Real-Time Optimal Synthetic Inversion Recovery Image Selection (RT-OSIRIS) for Deep Brain Stimulation Targeting.

Deep brain stimulation (DBS) is a method of electrical neuromodulation used to treat a variety of neuropsychiatric conditions including essential tremor, Parkinson's disease, epilepsy, and obsessive-compulsive disorder. The procedure requires precise placement of electrodes such that the electrical contacts lie within or in close proximity to specific target nuclei and tracts located deep within the brain. DBS electrode trajectory planning has become increasingly dependent on direct targeting with the need for precise visualization of targets. MRI is the primary tool for direct visualization, and this has led to the development of numerous sequences to aid in visualization of different targets. Synthetic inversion recovery images, specified by an inversion time parameter, can be generated from T1 relaxation maps, and this represents a promising method for modifying the contrast of deep brain structures to accentuate target areas using a single acquisition. However, there is currently no accessible method for dynamically adjusting the inversion time parameter and observing the effects in real-time in order to choose the optimal value. In this work, we examine three different approaches to implementing an application for real-time optimal synthetic inversion recovery image selection and evaluate them based on their ability to display continually-updated synthetic inversion recovery images as the user modifies the inversion time parameter. These methods include continuously computing the inversion recovery equation at each voxel in the image volume, limiting the computation only to the voxels of the orthogonal slices currently displayed on screen, or using a series of lookup tables with precomputed solutions to the inversion recovery equation. We find the latter implementation provides for the quickest display updates both when modifying the inversion time and when scrolling through the image. We introduce a publicly available cross-platform application built around this conclusion. We also briefly discuss other details of the implementations and considerations for extensions to other use cases.

Enhancing outcomes in deep brain stimulation: a comparative study of direct targeting using 7T versus 3T MRI.

OBJECTIVE: The aim of this study was to compare outcomes of direct targeting in deep brain stimulation (DBS) for essential tremor using 7T MRI versus 3T MRI. The authors hypothesized that 7T MRI direct targeting would be noninferior to 3T MRI in early tremor outcomes. METHODS: A retrospective study was conducted on patients undergoing unilateral thalamic DBS for essential tremor between 2021 and 2023. Two matched cohorts were assessed, one using 7T MRI and the other using 3T MRI for surgical planning. The primary endpoint was the percentage improvement in the Fahn-Tolosa-Marin Tremor Rating Scale (TRS) scores. Additionally, the authors assessed optimized programming settings and variance in electrode position on postoperative imaging. Demographic and clinical data were compared using the nonparametric Mann-Whitney U-test. The squared Euclidean distance of each contact from the group mean centroid was calculated and averaged across the entire cohort to provide the variance (i.e., the mean squared distance) of electrode contact position. RESULTS: A total of 34 patients were analyzed, with 17 in each cohort. There were no significant differences in demographic information or mean surgical dates between the groups. There were no differences in intraoperative target repositioning or adverse events. The 7T group had a significantly greater TRS improvement than the 3T group (64.9% ± 11.4% vs 50.9% ± 16.4%, p = 0.004). Patients in the 7T cohort also had a lower mean stimulation current compared with those in the 3T cohort (2.0 ± 0.8 mA vs 2.7 ± 0.9 mA, p = 0.01). Image evaluation revealed that although the mean electrode position was comparable between 7T and 3T, the 7T electrode positioning was more clustered, indicating a lower variance in the final electrode location. The mean Euclidean distance between the individual electrode tips and the group centroid was significantly less at 7T than at 3T (1.82 ± 0.68 mm vs 2.75 ± 0.81 mm, p = 0.001). CONCLUSIONS: Despite concerns for increased artifacts and distortions at 7T, the authors show that these effects can be mitigated with an appropriate workflow, leading to improved surgical outcomes with direct targeting using 7T MRI. Their results suggest similar accuracy but greater precision in targeting with 7T MRI compared with 3T MRI, resulting in lower stimulation currents and improved tremor reduction. Future studies are needed to assess outcomes related to 7T MRI in targeting other subcortical structures.

7 T Lesion-Attenuated Magnetization-Prepared Gradient Echo Acquisition for Detection of Posterior Fossa Demyelinating Lesions in Multiple Sclerosis.

OBJECTIVES: Detection of infratentorial demyelinating lesions in multiple sclerosis (MS) presents a challenge in magnetic resonance imaging (MRI), a difficulty that is further heightened in 7 T MRI. This study aimed to assess the efficacy of a novel MRI approach, lesion-attenuated magnetization-prepared gradient echo acquisition (LAMA), for detecting demyelinating lesions within the posterior fossa and upper cervical spine on 7 T MRI and contrast its performance with conventional double-inversion recovery (DIR) and T2-weighted turbo spin echo sequences. MATERIALS AND METHODS: We conducted a retrospective cross-sectional study in 42 patients with a confirmed diagnosis of MS. All patients had 7 T MRI that incorporated LAMA, 3D DIR, and 2D T2-weighted turbo spin echo sequences. Three readers assessed lesion count in the brainstem, cerebellum, and upper cervical spinal cord using both DIR and T2-weighted images in one session. In a separate session, LAMA was analyzed alone. Contrast-to-noise ratio was also compared between LAMA and the conventional sequences. Lesion counts between methods were assessed using nonparametric Wilcoxon signed rank test. Interrater agreement in lesion detection was estimated by intraclass correlation coefficients. RESULTS: LAMA identified a significantly greater number of lesions than DIR + T2 (mean 6.4 vs 3.0; P < 0.001). LAMA also exhibited better interrater agreement (intraclass correlation coefficient [95% confidence interval], 0.75 [0.41-0.88] vs 0.61 [0.35-0.78]). The contrast-to-noise ratio for LAMA (3.7 ± 0.9) significantly exceeded that of DIR (1.94 ± 0.7) and T2 (1.2 ± 0.7) (all P's < 0.001). In cases with no lesions detected using DIR + T2, at least 1 lesion was identified in 83.3% with LAMA. Across all analyzed brain regions, LAMA consistently detected more lesions than DIR + T2. CONCLUSIONS: LAMA significantly improves the detection of infratentorial demyelinating lesions in MS patients compared with traditional methods. Integrating LAMA with standard magnetization-prepared 2 rapid acquisition gradient echo acquisition provides a valuable tool for accurately characterizing the extent of MS disease.

Diagnostic utility of 7T neuromelanin imaging of the substantia nigra in Parkinson's disease.

Parkinson's disease (PD) is a prevalent neurodegenerative disorder that presents a diagnostic challenge due to symptom overlap with other disorders. Neuromelanin (NM) imaging is a promising biomarker for PD, but adoption has been limited, in part due to subpar performance at standard MRI field strengths. We aimed to evaluate the diagnostic utility of ultra-high field 7T NM-sensitive imaging in the diagnosis of PD versus controls and essential tremor (ET), as well as NM differences among PD subtypes. A retrospective case-control study was conducted including PD patients, ET patients, and controls. 7T NM-sensitive 3D-GRE was acquired, and substantia nigra pars compacta (SNpc) volumes, contrast ratios, and asymmetry indices were calculated. Statistical analyses, including general linear models and ROC curves, were employed. Twenty-one PD patients, 13 ET patients, and 18 controls were assessed. PD patients exhibited significantly lower SNpc volumes compared to non-PD subjects. SNpc total volume showed 100% sensitivity and 96.8% specificity (AUC = 0.998) for differentiating PD from non-PD and 100% sensitivity and 95.2% specificity (AUC = 0.996) in differentiating PD from ET. Contrast ratio was not significantly different between PD and non-PD groups (p = 0.07). There was also significantly higher asymmetry index in SNpc volume in PD compared to non-PD cohorts (p < 0.001). NM signal loss in PD predominantly involved the inferior, posterior, and lateral aspects of SNpc. Akinetic-rigid subtype showed more significant NM signal loss compared to tremor dominant subtype (p < 0.001). 7T NM imaging demonstrates potential as a diagnostic tool for PD, including potential distinction between subtypes, allowing improved understanding of disease progression and subtype-related characteristics.

Synthetic Inversion Image Generation using MP2RAGE T1 Mapping for Surgical Targeting in Deep Brain Stimulation and Lesioning.

BACKGROUND: Advances in MRI technology have increased interest in direct targeting for deep brain stimulation (DBS). Various imaging sequences have been shown to provide increased contrast of numerous common DBS targets, such as T1-weighted, Fast Gray Matter Acquisition T1 Inversion Recovery (FGATIR), gray matter nulled, and Edge-Enhancing Gradient Echo (EDGE); however, the continual increase in the number of necessary sequences has led to an increase in imaging time, which is undesirable. Additionally, carefully timed inversion pulses can often lead to less-than-ideal contrast in some subjects, particularly in ultra-high field MRI, where B1+ field inhomogeneity can lead to substantial contrast variation. OBJECTIVES: This study proposes using 3D MP2RAGE-based T1 maps to retrospectively synthesize images of any desired inversion time, including T1-weighted, FGATIR, and EDGE contrasts, to visualize specific DBS targets at both 3T and 7T. METHOD: First, a systematic sequence optimization framework was applied to optimize MP2RAGE T1 mapping sequence parameters for the purpose of DBS planning. Next, we show that synthetic inversion-time images can be generated through a mathematical transformation of the T1 maps. The sequence was then applied to patients undergoing preoperative planning for DBS at 3T and 7T to generate synthetic contrasts used in surgical planning. RESULTS: We show that synthetic image contrasts can be generated across a full range of inversion times at 3T and 7T, including commonly used sequences for DBS targeting, such as T1-weighted, FGATIR, and EDGE. Acquisition through a single sequence shortens scan time compared to acquiring the sequences independently without affecting image quality or contrast. CONCLUSIONS: The generation of synthetic images for DBS targeting allows faster acquisition of many key sequences, as well as the ability to optimize contrast properties post-acquisition to account for the variable B1+ effects present in ultra-high field MRI. The proposed approach has the potential to reduce imaging time and improve the accuracy of DBS targeting at 1.5T, 3T, and 7T.

7T MRI in transient ischemic attacks: Have we only seen the tip of the iceberg?

Transient ischemic attack (TIA) has gained significant attention recently due to the increased incidence of subsequent stroke. However, there are many nonvascular clinical mimics of TIA, creating a need for improved biomarkers to identify a vascular origin. Following the recent approval of ultra-high field (UHF) 7T MRI in clinical practice, several clinical studies have highlighted its added utility in neuroimaging compared to lower-field 1.5T and 3T MRI, particularly in epilepsy and multiple sclerosis. Our case series of three patients with TIA illustrates that 7T MRI can depict small areas of intracortical microhemorrhages and microinfarctions, which could not be resolved with 3T or 1.5T MRI. There are currently no reports of intracortical localization of microhemorrhages in patients with TIA. This discovery may enhance our understanding and characterization of cerebrovascular abnormalities in TIAs. In addition, UHF imaging could potentially be utilized to distinguish transient neurological episodes secondary to cerebrovascular events from other differential considerations. Our cases highlight the underestimation of imaging abnormalities in cases of TIA and support the potential expanded application of clinical 7T to assess patients with TIA. Future studies are necessary at 7T redundant to determine the true incidence of such lesions in TIA and to examine the correlation between cortical microhemorrhages and subsequent ischemic stroke, hemorrhagic events, and neurocognitive impairment.

Optimization of MP2RAGE T1 mapping with radial view-ordering for deep brain stimulation targeting at 7 T MRI.

PURPOSE: Deep brain stimulation (DBS) is an effective treatment of various neurological disorders. Due to higher intrinsic signal, 7 T MRI can potentially improve delineation of DBS targets. However, the severe RF transmit field (B1+) inhomogeneity at 7 T can compromise the image contrast of traditional single-contrast sequences for DBS targeting, leading to sub-optimal target visualization. The Magnetization Prepared 2 Rapid Acquisition Gradient Echo (MP2RAGE)-based T1 mapping provides an alternative to the traditional single-contrast techniques by allowing retrospective synthesis of images at arbitrary inversion times to aid in visualization of various DBS targets. With this approach, optimization of sequence parameters to create T1 maps with low noise and low quantification bias is critical, as these characteristics directly affect the noise and uniformity of the synthetic images. In this work, we perform sequence optimization for MP2RAGE-based T1 mapping using a radial view-ordering technique to improve image quality, and demonstrate the clinical utility of T1 mapping approach for DBS targeting. METHODS: We first introduce a systematic sequence optimization framework for 7 T MP2RAGE T1 mapping by formulating it into a constrained, multi-dimensional optimization process considering the effect of B1+ inhomogeneity on image noise, T1 quantification bias, and image blurring. With this framework, we investigate the use of radial view-order approach for T1 mapping, in lieu of the conventional linear view-ordering. Bloch's equation-based simulations were performed to compare the T1 maps generated using different approaches. Images of healthy volunteer and patients were acquired on a clinical 7 T MRI scanner for validation and to demonstrate the utility of T1 mapping for DBS targeting. RESULTS: Numerical experiments demonstrated that the proposed framework allowed optimization of image SNR in T1 maps while controlling the quantification bias and image blurring, therefore facilitating the selection of optimal sequence parameters for visualizing DBS targets. The optimized sequence using radial view-ordering offered 40-60% noise reduction compared to the linear view-ordering. The improvement of SNR was confirmed in the in vivo examples. Clinical images showed that the synthetic images generated from the optimized T1 maps allowed clear visualization of DBS targets. CONCLUSION: We demonstrated the optimization of MP2RAGE T1 mapping with radial view-ordering technique for DBS targeting at 7 T and showed that the optimized sequence allows retrospective generation of synthetic inversion time images commonly utilized in DBS targeting, such as fast gray matter acquisition T1 inversion recovery (FGATIR) and edge-enhancing gradient echo (EDGE) sequences.

Measurement of enhanced vasa vasorum density in a porcine carotid model using photon counting detector CT.

Purpose: The onset of atherosclerosis is preceded by changes in blood perfusion within the arterial wall due to localized proliferation of the vasa vasorum. The purpose of this study was to quantify these changes in spatial density of the vasa vasorum using a research whole-body photon-counting detector CT (PCD-CT) scanner and a porcine model. Approach: Vasa vasorum angiogenesis was stimulated in the left carotid artery wall of anesthetized pigs ( n = 5 ) while the right carotid served as a control. After a 6-week recovery period, the animals were scanned on the PCD-CT prior to and after injection of iodinated contrast. Annular regions of interest were used to measure wall enhancement in the injured and control arteries. The exact Wilcoxon-signed rank test was used to determine whether a significant difference in contrast enhancement existed between the injured and control arterial walls. Results: The greatest arterial wall enhancement was observed following contrast recirculation. The wall enhancement measurements made over these time points revealed that the enhancement was greater in the injured artery for 13/16 scanned arterial regions. Using an exact Wilcoxon-signed rank test, a significantly increased enhancement ratio was found in injured arteries compared with control arteries ( p = 0.013 ). Vasa vasorum angiogenesis was confirmed in micro-CT scans of excised arteries. Conclusions: Whole-body PCD-CT scanners can be used to detect and quantify the increased perfusion occurring within the porcine carotid arterial wall resulting from an increased density of vasa vasorum.

Clinical application of ultra-high resolution compressed sensing time-of-flight MR angiography at 7T to detect small vessel pathology.

3D time-of-flight (TOF) MR angiography (MRA) benefits from ultra-high-field MRI (≥7 T) due to improved contrast and increased signal-to-noise ratio. However, high-resolution TOF MRA at 7T usually requires longer acquisition times. In addition, relatively higher specific absorption rate (SAR) at 7T limits the choice of optimal pulse sequence parameters, especially if venous saturation is employed. Here, we illustrate the clinical application of ultra-high resolution cerebral 7T TOF MRA using compressed sensing in cases of artery of Percheron and lacunar infarcts, which showed superior resolution and exquisite details pertinent to the clinical diagnosis. The technical challenges associated with high-resolution 7T imaging were alleviated by optimization of sequence parameters and utilization of compressed sensing acceleration.

Central Vein Sign in Multiple Sclerosis: A Comparison Study of the Diagnostic Performance of 3T versus 7T MRI.

BACKGROUND AND PURPOSE: An early and accurate diagnosis of multiple sclerosis remains challenging in clinical neurology. Established diagnostic methods have less than desirable sensitivity and specificity. An accurate, noninvasive diagnostic test for MS could have a major impact on diagnostic criteria. We compared the frequency of detection of the central vein sign (CVS) in white matter lesions of MS and controls on 7T T2*-weighted and SWI to 3T SWI. Additionally, we assessed the diagnostic performance of 7T T2*, 7T SWI, and 3T SWI for MS. MATERIALS AND METHODS: A retrospective case-control study was performed of patients with MS having both 7T MRI and 3T MRI. A control group of patients without MS was selected. Diagnosis of MS was established by board-certified neurologists with fellowship training in autoimmune neurology in line with the 2017 McDonald criteria. Percentage of lesions with a CVS was blindly measured for each technique. Diagnostic performance was computed by sensitivity, specificity, and positive and negative likelihood ratios (LRs). RESULTS: Sixty-one patients with MS (903 lesions) and 39 controls (1088 lesions) were included. 7T T2* showed significantly more CVS (87%) than both 7T SWI (73%) and 3T SWI (31%) (all P < .001). CVS was identified in the control group in ≤6% of lesions on all sequences. Using a threshold of >40% of lesions with CVS on 7T T2* and >15% on 7T SWI, both sequences had an accuracy = 100%, sensitivity = 100%, specificity = 100%, infinite positive LR, and zero negative LR. Using an optimal threshold of >12%, 3T SWI had an accuracy = 96.0%, sensitivity = 93.4%, specificity = 100%, infinite positive LR, and negative LR = 0.066. CONCLUSIONS: 7T MRI had 100% sensitivity and specificity for the diagnosis of MS and is superior to 3T. Future revisions to MS diagnostic criteria may consider recommendations for 7T MRI and inclusion of CVS as a biomarker.

Technical note: Evaluation of artificial 120-kilovolt computed tomography images for radiation therapy applications.

PURPOSE: The purpose of this work is to evaluate the scaled computed tomography (CT) number accuracy of an artificial 120 kV reconstruction technique based on phantom experiments in the context of radiation therapy planning. METHODS: An abdomen-shaped electron density phantom was scanned on a clinical CT scanner capable of artificial 120 kV reconstruction using different tube potentials from 70 to 150 kV. A series of tissue-equivalent phantom inserts (lung, adipose, breast, solid water, liver, inner bone, 30%/50% CaCO3 , cortical bone) were placed inside the phantom. Images were reconstructed using a conventional quantitative reconstruction kernel as well as the artificial 120 kV reconstruction kernel. Scaled CT numbers of inserts were measured from images acquired at different kVs and compared with those acquired at 120 kV, which were deemed as the ground truth. The relative error was quantified as the percentage deviation of scaled CT numbers acquired at different tube potentials from their ground truth values acquired at 120 kV. RESULTS: Scaled CT numbers measured from images reconstructed using the conventional reconstruction demonstrated a strong kV-dependence. The relative error in scaled CT numbers ranged from 0.6% (liver insert) to 31.1% (cortical bone insert). The artificial 120 kV reconstruction reduced the kV dependence, especially for bone tissues. The relative error in scaled CT number was reduced to 0.4% (liver insert) and 2.6% (30% CaCO3 insert) using this technique. When tube potential selection was limited to the range of 90 to 150 kV, the relative error was further restrained to <1.2% for all tissue types. CONCLUSION: Phantom results demonstrated that using the artificial 120 kV technique, it was feasible to acquire raw projection data at the desired tube potential and then reconstruct images with scaled CT numbers comparable to those obtained directly at 120 kV. In radiotherapy applications, this technique may allow optimization of tube potential without complicating clinical workflow by eliminating the necessity of maintaining multiple sets of CT calibration curves.

Optimization of fast gray matter acquisition T1 inversion recovery (FGATIR) on 7T MRI for deep brain stimulation targeting.

Deep brain stimulation (DBS) is an increasingly utilized treatment for multiple neurological disorders. Continued improvements in DBS outcome are, in part, related to increasing ability to directly visualize stimulation targets by MRI. However, it is challenging to image DBS targets with conventional MRI techniques due to limited contrast. Fast Gray Matter Acquisition T1 Inversion Recovery (FGATIR) is a commonly used MRI sequence that improves visualization of several key DBS targets by suppressing white matter (WM) signal to better reveal deep-brain gray matter (GM) structures. Due to increased signal level at high field strength, application of FGATIR on 7T MRI may allow higher spatial resolution and better DBS targeting accuracy. However, successful utilization of FGATIR requires meticulous sequence optimization involving multiple parameters to maximize GM signal while suppressing WM. This is further complicated by the transmit RF field (B1+) inhomogeneity on 7T, which can cause severe contrast degradation. In this work, we introduce a systematic approach to optimize FGATIR and to improve visualization of thalamic DBS targets on 7T. FGATIR optimization is cast into a constrained optimization problem whose objective function and constraints are designed to maximize the GM-WM contrast-to-noise ratio (CNR) while accounting for B1+ inhomogeneity. This approach allows a systematic search for optimal parameters across the multi-dimensional parametric space while limiting the negative effect of B1+ variation. Bloch equation simulations were performed to solve the proposed optimization problem and to compare the sequence derived from this method against the sequence optimized without considering B1+ inhomogeneity. The results showed that this approach can improve GM-WM CNR in the presence of B1+ inhomogeneity, especially in some high relative B1+ areas where several key thalamic DBS targets are located. Additionally, in vivo images were acquired on a clinical 7T MRI to further validate this approach. Severe contrast degradation in the thalamus was observed when B1+ effect was not considered in sequence optimization, while the proposed approach yielded improved image contrast in the thalamus with key DBS targets well-defined. These results demonstrated that the proposed method allowed optimization of FGATIR on 7T to better visualize thalamic DBS targets, which may lead to improved DBS targeting accuracy as well as treatment outcome.

Photon Counting CT: Clinical Applications and Future Developments.

The use of a photon counting detector in CT (PCD CT) is currently the subject of intense investigation and development. In this review article, we will describe potential clinical applications of this technology with a particular focus on the experience of our own institution with a prototype PCD CT scanner. PCDs have three primary advantages over conventional, energy integrating detectors (EIDs): they provide spectral information without need for a dedicated dual energy protocol; they are immune to electronic noise; and they can be made very high resolution without significant compromises to quantum efficiency. These advantages translate into several clinical applications. Metal artifacts, beam hardening artifacts, and noise streaks from photon starvation can be better mitigated using PCD CT. Certain incidental findings can be better characterized using the spectral information from PCD CT. High-contrast, high-resolution structures such as the temporal bone can be better visualized using PCD CT and at greatly reduced dose. We also discuss new possibilities on the horizon, including new contrast agents, and how anticipated improvements in PCD CT will translate to performance in these applications.

Implementation and experimental evaluation of Mega-voltage fan-beam CT using a linear accelerator.

BACKGROUND: Mega-voltage fan-beam Computed Tomography (MV-FBCT) holds potential in accurate determination of relative electron density (RED) and proton stopping power ratio (SPR) but is not widely available. OBJECTIVE: To demonstrate the feasibility of MV-FBCT using a medical linear accelerator (LINAC) with a 2.5 MV imaging beam, an electronic portal imaging device (EPID) and multileaf collimators (MLCs). METHODS: MLCs were used to collimate MV beam along z direction to enable a 1 cm width fan-beam. Projection data were acquired within one gantry rotation and preprocessed with in-house developed artifact correction algorithms before the reconstruction. MV-FBCT data were acquired at two dose levels: 30 and 60 monitor units (MUs). A Catphan 604 phantom was used to evaluate basic image quality. A head-sized CIRS phantom with three configurations of tissue-mimicking inserts was scanned and MV-FBCT Hounsfield unit (HU) to RED calibration was established for each insert configuration using linear regression. The determination coefficient ([Formula: see text]) was used to gauge the accuracy of HU-RED calibration. Results were compared with baseline single-energy kilo-voltage treatment planning CT (TP-CT) HU-RED calibration which represented the current standard clinical practice. RESULTS: The in-house artifact correction algorithms effectively suppressed ring artifact, cupping artifact, and CT number bias in MV-FBCT. Compared to TP-CT, MV-FBCT was able to improve the prediction accuracy of the HU-RED calibration curve for all three configurations of insert materials, with [Formula: see text] > 0.9994 and [Formula: see text] < 0.9990 for MV-FBCT and TP-CT HU-RED calibration curves of soft-tissue inserts, respectively. The measured mean CT numbers of blood-iodine mixture inserts in TP-CT drastically deviated from the fitted values but not in MV-FBCT. Reducing the radiation level from 60 to 30 MU did not decrease the prediction accuracy of the MV-FBCT HU-RED calibration curve. CONCLUSION: We demonstrated the feasibility of MV-FBCT and its potential in providing more accurate RED estimation.

Improved coronary calcification quantification using photon-counting-detector CT: an ex vivo study in cadaveric specimens.

OBJECTIVES: To compare the accuracy of coronary calcium quantification of cadaveric specimens imaged from a photon-counting detector (PCD)-CT and an energy-integrating detector (EID)-CT. METHODS: Excised coronary specimens were scanned on a PCD-CT scanner, using both the PCD and EID subsystems. The scanning and reconstruction parameters for EID-CT and PCD-CT were matched: 120 kV, 9.3-9.4 mGy CTDIvol, and a quantitative kernel (D50). PCD-CT images were also reconstructed using a sharper kernel (D60). Scanning the same specimens using micro-CT served as a reference standard for calcified volumes. Calcifications were segmented with a half-maximum thresholding technique. Segmented calcified volume differences were analyzed using the Friedman test and post hoc pairwise Wilcoxon signed rank test with the Bonferroni correction. Image noise measurements were compared between EID-CT and PCD-CT with a repeated-measures ANOVA test and post hoc pairwise comparison with the Bonferroni correction. A p < 0.05 was considered statistically significant. RESULTS: The volume measurements in 12/13 calcifications followed a similar trend: EID-D50 > PCD-D50 > PCD-D60 > micro-CT. The median calcified volumes in EID-D50, PCD-D50, PCD-D60, and micro-CT were 22.1 (IQR 10.2-64.8), 21.0 (IQR 9.0-56.5), 18.2 (IQR 8.3-49.3), and 14.6 (IQR 5.1-42.4) mm3, respectively (p < 0.05 for all pairwise comparisons). The average image noise in EID-D50, PCD-D50, and PCD-D60 was 60.4 (± 3.5), 56.0 (± 4.2), and 113.6 (± 8.5) HU, respectively (p < 0.01 for all pairwise comparisons). CONCLUSION: The PCT-CT system quantified coronary calcifications more accurately than EID-CT, and a sharp PCD-CT kernel further improved the accuracy. The PCD-CT images exhibited lower noise than the EID-CT images. KEY POINTS: • High spatial resolution offered by PCD-CT reduces partial volume averaging and consequently leads to better morphological depiction of coronary calcifications. • Improved quantitative accuracy for coronary calcification volumes could be achieved using high-resolution PCD-CT compared to conventional EID-CT. • PCD-CT images exhibit lower image noise than conventional EID-CT at matched radiation dose and reconstruction kernel.

Technical Note: kV-independent coronary calcium scoring: A phantom evaluation of score accuracy and potential radiation dose reduction.

PURPOSE: To determine the accuracy of CT number and calcium score of a kV-independent technique based on an artificial 120 kV reconstruction, and its potential to reduce radiation dose. METHODS: Anthropomorphic chest phantoms were scanned on a third-generation dual-source CT system equipped with the artificial 120 kV reconstruction. First, a phantom module containing a 20-mm diameter hydroxyapatite (HA) insert was scanned inside the chest phantoms at different tube potentials (70-140 kV) to evaluate calcium CT number accuracy. Next, three small HA inserts (diameter/length = 5 mm) were inserted into a pork steak and scanned inside the phantoms to evaluate calcium score accuracy at different kVs. Finally, the same setup was scanned using automatic exposure control (AEC) at 120 kV, and then with automatic kV selection (auto-kV). Phantoms were also scanned at 120 kV using a size-dependent mA chart. CT numbers of soft tissue and calcium were measured from different kV images. Calcium score of each small HA insert was measured using commercial software. RESULTS: The CT number difference from 120 kV was small with tube potentials from 90 to 140 kV for both soft tissue and calcium (maximal difference of 4/5 HU, respectively). Consistent calcium scores were obtained from images of different kVs compared to 120 kV, with a relative difference <8%. Auto-kV provided a 25-34% dose reduction compared to AEC alone. CONCLUSION: A kV-independent calcium scoring technique can produce artificial 120 kV images with consistent soft tissue and calcium CT numbers compared to standard 120 kV examinations. When coupled with auto-kV, this technique can reduce radiation by 25-34% compared to that with AEC alone, while providing consistent calcium scores as that of standard 120 kV examinations.

Deep-learning-based direct inversion for material decomposition.

PURPOSE: To develop a convolutional neural network (CNN) that can directly estimate material density distribution from multi-energy computed tomography (CT) images without performing conventional material decomposition. METHODS: The proposed CNN (denoted as Incept-net) followed the general framework of encoder-decoder network, with an assumption that local image information was sufficient for modeling the nonlinear physical process of multi-energy CT. Incept-net was implemented with a customized loss function, including an in-house-designed image-gradient-correlation (IGC) regularizer to improve edge preservation. The network consisted of two types of customized multibranch modules exploiting multiscale feature representation to improve the robustness over local image noise and artifacts. Inserts with various densities of different materials [hydroxyapatite (HA), iodine, a blood-iodine mixture, and fat] were scanned using a research photon-counting detector (PCD) CT with two energy thresholds and multiple radiation dose levels. The network was trained using phantom image patches only, and tested with different-configurations of full field-of-view phantom and in vivo porcine images. Furthermore, the nominal mass densities of insert materials were used as the labels in CNN training, which potentially provided an implicit mass conservation constraint. The Incept-net performance was evaluated in terms of image noise, detail preservation, and quantitative accuracy. Its performance was also compared to common material decomposition algorithms including least-square-based material decomposition (LS-MD), total-variation regularized material decomposition (TV-MD), and U-net-based method. RESULTS: Incept-net improved accuracy of the predicted mass density of basis materials compared with the U-net, TV-MD, and LS-MD: the mean absolute error (MAE) of iodine was 0.66, 1.0, 1.33, and 1.57 mgI/cc for Incept-net, U-net, TV-MD, and LS-MD, respectively, across all iodine-present inserts (2.0-24.0 mgI/cc). With the LS-MD as the baseline, Incept-net and U-net achieved comparable noise reduction (both around 95%), both higher than TV-MD (85%). The proposed IGC regularizer effectively helped both Incept-net and U-net to reduce image artifact. Incept-net closely conserved the total mass densities (i.e., mass conservation constraint) in porcine images, which heuristically validated the quantitative accuracy of its outputs in anatomical background. In general, Incept-net performance was less dependent on radiation dose levels than the two conventional methods; with approximately 40% less parameters, the Incept-net achieved relatively improved performance than the comparator U-net, indicating that performance gain by Incept-net was not achieved by simply increasing network learning capacity. CONCLUSION: Incept-net demonstrated superior qualitative image appearance, quantitative accuracy, and lower noise than the conventional methods and less sensitive to dose change. Incept-net generalized and performed well with unseen image structures and different material mass densities. This study provided preliminary evidence that the proposed CNN may be used to improve the material decomposition quality in multi-energy CT.

Noise reduction in CT image using prior knowledge aware iterative denoising.

The clinical demand for low image noise often limits the slice thickness used in many CT applications. However, a thick-slice image is more susceptible to longitudinal partial volume effects, which can blur key anatomic structures and pathologies of interest. In this work, we develop a prior knowledge aware iterative denoising (PKAID) framework that utilizes spatial data redundancy in the slice increment direction to generate low-noise, thin-slice images, and demonstrate its application in non-contrast head CT exams. The proposed technique takes advantage of the low noise of thicker images and exploits the structural similarity between the thick- and thin-slice images to reduce noise in the thin-slice image. Phantom data and patient cases (n= 3) of head CT were used to assess performance of this method. Images were reconstructed at clinically utilized slice thickness (5 mm) and thinner slice thickness (2 mm). PKAID was used to reduce image noise in 2 mm images using the 5 mm images as low-noise prior. Noise amplitude, noise power spectra (NPS), modulation transfer function (MTF), and slice sensitivity profiles (SSPs) of images before/after denoising were analyzed. The NPS and MTF analysis showed that PKAID preserved noise texture and resolution of the original thin-slice image, while reducing noise to the level of thick-slice image. The SSP analysis showed that the slice thickness of the original thin-slice image was retained. Patient examples demonstrated that PKAID-processed, thin-slice images better delineated brain structures and key pathologies such as subdural hematoma compared to the clinical 5 mm images, while additionally reducing image noise. To test an alternative PKAID utilization for dose reduction, a head exam with 40% dose reduction was simulated using projection-domain noise insertion. The image of 5 mm slice thickness was then denoised using PKAID. The results showed that the PKAID-processed reduced-dose images maintained similar noise and image quality compared to the full-dose images.

Multi-energy CT imaging for large patients using dual-source photon-counting detector CT.

Multi-energy CT imaging of large patients with conventional dual-energy (DE)-CT using an energy-integrating-detector (EID) is challenging due to photon starvation-induced image artifacts, especially in lower tube potential (80-100 kV) images. Here, we performed phantom experiments to investigate the performance of DECT for morbidly obese patients, using an iodine and water material decomposition task as an example, on an emulated dual-source (DS)-photon-counting-detector (PCD)-CT, and compared its performance with a clinical DS-EID-CT. An abdominal CT phantom with iodine inserts of different concentrations was wrapped with tissue-equivalent gel layers to emulate a large patient (50 cm lateral size). The phantom was scanned on a research whole-body single-source (SS)-PCD-CT (140 kV tube potential), a DS-PCD-CT (100/Sn140 kV; Sn140 indicates 140 kV with Sn filter), and a clinical DS-EID-CT (100/Sn140 kV) with the same radiation dose. Phantom scans were repeated five times on each system. The DS-PCD-CT acquisition was emulated by scanning twice on the SS-PCD-CT using different tube potentials. The multi-energy CT images acquired on each system were then reconstructed, and iodine- and water-specific images were generated using material decomposition. The root-mean-square-error (RMSE) between true and measured iodine concentrations were calculated for each system and compared. The images acquired on the DS-EID-CT showed severe artifacts, including ringing, reduced uniformity, and photon starvation artifacts, especially for low-energy images. These were largely reduced in DS-PCD-CT images. The CT number difference that was measured using regions-of-interest across field-of-view were reduced from 20.3 ± 0.9 (DS-EID-CT) to 2.5 ± 0.4 HU on DS-PCD-CT, showing improved image uniformity using DS-PCD-CT. Iodine RMSE was reduced from 3.42 ± 0.03 mg ml-1 (SS-PCD-CT) and 2.90 ± 0.03 mg ml-1 (DS-EID-CT) to 2.39 ± 0.05 mg ml-1 using DS-PCD-CT. DS-PCD-CT out-performed a clinical DS-EID-CT for iodine and water-based material decomposition on phantom emulating obese patients by reducing image artifacts and improving iodine quantification (RMSE reduced by 20%). With DS-PCD-CT, multi-energy CT can be performed on large patients that cannot be accommodated with current DECT.