E1801. Detection of Lung Nodules: Low-Dose CT versus Fast Spoiled Gradient Echo MR Images Acquired During PET/MR
  1. Joanna Kusmirek; University of Wisconsin
  2. Nevein Ibrahim; University of Wisconsin
  3. Minnie Kieler; University of Wisconsin
  4. Alan McMillan; University of Wisconsin
  5. Emmanuel Sampene; University of Wisconsin
  6. Elizabeth Sadowski; University of Wisconsin
MRI detection of small lung nodules has previously been shown to be a limitation of PET/MR with standard T1-weighted images used for body imaging. The goal of this study was to compare MR lung imaging using fast spoiled gradient echo (FSPGR) MR images to low-dose CT lung images obtained as a part of PET/CT examinations for detection of lung nodules.

Materials and Methods:
This was an IRB approved HIPAA compliant retrospective study. PET/MR exams performed between 5/2016 and 4/2019 were reviewed. Subject inclusion: PET/CT performed within 2 months of the PET/MR. MR imaging of the lungs was performed using a FSPGR MR images acquired during a single breathhold (24s), using the following parameters: TR=2.2; TE=1.0; NEX=partial; Phase=320; Freq=288; acceleration=2x1.5; FOV= 46cm; Slice=2.6mm; Flip=4; BW= 488Hz per pixel. CT lung imaging was performed with a low dose free breathing acquisition and following parameters: kV=140; smart mA/auto mA range 40-500; slice thickness=5mm; rotation time=0.5s; collimation=40x0.6 mm. PET/MR and PET/CT lung series were anonymized by assigning a random name to each series and uploaded to PACS. A radiologist blinded to the randomization reviewed the images on PACS and recorded all lung nodules. McNemar's chi square statistical analysis was performed to see if the proportion of detected nodules varied between the two modalities.

41 subjects were included. In 21 subjects no nodules were detected on MR or CT images, in 11 subjects the same nodules were seen on both MRI and CT images, in 3 subjects nodules were seen on CT but not MR (size <5mm), and in 6 subjects nodules were seen on MR but not on CT (size <5mm). McNemar's chi square statistic equaled 1.00 (p=0.3173), suggesting there is no difference in the detection of nodules between MR and CT images.

In this study, more lung nodules were seen on MR lung images obtained during PET/MR compared to low-dose CT lung images obtained during routine clinical PET/CT. Statistically, these exams are equivalent in the detection of lung nodules. This study would support using stand-alone PET/MR imaging for clinical staging and surveillance in populations where detection of lung nodules is relevant.