Originally published as JCO Early Release 10.1200/JCO.2008.20.2770 on February 2 2009

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Wagman, L. D. Search for Related Content PubMed Articles by Wagman, L. D. Social Bookmarking                            What's this?

In Reply

Center for Cancer Prevention and Treatment, St. Joseph Hospital, Orange, CA

Adam and Aloia write a passionate, thorough, and extensively referenced letter regarding the editorial I prepared for their article,¹ "Is Hepatic Resection Justified After Chemotherapy in Patients With Colorectal Liver Metastases and Lymph Node Involvement?" In fact, the article would be better titled, "Is Hepatic Resection With Portal Lymph Node Dissection After Response to Chemotherapy Justified?" as that is the conclusion that they spoke to. That aside, all conclusions must emanate from consistent, controlled data. I will reiterate my position that a 46-patient retrospective study performed over a 14-year period, during which diagnostic, surgical, and chemotherapeutic modalities dramatically changed, cannot be taken as "scientific proof." Outcome inferences affected by patterns of clinical approach (oncosurgical therapy, multiple hepatectomies, and so on) may serve as methods to generate hypotheses, but cannot be widely adopted as appropriate practice.

Let us consider our acceptance of even the most highly structured study, that of a phase III, prospective, randomized, double-blind clinical trial. When a post facto subset analysis is performed, we would use that subset analysis only as the jumping-off point for future clinical trials, not as a verification of the value of the studied intervention. The conclusions drawn from the type of retrospective study done here with subset analysis is a further deviation from standard, acceptable, statistically robust, or even statistically responsible conclusions.

I will address the third point specifically. Let us question the outcome in two other subsets in their patient population. The first is the large number of patients in the control (non–regional lymph node) group who had surgically unresected, but systemically treated, pedicular lymph nodes. Given that the incidence of pedicular lymph node involvement with colorectal cancer liver metastases has been consistently documented at 10% to 30%, how can the absence of florid nodal failure be explained in this group? And what about the 28 patients in this series whose lymph node positivity was identified intraoperatively? The authors cannot support the position that only the six "responders" underwent surgery. These confounding elements (uncertain baseline of positive pedicular nodes in non-RLN and selective intraoperative identification of positive nodes without a confirmed treatment response) make it difficult, from a surgical standpoint, to accept extirpation as an evidence-based component of the extended survival in those six patients. Maybe another way to address this lymphadenectomy question is to ask how many patients died with isolated pedicular involvement. More likely, they died of the uncontrolled systemic disease. Would the survival rates of patients with RLN have been any different with hepatic resection alone? Or would multiple cycles of chemotherapy given as sequential, on-off, intra-arterial, or chronobiologic regimens have resulted in an identical 5-year survival rate of 25%? Adam et al cannot answer this question with the current data.

It would be possible to address each of their points as outlined in the above letter; however, the main questions concern the selection of patients and the ability to generalize the results derived from retrospective subset analyses. The authors could accuse a colleague of a "closed attitude" only if that honest dissenter were unwilling to test the hypothesis(es) they generated.

The article clearly has scientific value. However, that value lies in hypothesis generation, not evidence-based medicine.

AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

REFERENCE

1. Adam R, de Haas RJ, Wicherts DA, et al: Is hepatic resection justified after chemotherapy in patients with colorectal liver metastases and lymph node involvement? J Clin Oncol 26:3672–3680, 2008.[Abstract/Free Full Text]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Wagman, L. D. Search for Related Content PubMed Articles by Wagman, L. D. Social Bookmarking                            What's this?