PT - JOURNAL ARTICLE AU - Sjoholm, Therese AU - Schaefferkoetter, Josh AU - Townsend, David TI - Ultralow dose CT attenuation correction for lung cancer PET/CT screening - a phantom evaluation study DP - 2016 May 01 TA - Journal of Nuclear Medicine PG - 308--308 VI - 57 IP - supplement 2 4099 - http://jnm.snmjournals.org/content/57/supplement_2/308.short 4100 - http://jnm.snmjournals.org/content/57/supplement_2/308.full SO - J Nucl Med2016 May 01; 57 AB - 308Objectives This work aims to evaluate ultralow dose CT attenuation correction (CTAC) for the purpose of lung cancer PET/CT screening.Methods As a first step, a heart/thorax phantom (RSD Alderson) incorporating fillable compartments for the lungs, liver and heart was scanned on Siemens Biograph mCT PET/CT scanner. A 12 mm diameter fillable sphere was inserted into one of the lungs to simulate a lung tumour and the phantom was filled with an F18 solution to achieve a 2:1 tumour to background ratio. A Perspex thorax overlay was used to simulate a standard sized patient. The phantom was scanned with a range of CTAC settings (120 kVp, 10-200 mAs). This gives a CT effective dose of 0.2 mSv for a 2 bed PET/CT scan at the lowest CTAC setting (Shrimpton, NRPB-PE/1/2004), which is a dose reduction of 83% compared to standard PET/CT lung scan setting at our centre (120 kVp, 60 mAs). CTAC image reconstruction was performed using Filtered Back Projection. For PET, datasets at net true counts of 1, 5, 10, 15 and 20 million counts were generated. 20 independent realizations were created for each count density. The PET datasets were iteratively reconstructed using the different CTAC scans (OP-OSEM 3i24s, Time of Flight and Point Spread Function correction, 2 mm pixel size, 2 mm Gaussian smoothing). Data generated from the highest quality CTAC scan (120 kVp, 200 mAs) and the 20 million net true counts PET data was set as reference. The percentage bias for sphere, lung and background compartment ROIs, and the variability of the background compartment, were assessed for the generated attenuation maps and reconstructed PET data. All analysis was performed in MATLAB (R2015a).Results Overall, small percentage bias and background variability were measured for the attenuation maps and PET data. The percentage bias for the lowest CTAC setting attenuation map (10 mAs), was marginally higher than the other settings, with percentage biases of 0.5%, -3.7% and 0.2% for the sphere, lung and background, respectively. This can be compared to percentage biases of 0.2%, -3.2% and 0.07% for the sphere, lung and background in the attenuation map created from the standard CTAC setting of 120 kVp and 60 mAs. The background variability was low for all attenuation maps (<0.2%). For the PET data, changing the CTAC setting did not change the bias or background variability. This was seen regardless of PET count density. Changing the PET count density was, as expected, observed to have a larger effect on the bias and background variability. The background bias and variability for PET data created from the 60 mAs CTAC was for example seen to go from 0.002% and 7.7% for 20 million net true counts, to 1.8% and 9.4% for 1 million net true counts.Conclusions Using a heart/thorax phantom, it was shown that ultralow dose CTAC can be used without introducing significant bias in the attenuation map and without loss of PET image quality. This is promising for achieving a total effective dose <1 mSv for a screening PET/CT scan.