Abstract
1463
Objectives Tumor hypoxia is associated with treatment resistance and a poor prognosis. Noninvasive assessment of intratumoral hypoxia using 18f-fluoromisonidazole (FMISO) PET can be used to stratify cancer patients who are eligible for modification of subsequent treatment to improve tumor control. However, respiratory motion during PET image acquisition introduces quantitative inaccuracies in non-small cell lung cancer (NSCLC). Although the solution for respiratory motion is respiratory gating, the optimum uptake time is 蠅4 hours after FMISO injection due to the slow blood clearance, which reduces radioactivity counts and degrades the image quality especially for gating. Recently, the time-of-flight (TOF) technique has improved the signal-to-noise ratio, making it possible to gate respiratory movement in FMISO-PET. We evaluated the effect of respiratory tumor movement on the errors of the quantification of hypoxia in non-gated acquisition compared to gated acquisition as the standard.
Methods FMISO-PET/CT and FDG-PET/CT were performed in 18 patients with pretreatment Stage I and II NSCLC. The interval between FDG- and FMISO-PET/CT was ≤2 days. FDG- and FMISO-PET images were scanned at an hour and 4 hours after injection, respectively. Respiratory gating was performed in both PET and CT. The respiratory signal was obtained using a pressure sensor integrated in an elastic belt placed around the patient’s upper abdomen. The respiratory cycle was divided into 5 phases and we used the third phase for analysis which was equivalent to expiratory phase. The 3-dimensional tumor location was measured at each phase using the gravity center of FDG uptake. The maximum distance (Dmax) between any 2 phases was considered as the respiratory movement distance of tumor. The errors of each parameter, defined as difference of gated and non-gated value (error = gated value - non-gated value), were calculated. We investigated 1) the errors in SUVmax, tumor-to-muscle ratio (TMR), tumor-to-blood ratio (TBR), and 2) the correlation of the Dmax with the errors of SUVmax, TMR, and TBR.
Results The tumor locations were the upper lobe (N=8), the middle lobe (N=2), and the lower lobe (N=8). The tumor diameter was 25.8 ± 7.3 mm, and the Dmax was 7.59 ± 4.95 mm. Although the Dmax was not associated with the tumor diameter (p=0.41), the Dmax was significantly longer in the lower lobe (10.7 ± 1.5mm) than in the upper and middle lobes (5.1 ± 1.3 mm, p=0.0135). SUVmax, TMR, and TBR were significantly higher in gated than in non-gated images (gated vs. non-gated; 1.75 ± 0.81 vs. 1.59 ± 0.72 (p=0.002), 1.35 ± 0.66 vs. 1.28 ± 0.60 (p=0.042), and 1.51 ± 0.65 vs. 1.38 ±0.58 (p=0.002), respectively). As the results, the errors were 0.15 ± 0.17 in SUVmax, 0.07 ± 0.16 in TMR, and 0.13 ± 0.15 in TBR. The error of SUVmax (R=0.52, p=0.027), TMR (R=0.50, p=0.036), and TBR (R=0.48, p=0.042) were all correlated with Dmax. In patients with Dmax > 5mm, non-gated images always produced smaller indices than gated images. When tumor SUVmax above the mean SUVmax + 2 SD of muscle (蠅1.67 SUV) was regarded as a hypoxic tumor, 3 (17%) patients were underdiagnosed as non-hypoxic in non-gated images.
Conclusions The distance of tumor movement by respiration was strongly correlated with the errors of quantification of non-gated FMISO-PET. The respiratory movement of NSCLC leads to underestimation on tumor hypoxia, especially in the lower lobe. The respiratory gate should be taken into consideration for FMISO-PET/CT on NSCLC to avoid overlooking hypoxic tumors.
The characteristics of patients
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.