Search

• None of the four tested imaging platforms provided acceptable results in all patients. • Segmentation accuracy, quantitative results and qualitative score are related. • High-density lung regions account for the largest differences in quantitative metrics. • Lung segmentation accuracy impacts differently on different quantitative metrics. To assess the impact of lung segmentation accuracy in an automatic pipeline for quantitative analysis of CT images. Four different platforms for automatic lung segmentation based on convolutional neural network (CNN), region-growing technique and atlas-based algorithm were considered. The platforms were tested using CT images of 55 COVID-19 patients with severe lung impairment. Four radiologists assessed the segmentations using a 5-point qualitative score (QS). For each CT series, a manually revised reference segmentation (RS) was obtained. Histogram-based quantitative metrics (QM) were calculated from CT histogram using lung segmentationsfrom all platforms and RS. Dice index (DI) and differences of QMs (ΔQMs) were calculated between RS and other segmentations. Highest QS and lower ΔQMs values were associated to the CNN algorithm. However, only 45% CNN segmentations were judged to need no or only minimal corrections, and in only 17 cases (31%), automatic segmentations provided RS without manual corrections. Median values of the DI for the four algorithms ranged from 0.993 to 0.904. Significant differences for all QMs calculated between automatic segmentations and RS were found both when data were pooled together and stratified according to QS, indicating a relationship between qualitative and quantitative measurements. The most unstable QM was the histogram 90th percentile, with median ΔQMs values ranging from 10HU and 158HU between different algorithms. None of tested algorithms provided fully reliable segmentation. Segmentation accuracy impacts differently on different quantitative metrics, and each of them should be individually evaluated according to the purpose of subsequent analyses. [ABSTRACT FROM AUTHOR]

• A new approach has been proposed to overcome limitations of quantitative metrics calculated from lung CT images. • New metrics, derived from physics assumptions and with physiological significance, are introduced. • New metrics show lower dependencies from CT-number and reduced inter and intra patient variability. • Quantitative metrics of lung COVID-19 diseases can be described by subdividing the organ into 24 sub-regions. Quantitative metrics in lung computed tomography (CT) images have been widely used, often without a clear connection with physiology. This work proposes a patient-independent model for the estimation of well-aerated volume of lungs in CT images (WAVE). A Gaussian fit, with mean (Mu.f) and width (Sigma.f) values, was applied to the lower CT histogram data points of the lung to provide the estimation of the well-aerated lung volume (WAVE.f). Independence from CT reconstruction parameters and respiratory cycle was analysed using healthy lung CT images and 4DCT acquisitions. The Gaussian metrics and first order radiomic features calculated for a third cohort of COVID-19 patients were compared with those relative to healthy lungs. Each lung was further segmented in 24 subregions and a new biomarker derived from Gaussian fit parameter Mu.f was proposed to represent the local density changes. WAVE.f resulted independent from the respiratory motion in 80% of the cases. Differences of 1%, 2% and up to 14% resulted comparing a moderate iterative strength and FBP algorithm, 1 and 3 mm of slice thickness and different reconstruction kernel. Healthy subjects were significantly different from COVID-19 patients for all the metrics calculated. Graphical representation of the local biomarker provides spatial and quantitative information in a single 2D picture. Unlike other metrics based on fixed histogram thresholds, this model is able to consider the inter- and intra-subject variability. In addition, it defines a local biomarker to quantify the severity of the disease, independently of the observer. [ABSTRACT FROM AUTHOR]

Purpose Resolution modeling (RM) of PET systems has been introduced in iterative reconstruction algorithms for oncologic PET. The RM recovers the loss of resolution and reduces the associated partial volume effect. While these methods improved the observer performance, particularly in the detection of small and faint lesions, their impact on quantification accuracy still requires thorough investigation. The aim of this study was to characterize the performances of the RM algorithms under controlled conditions simulating a typical 18 F-FDG oncologic study, using an anthropomorphic phantom and selected physical figures of merit, used for image quantification. Methods Measurements were performed on Biograph HiREZ (B_HiREZ) and Discovery 710 (D_710) PET/CT scanners and reconstructions were performed using the standard iterative reconstructions and the RM algorithms associated to each scanner: TrueX and SharpIR, respectively. Results RM determined a significant improvement in contrast recovery for small targets (≤17 mm diameter) only for the D_710 scanner. The maximum standardized uptake value (SUV max ) increased when RM was applied using both scanners. The SUV max of small targets was on average lower with the B_HiREZ than with the D_710. Sharp IR improved the accuracy of SUV max determination, whilst TrueX showed an overestimation of SUV max for sphere dimensions greater than 22 mm. The goodness of fit of adaptive threshold algorithms worsened significantly when RM algorithms were employed for both scanners. Conclusions Differences in general quantitative performance were observed for the PET scanners analyzed. Segmentation of PET images using adaptive threshold algorithms should not be undertaken in conjunction with RM reconstructions. [ABSTRACT FROM AUTHOR]

Purpose The purpose of this study was to evaluate the accuracy of the probabilistic tracking method of diffusion tensor imaging fiber tractography (DTI-FT), compared to imaging-represented electrophysiology data in epileptic patients who underwent stereoelectroncephalography (SEEG) for neurosurgical planning. Methods We identified 17 epileptic patients (mean age 26 ± 11) who underwent DTI acquisition on a 1.5T Philips Achieva scanner by using echo planar imaging and a Receive Coil Head SENSE 8 channels. Sequence details include: 64 non collinear diffusion gradient directions, b-value = 1000 s/mm2, TR/TE = 6522/86 ms, FA = 90°, SENSE factor = 2 and voxel-size = 2 × 2 × 2.6 mm. In addition, a three-dimensional T1-weighted Fast Field Echo (FFE) image was acquired with a TR/TE = 7.22/3.24 ms and FA = 8° and voxel size = 0.46 × 0.46 × 0.90 mm. These data were obtained with other brain MR sequences, used for SEEG planning. Depending on electrodes location, we recorded tibial and median somatosensory-evoked potentials (SEP) in 11 patients and flash-visual evoked potential (VEP) in 12. The thalamocortical (THC) and optical radiation (OR) tracts were blindly reconstructed using a probabilistic algorithm implemented in FSL 5.0.6, using two ROIs corresponding to the sensory thalamus and white matter (WM) lying underneath the somatosensory cortex for the THC, and the lateral geniculate body and WM lying next to the calcarine cortex for the OR. The 3D tractographies were then compared to the position of SEEG electrodes contacts, on which evoked potentials (EPs) were recorded. Results Comparison between DTI-FT and EPs shown that a superposition of the neurophysiological recording contacts and the tracts resulted in 15 out of 17 patients (88%). In particular, we obtained correspondence in the reconstruction of 9 out of 11 THC and 8 out of 12 OR. Conclusions This study is a preliminary one, suggesting that EPs recorded during SEEG monitoring can be considered as a valuable tool to validate the probabilistic tracking method in the reconstruction of THC and OR tracts. [ABSTRACT FROM AUTHOR]