Qualitative Review
An example of the images at low and ultra-low injected activities are shown in figures 1 and 2. There is decreasing noise in the liver and other tissues with increasing acquisition time but lesions are clearly visualized even at short acquisition times in these examples. Table 1 shows the results of the qualitative review across both patient cohorts. The qualitative review of the Low dose group image sets resulted in an average minimum time per scan of 2.6 minutes with an optimal time per scan of 3.3 minutes. The qualitative review of the Ultra-low dose group images sets resulted in an average minimum time per scan of 4 minutes with an optimal time per scan of 5.6 minutes. To accommodate the preferences of the reader who was least tolerant of noise (Reader 2), we elected to use 1.0 MBq/kg and set minimal and optimal acquisition times of 3 and 5 minutes, respectively, for routine clinical studies unless clinical circumstances warrant minimizing radiation dose in which case would use 0.5 MB/kg and use minimal and optimal acquisition times of 4 and 6 minutes, respectively.
Table 1: Reader Results for Low dose (1MBq/kg) and Ultra-low dose Images (0.5MBq/kg)
Quantitative Review
The results of the measurement of COV are shown below in figure 3. For the Low dose group (1MBq/kg) the COV for the liver for the physician-defined qualitative minimum scan time of 3 minutes was 8.1% and 6.5% for the optimum scan time of 5 minutes. For the Ultra-low dose group (0.5MBq/kg) the COV for the liver at the minimum time of 4 minutes was 10.3% and 8.6% for the optimum scan time of 6 minute.
Increasing the scan times past the optimal time point in both groups did not yield a significant reduction in the COV. However, reducing the scan times below the minimum acceptable time per scan resulted in a significant increase in the COV, consistent with the reduced image quality perceived by expert PET readers. The COV for the liver from both cohorts at even the minimum acquisition time was well below the recommended EARL criteria of below 15% for image noise.
At the consensus optima and minimum acquisition durations, the coefficient of variance in liver was 10% or less for both administered
Figure 3: Liver COV (coefficient of variation) for the Low dose (1MBq/kg) and Ultra-low dose groups (0.5MBq/kg). The solid lines indicate the COV at the optimal acquisition time and the dashed lines indicate the COV at the minimum acquisition times.
Ultra-High sensitivity scanning mode
As can be seen in figures 4 and 5, the UHS mode resulted in a reduction in the image noise as measured by the liver COV, particularly at shorter acquisition times, potentially allowing for minimum acquisition times of around 2 minutes or less, depending on administered activity while maintaining a hepatic COV <15%.
Figure 4: Coefficient of variation for the low dose group for High sensitivity (MRD85) and ultra-high sensitivity (MRD322) scanning modes using 1MBq/kg
Figure 5: Coefficient of variation for the ultra-low dose group for High sensitivity (MRD85) and ultra-high sensitivity (MRD322) scanning modes using 0.5MBq/kg.
Effective dose
A summary of the administered activities and estimated effective doses is provided in table 2.
For the Low dose group, the patients received an average administered activity of 84.3MBq of [18F]-FDG with an average weight of 82kg (1.02MBq/kg). The effective dose calculated for [18F]-FDG by the Siemens software gave an average dose of 1.6mSv with the average effective dose from the CT of 1.1mSv giving a total effective dose of 2.7mSv for the procedure.
For the ultra-low dose group, the patients received an average administered activity of 44.6MBq of [18F]-FDG with an average weight of 84kg (0.53MBq/kg). The effective dose calculated for [18F]-FDG gave an average dose of 0.8mSv with the average effective dose from the CT of 1.2mSv giving a total effective dose of 2mSv for the procedure.
Table 2: average weight, administered activity and dose data for Low dose (1MBq/kg) and Ultra-low dose groups (0.5MBq/kg)