By Kate O’Rourke
Should adjuvant chemotherapy be given to all patients undergoing hepatic resection for colorectal liver metastases? At the recent annual meeting of the Society of Surgical Oncology, a spirited pro–con debate was devoted to this topic. (The debaters, it should be noted, were assigned their positions.)
Surgery Alone Is Best
“Surgery alone for colorectal hepatic metastases is enough,” said Nicholas Petrelli, MD, medical director of the Helen F. Graham Cancer Center, Newark, Del. “We do not need chemotherapy. Hepatic resection is the only potentially curable treatment for colorectal liver metastases.” He argued that overall survival (OS) is “the end point that matters,” and because available drugs don’t improve OS in this patient population, individuals should not be exposed to the agents’ toxicities.
Chemotherapeutics and biologics can cause side effects from skin rash to myocardial infarction, neurotoxicity, steatohepatitis, cerebral hemorrhage, tracheal stenosis and bowel perforation, all of which negatively affect quality of life. For example, irinotecan causes steatohepatitis; 5-fluorouracil (5-FU) and irinotecan cause steatosis; oxaliplatin causes sinusoidal dilatation; and bevacizumab (Avastin, Genentech) is associated with liver degeneration and hemorrhage.
“These side effects can [cause] surgical mortality, especially with prolonged chemotherapy, in general six cycles or more,” said Dr. Petrelli, pointing to a study in Annals of Surgery (2006;243:1-7). A study in the Journal of Clinical Oncology also concluded that chemotherapy had a detrimental effect on 90-day mortality (2006;24:2065-2071).
Three prospective randomized trials demonstrate that chemotherapy doesn’t improve OS, said Dr. Petrelli. In a pooled analysis of two randomized trials evaluating complete resection of colorectal liver metastases, investigators did not identify any statistically significant difference in OS or progression-free survival (PFS) in patients who received adjuvant 5-FU-leucovorin chemotherapy postsurgery and those who received surgery alone (J Clin Oncol 2008;26:4906-4911).
In the EORTC (European Organisation for Research and Treatment of Cancer) 40983 trial, 364 patients with colorectal cancer and one to four resectable hepatic metastases were randomized to surgery alone or six cycles of FOLFOX4 (oxaliplatin, 5-FU and leucovorin) followed by surgery and six cycles of postoperative FOLFOX4 (Lancet 2008;371:1007-1016). At three years, patients who received chemotherapy had higher PFS rates (42.4% vs. 33.2%), but this did not translate into improved OS at 8.5 years, according to a study presented at the 2012 American Society of Clinical Oncology annual meeting (abstract 3503). Furthermore, said Dr. Petrelli, patients in the chemotherapy arm had higher rates of postoperative complications, including biliary fistula (8% vs. 4%), hepatic failure (7% vs. 5%), intraabdominal infection (7% vs. 2%) and the need for reoperation (3% vs. 2%) (P=0.04).
“How good are the agents that we use today? They are not good at all,” said Dr. Petrelli.
Even when patients receiving chemotherapy for colorectal liver metastases have a complete response on computed tomography (CT), this does not mean they are cured. In a recent study, 66 lesions disappeared on CT, but macroscopic cancer was seen in 20 of these patients at surgery (J Clin Oncol 2006;24:3939-3945). Sites of 15 initial liver metastases invisible at surgery were resected and 12 had viable tumors. The sites of 31 other initial liver metastases, invisible at surgery, were left in place during surgery, and at one year of follow-up, 23 of the 31 liver metastases considered in complete response had recurred in situ.
The last nail in the coffin, said Dr. Petrelli, is a study evaluating the effect of bevacizumab on future liver remnant hypertrophy after portal vein occlusion before major hepatectomy for colorectal liver metastases (Ann Surg Oncol 2008;16:1553-1559). Patients with a portal vein occlusion had a smaller increase in volume hypertrophy if they received bevacizumab (15% vs. 40%; P<0.05).
With no improvement in OS, said Dr. Petrelli, these agents are hard to justify, especially when you consider the cost. Eight-week courses of FOLFOX (oxaliplatin plus 5-FU; $11,889), FOLFOX plus bevacizumab ($21,033), and FOLFIRI (irinotecan, 5-FU and leucovorin) plus cetuximab ($30,790) are pricey (N Engl J Med 2004;351:317-319).
“Value is cost and quality of care. Chemotherapy increases patient cost and health care costs. It increases hepatic toxicity, systemic toxicity and poor quality of life, said Dr. Petrelli. “There is no improvement in overall survival, leading to no value.”
Chemotherapy Improves Outcomes
Mitchell Posner, MD, Thomas D. Jones Professor and chief of general surgery and surgical oncology, University of Chicago, argued the pro-chemotherapy position. With a five-year survival of 38% after resection of hepatic metastases (Clin Epidemiol 2012;4:283-301), patients with colorectal cancer have a substantial risk for recurrence, and chemotherapy can improve outcomes.
“When you take patients who are initially unresectab
le (R1), treat them with chemotherapy and then resect them, you see the [survival] curve closely approaches those patients who are initially resectable (R0),” said Dr. Posner, pointing to a study in Annals of Surgery (2008;248:626-637). This, he said, demonstrates the efficacy of chemotherapy in stage IV disease.
He then pointed out that in the previously mentioned pooled analysis of randomized trials of adjuvant chemotherapy after hepatic resection, a multivariate analysis revealed a statistically significant improvement in PFS (hazard ratio [HR], 1.39; P=0.026) and OS (HR, 1.39; P=0.046) with chemotherapy. Additionally, overall, there was a trend for improved median PFS (27.9 vs. 18.8 months; P=0.058) and median OS (62.2 vs. 47.3 months; P=0.095).
OS is not the only end point that should matter, argued Dr. Posner. Patients who received chemotherapy in the EORTC 40983 trial had improved PFS (HR, 0.73; P=0.025), an end point that many clinicians value. A clinical trial of 173 patients in France showed that adding chemotherapy after hepatic resection resulted in improved disease-free survival (DFS) rates (33.5% vs. 26.7%; P=0.028) and a trend for improved OS (51.1% vs. 41.9%; P=0.13) (J Clin Oncol 2006;24:4976-4982). In a French study of 1,471 patients with solitary, metachronous colorectal liver metastases, postoperative chemotherapy improved DFS and OS at five years (65% vs. 55%; P<0.01) (Ann Surg 2010;252:774-787).
According to Dr. Posner, clinicians are capable of managing treatment toxicities, and these do not increase mortality and cause only mild morbidity when chemotherapy is limited to a reasonable number of cycles. In patients with colorectal liver metastases who received preoperative chemotherapy before major hepatectomy, morbidity was roughly 10% in patients who received no chemotherapy, 20% in patients who received five cycles of chemotherapy or less, 50% in patients who received six to nine cycles, and more than 60% in those who received 10 cycles or more (Ann Surg 2006;244:897-907; Ann Surg 2008;247:118-124).
In rebuttal remarks, Dr. Petrelli said he completely agreed that using chemotherapy to make unresectable patients resectable was a good idea, but this was not the group of patients they were debating.
“We are talking about patients who present initially with resectable disease, and in those patients, the toxicity is sky high and there really is no benefit in terms of overall survival,” he said.
Dr. Posner, in his rebuttal, called Dr. Petrelli a “flip-flopper.” If OS is actually the only important end point, Dr. Posner wondered, then why is Dr. Petrelli, currently chair of the NSABP Gastrointestinal committee, involved in designing trials with primary end points other than OS? The NSABP C09 trial tested the addition of chemotherapy for hepatic resection using a PFS end point and the NSABP C11 trial, which examined perioperative versus postoperative chemotherapy for resectable hepatic metastases, also used recurrence-free survival as the primary end point, said Dr. Posner.