differing tumor biology of the two cancers. Up to 95% of patients with stage IV ocular melanoma develop hepatic metastasis. In contrast, this occurs in only 10% to 20% of patients with stage IV cutaneous melanoma, with the majority of metastases occurring primarily in distant lymph nodes or soft tissue.[18,19] Ocular melanoma, on the other hand, exhibits latency, with metastases appearing many years following removal of the primary disease. The liver is the sole site of metastasis in 60% to 80% of cases, with characteristic intrahepatic multiplicity.[21,22] Historically, prognosis is poor in patients with hepatic metastases of ocular melanoma, with life expectancy less than 6 months.
Overview of nonsurgical approaches
The presentation of stage IV melanoma with isolated liver metastasis has a historically poor prognosis, with a median survival of 4 to 6 months. Therapy for liver metastasis is limited: options include systemic and regional therapy, local destruction (ablation/radiation), and surgical resection.
Systemic options have historically included dacarbazine(Drug information on dacarbazine), temozolomide(Drug information on temozolomide) (Temodar), interleukin-2, paclitaxel, cisplatin, and carboplatin(Drug information on carboplatin).[11,12] Response rates have been low, and the effect on overall survival unclear at best. Within the past year, highly publicized reports presented at the annual meeting of the American Society of Clinical Oncology (ASCO) have suggested improvements in overall survival with two novel targeted agents—vemurafenib (PLX4032; Zelboraf) and ipilimumab (Yervoy).[8,9] In one trial (BRIM-3), vemurafenib was shown to improve response and survival (overall and progression-free) as a single agent compared with standard treatment (dacarbazine). The results were felt to be so striking that the trial was halted early (with a median follow-up of just over 3 months) to allow cross-over of the control arm. The efficacy of ipilimumab was confirmed in a second trial, also reported at ASCO's 2011 annual meeting.
Regional therapies, including isolated hepatic perfusion, have shown promising results in the treatment of unresectable hepatic metastases. Historically, hepatic perfusion has been associated with significant morbidity due to the complex nature of the technique; however, recent advances using percutaneous techniques have allowed for a resurgence in the evaluation of this method. Hepatic artery treatment with fotemustine(Drug information on fotemustine) has been promising, producing a 40% response rate and overall survival of 14 months. Phase I studies using hepatic arterial melphalan(Drug information on melphalan) infusion have shown the ability to achieve partial remission, although no significant effect on long-term survival has been demonstrated. Despite the benefits of this modality, no durable response has been achieved. Surgical resection in patients with isolated hepatic metastasis, while rarely performed, appears to offer the only real opportunity for cure.
Hepatectomy for metastatic melanoma: a review of the literature
The existing literature on the topic of hepatectomy for metastatic melanoma consists of small case reports of what appear to be highly selected patients. In a notable review of the prospectively maintained melanoma database at the John Wayne Cancer Institute, Rose et al analyzed 1750 patients who underwent treatment for melanoma metastases to the liver between 1971 and 1999. Meticulous selection of patients resulted in only 34 patients (2%) who met criteria for attempted surgical resection. Ten patients were deemed not resectable due to extent of disease at the time of surgery. Of the 24 remaining, 18 (75%) were rendered free of disease following resection. The median overall survival for the 18 patients with curative resection was 38 months. Median overall survival and median disease-free survival for the 24 patients with attempted resection were 28 months and 12 months, respectively, compared with a median survival of 4 months in the 10 patients who had exploration only. Prolonged disease-free interval, complete tumor resection, and lower tumor burden were all associated with improved outcomes.
In a report by Adam et al consisting of 1452 patients who underwent hepatic resection for noncolorectal, nonendocrine liver metastases, a subset of 44 patients with cutaneous melanoma liver metastases was described. The 5-year survival was 22%. Analysis of factors associated with poor prognosis revealed that these include age greater than 60 years, disease-free interval less than 12 months, extrahepatic metastases, R2 resection, and major hepatectomy. Similarly, Pawlik et al described their experience of 40 patients with hepatic melanoma, the primary tumor being cutaneous in 24 patients and ocular in 16 patients. Median disease-free interval was 63 months. Recurrence developed in 75% of patients after resection, with a median time to recurrence of 8.8 months for the ocular subset and 4.7 months for the cutaneous subset. The 5-year survival for patients with primary ocular melanoma was 20.5%; there were no 5-year survivors in the cutaneous melanoma group (Figure 1).
Much as with cutaneous melanoma, multiple reports have documented the benefit of surgical resection for hepatic metastases of ocular melanoma.[26-28] Rivoire et al followed 602 patients between 1983 and 1996 who were treated for ocular melanoma (Table 1). Using ultrasound, patients were evaluated to assess for development of hepatic metastases. Liver metastases developed in 63 patients (10.5%) following ocular treatment, with a median disease-free interval of 29 months. Twenty-eight patients (4.7%) met criteria for attempted surgical resection, with 14 patients undergoing R0 surgery. Median survival was 25 months for patients with R0 surgery, 16 months for R2 resection (14 patients), and 11 months for the 35 patients treated with palliative chemotherapy. When the data were subjected to univariate analysis, age < 70 years and completeness of surgical resection (R0) were associated with improved outcomes.
A series by Mariani et al retrospectively reviewed 3873 patients with ocular melanoma between 1991 and 2007. Hepatic metastases developed in 798 patients (21%), with a median time interval between ocular tumor diagnosis and treatment of hepatic metastases of 68 months. Of the patients who developed liver metastases, 255 (6.6%) underwent attempted surgical resection. R0, R1, and R2 surgery was performed in 76, 22, and 157 patients, respectively. Median overall survival was 27 months for R0 surgery, 17 months for R1 surgery, and 11 months for R2 surgery. Multivariate analysis revealed that completeness of surgical resection (R0 surgery), number of metastases resected (< 4), and disease-free interval following primary tumor diagnosis > 24 months were associated with improved outcomes.
More recently, Frenkel et al evaluated 558 patients with ocular melanoma, of whom 74 (13%) developed metastases. The median disease-free interval was 35 months from initial diagnosis. Only 35 patients underwent hepatectomy. Median survival for patients with fewer than 5 metastases was 55.2 months. Patients with > 5 metastases or miliary metastases had a median survival of 25.6 months. The survival of patients with an R0 resection was four times longer than the survival of those with an R1 or R2 resection (Figure 2).
Selection of patients for resection
Selection of the optimal patient for resection of melanoma liver metastases is based on multiple factors. A long disease-free interval, the ability to achieve an R0 resection, low tumor multiplicity, and no or limited extrahepatic disease increase the durability of surgical resection. This point was more eloquently made by Coit et al, who in referring to metastatic melanoma stated: “regardless of the extent of the operative procedure, resection of metastases in patients whose disease recurs early after the treatment of the primary tumor, in those who present with multiple lesions, and in those who present with disease that cannot be completely resected will only rarely be associated with subsequent long-term survival.”
It would be easy to be nihilistic if not for the real prospect of long-term survival seen in practices where aggressive approaches are taken. Although the analogy is imperfect, the concept of “variable interval reinforcements” is likely operative in such practices, with the occasional successes reinforcing enthusiasm for aggressive treatments. Figure 3 depicts one example of such a success—a 48-year-old male who presented to our institution for the treatment of metastatic melanoma. The patient initially presented with fatigue and anemia (hemoglobin level of 3.5 g/dL). Colonoscopy revealed an ileal lesion with biopsy-proven melanoma of unknown primary. Evaluation included a computed tomography (CT)/positron emission tomography (PET) scan, which identified an additional lesion in the liver. The patient underwent a small bowel resection from which he recovered unremarkably; he then started interleukin-2 therapy. Two years later he developed rapid progression of disease isolated to the right liver and was at that time referred for possible liver resection. The patient underwent an uneventful right hepatectomy and is currently disease free 6 years later.