The future of CT: deep learning reconstruction

New liver window width in detecting hepatocellular carcinoma on dynamic contrast-enhanced computed tomography with deep learning reconstruction

Changing a window width (WW) alters appearance of noise and contrast of CT images. The aim of this study was to investigate the impact of adjusted WW for deep learning reconstruction (DLR) in detecting hepatocellular carcinomas (HCCs) on CT with DLR. This retrospective study included thirty-five patients who underwent abdominal dynamic contrast-enhanced CT. DLR was used to reconstruct arterial, portal, and delayed phase images. The investigation of the optimal WW involved two blinded readers. Then, five other blinded readers independently read the image sets for detection of HCCs and evaluation of image quality with optimal or conventional liver WW. The optimal WW for detection of HCC was 119 (rounded to 120 in the subsequent analyses) Hounsfield unit (HU), which was the average of adjusted WW in the arterial, portal, and delayed phases. The average figures of merit for the readers for the jackknife alternative free-response receiver operating characteristic analysis to detect HCC were 0.809 (reader 1/2/3/4/5, 0.765/0.798/0.892/0.764/0.827) in the optimal WW (120 HU) and 0.765 (reader 1/2/3/4/5, 0.707/0.769/0.838/0.720/0.791) in the conventional WW (150 HU), and statistically significant difference was observed between them (p < 0.001). Image quality in the optimal WW was superior to those in the conventional WW, and significant difference was seen for some readers (p < 0.041). The optimal WW for detection of HCC was narrower than conventional WW on dynamic contrast-enhanced CT with DLR. Compared with the conventional liver WW, optimal liver WW significantly improved detection performance of HCC.

Impact of deep learning image reconstruction on volumetric accuracy and image quality of pulmonary nodules with different morphologies in low-dose CT

BACKGROUND

This study systematically compares the impact of innovative deep learning image reconstruction (DLIR, TrueFidelity) to conventionally used iterative reconstruction (IR) on nodule volumetry and subjective image quality (IQ) at highly reduced radiation doses. This is essential in the context of low-dose CT lung cancer screening where accurate volumetry and characterization of pulmonary nodules in repeated CT scanning are indispensable.

MATERIALS AND METHODS

A standardized CT dataset was established using an anthropomorphic chest phantom (Lungman, Kyoto Kaguku Inc., Kyoto, Japan) containing a set of 3D-printed lung nodules including six diameters (4 to 9 mm) and three morphology classes (lobular, spiculated, smooth), with an established ground truth. Images were acquired at varying radiation doses (6.04, 3.03, 1.54, 0.77, 0.41 and 0.20 mGy) and reconstructed with combinations of reconstruction kernels (soft and hard kernel) and reconstruction algorithms (ASIR-V and DLIR at low, medium and high strength). Semi-automatic volumetry measurements and subjective image quality scores recorded by five radiologists were analyzed with multiple linear regression and mixed-effect ordinal logistic regression models.

RESULTS

Volumetric errors of nodules imaged with DLIR are up to 50% lower compared to ASIR-V, especially at radiation doses below 1 mGy and when reconstructed with a hard kernel. Also, across all nodule diameters and morphologies, volumetric errors are commonly lower with DLIR. Furthermore, DLIR renders higher subjective IQ, especially at the sub-mGy doses. Radiologists were up to nine times more likely to score the highest IQ-score to these images compared to those reconstructed with ASIR-V. Lung nodules with irregular margins and small diameters also had an increased likelihood (up to five times more likely) to be ascribed the best IQ scores when reconstructed with DLIR.

CONCLUSION

We observed that DLIR performs as good as or even outperforms conventionally used reconstruction algorithms in terms of volumetric accuracy and subjective IQ of nodules in an anthropomorphic chest phantom. As such, DLIR potentially allows to lower the radiation dose to participants of lung cancer screening without compromising accurate measurement and characterization of lung nodules.

Comparison of iterative reconstruction implementations for multislice helical CT

The most mature image reconstruction algorithms in multislice helical computed tomography are based on analytical and iterative methods. Over the past decades, several methods have been developed for iterative reconstructions that improve image quality by reducing noise and artifacts. In the regularization step of iterative reconstruction, noise can be significantly reduced, thereby making low-dose CT. The quality of the reconstructed image can be further improved by using model-based reconstructions. In these reconstructions, the main focus is on modeling the data acquisition process, including the behavior of the photon beams, the geometry of the system, etc. In this article, we propose two model-based reconstruction algorithms using a virtual detector for multislice helical CT. The aim of this study is to compare the effect of using a virtual detector on image quality for the two proposed algorithms with a model-based iterative reconstruction using the original detector model. Since the algorithms are implemented using multiple GPUs, the merging of separately reconstructed volumes can significantly affect image quality. This issue is often referred to as the "long object" problem, for which we also present a solution that plays an important role in the proposed reconstruction processes. The algorithms were evaluated using mathematical and physical phantoms, as well as patient cases. The SSIM, MS-SSIM and L1 metrics were utilized to evaluate the image quality of the mathematical phantom case. To demonstrate the effectiveness of the algorithms, we used the CatPhan 600 phantom. Additionally, anonymized patient scans were used to showcase the improvements in image quality on real scan data.

Quantitative and qualitative evaluation of three MSCT for high resolution bone imaging

INTRODUCTION

Strategies for achieving high resolution varies between manufacturers. In CT, the helical mode with narrow collimation has long been considered as the gold standard for high-resolution imaging. More recently, incremental modes with small dexels and focal spot, have been developed but have not been compared with helical acquisitions under optimal conditions. The aim of this work is to compare the high-resolution acquisition strategies currently proposed by recent MSCT.

METHODS

Three CT systems were compared. A phantom was used to evaluate geometric accuracy, uniformity, scan slice geometry, and spatial resolution. Human dry bones were used to test different protocols on real bone architecture. A blind visual analysis was conducted by trained CT users for classifying the different acquisitions (p-values).

RESULTS

All systems give satisfactory results in terms of geometric accuracy and uniformity. The in-plane MTF at 5% were respectively 13.4, 15.9 and 18.1 lp/cm. Dry-bones evaluation confirms that acquisition#3 is considered as the best.

CONCLUSIONS

The incremental acquisition coupled with à small focal spot, and a high-sampling detector, overpasses the reference of low-pitch helical acquisitions for high-resolution imaging. Cortical bone, bony vessels, and tumoral matrix analysis are the very next challenges that will have to be managed to improve normal and pathologic bone imaging thanks to the availability UHR-CT systems.

How auto-differentiation can improve CT workflows: classical algorithms in a modern framework

Many of the recent successes of deep learning-based approaches have been enabled by a framework of flexible, composable computational blocks with their parameters adjusted through an automatic differentiation mechanism to implement various data processing tasks. In this work, we explore how the same philosophy can be applied to existing "classical" (i.e., non-learning) algorithms, focusing on computed tomography (CT) as application field. We apply four key design principles of this approach for CT workflow design: end-to-end optimization, explicit quality criteria, declarative algorithm construction by building the forward model, and use of existing classical algorithms as computational blocks. Through four case studies, we demonstrate that auto-differentiation is remarkably effective beyond the boundaries of neural-network training, extending to CT workflows containing varied combinations of classical and machine learning algorithms.

A CT deep learning reconstruction algorithm: Image quality evaluation for brain protocol at decreasing dose indexes in comparison with FBP and statistical iterative reconstruction algorithms

PURPOSE

To characterise the impact of Precise Image (PI) deep learning reconstruction algorithm on image quality, compared to filtered back-projection (FBP) and iDose4 iterative reconstruction for brain computed tomography (CT) phantom images.

METHODS

Catphan-600 phantom was acquired with an Incisive CT scanner using a dedicated brain protocol, at six different dose levels (volume computed tomography dose index (CTDIvol): 7/14/29/49/56/67 mGy). Images were reconstructed using FBP, levels 2/5 of iDose4, and PI algorithm (Sharper/Sharp/Standard/Smooth/Smoother). Image quality was assessed by evaluating CT numbers, image histograms, noise, image non-uniformity (NU), noise power spectrum, target transfer function, and detectability index.

RESULTS

The five PI levels did not significantly affect the mean CT number. For a given CTDIvol using Sharper-to-Smoother levels, the spatial resolution for all the investigated materials and the detectability index increased while the noise magnitude decreased, slightly affecting noise texture. For a fixed PI level increasing the CTDIvol the detectability index increased, the noise magnitude decreased. From 29 mGy, NU values converged within 1 Hounsfield Unit from each other without a substantial improvement at higher CTDIvol values.

CONCLUSIONS

The improved performances of intermediate PI levels in brain protocols compared to conventional algorithms seem to suggest a potential reduction of CTDIvol.

Ultra-low dose chest CT with silver filter and deep learning reconstruction significantly reduces radiation dose and retains quantitative information in the investigation and monitoring of lymphangioleiomyomatosis (LAM)

PURPOSE

Frequent CT scans to quantify lung involvement in cystic lung disease increases radiation exposure. Beam shaping energy filters can optimize imaging properties at lower radiation dosages. The aim of this study is to investigate whether use of SilverBeam filter and deep learning reconstruction algorithm allows for reduced radiation dose chest CT scanning in patients with lymphangioleiomyomatosis (LAM).

MATERIAL AND METHODS

In a single-center prospective study, 60 consecutive patients with LAM underwent chest CT at standard and ultra-low radiation doses. Standard dose scan was performed with standard copper filter and ultra-low dose scan was performed with SilverBeam filter. Scans were reconstructed using a soft tissue kernel with deep learning reconstruction (AiCE) technique and using a soft tissue kernel with hybrid iterative reconstruction (AIDR3D). Cyst scores were quantified by semi-automated software. Signal-to-noise ratio (SNR) was calculated for each reconstruction. Data were analyzed by linear correlation, paired t-test, and Bland-Altman plots.

RESULTS

Patients averaged 49.4 years and 100% were female with mean BMI 26.6 ± 6.1 kg/m2. Cyst score measured by AiCE reconstruction with SilverBeam filter correlated well with that of AIDR3D reconstruction with standard filter, with a 1.5% difference, and allowed for an 85.5% median radiation dosage reduction (0.33 mSv vs. 2.27 mSv, respectively, p < 0.001). Compared to standard filter with AIDR3D, SNR for SilverBeam AiCE images was slightly lower (3.2 vs. 3.1, respectively, p = 0.005).

CONCLUSION

SilverBeam filter with deep learning reconstruction reduces radiation dosage of chest CT, while maintaining accuracy of cyst quantification as well as image quality in cystic lung disease.

CLINICAL RELEVANCE STATEMENT

Radiation dosage from chest CT can be significantly reduced without sacrificing image quality by using silver filter in combination with a deep learning reconstructive algorithm.

KEY POINTS

• Deep learning reconstruction in chest CT had no significant effect on cyst quantification when compared to conventional hybrid iterative reconstruction. • SilverBeam filter reduced radiation dosage by 85.5% compared to standard dose chest CT. • SilverBeam filter in coordination with deep learning reconstruction maintained image quality and diagnostic accuracy for cyst quantification when compared to standard dose CT with hybrid iterative reconstruction.

Pneumonia detection based on RSNA dataset and anchor-free deep learning detector

Pneumonia is a highly lethal disease, and research on its treatment and early screening tools has received extensive attention from researchers. Due to the maturity and cost reduction of chest X-ray technology, and with the development of artificial intelligence technology, pneumonia identification based on deep learning and chest X-ray has attracted attention from all over the world. Although the feature extraction capability of deep learning is strong, existing deep learning object detection frameworks are based on pre-defined anchors, which require a lot of tuning and experience to guarantee their excellent results in the face of new applications or data. To avoid the influence of anchor settings in pneumonia detection, this paper proposes an anchor-free object detection framework and RSNA dataset based on pneumonia detection. First, a data enhancement scheme is used to preprocess the chest X-ray images; second, an anchor-free object detection framework is used for pneumonia detection, which contains a feature pyramid, two-branch detection head, and focal loss. The average precision of 51.5 obtained by Intersection over Union (IoU) calculation shows that the pneumonia detection results obtained in this paper can surpass the existing classical object detection framework, providing an idea for future research and exploration.

A Systematic Literature Review of 3D Deep Learning Techniques in Computed Tomography Reconstruction

Computed tomography (CT) is used in a wide range of medical imaging diagnoses. However, the reconstruction of CT images from raw projection data is inherently complex and is subject to artifacts and noise, which compromises image quality and accuracy. In order to address these challenges, deep learning developments have the potential to improve the reconstruction of computed tomography images. In this regard, our research aim is to determine the techniques that are used for 3D deep learning in CT reconstruction and to identify the training and validation datasets that are accessible. This research was performed on five databases. After a careful assessment of each record based on the objective and scope of the study, we selected 60 research articles for this review. This systematic literature review revealed that convolutional neural networks (CNNs), 3D convolutional neural networks (3D CNNs), and deep learning reconstruction (DLR) were the most suitable deep learning algorithms for CT reconstruction. Additionally, two major datasets appropriate for training and developing deep learning systems were identified: 2016 NIH-AAPM-Mayo and MSCT. These datasets are important resources for the creation and assessment of CT reconstruction models. According to the results, 3D deep learning may increase the effectiveness of CT image reconstruction, boost image quality, and lower radiation exposure. By using these deep learning approaches, CT image reconstruction may be made more precise and effective, improving patient outcomes, diagnostic accuracy, and healthcare system productivity.

Review of Deep Learning Approaches for Interleaved Photoacoustic and Ultrasound (PAUS) Imaging

Photoacoustic (PA) imaging provides optical contrast at relatively large depths within the human body, compared to other optical methods, at ultrasound (US) spatial resolution. By integrating real-time PA and US (PAUS) modalities, PAUS imaging has the potential to become a routine clinical modality bringing the molecular sensitivity of optics to medical US imaging. For applications where the full capabilities of clinical US scanners must be maintained in PAUS, conventional limited view and bandwidth transducers must be used. This approach, however, cannot provide high-quality maps of PA sources, especially vascular structures. Deep learning (DL) using data-driven modeling with minimal human design has been very effective in medical imaging, medical data analysis, and disease diagnosis, and has the potential to overcome many of the technical limitations of current PAUS imaging systems. The primary purpose of this article is to summarize the background and current status of DL applications in PAUS imaging. It also looks beyond current approaches to identify remaining challenges and opportunities for robust translation of PAUS technologies to the clinic.

EANM procedural recommendations for managing the paediatric patient in diagnostic nuclear medicine

PURPOSE

The manuscript aims to characterize the principles of best practice in performing nuclear medicine procedures in paediatric patients. The paper describes all necessary technical skills that should be developed by the healthcare professionals to ensure the best possible care in paediatric patients, as it is particularly challenging due to psychological and physical conditions of children.

METHODS

We performed a comprehensive literature review to establish the most relevant elements of nuclear medicine studies in paediatric patients. We focused the attention to the technical aspects of the study, such as patient preparation, imaging protocols, and immobilization techniques, that adhere to best practice principles. Furthermore, we considered the psychological elements of working with children, including comforting and distraction strategies.

RESULTS

The extensive literature review combined with practical conclusions and recommendations presented and explained by the authors summarizes the most important principles of the care for paediatric patient in the nuclear medicine field.

CONCLUSION

Nuclear medicine applied to the paediatric patient is a very special and challenging area, requiring proper education and experience in order to be performed at the highest level and with the maximum safety for the child.

Prediction of Anemia From Cerebral Venous Sinus Attenuation on Deep-Learning Reconstructed Brain Computed Tomography Images

OBJECTIVE

The aim of the study is to evaluate whether the prediction of anemia is possible using quantitative analyses of unenhanced cranial computed tomography (CT) with deep learning reconstruction (DLR) compared with conventional methods.

METHODS

This cross-sectional retrospective study included 116 participants (76 males; mean age, 66.7) who had hemoglobin (Hb) levels obtained within 24 hours of unenhanced cranial CT, which included 2 reconstruction methods: DLR and hybrid iterative reconstruction. Regions of interest were the confluence of sinuses (CoS) and the right and left transverse sinuses. In addition, edge rise distance of cerebrospinal fluid and venous was measured.

RESULTS

Spearman rank correlation coefficient demonstrated a positive association between Hb levels and sinus attenuation values. Among these, the CoS in DLR had the best correlation ( r = 0.703, P < 0.001). For the prediction of anemia (Hb < 11 g/dL), the area under the curve of CoS in DLR (area under the curve = 0.874; 95% confidence interval, 0.798-0.949; P < 0.001) was the highest; however, there were no significant differences among reconstruction method and sinus. The attenuation values of DLR were significantly higher than those of hybrid iterative reconstruction ( P < 0.001, paired t test), and the differences between the 2 methods were 4.1 (standard deviation [SD], 1.6) for CoS, 5.2 (SD, 2.2) for right transverse sinuses, and 5.8 (SD, 2.4) for left transverse sinuses. The signal-to-noise ratio ( P < 0.001, paired t test) and edge rise distance ( P < 0.001, Wilcoxon signed rank test) of DLR was significantly higher.

CONCLUSIONS

Higher CT attenuation values should be considered for predicting anemia based on brain DLR images.

Image Quality Improvement in Deep Learning Image Reconstruction of Head Computed Tomography Examination

In this study, we assess image quality in computed tomography scans reconstructed via DLIR (Deep Learning Image Reconstruction) and compare it with iterative reconstruction ASIR-V (Adaptive Statistical Iterative Reconstruction) in CT (computed tomography) scans of the head. The CT scans of 109 patients were subjected to both objective and subjective evaluation of image quality. The objective evaluation was based on the SNR (signal-to-noise ratio) and CNR (contrast-to-noise ratio) of the brain's gray and white matter. The regions of interest for our study were set in the BGA (basal ganglia area) and PCF (posterior cranial fossa). Simultaneously, a subjective assessment of image quality, based on brain structure visibility, was conducted by experienced radiologists. In the assessed scans, we obtained up to a 54% increase in SNR for gray matter and a 60% increase for white matter using DLIR in comparison to ASIR-V. Moreover, we achieved a CNR increment of 58% in the BGA structures and 50% in the PCF. In the subjective assessment of the obtained images, DLIR had a mean rating score of 2.8, compared to the mean score of 2.6 for ASIR-V images. In conclusion, DLIR shows improved image quality compared to the standard iterative reconstruction of CT images of the head.

Computed Tomography 2.0: New Detector Technology, AI, and Other Developments

ABSTRACT

Computed tomography (CT) dramatically improved the capabilities of diagnostic and interventional radiology. Starting in the early 1970s, this imaging modality is still evolving, although tremendous improvements in scan speed, volume coverage, spatial and soft tissue resolution, as well as dose reduction have been achieved. Tube current modulation, automated exposure control, anatomy-based tube voltage (kV) selection, advanced x-ray beam filtration, and iterative image reconstruction techniques improved image quality and decreased radiation exposure. Cardiac imaging triggered the demand for high temporal resolution, volume acquisition, and high pitch modes with electrocardiogram synchronization. Plaque imaging in cardiac CT as well as lung and bone imaging demand for high spatial resolution. Today, we see a transition of photon-counting detectors from experimental and research prototype setups into commercially available systems integrated in patient care. Moreover, with respect to CT technology and CT image formation, artificial intelligence is increasingly used in patient positioning, protocol adjustment, and image reconstruction, but also in image preprocessing or postprocessing. The aim of this article is to give an overview of the technical specifications of up-to-date available whole-body and dedicated CT systems, as well as hardware and software innovations for CT systems in the near future.

Deep Learning Convolutional Neural Network Reconstruction and Radial k-Space Acquisition MR Technique for Enhanced Detection of Retropatellar Cartilage Lesions of the Knee Joint

OBJECTIVES

To assess diagnostic performance of standard radial k-space (PROPELLER) MRI sequences and compare with accelerated acquisitions combined with a deep learning-based convolutional neural network (DL-CNN) reconstruction for evaluation of the knee joint.

METHODS

Thirty-five patients undergoing MR imaging of the knee at 1.5 T were prospectively included. Two readers evaluated image quality and diagnostic confidence of standard and DL-CNN accelerated PROPELLER MR sequences using a four-point Likert scale. Pathological findings of bone, cartilage, cruciate and collateral ligaments, menisci, and joint space were analyzed. Inter-reader agreement (IRA) for image quality and diagnostic confidence was assessed using intraclass coefficients (ICC). Cohen's Kappa method was used for evaluation of IRA and consensus between sequences in assessing different structures. In addition, image quality was quantitatively evaluated by signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) measurements.

RESULTS

Mean acquisition time of standard vs. DL-CNN sequences was 10 min 3 s vs. 4 min 45 s. DL-CNN sequences showed significantly superior image quality and diagnostic confidence compared to standard MR sequences. There was moderate and good IRA for assessment of image quality in standard and DL-CNN sequences with ICC of 0.524 and 0.830, respectively. Pathological findings of the knee joint could be equally well detected in both sequences (κ-value of 0.8). Retropatellar cartilage could be significantly better assessed on DL-CNN sequences. SNR and CNR was significantly higher for DL-CNN sequences (both p < 0.05).

CONCLUSIONS

In MR imaging of the knee, DL-CNN sequences showed significantly higher image quality and diagnostic confidence compared to standard PROPELLER sequences, while reducing acquisition time substantially. Both sequences perform comparably in the detection of knee-joint pathologies, while DL-CNN sequences are superior for evaluation of retropatellar cartilage lesions.

Artificial intelligence in coronary artery calcium measurement: Barriers and solutions for implementation into daily practice

Coronary artery calcification (CAC) measurement is a valuable predictor of cardiovascular risk. However, its measurement can be time-consuming and complex, thus driving the desire for artificial intelligence (AI)-based approaches. The aim of this review is to explore the current status of CAC volume measurement using AI-based systems for the automated prediction of cardiovascular events. We also make proposals for the implementation of these systems into clinical practice. Research to date on applying AI to CAC scoring has shown the potential for automation and risk stratification, and, overall, efficacy and a high level of agreement with categorisation by trained clinicians have been demonstrated. However, research in this field has not been uniform or directed. One contributing factor may be a lack of integration and communication between computer scientists and cardiologists. Clinicians, institutions, and organisations should work together towards applying this technology to improve processes, preserve healthcare resources, and improve patient outcomes.

Advanced Neuroimaging With Photon-Counting Detector CT

ABSTRACT

Photon-counting detector computed tomography (PCD-CT) is an emerging technology and promises the next step in CT evolution. Photon-counting detectors count the number of individual incoming photons and assess the energy level of each of them. These mechanisms differ substantially from conventional energy-integrating detectors. The new technique has several advantages, including lower radiation exposure, higher spatial resolution, reconstruction of images with less beam-hardening artifacts, and advanced opportunities for spectral imaging. Research PCD-CT systems have already demonstrated promising results, and recently, the first whole-body full field-of-view PCD-CT scanners became clinically available. Based on published studies of preclinical systems and the first experience with clinically approved scanners, the performance can be translated to valuable neuroimaging applications, including brain imaging, intracranial and extracranial CT angiographies, or head and neck imaging with detailed assessment of the temporal bone. In this review, we will provide an overview of the current status in neuroimaging with upcoming and potential clinical applications.

Feasibility of dynamic chest radiography to calculate lung volumes in adult people with cystic fibrosis: a pilot study

INTRODUCTION

Dynamic chest radiography (DCR) is a novel, low-dose, real-time digital imaging system where software identifies moving thoracic structures and can automatically calculate lung areas. In an observational, prospective, non-controlled, single-centre pilot study, we compared it with whole-body plethysmography (WBP) in the measurement of lung volume subdivisions in people with cystic fibrosis (pwCF).

METHODS

Lung volume subdivisions were estimated by DCR using projected lung area (PLA) during deep inspiration, tidal breathing and full expiration, and compared with same-day WBP in 20 adult pwCF attending routine review. Linear regression models to predict lung volumes from PLA were developed.

RESULTS

Total lung area (PLA at maximum inspiration) correlated with total lung capacity (TLC) (r=0.78, p<0.001), functional residual lung area with functional residual capacity (FRC) (r=0.91, p<0.001), residual lung area with residual volume (RV) (r=0.82, p=0.001) and inspiratory lung area with inspiratory capacity (r=0.72, p=0.001). Despite the small sample size, accurate models were developed for predicting TLC, RV and FRC.

CONCLUSION

DCR is a promising new technology that can be used to estimate lung volume subdivisions. Plausible correlations between plethysmographic lung volumes and DCR lung areas were identified. Further studies are needed to build on this exploratory work in both pwCF and individuals without CF.

TRIAL REGISTRATION NUMBER

ISRCTN64994816.

Artificial Intelligence in CT and MR Imaging for Oncological Applications

Cancer care increasingly relies on imaging for patient management. The two most common cross-sectional imaging modalities in oncology are computed tomography (CT) and magnetic resonance imaging (MRI), which provide high-resolution anatomic and physiological imaging. Herewith is a summary of recent applications of rapidly advancing artificial intelligence (AI) in CT and MRI oncological imaging that addresses the benefits and challenges of the resultant opportunities with examples. Major challenges remain, such as how best to integrate AI developments into clinical radiology practice, the vigorous assessment of quantitative CT and MR imaging data accuracy, and reliability for clinical utility and research integrity in oncology. Such challenges necessitate an evaluation of the robustness of imaging biomarkers to be included in AI developments, a culture of data sharing, and the cooperation of knowledgeable academics with vendor scientists and companies operating in radiology and oncology fields. Herein, we will illustrate a few challenges and solutions of these efforts using novel methods for synthesizing different contrast modality images, auto-segmentation, and image reconstruction with examples from lung CT as well as abdome, pelvis, and head and neck MRI. The imaging community must embrace the need for quantitative CT and MRI metrics beyond lesion size measurement. AI methods for the extraction and longitudinal tracking of imaging metrics from registered lesions and understanding the tumor environment will be invaluable for interpreting disease status and treatment efficacy. This is an exciting time to work together to move the imaging field forward with narrow AI-specific tasks. New AI developments using CT and MRI datasets will be used to improve the personalized management of cancer patients.

"Image quality evaluation of the Precise image CT deep learning reconstruction algorithm compared to Filtered Back-projection and iDose4: a phantom study at different dose levels"

PURPOSE

To characterize the performance of the Precise Image (PI) deep learning reconstruction (DLR) algorithm for abdominal Computed Tomography (CT) imaging.

METHODS

CT images of the Catphan-600 phantom (equipped with an external annulus) were acquired using an abdominal protocol at four dose levels and reconstructed using FBP, iDose⁴ (levels 2,5) and PI ('Soft Tissue' definition, levels 'Sharper','Sharp','Standard','Smooth','Smoother'). Image noise, image non-uniformity, noise power spectrum (NPS), target transfer function (TTF), detectability index (d'), CT numbers accuracy and image histograms were analyzed.

RESULTS

The behavior of the PI algorithm depended strongly on the selected level of reconstruction. The phantom analysis suggested that the PI image noise decreased linearly by varying the level of reconstruction from Sharper to Smoother, expressing a noise reduction up to 80% with respect to FBP. Additionally, the non-uniformity decreased, the histograms became narrower, and d' values increased as PI reconstruction levels changed from Sharper to Smoother. PI had no significant impact on the average CT number of different contrast objects. The conventional FBP NPS was deeply altered only by Smooth and Smoother levels of reconstruction. Furthermore, spatial resolution was found to be dose- and contrast-dependent, but in each analyzed condition it was greater than or comparable to FBP and iDose⁴ TTFs.

CONCLUSIONS

The PI algorithm can reduce image noise with respect to FBP and iDose⁴; spatial resolution, CT numbers and image uniformity are generally preserved by the algorithm but changes in NPS for the Smooth and Smoother levels need to be considered in protocols implementation.

Application of the Model of Spots for Inverse Problems

This article proposes the application of a new mathematical model of spots for solving inverse problems using a learning method, which is similar to using deep learning. In general, the spots represent vague figures in abstract "information spaces" or crisp figures with a lack of information about their shapes. However, crisp figures are regarded as a special and limiting case of spots. A basic mathematical apparatus, based on L4 numbers, has been developed for the representation and processing of qualitative information of elementary spatial relations between spots. Moreover, we defined L4 vectors, L4 matrices, and mathematical operations on them. The developed apparatus can be used in Artificial Intelligence, in particular, for knowledge representation and for modeling qualitative reasoning and learning. Another application area is the solution of inverse problems by learning. For example, this can be applied to image reconstruction using ultrasound, X-ray, magnetic resonance, or radar scan data. The introduced apparatus was verified by solving problems of reconstruction of images, utilizing only qualitative data of its elementary relations with some scanning figures. This article also demonstrates the application of a spot-based inverse Radon algorithm for binary image reconstruction. In both cases, the spot-based algorithms have demonstrated an effective denoising property.

From 2D projections to the 3D rotation matrix: an attempt for finding a machine learning approach for the efficient evaluation of mechanical joining elements in X-ray computed tomography volume data

Destructive and predominantly manual procedures are commonly used in the automotive industry for the testing of mechanical joints, such as rivets or screws. Combining X-ray computed tomography (CT) and machine learning (ML) bears the potential of a non-destructive and largely automated methodology. Assuming the desired result is a comprehensible and documentable evaluation, three basic steps need to be automatized: First, a joint must be detected and identified as such in a CT scan of the joined parts. Second, the detected region containing the joint is rotated to a predefined orientation. Third, key measures in cross-sections from the newly oriented joint are dimensioned and documented. This work deals only with the second step, the rotation. On the one hand, we present a methodology for creating a well-curated data set for the contextual machine learning application. On the other, we evaluate its performance on the well-known ResNet50. More concretely, we investigate if it is possible for a deep convolutional neural network (CNN) to learn the respective rotation matrix from three volume projections that are perpendicular to each other. Two scenarios are investigated: In one scenario we assume that future data that is presented to the network has similar rivet demographics to historic data. We therefore do not employ hold-out sets for the network evaluation. In the other scenario we assume the opposite and therefore evaluating the networks performance with hold-out sets. We show that from a machine learning point of view, a CNN like ResNet50 is well able to learn this relationship with acceptable accuracy. In most cases the validation loss dropped below 0.1 after only a couple of epochs. In one particular case, we even reached both mean and median errors lower than 0.2 for approximately 80% of the entire test set of 1600 examples using our methodology. From an application point of view, however, these low test set errors should be treated with caution since small deviations from the intended rotation matrix can cause volume warping and translation. In another case, in which we used a hold-out set, only a fraction of the median errors were below 0.2.

A novel denoising method for CT images based on U-net and multi-attention

Reducing the radiation dose may lead to increased noise in medical computed tomography (CT), which can adversely affect the radiologists' judgment. Many efforts have been devoted to the denoising of low-dose CT (LDCT) images. However, it is often observed that denoised medical images usually lose some important clinical lesion edge information and may affect doctors' clinical diagnosis. For a denoising neural network, it is expected that the neural network can well retain the detailed features and make the network more anthropomorphic, and to simulate the attention mechanism of observation, being a valuable feature of the thinking process of human brain. Based on U-network (U-Net) and multi-attention mechanism, a novel denoising method for medical CT images is proposed in this study. To obtain different feature information in CT images, three attention modules are proposed in our method. The local attention module is developed to localize the surrounding information of the feature map and calculate each pixel from the context extracted from the feature map. The multi-feature channel attention module can automatically learn and extract features, suppress some invalid information and add different weights to each channel in the feature map according to different tasks. The hierarchical attention module allows the deep neural network to extract a large amount of feature information. This study also introduces an enhanced learning module to learn and retain the detail information in the image by stacking multi-layer convolution layer, batch normalization (BN) layer and activation function layer to increase the network depth. Experimental studies are conducted, and comparisons with the state-of-the-art networks are made, and the results demonstrate that the developed method can effectively remove the noise in CT images and improve the image quality in the evaluation metrics of peak signal to noise ratio (PSNR) and structural similarity (SSIM). Our method achieved 34.7329 of PSNR and 0.9293 of SSIM for σ = 10 on the QIN_LUNG_CT dataset, and achieved 28.9163 of PSNR and 0.8602 of SSIM on the Mayo Clinic LDCT Grand Challenge dataset.

Artificial Intelligence in Emergency Radiology: Where Are We Going?

Emergency Radiology is a unique branch of imaging, as rapidity in the diagnosis and management of different pathologies is essential to saving patients' lives. Artificial Intelligence (AI) has many potential applications in emergency radiology: firstly, image acquisition can be facilitated by reducing acquisition times through automatic positioning and minimizing artifacts with AI-based reconstruction systems to optimize image quality, even in critical patients; secondly, it enables an efficient workflow (AI algorithms integrated with RIS-PACS workflow), by analyzing the characteristics and images of patients, detecting high-priority examinations and patients with emergent critical findings. Different machine and deep learning algorithms have been trained for the automated detection of different types of emergency disorders (e.g., intracranial hemorrhage, bone fractures, pneumonia), to help radiologists to detect relevant findings. AI-based smart reporting, summarizing patients' clinical data, and analyzing the grading of the imaging abnormalities, can provide an objective indicator of the disease's severity, resulting in quick and optimized treatment planning. In this review, we provide an overview of the different AI tools available in emergency radiology, to keep radiologists up to date on the current technological evolution in this field.

A Review of Deep Learning CT Reconstruction: Concepts, Limitations, and Promise in Clinical Practice

Purpose of Review

Deep Learning reconstruction (DLR) is the current state-of-the-art method for CT image formation. Comparisons to existing filter back-projection, iterative, and model-based reconstructions are now available in the literature. This review summarizes the prior reconstruction methods, introduces DLR, and then reviews recent findings from DLR from a physics and clinical perspective.

Recent Findings

DLR has been shown to allow for noise magnitude reductions relative to filtered back-projection without suffering from “plastic” or “blotchy” noise texture that was found objectionable with most iterative and model-based solutions. Clinically, early reader studies have reported increases in subjective quality scores and studies have successfully implemented DLR-enabled dose reductions.

Summary

The future of CT image reconstruction is bright; deep learning methods have only started to tackle problems in this space via addressing noise reduction. Artifact mitigation and spectral applications likely be future candidates for DLR applications.

Innovations in thoracic imaging: CT, radiomics, AI and x-ray velocimetry

In recent years, pulmonary imaging has seen enormous progress, with the introduction, validation and implementation of new hardware and software. There is a general trend from mere visual evaluation of radiological images to quantification of abnormalities and biomarkers, and assessment of ‘non visual’ markers that contribute to establishing diagnosis or prognosis. Important catalysts to these developments in thoracic imaging include new indications (like computed tomography [CT] lung cancer screening) and the COVID‐19 pandemic. This review focuses on developments in CT, radiomics, artificial intelligence (AI) and x‐ray velocimetry for imaging of the lungs. Recent developments in CT include the potential for ultra‐low‐dose CT imaging for lung nodules, and the advent of a new generation of CT systems based on photon‐counting detector technology. Radiomics has demonstrated potential towards predictive and prognostic tasks particularly in lung cancer, previously not achievable by visual inspection by radiologists, exploiting high dimensional patterns (mostly texture related) on medical imaging data. Deep learning technology has revolutionized the field of AI and as a result, performance of AI algorithms is approaching human performance for an increasing number of specific tasks. X‐ray velocimetry integrates x‐ray (fluoroscopic) imaging with unique image processing to produce quantitative four dimensional measurement of lung tissue motion, and accurate calculations of lung ventilation.

See relatedEditorial

Can deep learning improve image quality of low-dose CT: a prospective study in interstitial lung disease

OBJECTIVES

To investigate whether deep learning reconstruction (DLR) could keep image quality and reduce radiation dose in interstitial lung disease (ILD) patients compared with HRCT reconstructed with hybrid iterative reconstruction (hybrid-IR).

METHODS

Seventy ILD patients were prospectively enrolled and underwent HRCT (120 kVp, automatic tube current) and LDCT (120 kVp, 30 mAs) scans. HRCT images were reconstructed with hybrid-IR (Adaptive Iterative Dose Reduction 3-Dimensional [AIDR3D], standard-setting); LDCT images were reconstructed with DLR (Advanced Intelligence Clear-IQ Engine [AiCE], lung/bone, mild/standard/strong setting). Image noise, streak artifact, overall image quality, and visualization of normal and abnormal features of ILD were evaluated.

RESULTS

The mean radiation dose of LDCT was 38% of HRCT. Objective image noise of reconstructed LDCT images was 33.6 to 111.3% of HRCT, and signal-to-noise ratio (SNR) was 0.9 to 3.1 times of the latter (p < 0.001). LDCT-AiCE was not significantly different from or even better than HRCT in overall image quality and visualization of normal lung structures. LDCT-AiCE (lung, mild/standard/strong) showed progressively better recognition of ground glass opacity than HRCT-AIDR3D (p < 0.05, p < 0.01, p < 0.001), and LDCT-AiCE (lung, mild/standard/strong; bone, mild) was superior to HRCT-AIDR3D in visualization of architectural distortion (p < 0.01, p < 0.01, p < 0.01; p < 0.05). LDCT-AiCE (bone, strong) was better than HRCT-AIDR3D in the assessment of bronchiectasis and/or bronchiolectasis (p < 0.05). LDCT-AiCE (bone, mild/standard/strong) was significantly better at the visualization of honeycombing than HRCT-AIDR3D (p < 0.05, p < 0.05, p < 0.01).

CONCLUSION

Deep learning reconstruction could effectively reduce radiation dose and keep image quality in ILD patients compared to HRCT with hybrid-IR.

KEY POINTS

• Deep learning reconstruction was a novel image reconstruction algorithm based on deep convolutional neural networks. It was applied in chest CT studies and received auspicious results. • HRCT plays an essential role in the whole process of diagnosis, treatment efficacy evaluation, and follow-ups for interstitial lung disease patients. However, cumulative radiation exposure could increase the risks of cancer. • Deep learning reconstruction method could effectively reduce the radiation dose and keep the image quality compared with HRCT reconstructed with hybrid iterative reconstruction in patients with interstitial lung disease.

Conversion from dose-length product to effective dose in computed tomography venography of the lower extremities

For radiation dose assessment of computed tomography (CT), effective dose (ED) is often estimated by multiplying the dose-length product (DLP), provided automatically by the CT scanner, by a conversion factor. We investigated such conversion in CT venography of the lower extremities performed in conjunction with CT pulmonary angiography. The study subjects consisted of eight groups imaged using different scanners and different imaging conditions (five and three groups for the GE and Siemens scanners, respectively). Each group included ten men and ten women. The scan range was divided into four anatomical regions (trunk, proximal thigh, knee and distal leg), and DLP was calculated for each region (regional DLP). Regional DLP was multiplied by a conversion factor for the respective region, to convert it to ED. The sum of the ED values for the four regions was obtained as standard ED. Additionally, the sum of the four regional DLP values, an approximate of the scanner-derived DLP, was multiplied by the conversion factor for the trunk (0.015 mSv mGy cm^-1), as a simplified method to obtain ED. When using the simplified method, ED was overestimated by 32.3%-70.2% and 56.5%-66.2% for the GE and Siemens scanners, respectively. The degree of overestimation was positively and closely correlated with the contribution of the middle and distal portions of the lower extremities to total radiation exposure. ED/DLP averaged within each group, corresponding to the conversion factor, was 0.0089-0.0114 and 0.0091-0.0096 mSv mGy cm^-1for the GE and Siemens scanners, respectively. In CT venography of the lower extremities, ED is greatly overestimated by multiplying the scanner-derived DLP by the conversion factor for the trunk. The degree of overestimation varies widely depending on the imaging conditions. It is recommended to divide the scan range and calculate ED as a sum of regional ED values.

Impact of deep learning-based image reconstruction on image quality compared with adaptive statistical iterative reconstruction-Veo in renal and adrenal computed tomography

OBJECTIVE

To evaluate image quality of deep learning-based image reconstruction (DLIR) in contrast-enhanced renal and adrenal computed tomography (CT) compared with adaptive statistical iterative reconstruction-Veo (ASiR-V).

METHODS

We prospectively recruited 52 patients. All images were reconstructed with ASiR-V 30%, ASiR-V 70%, and DLIR at low, medium, and high reconstruction strengths. CT number, noise, noise reduction rate, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were measured and calculated within the region of interest (ROI) on subcutaneous fat, bilateral renal cortices, renal medulla, renal arteries, and adrenal glands. For qualitative analyses, the differentiation of the renal cortex and medulla, conspicuity of the adrenal gland boundary, sharpness, artifacts, and subjective noise were assessed. The overall image quality was calculated on a scale from 0 (worst) to 15 (best) based on the five values above and the score≥9 was acceptable.

RESULTS

CT number does not significantly differ between the reconstruction datasets. Noise does not significantly differ between ASiR-V 30% and DLIR-L, but it is significantly lower using ASiR-V 70%, DLIR-M, and DLIR-H. The noise reduction rate relative to ASiR-V 30% is significantly different between the DLIR groups and ASiR-V 70%, and DLIR-H yields the highest noise reduction rate (61.6%). SNR and CNR are higher for DLIR-M, DLIR-H, and ASiR-V 70% than for ASiR-V 30% and DLIR-L. DLIR-H shows the best SNR and CNR. The overall image quality yields the same pattern for DLIR-H, with the highest score. Percentages of cases with overall image quality score≥9 are 100% (DLIR-H), 94.23% (DLIR-M), 90.38% (ASiR-V70%), 67.31% (DLIR-L), and 63.46% (ASiR-V30%), respectively.

CONCLUSIONS

DLIR significantly improved the objective and subjective image quality of renal and adrenal CTs, yielding superior noise reduction compared with ASiR-V.

Radiation Dose Modulation of Computed Tomography Component in Positron Emission Tomography/Computed Tomography

In oncology practice, the CT component of PET/CT may be used for attenuation correction, lesion localization, and CT diagnosis, and significantly enhances the clinical benefit of PET. However, acquisition of CT covering the whole body increases radiation dose and consequently the risk of cancer induction, and optimization should be pursued. In CT, radiation dose is a major determinant of image quality, and is mainly adjusted by modulation of tube current. Automatic exposure control (AEC) is widely used for tube current modulation, and increases tube current in a large patient and in strongly attenuating regions of a given patient to preserve image quality despite strong X-ray attenuation. Radiation dose determined by AEC depends on various factors, such as the type of AEC software, scout imaging direction, arm positioning, and patient centering. Because radiation dose reduction increases image noise and may degrade clinical utility, image quality should be assessed together with radiation dose in the process of optimization. Clinical demands for image quality vary largely depending on the aim of the CT component, with lower quality being sufficient for lesion localization than for CT diagnosis. Therefore, optimal radiation dose differs according to the aim. Determining optimal dose is a somewhat subjective and difficult task, and use of the diagnostic reference level, determined based on national or regional survey, is recommended to recognize need for optimization. The volume CT dose index and dose-length product are used as indices of CT radiation dose, and effective dose may also be calculated for comparison of stochastic effects among different radiation sources and among different imaging procedures. Wide coverage from the head to the lower extremities causes problems in estimating these indices in whole-body PET/CT. CT definitely enhances clinical benefits of PET but simultaneously increases potential detriments due to radiation exposure. In the era of hybrid imaging, nuclear medicine practitioners should be aware of the technology and radiation dose management of CT.

Assessment of gastric wall structure using ultra-high-resolution computed tomography

PURPOSE

To evaluate the image quality of ultra-high-resolution CT (U-HRCT) in the comparison among four different reconstruction methods, focusing on the gastric wall structure, and to compare the conspicuity of a three-layered structure of the gastric wall between conventional HRCT (C-HRCT) and U-HRCT.

METHOD

Our retrospective study included 48 patients who underwent contrast-enhanced U-HRCT. Quantitative analyses were performed to compare image noise of U-HRCT between deep-learning reconstruction (DLR) and other three methods (filtered back projection: FBP, hybrid iterative reconstruction: Hybrid-IR, and Model-based iterative reconstruction: MBIR). The mean overall image quality scores were also compared between the DLR and other three methods. In addition, the mean conspicuity scores for the three-layered structure of the gastric wall at five regions were compared between C-HRCT and U-HRCT.

RESULTS

The mean noise of U-HRCT with DLR was significantly lower than that with the other three methods (P < 0.001). The mean overall image quality scores with DLR images were significantly higher than those with the other three methods (P < 0.001). Regarding the comparison between C-HRCT and U-HRCT, the mean conspicuity scores for the three-layered structure of the gastric wall on U-HRCT were significantly better than those on C-HRCT in the fornix (5 [5-5] vs. 3.5 [3-4], P < 0.001), body (4 [3.25-5] vs. 4 [3-4], P = 0.039), angle (5 [4-5] vs. 3 [2-4], P < 0.001), and antral posterior (4 [3.25-5] vs. 2 [2-4], P < 0.001), except for antral anterior (4 [3-5] vs. 3 [3-4], P = 0.230) CONCLUSION: U-HRCT using DLR improved the image noise and overall image quality of the gastric wall as well as the conspicuity of the three-layered structure, suggesting its utility for the evaluation of the anatomical details of the gastric wall structure.