Mediaboss Marketing

This retrospective study was approved by the Osaka University Clinical Research Review Committee, and the requirement for informed consent was waived by the Osaka University Clinical Research Review Committee. All methods were carried out in accordance with relevant guidelines and regulations. Patients who underwent CT for evaluation of aortic or cardiac disease were eligible for inclusion in this study because we obtained a single scan in one breath-hold from the supraclavicular area to the symphysis pubis in these patients, whereas separate scans were obtained for the chest and abdominopelvic regions in other patients. Enrolled were 73 consecutive patients who underwent CT between January and February 2019 or between December 2020 and January 2021 (50 men and 23 women; age range, 2591years; mean age, 72.9years). The clinical indications for CT in these patients are listed in Table 1.

CT was performed using a 160- or 320-slice CT scanner (Aquilion Precision, Canon Medical Systems, Otawara, Japan, n=34, or Aquilion ONE GENESIS Edition, Canon Medical Systems, n=39). A pre-contrast scan was performed in all patients from the supraclavicular area to the symphysis pubis during a single breath hold. Tube current was adjusted individually using an auto-exposure control technique with a standard deviation setting of 15. The remaining scan parameters were as follows: tube voltage, 120kVp; rotation time, 0.5s; helical pitch, 0.83. Although post-contrast scans were also acquired in 31 patients, only the pre-contrast images were used in this study.

From the raw data of each patient, two sets of axial images were reconstructed, with a slice thickness/interval of 4/4 and 1/1mm. A hybrid iterative reconstruction algorithm (AIDR 3D, Canon Medical Systems) with a weak strength setting was applied. The remaining reconstruction parameters were as follows: kernel, FC03; reconstruction field of view, 350mm (pixel size, 0.680.68mm).

VTS is a conditional-GAN based algorithm. Thick-slice images with slice thickness/intervals of 310mm were randomly simulated from real thin-slice images by down-sampling with Gaussian smoothing. A pair of original thin-slice images and simulated thick-slice images were used to train the VTS generator in the GAN framework (Fig.1). The generator is an encoder-decoder type architecture with skip connections inspired by U-Net to reconstruct high resolution images. The role of the discriminator is to enable the generator to output virtual thin-slice images that are hard to distinguish from real ones. Both the generator and the discriminator are composed of 3D Convolutional Neural Networks. The conditioning labels (e.g. slice interval) associated with input thick images are fed into the discriminator to improve the accuracies of super resolution. While generator training, L1 loss was calculated in addition to adversarial loss, to minimize the pixel-wise intensity difference between the original (ground truth) and the generated thin-slice images, as these should be as close as possible. VTS software is a function of the PACS viewer (SYNAPSE SAI Viewer Version 1.0, FUJIFILM, Tokyo, Japan), which has regulatory approval in Japan. The training CT data for this software contained CT images of various body parts (head, chest, abdomen, and legs) obtained with scanners of various manufacturers. Thus, the software can be applied to any part of the body. The generated VTS images were isotropic with voxel size of 111mm. The details of the VTS technique have been presented at a previous conference, and the manuscript is available for reference on the preprint server14. VTS software was applied to the 4-mm-thick data set of each patient to generate 1-mm-thick VTS images.

Adversarial training framework for thickthin slice translation of CT images.

Two radiologists familiar with abdominal radiology (9 and 6years experience) independently reviewed the sagittal images reformatted from 4-mm-thick images and the VTS images and evaluated the visibility of the intervertebral spaces in each of four regions: cervical, upper thoracic, lower thoracic, and lumbar spine. They reviewed these images on a commercially available workstation (SYNAPSE VINCENT version 5.3.001, FUJIFILM), and assigned a score using the following 4-point scale: 4, all intervertebral spaces are visible; 3, most intervertebral spaces are visible but some are unclear; 2, most intervertebral spaces are unclear; 1, no intervertebral spaces are visible. The radiologists were informed that the images for evaluation were either 4-mm-thick or VTS images, but were blinded to the patients identity, medical background, and the reconstruction protocol used.

Two radiologists familiar with abdominal radiology (16 and 9years experience), different to the radiologists who performed the qualitative assessment, independently measured the height of the first thoracic (Th1) and first lumbar (L1) vertebrae on sagittal reformatted images made from each of the 4-mm-thick, true 1-mm-thick, and VTS data sets. Height was measured at the anterior border of each of these vertebrae. The absolute values of the difference between the measured heights on the 4-mm-thick and true 1-mm-thick images (D1) were calculated, as well as the absolute values of the difference between the measured heights on VTS and true 1-mm-thick images (D2). The absolute percentage errors between the measured heights on the 4-mm-thick and true 1-mm-thick images (%Error1) was also calculated by dividing D1 by the measured height on true 1-mm-thick images, as well as the absolute percentage errors between the measured heights on VTS and true 1-mm-thick images (%Error2). Measurements were performed using a workstation (SYNAPSE VINCENT version 5.3.001).

The same two radiologists who performed the qualitative assessment also independently evaluated the possible presence of compression fracture using the sagittal reformatted images constructed from each of the 4-mm-thick images and the VTS images. They classified the likelihood of compression fracture in all vertebrae using the following 4-point confidence score scale: 1, probably no fracture present; 2, indefinite presence of fracture; 3, fracture probably present; and 4, fracture definitely present. Before the assessment, they were informed that a confidence level of 3 or 4 would be considered a positive finding for the calculation of sensitivity and positive predictive value (PPV). The criteria for compression fracture used in this study were: 1, ratio of the anterior height of the vertebra (AH) to the posterior height (PH) <0.75; 2, ratio of the central height of the vertebrae (CH) to AH or PH<0.8; 3, height of a vertebra reduced by >20% compared with those above and below15. The reference standard was determined by two other radiologists (16 and 9years experience) who evaluated the presence or absence of compression fracture on sagittal images reformatted from the true 1-mm-thick images, in consensus.

Visual scores regarding the visibility of intervertebral spaces were compared using Wilcoxon signed rank test. The absolute values of the difference in measured vertebral heights (D1 and D2) were compared using paired t-test. The absolute percentage errors of the measured vertebral heights (%Error1 and %Error2) were also compared using paired t-test. Interobserver agreement for each of D1 and D2 was evaluated by intraclass correlation coefficient (ICC). To analyze diagnostic performance for detecting compression fracture, jackknife free-response receiver-operating characteristic (JAFROC) analysis was performed using JAFROC software (JAFROC Version 4.2.1, http://www.devchakraborty.com). This software computes the figure of merit (FOM), which is defined as the probability that a lesion is rated higher than the highest rated non-lesion on a normal image16. In the present study, JAFROC1 was used rather than JAFROC or JAFROC2 because of its high statistical power for human observers17. For all tests, a P value less than 0.05 was considered significant.

Go here to read the rest:
Three-dimensional conditional generative adversarial network-based virtual thin-slice technique for the morphological evaluation of the spine |...