18F-EF5 PET Is Predictive of Response to Fractionated Radiotherapy in Preclinical Tumor Models

Abstract

We evaluated the relationship between pre-treatment positron emission tomography (PET) using the hypoxic tracer 18F-[2-(2-nitro-1-H-imidazol-1-yl)-N-(2,2,3,3,3- pentafluoropropyl) acetamide] (18F-EF5) and the response of preclinical tumor models to a range of fractionated radiotherapies. Subcutaneous HT29, A549 and RKO tumors grown in nude mice were imaged using 18F-EF5 positron emission tomography (PET) in order to characterize the extent and heterogeneity of hypoxia in these systems. Based on these results, 80 A549 tumors were subsequently grown and imaged using 18F-EF5 PET, and then treated with one, two, or four fraction radiation treatments to a total dose of 10-40 Gy. Response was monitored by serial caliper measurements of tumor volume. Longitudinal post-treatment 18F-EF5 PET imaging was performed on a subset of tumors. Terminal histologic analysis was performed to validate 18F-EF5 PET measures of hypoxia. EF5-positive tumors responded more poorly to low dose single fraction irradiation relative to EF5-negative tumors, however both groups responded similarly to larger single fraction doses. Irradiated tumors exhibited reduced 18F-EF5 uptake one month after treatment compared to control tumors. These findings indicate that pre- treatment 18F-EF5 PET can predict the response of tumors to single fraction radiation treatment. However, increasing the number of fractions delivered abrogates the difference in response between tumors with high and low EF5 uptake pre-treatment, in agreement with traditional radiobiology.

Fig 1. (A) Representative axial slices from microCT (top row), ¹⁸F-EF5 microPET (middle row), and anti-EF5 immunohistochemistry on (bottom row) of HT29, A549 and RKO subcutaneous tumors. CT image intensities represent Hounsfield Units (HU). PET images have been normalized by the mean background muscle %ID/g to yield tumor/muscle ratio (T/M) images. T on microCT and microPET images denotes tumor location. Brown stained areas in IHC images denote regions of bound EF5. (B) Scatterplot of the distribution of ¹⁸F-EF5 uptake in individual HT29, A549 and RKO subcutaneous tumors using three different metrics, T/M measured using CT-derived ROIs (T/MCT), T/M measured using PET-derived ROIs (T/MPET) and %ID/g values measured using CT-derived ROIs (%ID/gCT). (C) Scatterplot of the relationship between T/MCT for HT29, A549 and RKO tumors and their volume.

Fig 2. (A) Scatterplot of the relationship between ¹⁸F-EF5 T/M values for individual tumors when quantified using either PET and CT derived ROIs (T/MPET and T/MCT respectively), with a linear regression line superimposed. The four quadrants represent regions where both T/MPET and T/MCT are below the designated hypoxia threshold (EF5 -), or both are above the threshold (EF5 +), or only one is above the threshold (EF5CT+/EF5PET- or EF5CT-/EF5PET+). (B) Scatterplot of ¹⁸F-EF5 tumor uptake vs tumor volume.

The gray region denotes EF5 negative tumors (defined as T/MPET < 2.5), and the red region denotes EF5 positive tumors (T/MPET > 2.5).

Fig 3. Plots of response of more and less hypoxic tumors to (A) single fraction and (B) multiple fraction treatments.

Volumes are normalized to pre-treatment volumes.

Fig 4. Effect of radiation on tumor ¹⁸F-EF5 uptake.

(A) Representative axial ¹⁸F-EF5 PET images of two tumors at 1 day pre-treatment and 13 / 32 days post-treatment. Solid arrow denotes unirradiated control tumor, and dashed arrow denotes tumor which was treated with 20 Gy. (B,C) Plots showing changes in ¹⁸F-EF5 uptake and volume, respectively, in control (n = 6) and irradiated (n = 6) tumors during the treatment timecourse.