Patient population
Among 95 consecutive patients who underwent lung perfusion SPECT and lung CT within 14 days because of suspected PE between June 2019 and August 2020 in department of cardiovascular medicine, we identified 28 patients (73 ± 10 years) who had normal pulmonary artery on contrast lung CT (Fig. 1). The institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived (M20076).
Lung planar scan and perfusion SPECT
Lung planar and lung perfusion SPECT was performed using a dual-head gamma camera (Infinia, GE Healthcare, Buckinghamshire, UK) equipped with low-energy, high-resolution collimators. Planar images were recorded for 10 s each in the anterior and posterior views at deep-intake and free-breathing conditions after injecting 370 MBq of technetium-99m-macroaggregated albumin (99mTc-MAA) with the patient in a supine position. Subsequently, SPECT data were acquired with 72 projections over an orbit of 360 degrees per step and 15 s per projection in free-breathing conditions. The image matrix size was 128 × 128. The image reconstruction was performed using filtered back projection and processed with Butterworth prefiltering (critical frequency 0.5, power 10.0).
Chest CT
Data from chest CT performed within 14 days before and after lung perfusion SPECT were used (Aquilion Precision; Canon medical systems, Tokyo, Japan). Patients were scanned in the supine position with the following acquisition parameters: tube voltage, 100 kV at auto mA; rotation time, 0.5 s; collimation, 0.25 × 160 mm; and pitch, 1.381. Patients received IV injections of 80 ml Omnipaque-350 contrast at 3.5–4.5 ml/s via IV access, followed by a 40 ml saline flush. Individual contrast optimization was achieved by using a 20 ml test bolus in the right ventricle with a trigger level of 150 HU. An additional delay of 11 s was added before image acquisition in every examination. All scans were reconstructed as 2.0-mm-thick slices with an increment of 2.0 mm.
Breathing synchronization software
Step 1
In this step, the original anterior and posterior planar images were summed into one planar image to perform in the step that followed. The summed planar image at deep intake (A) was created from the original anterior and posterior planar images. The summed planar image at free breathing (B) was created from the original anterior and posterior planar images (Fig. 2 Step 1).
Step 2
In this step, planar images matched the pixel and matrix size with that of the original SPECT coronal image by aligning the images. The aligned planar image at deep intake (C) was created by aligning image A with the original SPECT coronal image as a reference, using rigid-body transformation with the cost function of mean squared error. The aligned planar image at free breathing (D) was created from image B with the same parameters as rigid-body transformation (Fig. 2 Step 2).
Step 3
In this step, the two optimized variables (X and Y) were calculated for scaling from the aligned planar image at free breathing to the aligned planar image at deep intake. The scaled planar image (E) was created from image D and deformed using the following two scaling parameters: X, which is the baseline position to scale image D, and Y, which is the magnification value to scale image D below X. The two parameters (X and Y) were calculated to maximize the agreement between image C as a reference and image E using normalized mutual information (NMI) (Fig. 2 Step 3).
Step 4
The scaled SPECT image was created from the original SPECT image and deformed using the two parameters (X, Y) (Fig. 2 Step 4).
Assessment of conformity between lung perfusion SPECT and lung CT and statistical analysis
Lung perfusion SPECT images and chest CT images were analyzed using a workstation (SYNAPSE VINCENT; FUJIFILM Medical Co., Ltd, Tokyo, Japan). The lungs volume was calculated from the original and scaled lung perfusion SPECT. The scaled lung perfusion SPECT was obtained using the breathing synchronization software. The edge of the lung perfusion SPECT was determined at 5–20% of the maximum counts using a histogram (Fig. 3). Lung volume as the gold standard was defined as the lung volume seen on chest CT. All lung volumes were calculated using the workstation. We compared the lung volumes, including those calculated from the original lung perfusion SPECT and the lung perfusion SPECT using the breathing synchronization software (the scaled SPECT image in each cutoff of the maximum counts) and lung CT using Dunnett’s test. Visual conformity between the lung SPECT images obtained with and without the use of the software and the lung CT image was scored 0–4 (0: 0–25%, 1: 25–50%, 2: 50–75%, 3: 75–90%, 4: > 90%) by two specialists in nuclear medicine and assessed. The Mann–Whitney U test was used to compare the scores between the original lung perfusion SPECT and the lung perfusion SPECT performed using the breathing synchronization software, the scaled SPECT image.