For resectable liver mets: Preop chemotherapy or not?

February 1, 2008

When colorectal cancer metastasizes to the liver, hepatic resection can offer a survival benefit and even a "cure" in a fraction of patients. Five-year overall survival in some recent series approaches 60%. But the role of neoadjuvant chemotherapy in this group of patients has not been well established

ORLANDO—When colorectal cancer metastasizes to the liver, hepatic resection can offer a survival benefit and even a "cure" in a fraction of patients. Five-year overall survival in some recent series approaches 60%. But the role of neoadjuvant chemotherapy in this group of patients has not been well established, and the issue was debated by two specialists at the invitation of the organizers of the 2008 Gastrointestinal Cancer Symposium. John L. Marshall, MD, chief of hematology and oncology, Georgetown University Lombardi Comprehensive Cancer Center, was assigned to the "pro" side of the debate, and Kenneth K. Tanabe, MD, chief of surgical oncology, Massachusetts General Hospital Cancer Center and Harvard Medical School, was given the "con" side.

Preoperative chemotherapy, Dr. Marshall said, is an essential element of "curative management" in patients who are clearly resectable from the outset, mainly because it treats occult disease that cannot be attacked by the surgeon's knife.

But Dr. Tanabe pointed out that preoperative chemotherapy can complicate the surgeon's mission, delay a potentially curative operation, and raise the risk of hepatotoxicity. "Don't assume that a response to chemotherapy will always facilitate liver resection," he said.

Both speakers emphasized that this debate is relevant only to patients who are "definitely resectable" and not to those with borderline or potentially resectable tumors (where there is little controversy) or unresectable tumors (where there is no controversy at all).

The definition of resectability, they added, has changed recently, mainly in that the number of metastases is no longer a deciding factor. "Complex operations are performed routinely, and long-term survivors are observed in each group," Dr. Tanabe noted

NCCN guidelines

In the treatment of this population, guidelines from the National Comprehensive Cancer Network (NCCN) for patients with colon cancer and synchronous liver metastases offer several options but no clear answers. Options include:

(1) Colectomy with synchronous or subsequent liver resection.

(2) Neoadjuvant therapy followed by synchronous or staged colectomy and resection of metastatic disease.

(3) Colectomy followed by chemotherapy and staged resection of metastases.

"The surgeon is still doing most of the curing," acknowledged Dr. Marshall, the pro chemotherapy speaker. "Adjuvant therapy fails most of the time. Neoadjuvant therapy is the same as adjuvant, but we have our eyes open. We can see if we are having an impact on the tumor, which is a great advantage."

More important, one wants to affect the biology of the tumor, he said. Whether this occurs, or whether chemotherapy simply alters the anatomy of the metastasis, has not been determined, although there is a suggestion of wider benefit, Dr. Marshall said.

The EORTC 40983 trial (Nordinger et al: ASCO 2007, abstract LBA5)—which randomized potentially resectable patients to surgery or to preoperative FOLFOX4 (6 cycles), surgery, then adjuvant FOLFOX 4 (6 cycles)—found that neoadjuvant therapy downstaged lesion size by 30% and improved progression-free survival by 23%.

"Preoperative chemotherapy clearly changed the anatomy" of the tumors and possibly affected biology, since it slowed down the disease process, Dr. Marshall said.

Additionally, the CRYSTAL trial (van Custem E et al: ASCO 2007, abstract 4000) and the OPUS trial (Bokemeyer C et al: ECCO 2007, abstract 3004), which included cetuximab (Erbitux) preoperatively, showed improved resection rates with neoadjuvant therapy.

This gives "further ammunition" to the pro argument, Dr. Marshall said. "As we increase response rates with multidrug chemotherapy, I think we will probably influence the tumor biology," he suggested.

No 'isolated' events

Dr. Marshall emphasized that preoperative chemotherapy targets hidden disease.

"The biology of metastases and the metastatic process is complicated. When we observe an isolated metastasis in the liver, our first thought is that this is almost certainly not an isolated event," he said. "Simply playing the odds when one is dealing with billions of cancer cells, banking on the chance that only one colony of cells would become established, makes the concept of a true isolated metastasis hard to fathom."

While the occasional patient may be cured by the resection of such a metastasis, in most who are deemed eligible for curative resection, an isolated hepatic metastasis does not exist, he said.

"Instead, these patients have a series of metastatic lesions, often in more than one lobe or in more than one organ site. Clearly, the biology of this tumor, with multiple sites of metastases, is more in line with our logic that this is a systemic process," he observed.

In this case, he said, chemotherapy must play a critical role, "as resection will only treat those metastases that are visible at the time of the operation and do nothing for the other likely sites of disease that are subclinical at the time of intervention."

Problems with preop chemo

While agreeing that preoperative chemotherapy may be the most timely way of delivering the drugs, Dr. Tanabe reminded listeners of the "known problems" with this approach.

Preoperative chemotherapy hinders the detection of occult metastases and the localization of known tumor sites, can produce hepatotoxicity, and can delay potentially curative resection. Additionally, patients who become resistant to first-line therapy (as evidenced by metastases) are likely to be unresponsive to second-line chemotherapy, he pointed out.

A radiographic complete response is not equivalent to a pathologic complete response, Dr. Tanabe noted.

"It is important to remove these lesions. 'Blind removal' sounds easy in concept, but it is not," he emphasized. Following a radiographic complete response, residual tumor is identified in more than three-fourths of samples, and chemotherapy reduces the sensitivity of PET in detecting these, he said.

"Does response to chemotherapy facilitate resections? Sometimes. Can response to chemotherapy complicate resections? Yes," Dr. Tanabe remarked.

He added that while chemotherapy response is a useful means of assessing the tumor biology, "it's not as if we don't have a robust set of other tools," including number and size of metastases, synchronous vs metachronous lesions at presentation, tumor differentiation, lymph node status, CEA levels, and so forth.

Dr. Tanabe also emphasized that response rates of greater than 50% in the first-line setting are reduced to 10% to 25% second line. "We recognize that in patients refractory to first-line chemotherapy, a robust response in the second line is not likely," he said.

The bottom line is that the benefit of preoperative chemotherapy is just not very impressive, Dr. Tanabe said.

In the EORTC trial, also described by Dr. Marshall, the 8% absolute benefit in 3-year progression-free survival was "barely significant" (P = .041), he noted, and not in line with chemotherapy trials in metastatic patients. These include the meta-analysis of chemotherapy vs best supportive care (Br Med J 321:531, 2000) and the study by de Gramont et al (J Clin Oncol 18:2938, 2000) showing added benefit when oxaliplatin (Eloxatin) was given with standard chemotherapy.

The hazard ratios in these trials—1.54 and 1.71, respectively—were "substantially more robust" than the 1.29 hazard ratio of the EORTC trial in chemotherapy-naïve patients. "The observed benefit did not match the expectation," Dr. Tanabe said.

Importantly, should chemotherapy fail, there may be a "lost window of opportunity" for a timely resection. "Tumor growth in a critical area may render metastases unresectable," he said.

Finally, Dr. Tanabe said, one has to be concerned about chemotherapy-associated toxicities and complications. Arterial thromboembolic events—seen in 4.4% of patients receiving bevacizumab (Avastin) plus chemotherapy in trials—can render patients unfit for resection, and hepatotoxicity is not uncommon, including steatohepatitis, inflammation, hepatocyte ballooning, and vascular lesions.

Hepatotoxicity may not only limit the extent of the resection but also increase the risk of perioperative morbidity and mortality. In EORTC 40983, the perioperative complication rates were doubled with the addition of chemotherapy (21% vs 10%), although the relationship between toxicity and particular agents and regimens remains unclear, he said.

A balancing act

In the end, Dr. Tanabe said, resection of liver metastases is a "complicated balancing act" that involves many factors: likelihood of recurrence, comorbidities, complexity of the resection, operative morbidity and mortality risks, quantity and quality of the residual liver, quality of the preop imaging, and patient age.

"Once chemotherapy is given, most of these factors can change," he said.

While acknowledging that some patients could benefit from preoperative chemotherapy, Dr. Tanabe described those patients he would always take straight to surgery:

• Those with good tumor biology who can undergo a straightforward liver resection, who have a critically located tumor that would become unresectable should it progress, or who cannot tolerate chemotherapy.

• Those with underlying obesity or diabetes who can undergo a standard liver resection.

• Those proven to be refractory to first-line chemotherapy.

While there is no established "winning" strategy at this time, both speakers agreed that a few things are indisputable. A multidisciplinary team should plan the treatment, an expert surgeon should wield the knife, and candidates for resection should be carefully selected. In this respect, "the bar is moving daily," Dr. Marshall said.