FDG-PET/CT: Is ultra-HD-PET (combination of TOF-acquisition and HD-reconstruction) and/or HD-reconstruction alone superior to iterative reconstruction in the detection of malignant lung lesions, especially in obese patients?

Objectives The aim was to evaluate if ultra-HD-PET (HD-PET + time of flight (TOF)) and/or high definition reconstruction (HD-PET) is superior to iterative reconstruction (IR) in the detection of malignant lung lesions by FDG-PET/CT in patients with bronchial carcinoma (BC) and in consideration of the body mass index (BMI) >25 kg/m2.

Methods 22 patients with BC were prospectively examined with FDG-PET/CT (Biograph mCT40, Siemens). A mean of 226±14 MBq FDG were applied i.v., and the acquisition was initiated 79±10 min p.i. Pulmonal lesions with hypermetabolism were visually detected and SUVmax-values were calculated by VOI analysis. The number of pulmonal lesions with hypermetabolism and SUVmax were compared for all 3 different reconstruction methods. SUVmax/background-ratios were calculated by drawing identical VOIs in the surrounding, physiological lung tissue and were used to define the proportion of radiotracer uptake. Furthermore, SUVmax of patients with a BMI > 25 kg/m2 were compared with normal weight patients. The significance level was defined as p ≤ 0,05.

Results In our preliminary results, 89 (100%) lung lesions with glucose hypermetabolism were detected by ultra-HD-PET, 78 (87,6%) by HD-PET without TOF and 61 (69%) by IR (p<0,001). In a lesion-to-lesion analysis, SUVmax was significantly higher in ultra-HD-PET (p=0,029) and HD-PET (p=0,01) than IR. The SUVmax/background-ratio was significantly higher in ultra-HD-PET than in HD-PET and IR (p=0,01). 9 of 22 patients (41%) had a BMI > 25kg/m2. However, the BMI did not significantly influence the findings in any reconstruction mode.

Conclusions In malignant BC lesions, ultra-HD-PET is superior to HD-PET which is again superior to IR with regard to sensitivity, SUVmax-values and SUVmax/background-ratios. In our small patient collective, no significant influence of the BMI with a cut-off value of >25 kg/m2 was demonstrated between the different reconstruction algorithms