LETTERS TO THE EDITOR

To the Editor:

We are writing to respond to the editorial by Brennan ¹ appearing in Annals of Surgery in the same issue as our paper on the survival of patients staged by FDG-PET before resection of hepatic metastases from colorectal cancer. ²

Dr. Brennan levels two main criticisms at our work. The first is that we claimed priority for the finding that about 25% of patients who undergo FDG-PET after conventional staging with CT are discovered to have more extensive disease than originally suspected. In fact, the introduction of our paper reviews, in some detail, the prior studies that established this fact, including one from Dr. Brennan’s own institution. ³ The meta-analysis ⁴ cited in the editorial was published after our paper was in press.

The second criticism is that we reportedly claimed that FDG-PET increases the survival rate of patients with colorectal cancer metastases to the liver. In fact, we were quite careful to state the following in our discussion: “...use of FDG-PET is not improving survival by itself, but is allowing surgical techniques to be applied with greater likelihood of benefit to patients. The result is better survival of patients who do receive surgery because the target population for surgery has changed, rather than because FDG-PET and surgery produce longer survival times for the same sort of patients represented in previously published studies.” Therefore, neither criticism is justified on careful reading of our paper.

Concern was also expressed regarding use of the terms “overall survival” and “resectability rates.” The term “overall survival” was used to contrast with “disease-free survival” in the resected patients. The paper is clear that the survival referred to was that in the patients who underwent resection. Regarding operability and resectability rates, these terms were used in conformity with commonly accepted definitions in the surgical literature. Again, it is clear in the test and figures that “resectability rate” referred to the frequency of resection in patients who came to laparotomy after FDG-PET.

I believe that Dr. Brennan missed the opportunity to set the course in this important area. An editorial is by definition the opinion of the journal and its editor, although today an editor often selects a surrogate to represent him or her in a specialized area. The editorialist has at least three responsibilities to the journal’s readers. The first is to inform why the article was considered sufficiently meritorious and interesting to be selected for publication (and for editorial comment) and in what way it affects the field. The second is to discuss the shortcomings of the research, and the third is to outline questions unanswered and work yet to be done. This editorial falls short in each of these areas. Instead, it tells us that all three papers on the subject of PET scanning published in the March 2001 issue of Annals of Surgery^1,5,6 have little merit and that randomized controlled trials should have been performed.

Randomized controlled trials are, of course, the gold standard toward which we should aim. However, they are costly and often lengthy. As a result, selectivity is required in deciding which questions are of sufficient promise and merit to warrant study via this process. Nonrandomized, single-institution studies of an emerging technology like FDG-PET, such as our own study and that published by the editorialist’s own institution, ³ often provide the rationale for a randomized trial. Therefore, they may have a very important role in this respect alone.

We have carefully considered whether a randomized trial could be done to confirm the findings of our study and we have presented our data as the basis for the planning of a multi-center study by the American College of Surgeons Oncology Group (ACOSOG). The data were judged by the ACOSOG hepatobiliary organ site committee to be so convincing that a Phase II study, rather than a randomized Phase III trial, was recommended. There was a consensus that sufficient equipoise would not exist in the surgical community and among patients to have a control group in a multi-center trial of the use of FDG-PET in metastatic colorectal cancer. In effect, it was predicted that a randomized, controlled trial would fail because of lack of accrual.

Furthermore, it was felt that a Phase II trial would be sufficiently conclusive as to the value of FDG-PET if it showed that FDG-PET detected new disease after conventional staging, particularly if this new disease would not have been discovered at laparotomy. This trial is now being put forward through the auspices of the ACOSOG. For these same reasons, two currently open ACOSOG studies of the utility of FDG-PET in lung and esophageal cancer, respectively, are also both Phase II trials.

Dr. Brennan laments the lack of randomized controlled trials of FDG-PET in cancer, but doesn’t provide the journal’s readers with any insight regarding how he believes these could be accomplished.

Steven M. Strasberg, MD
Barry A. Siegal, MD