Objectives/hypothesis: Molecular analysis of surgical margins is playing an increasingly important role in establishing surgical margins. Most markers lack the sensitivity and ease of applicability for effective clinical use. To date, the proto-oncogene eIF4E (4E) is elevated in 100% of head and neck squamous cell carcinoma tumors and is of prognostic value in predicting recurrence. In a retrospective study, 4E overexpression in the margins appeared to be a more sensitive predictor of recurrence when compared with p53. The goal was to confirm this finding in a prospective study and also to compare the expression of matrix metalloproteinase-9 (MMP-9) to 4E expression in tumors and margins. Other objectives were to determine which of these markers have prognostic significance in predicting recurrence and elucidate whether there is any additional benefit to analysis of surgical margins with a combination of the three molecular markers.
Study design: A prospective study was performed on all patients who consecutively underwent primary surgical resection between 1998 and 1999 for head and neck squamous cell carcinoma. Patient and tumor characteristics were reviewed, and time to recurrence was noted.
Methods: Paraffin-embedded sections of tumors and all histologically tumor-free margins were analyzed for the presence or absence of 4E, p53, and MMP-9 with immunohistochemical analysis. Patients were followed according to the institution's head and neck cancer protocol, and time to recurrence was noted.
Results: Ninety-eight percent of tumors overexpressed 4E, 65% overexpressed p53, and 92% overexpressed MMP-9. Of the 52 patients with tumor-free margins, 52%, 46%, and 54% had positive margins for 4E, p53, and MMP-9, respectively. Although no significant correlation between 4E and p53 expression was seen in the margins (P =.16), a significant correlation between 4E and MMP-9 expression was noted (P =.0002). However, when expression of 4E and p53 in the margins of only the patients who overexpressed p53 in the tumors was compared (n = 34), there was a significant correlation (P =.04). There was also a significant difference in the disease-free interval between patients with 4E-positive and 4E-negative margins (P =.003). This difference in time to recurrence was not significant for the p53-positive versus the p53-negative group (P =.18) but approached significance when MMP-9 was used as a marker (P =.07). Although the univariate analysis showed that stage, nodal disease, grade, and 4E expression in the margins were significantly associated with disease-free interval, in the Cox multiple regression analysis, only 4E expression in the margin was significantly associated with disease-free interval (P =.01).
Conclusions: The era for molecular analysis of surgical margins is here. Although a significant correlation was seen between 4E and MMP-9, overexpression of 4E appears to be a significant predictor of recurrence when compared with the well-studied tumor suppressor gene p53 and a relatively novel marker, MMP-9.