NEOADJUVANT THERAPY FOR PANCREATIC CANCER

May 6, 2010 — Pancreatic cancer patients with locally nonresectable tumors should be included in neoadjuvant protocols, according to the results of a new meta-analysis. The patients should then be reevaluated for resection, which appears to be a possibility for a "relevant" number of these individuals.

The authors found that a third of patients who were initially judged "unresectable" were able to undergo resection after neoadjuvant therapy. Their average survival time was 20.5 months, compared with 23.3 months for patients whose tumors were considered resectable on preoperative examination.

The study is published in the April issue of PLoS Medicine.

The authors, led by Jörg Kleeff, MD, from the Technische Universität München in Germany, note that their most important findings are "that in the group of resectable tumor patients, resection and survival rates after neoadjuvant therapy are similar to the ones observed in primarily resected tumor that are treated by adjuvant therapy." This suggests that for patients with resectable tumors, neoadjuvant therapy will not provide any clinical benefit.

Conversely, approximately one third of patients with tumors initially considered to be unresectable and who were able to undergo resection after neoadjuvant therapy had a similar survival rate to patients judged resectable before neoadjuvant treatment, they note.

Even though data regarding the optimal chemotherapeutic and radiotherapeutic regimen cannot be extrapolated because of the heterogeneity of the trials, the "data suggest that combination chemotherapies result in higher response rates, which are reflected by higher resection rates, at least in the group of initially nonresectable tumor patients," they write. "But the available data strongly suggest that patients with locally advanced/unresectable tumors be included in neoadjuvant protocols and subsequently be reevaluated for resection."

Interpreting the Results

However, Alexander S. Rosemurgy, II, MD, professor of surgery and medicine at the University of South Florida in Tampa, urged caution in interpreting the results. "They only reported the data that others have presented," said Dr. Rosemurgy, who was not involved in the study. "The question is: How many of these patients were really unresectable at their presentation?"

"Nonetheless, they received neoadjuvant therapy, and 'only' 33% were resected," he added.

Daniel Abbott, MD, from the Department of Surgery, NorthShore University Health Systems/University of Chicago, Illinois, agrees.

"First and foremost, we must keep in mind that there are no prospective randomized studies that thoroughly evaluate the use of neoadjuvant therapy in pancreas cancer," he told Medscape Oncology. "Specifically, there are no data that address, in a scientific fashion, preoperative vs postoperative chemotherapy or radiation with respect to extended survival."

"In the end, the heterogeneity of existing studies with respect to study cohort and intervention make existing data difficult to interpret," said Dr. Abbott, who was not involved in the study.

In their paper, Dr. Kleeff and colleagues point out that even though neoadjuvant therapy for pancreatic cancer has been proposed for more than 2 decades, and although there is strong evidence of its benefit in other tumor types, there has been no compelling evidence for its use in pancreatic cancer.

"Certain phase 2 studies suggest that the use of neoadjuvant therapy is beneficial," explained Dr. Abbott, "and large series from high-volume, tertiary-care centers would argue that neoadjuvant therapy for borderline resectable disease provides, at the very least, patients with previously inoperable pancreas cancer a chance to be reevaluated radiographically, and perhaps downstaged."

This approach can allow for patient selection; those who display disease progression despite nonsurgical therapy can avoid an operation with high morbidity, he noted. "While not all of these patients will benefit from such therapies, a select few will be afforded an operation with higher margin-negative rate."

Study Details

Thus far, the authors reiterate, the only chance for curative treatment of pancreatic cancer is complete resection of disease. But even then, only approximately 10% to 20% of patients are considered candidates for curative resection. And after a potentially curative resection, the rate of local recurrence and/or distant metastasis remains high.

Adjuvant and neoadjuvant therapy strategies have been investigated, but the optimal regimen has yet to be determined. However, there is a strong rationale for a neoadjuvant approach, according to the researchers, because a relevant percentage of pancreatic cancer patients are diagnosed with nonmetastatic but locally advanced disease, and microscopic incomplete resections are common.

However, data from randomized phase 3 trials regarding the role of neoadjuvant therapy for pancreatic cancer remain unavailable.

A total of 111 studies, involving 4394 pancreatic cancer patients, were included in the analysis, and the authors evaluated the effects of neoadjuvant chemotherapy and/or radiotherapy on tumor response, tumor resectability, and patient survival. They subdivided the studies into 2 groups: patients with tumors that were considered resectable on preoperative examination (the resectable group), and patients whose tumors were borderline resectable or unresectable (the unresectable group).

In the majority of studies (96.4%), neoadjuvant chemotherapy was administered, with the main agents being gemcitabine (Gemzar, Lily), 5-FU (and oral analogues), mitomycin C, and platinum compounds. Neoadjuvant radiotherapy was applied in 93.7% of the studies, with doses ranging from 24 to 63 Gy.

For the entire cohort, the estimated complete response was 3.9% (95% confidence interval [CI], 3% - 4.9%), and partial response was 29.1% (95% CI, 24.5% - 34%). Stable disease averaged 43.9% (95% CI, 37.9% - 50%) for all patients in the study population, and tumor progression for those undergoing therapy was estimated to be 20.8% (95% CI, 17.3% - 24.6%). The authors note that the pooled percentages varied little among the 2 groups.

In the resectable group, resectability was estimated at 73.6% (95% CI, 65.9% - 80.6%), whereas in the unresectable group, it was 33.2% (95% CI, 25.8% - 41.1%).

Of the patients considered to have resectable disease prior to treatment, 88.1% (95% CI, 82.9% - 92.4%) were explored after restaging and, of those, 85.7% (95% CI, 78.9% - 91.2%) could be resected. In the unresectable group, 46.9% (95% CI, 36.9% - 57.1%) of the patients were explored after restaging and, of this subgroup, 69.9% (95% CI, 61.2% - 77.9%) could be successfully resected.

The estimated median survival for the entire group of patients who underwent resection was 22.4 months (range, 9 to 62 months). As expected, the median survival of patients who did not undergo resection was shorter — a median of 9.5 months (range, 6 to 21 months). The estimated 1- and 2-year probabilities of survival for resected patients in the resectable group were 77.9% and 47.4%, respectively, and in the unresectable group were 79.8% and 50.1%, respectively.

Morbidity data were available in 50 of the studies, and mortality data in 85 of the studies. For all patients, perioperative morbidity was estimated at 34.2% (95% CI, 28.3% - 40.4%), whereas in-hospital mortality after neoadjuvant treatment and tumor resection was estimated at 5.3% (95% CI, 4.1% - 6.8%).

The authors conclude that randomized trials are now needed to confirm these findings, and to "clearly establish the role of neoadjuvant therapy specifically in locally advanced/unresectable tumors."

Dr. Abbott concurred that these data suggest that further systematic, multicenter prospective studies are required "to address this clinical dilemma."

"Without more rigorous investigation, our counseling of patients with pancreatic cancer will rely on suggestive yet inconclusive evidence-based data," he said. "Hopefully, this realization will provide the impetus for more sophisticated, collaborative efforts."

No direct funding was received for this study. The authors have disclosed no relevant financial relationships.

PLoS Med. 2010;7: e1000267. Abstract