In this excellent overview of the chemotherapy available for pancreatic cancer, Drs. Berlin and Rothenberg have correctly emphasized the largely palliative nature of such therapy. In metastatic disease, chemotherapy has no curative potential, so the objective should be to maximize quality of life for as long as possible.

The percentage of pancreatic cancer patients who achieve a symptomatic improvement from chemotherapy is substantially higher than the percentage of those who achieve an objective response.[1] Objective response understates the effectiveness of chemotherapy because of the florid desmoplastic response accompanying ductal carcinoma of the pancreas. This scar tissue usually does not shrink rapidly, even in response to effective chemotherapy. Imaging this mass may not reveal the true response to therapy, which has led to the adoption of various systems for measuring "clinical benefit response" (pioneered by Dr. Rothenberg[2]) as end points in pancreatic cancer trials.

5-FU and Gemcitabine

Gemcitabine (Gemzar) has replaced fluorouracil (5-FU) as the most widely used drug in advanced pancreatic cancer, largely due to the results of a single pivotal trial.[3] Our only disagreement with this article is that it does not address the theory, implicit in that trial, that one schedule of 5-FU is as good as another. In the pivotal trial of gemcitabine in pancreatic cancer, the control arm received an ineffective schedule of weekly bolus 5-FU, which had been largely abandoned in gastrointestinal cancer for over two decades. Although the results of that trial were presented as demonstrating the superiority of gemcitabine over 5-FU, the trial was not structured to make such a comparison with clinically relevant schedules of 5-FU.

Even after more than 40 years of experience with 5-FU, it is still difficult to accurately define its effectiveness in pancreatic cancer for several reasons. Early trials overstated response rates because of nonobjective response criteria. Tumor response measured by conventional criteria is inadequate in evaluating palliative effect in pancreatic cancer for the reasons noted above. In addition, most trials of 5-FU in pancreatic cancer have enrolled very small numbers of patients.

It would seem that survival would be a better end point than response, but survival data present another major problem: Most trials include heterogeneous mixes of patients with metastatic vs locally advanced disease. Median survival in the former group is usually less than half of that in the latter group, so patient mix can greatly influence survival outcome in individual trials. Bearing in mind all of these limitations, the activity of 5-FU and gemcitabine in a few recent trials in pancreatic cancer is summarized in Table 1.[3-13]

A reasonable conclusion is that bolus 5-FU is essentially inactive in pancreatic cancer and that bolus 5-FU with leucovorin is not much better. Infusional schedules of 5-FU appear to have activity comparable to or perhaps even superior to gemcitabine, but have never been compared to the newer agent prospectively. But what is most important is that because the toxicities of infusional 5-FU and gemcitabine do not overlap, they can be used in combination at essentially full doses, as demonstrated by Hidalgo et al.[14] This combination, with an objective response rate of 19% and a symptomatic response rate of 45%, resulted in a median survival of 10.3 months (almost 70% longer than has been seen with either drug alone) in a population that was 65% metastatic.
Pages: 1  2  3