Association between FDG SUVmax and EGFR/KRAS mutation status in lung adenocarcinomas – An analysis of 80 cases

Abstract

1761

Objectives: Preoperative FDG SUVmax in untreated patients is prognostically important. We looked for an association between the presence of a EGFR/KRAS mutation in lung adenocarcinomas and FDG SUVmax, as EGFR mutation predicts sensitivity to erlotinhib.

Methods: Patients who had a preoperative MSKCC FDG PET scan without prior induction therapy, and lung adenocarcinomas tested for the presence of either an EGFR(exons 19 & 21)or KRAS(exon 2) mutation formed the data set. Tumor DNA was extracted, EGFR mutations were detected using PCR-based assays, KRAS(exon 2) mutations were detected by direct sequencing. The SUVmax within a volumetric ROI surrounding the tumor was measured on attenuation corrected, iteratively reconstructed FDG PET images using using AWS Centricity PACS(GEMS). A rank correlation of tumor size to SUVmax assessed the impact of tumor size on SUVmax. Kruskal-Wallis log rank test determined if there was any association between the SUV and the presence of a mutation, the p values of the log rank test were adjusted using an analysis of variance. In order to fully eliminate the effect of partial volume averaging we limited our analysis to the 56 tumors that had bi-dimensional measures greater than 1.5cm on the lung windows of a recent or concurrent CT image set.

Results: 36 men & 44 women met study criteria, of these 16 were oligo or never smokers (<15 pack years). There was a strong correlation between SUV and the size of the tumor (rank correlation = 0.52), using analysis of variance in log scale to correct for tumor size, the p value of the log rank test was reduced from 0.2 to 0.072. Restricting the analysis to lesions >1.5 cm left 27 patients without a mutation, 13 with an EGFR mutation, and 24 with a KRAS mutation. Even without correcting for tumor size there was evidence of a statistically significant association between the SUVmax and the presence or absence of an EGFR mutation, p=0.035. After correction for tumor size this association became stronger p=0.026.

Conclusions: Adenocarcinomas with an EGFR mutation have a lower SUVmax than adenocarcinomas with a KRAS mutation or with neither of these two mutations. Recognition of the importance of and correction for tumor size is essential to uncover the true biological information within FDG PET images.

SUVmax vs. Mutation Status