Correlation Between Apparent Diffusion Coefficients and Standardized Uptake Values in Hybrid (18)F-FDG PET/MR: Preliminary Results in Rectal Cancer

Abstract

Purpose: Fluorine-18-fluorodeoxyglucose ((18)F-FDG) positron emission tomography (PET) and diffusion-weighted magnetic resonance imaging (DWI) share the same role in clinical oncology and it is feasible to obtain the standardized uptake value (SUV) and apparent diffusion coefficient (ADC) simultaneously by emerging the hybrid positron emission tomography/magnetic resonance (PET/MR). This study investigated the correlation between the ADCs of rectal cancer lesions and their SUVs derived from hybrid PET/MR.

Methods: Nine patients with histologically proven rectal adenocarcinoma (5 men, 4 women; mean age, 70 ± 15.91 years) underwent torso (18)F-FDG PET/CT and regional hybrid (18)F-FDG PET/MR sequentially. A fixed threshold value of 40 % of maximum uptake was used to determine tumor volume of interest (VOI) on PET image; SUVmax, SUVpeak, and SUVmean were calculated automatically. A single freehand region of interest (ROI) was drawn on high b-value (b1000) DWI image and copied to corresponding ADC map to determine the ADCmean of rectal cancer lesion. Spearman's rank correlation coefficient (ρ) was calculated to determine the correlation between SUVs and ADC values.

Results: SUVmax, SUVpeak, and SUVmean derived by hybrid PET/MR were 12.35 ± 4.66 (mean ± standard deviation), 9.66 ± 3.15 and 7.41 ± 2.54, respectively. The ADCmean value of rectal cancer lesions was 1.02 ± 0.08 × 10(-3)mm(2)/s. ADCmean was significantly and inversely correlated with SUV values (SUVmax, ρ = -0.95, p < 0.001; SUVpeak, ρ = -0.93, p < 0.001; SUVmean, ρ = -0.91, p = 0.001).

Conclusions: This preliminary hybrid PET/MR study demonstrates a significant inverse correlation exists between metabolic activity on (18)F-FDG PET and water diffusion on DWI in rectal cancer.

Keywords: Diffusion magnetic resonance imaging; Fluorodeoxyglucose F18; Multimodal Imaging; Positron-emission tomography; Rectal Neoplasms.

Fig. 1

An 82-year-old man with rectal adenocarcinoma. Axial T2-weighted SPACE image (a), fused ¹⁸ F-FDG PET and axial T2-weighted SPACE images (b), b1000 diffusion image (c), and ADC map (d) all acquired by hybrid PET/MR. ADC value was measured by drawing a freehand ROI along the high signal intensity tumor border on a single slice b1000 image containing the largest tumor area (c). The ROI was copied to the ADC map (d) to calculate the ADCmean. ADC map was automatically generated in grayscale by the operating system, using a mono-exponential decay model

Fig. 2

Correlation analyses of SUVmax (a), SUVpeak (b), and SUVmean (c) of rectal cancer lesions obtained by PET/CT and subsequent hybrid PET/MR. High correlations were found between SUV values from two modalities. ρ = Spearman’s rank correlation coefficient

Fig. 3

A 58-year-old man with rectal adenocarcinoma. Axial T2-weighted SPACE image (a), fused ¹⁸ F-FDG PET image and axial T2-weighted SPACE image (b) obtained by hybrid PET/MR showing FDG avid rectal wall thickening at the left lateral wall. SUVmax, SUVpeak, and SUVmean of the tumor were 8.14, 6.76, and 5.21, respectively. DWI (c) depicted the tumor as a high signal intensity lesion, and the corresponding ADC map (d) showed reduced ADC in the tumor (ADCmean was 1.04 × 10⁻³ mm²/s)

Fig. 4

Correlations between ADC and SUV values assessed by hybrid PET/MR. Scatter plots showed the strong inverse correlations between ADCmean and SUV values. ADCmean was found to be significantly and negatively correlated with SUVmax (a), SUVpeak (b), and SUVmean (c) by Spearman’s rank correlation analyses (SUVmax, ρ = −0.95, p < 0.001; SUVpeak, ρ = −0.93, p < 0.001; SUVmean, ρ = −0.91, p = 0.001). ρ = Spearman’s rank correlation coefficient