| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Editorial |
University of Wisconsin Hospital, 600 Highland Avenue, K4/764 CSC, Madison, Wisconsin 53792
Correspondence: Address correspondence and reprint requests to: Sharon M. Weber, MD; E-mail: webers{at}surgery.wisc.edu.
Defining the ability of histological or molecular markers to predict outcome for cancer patients is crucial. Because pancreatic adenocarcinoma is currently associated with such a poor prognosis, understanding the significance of such markers could have a profound effect on patient care, both in predicting prognosis and in guiding therapy.
In this issue of Annals of Surgical Oncology, White et al.1 from Duke University take an important first step toward defining the role of histological and molecular markers for patients with pancreatic adenocarcinoma. All of the patients evaluated in this study were treated with neoadjuvant chemoradiotherapy (CRT) followed by surgical resection. Histological evaluation of 54 post-CRT pancreatic tumor specimens was performed to correlate these findings with overall survival. Specifically, tumor fibrosis, tumor load (the percentage of tumor composed of viable tumor cells), and the presence of tumor necrosis were analyzed. Multivariate analysis revealed that in addition to the routine histological prognosticators of lymph node positivity and tumor differentiation, the presence of a large residual tumor load and, paradoxically, the presence of tumor necrosis was associated with a poor prognosis. Although immunohistochemistry was also performed on a subset of these patients to evaluate the significance of epidermal growth factor receptor and p53 expression, there was no correlation between the expression of these markers and overall survival. The unexpected finding of tumor necrosis as a negative predictive factor is difficult to explain, although the authors suggest that this may be a surrogate marker for rapidly dividing tumor. Clearly these findings address an important issue—defining which patients have an improved prognosis after neoadjuvant therapy for pancreatic adenocarcinoma.
Other investigators have attempted to characterize the genetic changes that occur in pancreatic cancer tissue, to correlate these findings with prognosis. In various studies, bcl-2,2 p53,2 transforming growth factor ß,3 erbB-3,4 vascular endothelial growth factor,5 and activators of matrix metalloproteinases6 present in pancreatic cancer tissue have all been correlated with outcome. Unfortunately, all of these findings are preliminary. Thus, the clinical relevance is still being investigated. Ideally, molecular markers for patients with cancer would enhance our ability to determine prognosis and permit us to individualize therapy. Unfortunately, the paucity of data examining the relevance of molecular markers makes it clear that we are not yet able to make this leap for patients with pancreatic cancer.
One of the most frustrating issues for physicians who treat patients with pancreatic cancer is our lack of ability to define the ideal adjuvant or neoadjuvant therapy. The role of adjuvant therapy for patients with pancreatic cancer is difficult to determine according to the limited and flawed data available from randomized prospective trials.7–9 Although many investigators are now routinely using neoadjuvant treatment, there are no randomized data evaluating its use. Supporters of neoadjuvant therapy, including the authors, give many intriguing and compelling reasons to use it routinely for patients with resectable pancreatic cancer. The problem is that studies evaluating the use of neoadjuvant therapy have compared survival only with historical controls. Along with the usual problems in using these retrospective series for comparison of outcome, studies evaluating neoadjuvant therapy are also hindered by the calculation of survival from the date of tissue diagnosis (which, in the authors’ own series, added a median of 3.5 months before surgical resection)10: this makes a survival comparison with historical controls even more difficult to interpret. Thus, not only are there few consistent data regarding the relevance of molecular markers for pancreatic cancer, but it is also difficult to envision that it will help us guide therapy in the near future.
Unfortunately, performing a truly impartial national trial is unlikely because of the potentially insurmountable biases at many institutions toward treating patients with neoadjuvant therapy. Unfortunately, it seems that in the year 2004, we still have not defined whether chemotherapy or radiotherapy (or both) is necessary to treat patients with resectable pancreatic cancer, nor have we defined the ideal timing to administer these therapies.
The authors present interesting data that correlate histological factors with outcome. However, the patients in this study had already received neoadjuvant therapy when the resected tissue was evaluated. Although defining which patients have an improved prognosis after neoadjuvant therapy is valuable, perhaps a more clinically relevant question might be whether histological or molecular markers present on preoperative biopsy predict whether a given patient has a greater chance of responding to neoadjuvant therapy. In other words, can we learn something from the findings of White et al. that might help us to improve patient selection for neoadjuvant treatment? In institutions where neoadjuvant therapy is used routinely, this issue might be addressed by obtaining a preoperative core biopsy and comparing the pre- and post-CRT tissue markers. These results could then be correlated with response to CRT and, more importantly, overall and disease-free survival.
Defining the ideal treatment for patients with pancreatic adenocarcinoma is difficult. The findings of White et al. are an important early step toward defining which patients may have an improved prognosis after neoadjuvant therapy for pancreatic cancer. However, to further define the role of histological and molecular markers for patients with pancreas cancer, it is essential that future clinical trials collect tissue to correlate these markers with outcome. It is only in this way that we have a chance to improve therapeutic decisions and, thus, outcome in these patients.
Received for publication October 27, 2004. Accepted for publication November 19, 2004.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
