Computer-aided detection (CAD) of solid pulmonary nodules in chest x-ray equivalent ultralow dose chest CT - first in-vivo results at dose levels of 0.13mSv

Imaging of Solid Pulmonary Nodules

Early detection with accurate classification of solid pulmonary nodules is critical in reducing lung cancer morbidity and mortality. Computed tomography (CT) remains the most widely used imaging examination for pulmonary nodule evaluation; however, other imaging modalities, such as PET/CT and MRI, are increasingly used for nodule characterization. Current advances in solid nodule imaging are largely due to developments in machine learning, including automated nodule segmentation and computer-aided detection. This review explores current multi-modality solid pulmonary nodule detection and characterization with discussion of radiomics and risk prediction models.

Absolute ground truth-based validation of computer-aided nodule detection and volumetry in low-dose CT imaging

PURPOSE

To validate the performance of computer-aided detection (CAD) and volumetry software using an anthropomorphic phantom with a ground truth (GT) set of 3D-printed nodules.

METHODS

The Kyoto Kaguku Lungman phantom, containing 3D-printed solid nodules including six diameters (4 to 9 mm) and three morphologies (smooth, lobulated, spiculated), was scanned at varying CTDIvol levels (6.04, 1.54 and 0.20 mGy). Combinations of reconstruction algorithms (iterative and deep learning image reconstruction) and kernels (soft and hard) were applied. Detection, volumetry and density results recorded by a commercially available AI-based algorithm (AVIEW LCS + ) were compared to the absolute GT, which was determined through µCT scanning at 50 µm resolution. The associations between image acquisition parameters or nodule characteristics and accuracy of nodule detection and characterization were analyzed with chi square tests and multiple linear regression.

RESULTS

High levels of detection sensitivity and precision (minimal 83 % and 91 % respectively) were observed across all acquisitions. Neither reconstruction algorithm nor radiation dose showed significant associations with detection. Nodule diameter however showed a highly significant association with detection (p < 0.0001). Volumetric measurements for nodules > 6 mm were accurate within 10 % absolute range from volumeGT, regardless of dose and reconstruction. Nodule diameter and morphology are major determinants of volumetric accuracy (p < 0.001). Density assignment was not significantly influenced by any parameters.

CONCLUSIONS

Our study confirms the software's accurate performance in nodule volumetry, detection and density characterization with robustness for variations in CT imaging protocols. This study suggests the incorporation of similar phantom setups in quality assurance of CAD tools.

Superior metal artifact reduction of tin-filtered low-dose CT in imaging of lumbar spinal instrumentation compared to conventional computed tomography

OBJECTIVE

To compare the image quality of low-dose CT (LD-CT) with tin filtration of the lumbar spine after metal implants to standard clinical CT, and to evaluate the potential for metal artifact and dose reduction.

MATERIALS AND METHODS

CT protocols were optimized in a cadaver torso. Seventy-four prospectively included patients with metallic lumbar implants were scanned with both standard CT (120 kV) and tin-filtered LD-CT (Sn140kV). CT dose parameters and qualitative measures (1 = worst,4 = best) were compared. Quantitative measures included noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and the width and attenuation of the most prominent hypodense metal artifact. Standard CT and LD-CT were assessed for imaging findings.

RESULTS

Tin-filtered LD-CT was performed with 60% dose saving compared to standard CT (median effective dose 3.22 mSv (quartile 1-3: 2.73-3.49 mSv) versus 8.02 mSv (6.42-9.27 mSv; p < .001). Image quality of CT and tin-filtered low-dose CT was good with excellent depiction of anatomy, while image noise was lower for CT and artifacts were weaker for tin-filtered LD-CT. Quantitative measures also revealed increased noise for tin-filtered low-dose CT (41.5HU), lower SNR (2) and CNR (0.6) compared to CT (32HU,3.55,1.03, respectively) (all p < .001). However, tin-filtered LD-CT performed superior regarding the width and attenuation of hypodense metal artifacts (2.9 mm and -767.5HU for LD-CT vs. 4.1 mm and -937HU for CT; all p < .001). No difference between methods was observed in detection of imaging findings.

CONCLUSION

Tin-filtered LD-CT with 60% dose saving performs comparable to standard CT in detection of pathology and surgery related complications after lumbar spinal instrumentation, and shows superior metal artifact reduction.

Framework for dual-energy-like chest radiography image synthesis from single-energy computed tomography based on cycle-consistent generative adversarial network

BACKGROUND

Dual-energy (DE) chest radiography (CXR) enables the selective imaging of two relevant materials, namely, soft tissue and bone structures, to better characterize various chest pathologies (i.e., lung nodule, bony lesions, etc.) and potentially improve CXR-based diagnosis. Recently, deep-learning-based image synthesis techniques have attracted considerable attention as alternatives to existing DE methods (i.e., dual-exposure-based and sandwich-detector-based methods) because software-based bone-only and bone-suppression images in CXR could be useful.

PURPOSE

The objective of this study was to develop a new framework for DE-like CXR image synthesis from single-energy computed tomography (CT) based on a cycle-consistent generative adversarial network.

METHODS

The core techniques of the proposed framework are divided into three categories: (1) data configuration from the generation of pseudo CXR from single energy CT, (2) learning of the developed network architecture using pseudo CXR and pseudo-DE imaging using a single-energy CT, and (3) inference of the trained network on real single-energy CXR. We performed a visual inspection and comparative evaluation using various metrics and introduced a figure of image quality (FIQ) to consider the effects of our framework on the spatial resolution and noise in terms of a single index through various test cases.

RESULTS

Our results indicate that the proposed framework is effective and exhibits potential synthetic imaging ability for two relevant materials: soft tissue and bone structures. Its effectiveness was validated, and its ability to overcome the limitations associated with DE imaging techniques (e.g., increase in exposure dose owing to the requirement of two acquisitions, and emphasis on noise characteristics) via an artificial intelligence technique was presented.

CONCLUSIONS

The developed framework addresses X-ray dose issues in the field of radiation imaging and enables pseudo-DE imaging with single exposure.

Ultra-low dose computed tomography protocols using spectral shaping for lung cancer screening: Comparison with low-dose for volumetric LungRADS classification

PURPOSE

To compare Low-Dose Computed Tomography (LDCT) with four different Ultra-Low-Dose Computed Tomography (ULDCT) protocols for PN classification according to the Lung Reporting and Data System (LungRADS).

METHODS

Three hundred sixty-one participants of an ongoing lung cancer screening (LCS) underwent single-breath-hold double chest Computed Tomography (CT), including LDCT (120kVp, 25mAs; CTDIvol 1,62 mGy) and one ULDCT among: fully automated exposure control ("ULDCT₁"); fixed tube-voltage and current according to patient size ("ULDCT₂"); hybrid approach with fixed tube-voltage ("ULDCT₃") and tube current automated exposure control ("ULDCT₄"). Two radiologists (R1, R2) assessed LungRADS 2022 categories on LDCT, and then after 2 weeks on ULDCT using two different kernels (R1: Qr49_{ADMIRE 4}; R2: Br49_{ADMIRE 3}). Intra-subject agreement for LungRADS categories between LDCT and ULDCT was measured by the k-Cohen Index with Fleiss-Cohen weights.

RESULTS

LDCT-dominant PNs were detected in ULDCT in 87 % of cases on Qr49_{ADMIRE 4} and 88 % on Br49_{ADMIRE 3}. The intra-subject agreement was: κULDCT₁ = 0.89 [95 %CI 0.82-0.96]; κULDCT₂ = 0.90 [0.81-0.98]; κULDCT₃ = 0.91 [0.84-0.99]; κULDCT₄ = 0.88 [0.78-0.97] on Qr49_{ADMIRE 4}, and κULDCT₁ = 0.88 [0.80-0.95]; κULDCT₂ = 0.91 [0.86-0.96]; κULDCT₃ = 0.87 [0.78-0.95]; and κULDCT₄ = 0.88 [0.82-0.94] on Br49_{ADMIRE 3}. LDCT classified as LungRADS 4B were correctly identified as LungRADS 4B at ULDCT₃, with the lowest radiation exposure among the tested protocols (median effective doses were 0.31, 0.36, 0.27 and 0.37 mSv for ULDCT₁, ULDCT₂, ULDCT₃, and ULDCT₄, respectively).

CONCLUSIONS

ULDCT by spectral shaping allows the detection and characterization of PNs with an excellent agreement with LDCT and can be proposed as a feasible approach in LCS.

An IoT-Based Deep Learning Framework for Real-Time Detection of COVID-19 through Chest X-ray Images

Over the next decade, Internet of Things (IoT) and the high-speed 5G network will be crucial in enabling remote access to the healthcare system for easy and fast diagnosis. In this paper, an IoT-based deep learning computer-aided diagnosis (CAD) framework is proposed for online and real-time COVID-19 identification. The proposed work first fine-tuned the five state-of-the-art deep CNN models such as Xception, ResNet50, DenseNet201, MobileNet, and VGG19 and then combined these models into a majority voting deep ensemble CNN (DECNN) model in order to detect COVID-19 accurately. The findings demonstrate that the suggested framework, with a test accuracy of 98%, outperforms other relevant state-of-the-art methodologies in terms of overall performance. The proposed CAD framework has the potential to serve as a decision support system for general clinicians and rural health workers in order to diagnose COVID-19 at an early stage.

Semi-automated volumetry of pulmonary nodules: Intra-individual comparison of standard dose and chest X-ray equivalent ultralow dose chest CT scans

PURPOSE

To assess the performance of semi-automated volumetry of solid pulmonary nodules on single-energy tin-filtered ultralow dose (ULD) chest CT scans at a radiation dose equivalent to chest X-ray relative to standard dose (SD) chest CT scans and assess the impact of kernel and iterative reconstruction selection.

METHODS

Ninety-four consecutive patients from a prospective single-center study were included and underwent clinically indicated SD chest CT (1.9 ± 0.8 mSv) and additional ULD chest CT (0.13 ± 0.01 mSv) in the same session. All scans were reconstructed with a soft tissue (Br40) and lung (Bl64) kernel as well as with Filtered Back Projection (FBP) and Iterative Reconstruction (ADMIRE-3 and ADMIRE-5). One hundred and forty-eight solid pulmonary nodules were identified and analysed by semi-automated volumetry on all reconstructions. Nodule volumes were compared amongst all reconstructions thereby focusing on the agreement between SD and ULD scans.

RESULTS

Nodule volumes ranged from 58.5 (28.8-126) mm3 for ADMIRE-5 Br40 ULD reconstructions to 72.5 (39-134) mm3 for FBP Bl64 SD reconstructions with significant differences between reconstructions (p < 0.001). Interscan agreement of volumes between two given reconstructions ranged from ICC = 0.605 to ICC = 0.999. Between SD and ULD scans, agreement of nodule volumes was highest for FBP Br40 (ICC = 0.995), FBP Bl64 (ICC = 0.939) and ADMIRE-5 Bl64 (ICC = 0.994) reconstructions. ADMIRE-3 reconstructions exhibited reduced interscan agreement of nodule volumes (ICCs from 0.788 - 0.882).

CONCLUSIONS

The interscan agreement of node volumes between SD and ULD is high depending on the choice of kernel and reconstruction algorithm. However, caution should be exercised when comparing two image series that were not identically reconstructed.

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

Deep learning-based approach for detecting COVID-19 in chest X-rays

Today, 2019 Coronavirus (COVID-19) infections are a major health concern worldwide. Therefore, detecting COVID-19 in X-ray images is crucial for diagnosis, evaluation, and treatment. Furthermore, expressing diagnostic uncertainty in a report is a challenging duty but unavoidable task for radiologists. This study proposes a novel CNN (Convolutional Neural Network) model for automatic COVID-19 identification utilizing chest X-ray images. The proposed CNN model is designed to be a reliable diagnostic tool for two-class categorization (COVID and Normal). In addition to the proposed model, different architectures, including the pre-trained MobileNetv2 and ResNet50 models, are evaluated for this COVID-19 dataset (13,824 X-ray images) and our suggested model is compared to these existing COVID-19 detection algorithms in terms of accuracy. Experimental results show that our proposed model identifies patients with COVID-19 disease with 96.71 percent accuracy, 91.89 percent F1-score. Our proposed approach CNN’s experimental results show that it outperforms the most advanced algorithms currently available. This model can assist clinicians in making informed judgments on how to diagnose COVID-19, as well as make test kits more accessible.

Differential Diagnosis of Preinvasive Lesions in Small Pulmonary Nodules by Dual Source Computed Tomography Imaging

This study was aimed to explore the differential diagnosis value of preinvasive lesions/minimally invasive adenocarcinoma and invasive adenocarcinoma manifesting as small pulmonary nodules under dual source computed tomography (DSCT) imaging. The patients with nodular manifestations of adenocarcinoma in situ (AIS)/microinfiltrating adenocarcinoma (MIA) were selected as group X, including 14 cases. A total of 31 cases with nodular infiltrating adenocarcinoma were selected as group Y. The enhanced dual-energy image obtained by DSCT dual-energy scan was transferred to the software to obtain the energy image and iodine distribution map. SPSS 18.0 was used for statistical analysis. P < 0.05 was considered statistically significant. All measurements were labeled as mean x͞±S standard deviation. In the CT findings of microinfiltrating adenocarcinoma and infiltrating adenocarcinoma, lobulation sign, burr sign, vacuole sign, and pleural depression sign can help the diagnosis of infiltrating adenocarcinoma. The results showed that lobulation sign, burr sign, vacuole sign, and pleural depression sign could be used as the distinguishing feature of preinvasive lesion/microinvasive adenocarcinoma and invasive adenocarcinoma. Receiver-operating characteristic (ROC) curve analysis showed that the critical value, sensitivity, and specificity of lesion diameter ≥1.4 cm and CT value ≥14.14HU for diagnosis of invasive lung adenocarcinoma were 1.32 and 14.14, 88.4% and 94.4%, and 67.3% and 75.8%, respectively. There were substantial differences in CT values between the two groups under low energy level (42-99 kev) (P < 0.05). DSCT dual-energy imaging can quantitatively identify preinvasive pulmonary nodules with multiple parameters.

Diagnostic validation of a deep learning nodule detection algorithm in low-dose chest CT: determination of optimized dose thresholds in a virtual screening scenario

OBJECTIVES

This study was conducted to evaluate the effect of dose reduction on the performance of a deep learning (DL)-based computer-aided diagnosis (CAD) system regarding pulmonary nodule detection in a virtual screening scenario.

METHODS

Sixty-eight anthropomorphic chest phantoms were equipped with 329 nodules (150 ground glass, 179 solid) with four sizes (5 mm, 8 mm, 10 mm, 12 mm) and scanned with nine tube voltage/current combinations. The examinations were analyzed by a commercially available DL-based CAD system. The results were compared by a comparison of proportions. Logistic regression was performed to evaluate the impact of tube voltage, tube current, nodule size, nodule density, and nodule location.

RESULTS

The combination with the lowest effective dose (E) and unimpaired detection rate was 80 kV/50 mAs (sensitivity: 97.9%, mean false-positive rate (FPR): 1.9, mean CTDIvol: 1.2 ± 0.4 mGy, mean E: 0.66 mSv). Logistic regression revealed that tube voltage and current had the greatest impact on the detection rate, while nodule size and density had no significant influence.

CONCLUSIONS

The optimal tube voltage/current combination proposed in this study (80 kV/50 mAs) is comparable to the proposed combinations in similar studies, which mostly dealt with conventional CAD software. Modification of tube voltage and tube current has a significant impact on the performance of DL-based CAD software in pulmonary nodule detection regardless of their size and composition.

KEY POINTS

• Modification of tube voltage and tube current has a significant impact on the performance of deep learning-based CAD software. • Nodule size and composition have no significant impact on the software's performance. • The optimal tube voltage/current combination for the examined software is 80 kV/50 mAs.

Automated detection of COVID-19 cases from chest X-ray images using deep neural network and XGBoost

Introduction

In late 2019 and after the COVID-19 pandemic in the world, many researchers and scholars tried to provide methods for detecting COVID-19 cases. Accordingly, this study focused on identifying patients with COVID-19 from chest X-ray images.

Methods

In this paper, a method for diagnosing coronavirus disease from X-ray images was developed. In this method, DenseNet169 Deep Neural Network (DNN) was used to extract the features of X-ray images taken from the patients’ chests. The extracted features were then given as input to the Extreme Gradient Boosting (XGBoost) algorithm to perform the classification task.

Results

Evaluation of the proposed approach and its comparison with the methods presented in recent years revealed that this method was more accurate and faster than the existing ones and had an acceptable performance for detecting COVID-19 cases from X-ray images. The experiments showed 98.23% and 89.70% accuracy, 99.78% and 100% specificity, 92.08% and 95.20% sensitivity in two and three-class problems, respectively.

Conclusion

This study aimed to detect people with COVID-19, focusing on non-clinical approaches. The developed method could be employed as an initial detection tool to assist the radiologists in more accurate and faster diagnosing the disease.

Implication for practice

The proposed method's simple implementation, along with its acceptable accuracy, allows it to be used in COVID-19 diagnosis. Moreover, the gradient-based class activation mapping (Grad-CAM) can be used to represent the deep neural network's decision area on a heatmap. Radiologists might use this heatmap to evaluate the chest area more accurately.

First Performance Evaluation of an Artificial Intelligence-Based Computer-Aided Detection System for Pulmonary Nodule Evaluation in Dual-Source Photon-Counting Detector CT at Different Low-Dose Levels

OBJECTIVE

The aim of this study was to evaluate the image quality (IQ) and performance of an artificial intelligence (AI)-based computer-aided detection (CAD) system in photon-counting detector computed tomography (PCD-CT) for pulmonary nodule evaluation at different low-dose levels.

MATERIALS AND METHODS

An anthropomorphic chest-phantom containing 14 pulmonary nodules of different sizes (range, 3-12 mm) was imaged on a PCD-CT and on a conventional energy-integrating detector CT (EID-CT). Scans were performed with each of the 3 vendor-specific scanning modes (QuantumPlus [Q+], Quantum [Q], and High Resolution [HR]) at decreasing matched radiation dose levels (volume computed tomography dose index ranging from 1.79 to 0.31 mGy) by adapting IQ levels from 30 to 5. Image noise was measured manually in the chest wall at 8 different locations. Subjective IQ was evaluated by 2 readers in consensus. Nodule detection and volumetry were performed using a commercially available AI-CAD system.

RESULTS

Subjective IQ was superior in PCD-CT compared with EID-CT (P < 0.001), and objective image noise was similar in the Q+ and Q-mode (P > 0.05) and superior in the HR-mode (PCD 55.8 ± 11.7 HU vs EID 74.8 ± 5.4 HU; P = 0.01). High resolution showed the lowest image noise values among PCD modes (P = 0.01). Overall, the AI-CAD system delivered comparable results for lung nodule detection and volumetry between PCD- and dose-matched EID-CT (P = 0.08-1.00), with a mean sensitivity of 95% for PCD-CT and of 86% for dose-matched EID-CT in the lowest evaluated dose level (IQ5). Q+ and Q-mode showed higher false-positive rates than EID-CT at lower-dose levels (IQ10 and IQ5). The HR-mode showed a sensitivity of 100% with a false-positive rate of 1 even at the lowest evaluated dose level (IQ5; CDTIvol, 0.41 mGy).

CONCLUSIONS

Photon-counting detector CT was superior to dose-matched EID-CT in subjective IQ while showing comparable to lower objective image noise. Fully automatized AI-aided nodule detection and volumetry are feasible in PCD-CT, but attention has to be paid to false-positive findings.

Impact of dose reduction and iterative reconstruction algorithm on the detectability of pulmonary nodules by artificial intelligence

PURPOSE

The purpose of this study was to assess whether the performances of an automated software for lung nodule detection with computed tomography (CT) are affected by radiation dose and the use of iterative reconstruction algorithm.

MATERIALS AND METHODS

A chest phantom (Multipurpose Chest Phantom N1; Kyoto Kagaku Co. Ltd, Kyoto, Japan) with 15 pulmonary nodules was scanned with a total of five CT protocol settings with up to 20-fold dose reduction. All CT examinations were reconstructed with iterative reconstruction algorithms ADMIRE 3 and ADMIRE 5 and were then analyzed for the presence of pulmonary nodules with a fully automated computer aided detection software system (InferRead^TM CT Lung, Infervision), which is based on deep neural networks.

RESULTS

The sensitivity of fully automated pulmonary nodule detection for ground-glass nodules at standard dose CT was greater (70.0%; 14/20; 95% CI: 51.6-88.4%) than at 10-fold and 20-fold dose reduction (30.0%; 6/20; 95% CI: 0.0%-62.5%). There were less false positive findings when ADMIRE 5 reconstruction was used (4.0 ± 2.8 [SD]; range: 2-6) instead of ADMIRE 3 reconstruction (25.0 ± 15.6 [SD]; range: 14-36). There was no difference in the sensitivity of detection of solid and subsolid nodules between standard dose (100%; 95% CI: 100-100%) and 10- and 20-fold reduced dose CT (92.5%; 95% CI: 83.8-100.0%). Image noise was significantly greater with ADMIRE 3 (81 ± 2 [SD] [range: 79-84]; 104 ± 3 [SD] [range: 101-107]; 114 ± 5 [SD] [range: 110-119]; 193 ± 10 [SD] [range: 183-203]; 220 ± 16 [SD] [range: 210-238]) compared to ADMIRE 5 (44 ± 2 [SD] [range: 42-46]; 60 ± 2 [SD] [range: 57-61]; 66 ± 1 [SD] [range: 65-67]; 103 ± 4 [SD] [range: 98-106]; 110 ± 1 [SD] [range: 109-111]), respectively in each of the five CT protocols.

CONCLUSION

This phantom study suggests that dose reduction and iterative reconstruction settings have an impact on detectability of pulmonary nodules by artificial intelligence software and we therefore encourage adaption of dose levels and reconstruction methods prior to widespread implementation of fully automatic nodule detection software for lung cancer screening purposes.

Pneumonia Detection in Chest X-Ray Dose-Equivalent CT: Impact of Dose Reduction on Detectability by Artificial Intelligence

RATIONALE AND OBJECTIVES

There has been a significant increase of immunocompromised patients in recent years due to new treatment modalities for previously fatal diseases. This comes at the cost of an elevated risk for infectious diseases, most notably pathogens affecting the respiratory tract. Because early diagnosis and treatment of pneumonia can help reducing morbidity and mortality, we assessed the performance of a deep neural network in the detection of pulmonary infection in chest X-ray dose-equivalent computed tomography (CT).

MATERIALS AND METHODS

The 100 patients included in this retrospective study were referred to our department for suspicion of pulmonary infection and/or follow-up of known pulmonary nodules. Every patient was scanned with a standard dose (1.43 ± 0.54 mSv) and a 20 times dose-reduced (0.07 ± 0.03 mSv) CT protocol. We trained a deep neural network to perform binary classification (pulmonary consolidation present or not) and assessed diagnostic performance on both standard dose and reduced dose CT images.

RESULTS

The areas under the curve of the deep learning algorithm for the standard dose CT was 0.923 (confidence interval [CI] 95%: 0.905-0.941) and significantly higher than the areas under the curve (0.881, CI 95%: 0.859-0.903) of the reduced dose CT (p = 0.001). Sensitivity and specificity of the standard dose CT was 82.9% and 93.8%, and of the reduced dose CT 71.0% and 93.3%.

CONCLUSION

Pneumonia detection with X-ray dose-equivalent CT using artificial intelligence is feasible and may contribute to a more robust and reproducible diagnostic performance. Dose reduction lowered the performance of the deep neural network, which calls for optimization and adaption of CT protocols when using AI algorithms at reduced doses.

Pelvic bone CT: can tin-filtered ultra-low-dose CT and virtual radiographs be used as alternative for standard CT and digital radiographs?

Objectives

To compare ultra-low-dose CT (ULD-CT) of the osseous pelvis with tin filtration to standard clinical CT (CT), and to assess the quality of computed virtual pelvic radiographs (VRs).

Methods

CT protocols were optimized in a phantom and three pelvic cadavers. Thirty prospectively included patients received both standard CT (automated tube voltage selection and current modulation) and tin-filtered ULD-CT of the pelvis (Sn140kV/50mAs). VRs of ULD-CT data were computed using an adapted cone beam–based projection algorithm and were compared to digital radiographs (DRs) of the pelvis. CT and DR dose parameters and quantitative and qualitative measures (1 = worst, 4 = best) were compared. CT and ULD-CT were assessed for osseous pathologies.

Results

Dose reduction of ULD-CT was 84% compared to CT, with a median effective dose of 0.38 mSv (quartile 1–3: 0.37–0.4 mSv) versus 2.31 mSv (1.82–3.58 mSv; p < .001), respectively. Mean dose of DR was 0.37 mSv (± 0.14 mSv). The median signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of bone were significantly higher for CT (64.3 and 21.5, respectively) compared to ULD-CT (50.4 and 18.8; p ≤ .01), while ULD-CT was significantly more dose efficient (figure of merit (FOM) 927.6) than CT (FOM 167.6; p < .001). Both CT and ULD-CT were of good image quality with excellent depiction of anatomy, with a median score of 4 (4–4) for both methods (p = .1). Agreement was perfect between both methods regarding the prevalence of assessed osseous pathologies (p > .99). VRs were successfully calculated and were equivalent to DRs.

Conclusion

Tin-filtered ULD-CT of the pelvis at a dose equivalent to standard radiographs is adequate for assessing bone anatomy and osseous pathologies and had a markedly superior dose efficiency than standard CT.

Key Points

• Ultra-low-dose pelvic CT with tin filtration (0.38 mSv) can be performed at a dose of digital radiographs (0.37 mSv), with a dose reduction of 84% compared to standard CT (2.31 mSv).

• Tin-filtered ultra-low-dose CT had lower SNR and CNR and higher image noise than standard CT, but showed clear depiction of anatomy and accurate detection of osseous pathologies.

• Virtual pelvic radiographs were successfully calculated from ultra-low-dose CT data and were equivalent to digital radiographs.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00330-021-07824-x.

Radiomics in the evaluation of lung nodules: Intrapatient concordance between full-dose and ultra-low-dose chest computed tomography

PURPOSE

The purpose of this study was to retrospectively evaluate the quantitative and qualitative intrapatient concordance of pulmonary nodule risk assessment by commercially available radiomics software between full-dose (FD) chest-CT and ultra-low-dose (ULD) chest CT.

MATERIALS AND METHODS

Between July 2013 and September 2015, 68 patients (52 men and16 women; mean age, 65.5±10.6 [SD] years; range: 35-87 years) with lung nodules≥5mm and<30mm who underwent the same day FD chest CT (helical acquisition; 120kV; automated tube current modulation) and ULD chest CT (helical acquisition; 135kV; 10mA fixed) were retrospectively included. Each nodule on each acquisition was assessed by a commercial radiomics software providing a similarity malignancy index (mSI), classifying it as "benign-like" (mSI<0.1); "malignant-like" (mSI>0.9) or "undetermined" (0.1≤mSI≤0.9). Intrapatient qualitative agreement was evaluated with weighted Cohen-Kappa test and quantitative agreement with intraclass correlation coefficient (ICC).

RESULTS

Ninety-nine lung nodules with a mean size of 9.14±4.3 (SD) mm (range: 5-25mm) in 68 patients (mean 1.46 nodule per patient; range: 1-5) were assessed; mean mSI was 0.429±0.331 (SD) (range: 0.001-1) with FD chest CT (22/99 [22%] "benign-like", 67/99 [68%] "undetermined" and 10/99 [10%] "malignant-like") and mean mSI was 0.487±0.344 (SD) (range: 0.002-1) with ULD chest CT (20/99 [20%] "benign-like", 59/99 [60%] "undetermined" and 20/99 [20%] "malignant-like"). Qualitative and quantitative agreement of FD chest CT with ULD chest CT were "good" with Kappa value of 0.60 (95% CI: 0.46-0.74) and ICC of 0.82 (95% CI: 0.73-0.87), respectively.

CONCLUSION

A good agreement in malignancy similarity index can be obtained between ULD chest CT and FD chest CT using radiomics software. However, further studies must be done with more case material to confirm our results and elucidate the diagnostic capabilities of radiomics software using ULD chest CT for lung nodule characterization by comparison with FD chest CT.

Early detection of lung cancer in Czech high-risk asymptomatic individuals (ELEGANCE): A study protocol

BACKGROUND

Lung cancer screening in high-risk population increases the proportion of patients diagnosed at a resectable stage.

AIMS

To optimize the selection criteria and quality indicators for lung cancer screening by low-dose CT (LDCT) in the Czech population of high-risk individuals. To compare the influence of screening on the stage of lung cancer at the time of the diagnosis with the stage distribution in an unscreened population. To estimate the impact on life-years lost according to the stage-specific cancer survival and stage distribution in the screened population. To calculate the cost-effectiveness of the screening program.

METHODS

Based on the evidence from large national trials - the National Lung Screening Trial in the USA (NLST), the NELSON study, the recent recommendations of the Fleischner society, the American College of Radiology, and I-ELCAP action group, we developed a protocol for a single-arm prospective study in the Czech Republic for the screening of high-risk asymptomatic individuals. The study commenced in August 2020.

RESULTS

The inclusion criteria are: age 55 to 74 years; smoking: ≥30 pack-years; smoker or ex-smoker <15 years; performance status (0-1). The screening timepoints are at baseline and 1 year. The LDCT acquisition has a target CTDIvol ≤0.5mGy and effective dose ≤0.2mSv for a standard-size patient. The interpretation of findings is primarily based on nodule volumetry, volume doubling time (and related risk of malignancy). The management includes follow-up LDCT, contrast enhanced CT, PET/CT, tissue sampling. The primary outcome is the number of cancers detected at a resectable stage, secondary outcomes include the average cost per diagnosis of lung cancer, the number, cost, complications of secondary examinations, and the number of potentially important secondary findings.

CONCLUSIONS

A study protocol for early detection of lung cancer in Czech high-risk asymptomatic individuals (ELEGANCE) study using LDCT has been described.

FCOD: Fast COVID-19 Detector based on deep learning techniques

The sudden COVID-19 pandemic has caused a serious global concern due to infections and mortality rates. It is a hazardous disease that has recently become the biggest crisis in the modern era. Due to the limitation of test kits and the need for screening and rapid diagnosis of patients, it is essential to perform a self-operating detection model as a fast recognition system to detect COVID-19 infection and prevent the spread among the people. In this paper, we propose a novel technique called Fast COVID-19 Detector (FCOD) to have a fast detection of COVID-19 using X-ray images. The FCOD is a deep learning model based on the Inception architecture that uses 17 depthwise separable convolution layers to detect COVID-19. Depthwise separable convolution layers decrease the computation costs, time, and they can have a reducing role in the number of parameters compared to the standard convolution layers. To evaluate FCOD, we used covid-chestxray-dataset, which contains 940 publicly available typical chest X-ray images. Our results show that FCOD can provide accuracy, F1-score, and AUC of 96%, 96%, and 0.95%, respectively in classifying COVID-19 during 0.014 s for each case. The proposed model can be employed as a supportive decision-making system to assist radiologists in clinics and hospitals to screen patients immediately.

Observer Performance for Detection of Pulmonary Nodules at Chest CT over a Large Range of Radiation Dose Levels

Background There is a wide variation in radiation dose levels that can be used with chest CT in order to detect indeterminate pulmonary nodules. Purpose To compare the performance of lower-radiation-dose chest CT with that of routine dose in the detection of indeterminate pulmonary nodules 5 mm or greater. Materials and Methods In this retrospective study, CT projection data from 83 routine-dose chest CT examinations performed in 83 patients (120 kV, 70 quality reference mAs [QRM]) were collected between November 2013 and April 2014. Reference indeterminate pulmonary nodules were identified by two nonreader thoracic radiologists. By using validated noise insertion, five lower-dose data sets were reconstructed with filtered back projection (FBP) or iterative reconstruction (IR; 30 QRM with FBP, 10 QRM with IR, 5 QRM with FBP, 5 QRM with IR, and 2.5 QRM with IR). Three thoracic radiologists circled pulmonary nodules, rating confidence that the nodule was a 5-mm-or-greater indeterminate pulmonary nodule, and graded image quality. Analysis was performed on a per-nodule basis by using jackknife alternative free-response receiver operating characteristic figure of merit (FOM) and noninferiority limit of -0.10. Results There were 66 indeterminate pulmonary nodules (mean size, 8.6 mm ± 3.4 [standard deviation]; 21 part-solid nodules) in 42 patients (mean age, 51 years ± 17; 21 men and 21 women). Compared with the FOM for routine-dose CT (size-specific dose estimate, 6.5 mGy ± 1.8; FOM, 0.86 [95% confidence interval: 0.80, 0.91]), FOM was noninferior for all lower-dose configurations except for 2.5 QRM with IR. The sensitivity for subsolid nodules at 70 QRM was 60% (range, 48%-72%) and was significantly worse at a dose of 5 QRM and lower, whether or not IR was used (P < .05). Diagnostic image quality decreased with decreasing dose (P < .001) and was better with IR at 5 QRM (P < .05). Conclusion CT images reconstructed at dose levels down to 10 quality reference mAs (size-specific dose estimate, 0.9 mGy) had noninferior performance compared with routine dose in depicting pulmonary nodules. Iterative reconstruction improved subjective image quality but not performance at low dose levels. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by White and Kazerooni in this issue.

Cascaded deep learning classifiers for computer-aided diagnosis of COVID-19 and pneumonia diseases in X-ray scans

Computer-aided diagnosis (CAD) systems are considered a powerful tool for physicians to support identification of the novel Coronavirus Disease 2019 (COVID-19) using medical imaging modalities. Therefore, this article proposes a new framework of cascaded deep learning classifiers to enhance the performance of these CAD systems for highly suspected COVID-19 and pneumonia diseases in X-ray images. Our proposed deep learning framework constitutes two major advancements as follows. First, complicated multi-label classification of X-ray images have been simplified using a series of binary classifiers for each tested case of the health status. That mimics the clinical situation to diagnose potential diseases for a patient. Second, the cascaded architecture of COVID-19 and pneumonia classifiers is flexible to use different fine-tuned deep learning models simultaneously, achieving the best performance of confirming infected cases. This study includes eleven pre-trained convolutional neural network models, such as Visual Geometry Group Network (VGG) and Residual Neural Network (ResNet). They have been successfully tested and evaluated on public X-ray image dataset for normal and three diseased cases. The results of proposed cascaded classifiers showed that VGG16, ResNet50V2, and Dense Neural Network (DenseNet169) models achieved the best detection accuracy of COVID-19, viral (Non-COVID-19) pneumonia, and bacterial pneumonia images, respectively. Furthermore, the performance of our cascaded deep learning classifiers is superior to other multi-label classification methods of COVID-19 and pneumonia diseases in previous studies. Therefore, the proposed deep learning framework presents a good option to be applied in the clinical routine to assist the diagnostic procedures of COVID-19 infection.

Chest X-ray Dose Equivalent Low-dose CT with Tin Filtration: Potential Role for the Assessment of Pectus Excavatum

RATIONALE AND OBJECTIVES

To determine the value of chest CT with tin filtration applying a dose equivalent to chest x-ray for the assessment of the Haller index for evaluation of pectus excavatum.

MATERIALS AND METHODS

Two hundred seventy-two patients from a prospective single center study were included and underwent a clinical standard dose chest CT (effective dose 1.8 ± 0.7 mSv) followed by a low-dose CT (0.13 ± 0.01 mSv) in the same session. Two blinded readers independently evaluated all data sets. Image quality for bony chest wall assessment was noted. Radiologists further assessed (a) transverse thoracic diameter, (b) anteroposterior thoracic diameter, and calculated (c) Haller index by dividing transverse diameter by anteroposterior diameter. The agreement of both readers in standard dose and low-dose CT was assessed using Lin's concordance correlation coefficient (p_c).

RESULTS

Subjective image quality was lower for low dose compared to standard dose CT images by both readers (p < 0.001). In total, 99% (n = 540) of low-dose CT scans were rated as diagnostic for bony chest wall assessment by both readers. There was a high agreement for assessment of transverse diameter, anteroposterior diameter and Haller index comparing both readers in standard dose and low-dose CT with p_c values indicating substantial agreement (i.e., 0.95> and ≤0.99) in 12/18 (67%) and almost perfect agreement (i.e., >0.99) in 6/18 (33%).

CONCLUSION

Our study suggests that low-dose CT with tin filtration applying a radiation dose equivalent to a plain chest X-ray is excellent for assessing the Haller index.

Optimization of acquisition parameters for reduced-dose thoracic CT: A phantom study

PURPOSE

The purpose of this study was to analyze the impact of different options for reduced-dose computed tomography (CT) on image noise and visibility of pulmonary structures in order to define the best choice of parameters when performing ultra-low dose acquisitions of the chest in clinical routine.

MATERIALS AND METHODS

Using an anthropomorphic chest phantom, CT images were acquired at four defined low dose levels (computed tomography dose index [CTDI_vol]=0.15, 0.20, 0.30 and 0.40mGy), by changing tube voltage, pitch factor, or rotation time and adapting tube current to reach the predefined CTDI_vol-values. Images were reconstructed using two different levels of iteration (adaptive statistical iterative reconstruction [ASIR®]-v70% and ASIR®-v100%). Signal-to-noise ratio (SNR) as well as contrast-to-noise ratio (CNR) was calculated. Visibility of pulmonary structures (bronchi/vessels) were assessed by two readers on a 5-point-Likert scale.

RESULTS

Best visual image assessments and CNR/SNR were obtained with high tube voltage, while lowest scores were reached with lower pitch factor followed by high tube current. Protocols favoring lower pitch factor resulted in decreased visibility of bronchi/vessels, especially in the periphery. Decreasing radiation dose from 0.40 to 0.30mGy was not associated with a significant decrease in visual scores (P<0.05), however decreasing radiation dose from 0.30mGy to 0.15mGy was associated with a lower visibility of most of the evaluated structures (P<0.001). While image noise could be significantly reduced when ASIR®-v100% instead of ASIR®-v70% was used, the visibility-scores of pulmonary structures did not change significantly.

CONCLUSION

Favoring high tube voltage is the best option for reduced-dose protocols. A decrease of SNR and CNR does not necessarily go along with reduced visibility of pulmonary structures.

Comparison of ultra-low dose chest CT scanning protocols for the detection of pulmonary nodules: a phantom study

PURPOSE

To test ultra-low-dose computed tomography (ULDCT) scanning protocols for the detection of pulmonary nodules (PN).

METHODS

A chest phantom containing 19 solid and 11 subsolid PNs was scanned on a third-generation dual-source computed tomography (CT) scanner. Five ULDCT scans (Sn100kVp and 120, 70, 50, 30, and 20 reference mAs, using tube current modulation), reconstructed with iterative reconstruction (IR) algorithm at strength levels 2, 3, 4, and 5, were compared with standard CT (120kVp, 150 reference mAs, using tube current modulation). PNs were subjectively assessed according to a 4-point scale: 0, nondetectable nodule; 1, detectable nodule, very unlikely to be correctly measured; 2, detectable nodule, likely to be correctly measured; 3, PN quality equal to standard of reference. PN scores were analysed according to the Lung Imaging Reporting and Data System (Lung-RADS), simulating detection of nodules at baseline and incidence screening round.

RESULTS

For the baseline round, there were 17 Lung-RADS 2, 4 Lung-RADS 3, 8 Lung-RADS 4A, and 1 Lung-RADS 4B PNs. They were detectable in any ULDCT protocol, with the exception of 1 nondetectable part-solid nodule in 1 scanning protocol (120 reference mAs; IR strength: 3). For the incidence round, there were 4 Lung-RADS 2, 14 Lung-RADS 3, 2 Lung-RADS 4A, and 10 Lung-RADS 4B PNs. Ten were nondetectable in at least one ULDCT dataset; however, they were at least detectable in ULDCT with 70 reference mAs (IR strength: 4 and 5).

CONCLUSIONS

ULDCT scanning protocols allowing the detection of PNs can be proposed for the purpose of lung cancer screening.

Pulmonary quantitative CT imaging in focal and diffuse disease: current research and clinical applications

The frenetic development of imaging technology-both hardware and software-provides exceptional potential for investigation of the lung. In the last two decades, CT was exploited for detailed characterization of pulmonary structures and description of respiratory disease. The introduction of volumetric acquisition allowed increasingly sophisticated analysis of CT data by means of computerized algorithm, namely quantitative CT (QCT). Hundreds of thousands of CTs have been analysed for characterization of focal and diffuse disease of the lung. Several QCT metrics were developed and tested against clinical, functional and prognostic descriptors. Computer-aided detection of nodules, textural analysis of focal lesions, densitometric analysis and airway segmentation in obstructive pulmonary disease and textural analysis in interstitial lung disease are the major chapters of this discipline. The validation of QCT metrics for specific clinical and investigational needs prompted the translation of such metrics from research field to patient care. The present review summarizes the state of the art of QCT in both focal and diffuse lung disease, including a dedicated discussion about application of QCT metrics as parameters for clinical care and outcomes in clinical trials.

Diagnostic accuracy of chest X-ray dose-equivalent CT for assessing calcified atherosclerotic burden of the thoracic aorta

OBJECTIVE

To determine the value of ultralow-dose chest CT for estimating the calcified atherosclerotic burden of the thoracic aorta using tin-filter CT and compare its diagnostic accuracy with chest direct radiography.

METHODS

A total of 106 patients from a prospective, IRB-approved single-centre study were included and underwent standard dose chest CT (1.7 ± 0.7 mSv) by clinical indication followed by ultralow-dose CT with 100 kV and spectral shaping by a tin filter (0.13 ± 0.01 mSv) to achieve chest X-ray equivalent dose in the same session. Two independent radiologists reviewed the CT images, rated image quality and estimated presence and extent of calcification of aortic valve, ascending aorta and aortic arch. Conventional radiographs were also reviewed for presence of aortic calcifications.

RESULTS

The sensitivity of ultralow-dose CT for the detection of calcifications of the aortic valve, ascending aorta and aortic arch was 93.5, 96.2 and 96.2%, respectively, compared with standard dose CT. The sensitivity for the detection of thoracic aortic calcification was significantly lower on chest X-ray (52.3%) compared with ultralow-dose CT (p < 0.001).

CONCLUSION

A reliable estimation of calcified atherosclerotic burden of the thoracic aorta can be achieved with modern tin-filter CT at dose values comparable to chest direct radiography. Advances in knowledge: Our findings suggest that ultralow-dose CT is an excellent tool for assessing the calcified atherosclerotic burden of the thoracic aorta with higher diagnostic accuracy than conventional chest radiography and importantly without the additional cost of increased radiation dose.

Emphysema quantification and lung volumetry in chest X-ray equivalent ultralow dose CT - Intra-individual comparison with standard dose CT

OBJECTIVES

To determine whether ultralow dose chest CT with tin filtration can be used for emphysema quantification and lung volumetry and to assess differences in emphysema measurements and lung volume between standard dose and ultralow dose CT scans using advanced modeled iterative reconstruction (ADMIRE).

METHODS

84 consecutive patients from a prospective, IRB-approved single-center study were included and underwent clinically indicated standard dose chest CT (1.7±0.6mSv) and additional single-energy ultralow dose CT (0.14±0.01mSv) at 100kV and fixed tube current at 70mAs with tin filtration in the same session. Forty of the 84 patients (48%) had no emphysema, 44 (52%) had emphysema. One radiologist performed fully automated software-based pulmonary emphysema quantification and lung volumetry of standard and ultralow dose CT with different levels of ADMIRE. Friedman test and Wilcoxon rank sum test were used for multiple comparison of emphysema and lung volume. Lung volumes were compared using the concordance correlation coefficient.

RESULTS

The median low-attenuation areas (LAA) using filtered back projection (FBP) in standard dose was 4.4% and decreased to 2.6%, 2.1% and 1.8% using ADMIRE 3, 4, and 5, respectively. The median values of LAA in ultralow dose CT were 5.7%, 4.1% and 2.4% for ADMIRE 3, 4, and 5, respectively. There was no statistically significant difference between LAA in standard dose CT using FBP and ultralow dose using ADMIRE 4 (p=0.358) as well as in standard dose CT using ADMIRE 3 and ultralow dose using ADMIRE 5 (p=0.966). In comparison with standard dose FBP the concordance correlation coefficients of lung volumetry were 1.000, 0.999, and 0.999 for ADMIRE 3, 4, and 5 in standard dose, and 0.972 for ADMIRE 3, 4 and 5 in ultralow dose CT.

CONCLUSIONS

Ultralow dose CT at chest X-ray equivalent dose levels allows for lung volumetry as well as detection and quantification of emphysema. However, longitudinal emphysema analyses should be performed with the same scan protocol and reconstruction algorithms for reproducibility.