Attention pyramid pooling network for artificial diagnosis on pulmonary nodules.

The development of automated tools using advanced technologies like deep learning holds great promise for improving the accuracy of lung nodule classification in computed tomography (CT) imaging, ultimately reducing lung cancer mortality rates. However, lung nodules can be difficult to detect and classify, from CT images since different imaging modalities may provide varying levels of detail and clarity. Besides, the existing convolutional neural network may struggle to detect nodules that are small or located in difficult-to-detect regions of the lung. Therefore, the attention pyramid pooling network (APPN) is proposed to identify and classify lung nodules. First, a strong feature extractor, named vgg16, is used to obtain features from CT images. Then, the attention primary pyramid module is proposed by combining the attention mechanism and pyramid pooling module, which allows for the fusion of features at different scales and focuses on the most important features for nodule classification. Finally, we use the gated spatial memory technique to decode the general features, which is able to extract more accurate features for classifying lung nodules. The experimental results on the LIDC-IDRI dataset show that the APPN can achieve highly accurate and effective for classifying lung nodules, with sensitivity of 87.59%, specificity of 90.46%, accuracy of 88.47%, positive predictive value of 95.41%, negative predictive value of 76.29% and area under receiver operating characteristic curve of 0.914.

Segmenting and classifying lung diseases with M-Segnet and Hybrid Squeezenet-CNN architecture on CT images.

Diagnosing lung diseases accurately and promptly is essential for effectively managing this significant public health challenge on a global scale. This paper introduces a new framework called Modified Segnet-based Lung Disease Segmentation and Severity Classification (MSLDSSC). The MSLDSSC model comprises four phases: "preprocessing, segmentation, feature extraction, and classification." Initially, the input image undergoes preprocessing using an improved Wiener filter technique. This technique estimates the power spectral density of the noisy and original images and computes the SNR assisted by PSNR to evaluate image quality. Next, the preprocessed image undergoes Segmentation to identify and separate the RoI from the background objects in the lung image. We employ a Modified Segnet mechanism that utilizes a proposed hard tanh-Softplus activation function for effective Segmentation. Following Segmentation, features such as MLDN, entropy with MRELBP, shape features, and deep features are extracted. Following the feature extraction phase, the retrieved feature set is input into a hybrid severity classification model. This hybrid model comprises two classifiers: SDPA-Squeezenet and DCNN. These classifiers train on the retrieved feature set and effectively classify the severity level of lung diseases.

Performance of glial fibrillary acidic protein (GFAP) and ubiquitin carboxy-terminal hydrolase L1 (UCH-L1) biomarkers in predicting CT scan results and neurological outcomes in children with traumatic brain injury (BRAINI-2 paediatric study): protocol of a European prospective multicentre study.

INTRODUCTION: In light of the burden of traumatic brain injury (TBI) in children and the excessive number of unnecessary CT scans still being performed, new strategies are needed to limit their use while minimising the risk of delayed diagnosis of intracranial lesions (ICLs). Identifying children at higher risk of poor outcomes would enable them to be better monitored. The use of the blood-based brain biomarkers glial fibrillar acidic protein (GFAP) and ubiquitin carboxy-terminal hydrolase-L1 (UCH-L1) could help clinicians in this decision. The overall aim of this study is to provide new knowledge regarding GFAP and UCH-L1 in order to improve TBI management in the paediatric population. METHODS AND ANALYSIS: We will conduct a European, prospective, multicentre study, the BRAINI-2 paediatric study, in 20 centres in France, Spain and Switzerland with an inclusion period of 30 months for a total of 2880 children and adolescents included. To assess the performance of GFAP and UCH-L1 used separately and in combination to predict ICLs on CT scans (primary objective), 630 children less than 18 years of age with mild TBI, defined by a Glasgow Coma Scale score of 13-15 and with a CT scan will be recruited. To evaluate the potential of GFAP and UCH-L1 in predicting the prognosis after TBI (secondary objective), a further 1720 children with mild TBI but no CT scan as well as 130 children with moderate or severe TBI will be recruited. Finally, to establish age-specific reference values for GFAP and UCH-L1 (secondary objective), we will include 400 children and adolescents with no history of TBI. ETHICS AND DISSEMINATION: This study has received ethics approval in all participating countries. Results from our study will be disseminated in international peer-reviewed journals. All procedures were developed in order to assure data protection and confidentiality. TRIAL REGISTRATION NUMBER: NCT05413499.

Exploring shared decision-making needs in lung cancer screening among high-risk groups and health care providers in China: a qualitative study.

BACKGROUND: The intricate balance between the advantages and risks of low-dose computed tomography (LDCT) impedes the utilization of lung cancer screening (LCS). Guiding shared decision-making (SDM) for well-informed choices regarding LCS is pivotal. There has been a notable increase in research related to SDM. However, these studies possess limitations. For example, they may ignore the identification of decision support and needs from the perspective of health care providers and high-risk groups. Additionally, these studies have not adequately addressed the complete SDM process, including pre-decisional needs, the decision-making process, and post-decision experiences. Furthermore, the East-West divide of SDM has been largely ignored. This study aimed to explore the decisional needs and support for shared decision-making for LCS among health care providers and high-risk groups in China. METHODS: Informed by the Ottawa Decision-Support Framework, we conducted qualitative, face-to-face in-depth interviews to explore shared decision-making among 30 lung cancer high-risk individuals and 9 health care providers. Content analysis was used for data analysis. RESULTS: We identified 4 decisional needs that impair shared decision-making: (1) LCS knowledge deficit; (2) inadequate supportive resources; (3) shared decision-making conceptual bias; and (4) delicate doctor-patient bonds. We identified 3 decision supports: (1) providing information throughout the LCS process; (2) providing shared decision-making decision coaching; and (3) providing decision tools. CONCLUSIONS: This study offers valuable insights into the decisional needs and support required to undergo LCS among high-risk individuals and perspectives from health care providers. Future studies should aim to design interventions that enhance the quality of shared decision-making by offering LCS information, decision tools for LCS, and decision coaching for shared decision-making (e.g., through community nurses). Simultaneously, it is crucial to assess individuals' needs for effective deliberation to prevent conflicts and regrets after arriving at a decision.

Predictive value of preoperative CT enhancement rate and CT perfusion parameters in colorectal cancer.

BACKGROUND: Angiogenesis is a critical step in colorectal cancer growth, progression and metastasization. CT are routine imaging examinations for preoperative clinical evaluation in colorectal cancer patients. This study aimed to investigate the predictive value of preoperative CT enhancement rate (CER) and CT perfusion parameters on angiogenesis in colorectal cancer, as well as the association of preoperative CER and CT perfusion parameters with serum markers. METHODS: This retrospective analysis included 42 patients with colorectal adenocarcinoma. Median of microvessel density (MVD) as the cut-off value, it divided 42 patients into high-density group (MVD ≥ 35/field, n = 24) and low-density group (MVD < 35/field, n = 18), and 25 patients with benign colorectal lesions were collected as the control group. Statistical analysis of CER, CT perfusion parameters, serum markers were performed in all groups. Receiver operating curves (ROC) were plotted to evaluate the diagnostic efficacy of relevant CT perfusion parameters for tumor angiogenesis; Pearson correlation analysis explored potential association between CER, CT perfusion parameters and serum markers. RESULTS: CER, blood volume (BV), blood flow (BF), permeability surface (PS) and carbohydrate antigen 19 - 9 (CA19-9), carbohydrate antigen 125 (CA125), carcinoembryonic antigen (CEA), trefoil factor 3 (TFF3), vascular endothelial growth factor (VEGF) in colorectal adenocarcinoma were significantly higher than those in the control group, the parameters in high-density group were significantly higher than those in the low-density group (P < 0.05); however, the time to peak (TTP) of patients in colorectal adenocarcinoma were significantly lower than those in the control group, and the high-density group showed a significantly lower level compared to the low-density group (P < 0.05). The combined parameters BF + TTP + PS and BV + BF + TTP + PS demonstrated the highest area under the curve (AUC), both at 0.991. Pearson correlation analysis showed that the serum levels of CA19-9, CA125, CEA, TFF3, and VEGF in patients showed positive correlations with CER, BV, BF, and PS (P < 0.05), while these indicators exhibited negative correlations with TTP (P < 0.05). CONCLUSIONS: Some single and joint preoperative CT perfusion parameters can accurately predict tumor angiogenesis in colorectal adenocarcinoma. Preoperative CER and CT perfusion parameters have certain association with serum markers.

HCA-DAN: hierarchical class-aware domain adaptive network for gastric tumor segmentation in 3D CT images.

BACKGROUND: Accurate segmentation of gastric tumors from CT scans provides useful image information for guiding the diagnosis and treatment of gastric cancer. However, automated gastric tumor segmentation from 3D CT images faces several challenges. The large variation of anisotropic spatial resolution limits the ability of 3D convolutional neural networks (CNNs) to learn features from different views. The background texture of gastric tumor is complex, and its size, shape and intensity distribution are highly variable, which makes it more difficult for deep learning methods to capture the boundary. In particular, while multi-center datasets increase sample size and representation ability, they suffer from inter-center heterogeneity. METHODS: In this study, we propose a new cross-center 3D tumor segmentation method named Hierarchical Class-Aware Domain Adaptive Network (HCA-DAN), which includes a new 3D neural network that efficiently bridges an Anisotropic neural network and a Transformer (AsTr) for extracting multi-scale context features from the CT images with anisotropic resolution, and a hierarchical class-aware domain alignment (HCADA) module for adaptively aligning multi-scale context features across two domains by integrating a class attention map with class-specific information. We evaluate the proposed method on an in-house CT image dataset collected from four medical centers and validate its segmentation performance in both in-center and cross-center test scenarios. RESULTS: Our baseline segmentation network (i.e., AsTr) achieves best results compared to other 3D segmentation models, with a mean dice similarity coefficient (DSC) of 59.26%, 55.97%, 48.83% and 67.28% in four in-center test tasks, and with a DSC of 56.42%, 55.94%, 46.54% and 60.62% in four cross-center test tasks. In addition, the proposed cross-center segmentation network (i.e., HCA-DAN) obtains excellent results compared to other unsupervised domain adaptation methods, with a DSC of 58.36%, 56.72%, 49.25%, and 62.20% in four cross-center test tasks. CONCLUSIONS: Comprehensive experimental results demonstrate that the proposed method outperforms compared methods on this multi-center database and is promising for routine clinical workflows.

A population-based study of the appearances of enteric Campylobacter and non-typhoidal Salmonella infections on computed tomography.

BACKGROUND: Swift identification and diagnosis of gastrointestinal infections are crucial for prompt treatment, prevention of complications, and reduction of the risk of hospital transmission. The radiological appearance on computed tomography could potentially provide important clues to the etiology of gastrointestinal infections. We aimed to describe features based on computed tomography of patients diagnosed with Campylobacter, Salmonella or Shigella infections in South Sweden. METHODS: This was a retrospective observational population-based cohort study conducted between 2019 and 2022 in Skåne, southern Sweden, a region populated by 1.4 million people. Using data from the Department of Clinical Microbiology combined with data from the Department of Radiology, we identified all patients who underwent computed tomography of the abdomen CTA two days before and up to seven days after sampling due to the suspicion of Campylobacter, Salmonella or Shigella during the study period. RESULTS: A total of 215 CTAs scans performed on 213 patients during the study period were included in the study. The median age of included patients was 45 years (range 11-86 years), and 54% (114/213) of the patients were women. Of the 215 CTAs, 80% (n = 172) had been performed due to Campylobacter and 20% (n = 43) due to Salmonella enteritis. CTA was not performed for any individual diagnosed with Shigella during the study period. There were no statistically significant differences in the radiological presentation of Campylobacter and Salmonella infections. CONCLUSION: The most common location of Campylobacter and Salmonella infections was the cecum, followed by the ascending colon. Enteric wall edema, contrast loading of the affected mucosa, and enteric fat stranding are typical features of both infections. The CTA characteristics of Campylobacter and Salmonella are similar, and cannot be used to reliably differentiate between different infectious etiologies.

Diagnostic value of one-stop CT energy spectrum and perfusion for angiogenesis in colon and rectum cancer.

OBJECTIVE: Evaluation of the predictive value of one-stop energy spectrum and perfusion CT parameters for microvessel density (MVD) in colorectal cancer cancer foci. METHODS: Clinical and CT data of 82 patients with colorectal cancer confirmed by preoperative colonoscopy or surgical pathology in our hospital from September 2019 to November 2022 were collected and analyzed retrospectively. Energy spectrum CT images were measured using the Protocols general module of the GSI Viewer software of the GE AW 4.7 post-processing workstation to measure the CT values of the arterial and venous phase lesions and the neighboring normal intestinal wall in a single energy range of 40 kev∼140 kev, and the slopes of the energy spectrum curves (λ) were calculated between 40 kev-90 kev; Iodine concentration (IC), Water concentration (WC), Effective-Z (Eff-Z) and Normalized iodine concentration (NIC) were measured by placing a region of interest (ROI) on the iodine concentration map and water concentration map at the lesion and adjacent to the normal intestinal wall.Perfusion CT images were scanned continuously and dynamically using GSI Perfusion software and analyzed by applying CT Perfusion 4.0 software.Blood volume (BV), blood flow (BF), surface permeability (PS), time to peak (TTP), and mean transit time (MTT) were measured respectively in the lesion and adjacent normal colorectal wall. Based on the pathological findings, the tumors were divided into a low MVD group (MVD < 35/field of view, n = 52 cases) and a high MVD group (MVD ≥ 35/field of view, n = 30 cases) using a median of 35/field of view as the MVD grouping criterion. The collected data were statistically analyzed, the subjects' operating characteristic curve (ROC) was plotted, and the area under curve (AUC), sensitivity, specificity, and Yoden index were calculated for the predicted efficacy of each parameter of the energy spectrum and perfusion CT and the combined parameters. RESULTS: The CT values, IC, NIC, λ, Eff-Z of 40kev∼140kev single energy in the arterial and venous phase of colorectal cancer in the high MVD group were higher than those in the low MVD group, and the differences were all statistically significant (p < 0.05). The AUC of each single-energy CT value in the arterial phase from 40 kev to 120 kev for determining the high or low MVD of colorectal cancer was greater than 0.8, indicating that arterial stage has a good predictive value for high or low MVD in colorectal cancer; AUC for arterial IC, NIC and IC + NIC were all greater than 0.9, indicating that in arterial colorectal cancer, both single and combined parameters of spectral CT are highly effective in predicting the level of MVD. The AUC of 40 kev to 90 kev single-energy CT values in the intravenous phase was greater than 0.9, and its diagnostic efficacy was more representative; The AUC of IC and NIC in venous stage were greater than 0.8, which indicating that the IC and NIC energy spectrum parameters in venous stage colorectal cancer have a very good predictive value for the difference between high and low MVDs, with the greatest diagnostic efficacy in IC.The values of BV and BF in the high MVD group were higher than those in the low MVD group, and the differences were statistically significant (P < 0.05), and the AUC of BF, BV, and BV + BF were 0.991, 0.733, and 0.997, respectively, with the highest diagnostic efficacy for determining the level of MVD in colorectal cancer by BV + BF. CONCLUSION: One-stop CT energy spectrum and perfusion imaging technology can accurately reflect the MVD in living tumor tissues, which in turn reflects the tumor angiogenesis, and to a certain extent helps to determine the malignancy, invasion and metastasis of living colorectal cancer tumor tissues based on CT energy spectrum and perfusion parameters.

Multi-institutional Protocol Guidance for Pediatric Photon-counting CT.

Performing CT in children comes with unique challenges such as greater degrees of patient motion, smaller and densely packed anatomy, and potential risks of radiation exposure. The technical advancements of photon-counting detector (PCD) CT enable decreased radiation dose and noise, as well as increased spatial and contrast resolution across all ages, compared with conventional energy-integrating detector CT. It is therefore valuable to review the relevant technical aspects and principles specific to protocol development on the new PCD CT platform to realize the potential benefits for this population. The purpose of this article, based on multi-institutional clinical and research experience from pediatric radiologists and medical physicists, is to provide protocol guidance for use of PCD CT in the imaging of pediatric patients.

Agreement of Fractional Flow Reserve Estimated by Computed Tomography With Invasively Measured Fractional Flow Reserve: A Systematic Review and Meta-Analysis.

BACKGROUND: Fractional flow reserve (FFR) is the ratio of blood pressure measured distal to a stenosis and pressure proximal to a stenosis. FFR can be estimated noninvasively using computed tomography (CT) although the usefulness of this technique remains controversial. This meta-analysis evaluated the agreement of FFR estimated by CT (FFR-CT) with invasively measured FFR. The study also evaluated the diagnostic accuracy of FFR-CT, defined as the ability of FFR-CT to classify lesions as hemodynamically significant (invasive FFR ≤0.8) or insignificant (invasive FFR >0.8). METHODS AND RESULTS: Forty-three studies reporting on 7291 blood vessels from 5236 patients were included. A moderate positive linear relationship between FFR-CT and invasively measured FFR was observed (Spearman correlation coefficient: 0.67). Agreement between the 2 measures increased as invasively measured FFR values approached 1. The overall diagnostic accuracy, sensitivity and specificity of FFR-CT were 82.2%, 80.9%, and 83.1%, respectively. Diagnostic accuracy of 90% could be demonstrated for FFR-CT values >0.90 and <0.49. The diagnostic accuracy of off-site tools was 79.4% and the diagnostic accuracy of on-site tools was 84.1%. CONCLUSIONS: The agreement between FFR-CT and invasive FFR is moderate although agreement is highest in vessels with FFR-CT >0.9. Diagnostic accuracy varies widely with FFR-CT value but is above 90% for FFR-CT values >0.90 and <0.49. Furthermore, on-site and off-site tools have similar performance. Ultimately, FFR-CT may be a useful adjunct to CT coronary angiography as a gatekeeper for invasive coronary angiogram.

Superiority of 18F-FAPI-42 PET/CT in the detection of primary tumor and management of appendiceal neoplasm to 18F-FDG PET/CT and CE-CT.

BACKGROUND: In the present study, we investigated the value of 18F-fibroblast-activation protein inhibitor (FAPI) positron emission tomography/computed tomography (18F-FAPI-42 PET/CT) to preoperative evaluations of appendiceal neoplasms and management for patients. METHODS: This single-center retrospective clinical study, including 16 untreated and 6 treated patients, was performed from January 2022 to May 2023 at Southern Medical University Nanfang Hospital. Histopathologic examination and imaging follow-up served as the reference standard. 18F-FAPI-42 PET/CT was compared to 18F-fluorodeoxyglucose (18F-FDG) PET/CT and contrast-enhanced CT (CE-CT) in terms of maximal standardized uptake value (SUVmax), diagnostic efficacy and impact on treatment decisions. RESULTS: The accurate detection of primary tumors and peritoneal metastases were improved from 28.6% (4/14) and 50% (8/16) for CE-CT, and 43.8% (7/16) and 85.0% (17/20) for 18F-FDG PET/CT, to 87.5% (14/16) and 100% (20/20) for 18F-FAPI-42 PET/CT. Compared to 18F-FDG PET/CT, 18F-FAPI-42 PET/CT detected more regions infiltrated by peritoneal metastases (108 vs. 43), thus produced a higher peritoneal cancer index (PCI) score (median PCI: 12 vs. 5, P < 0.01). 18F-FAPI-42 PET/CT changed the intended treatment plans in 35.7% (5/14) of patients compared to CE-CT and 25% (4/16) of patients compared to 18F-FDG PET/CT but did not improve the management of patients with recurrent tumors. CONCLUSIONS: The present study revealed that 18F-FAPI-42 PET/CT can supplement CE-CT and 18F-FDG PET/CT to provide a more accurate detection of appendiceal neoplasms and improved treatment decision making for patients.

Predicting balloon shape in percutaneous microcompression : an observational comparative analysis of Meckel's cave imaging and balloon morphology.

Achieving a pear-shaped balloon holds pivotal significance in the context of successful percutaneous microcompression procedures for trigeminal neuralgia. However, inflated balloons may assume various configurations, whether it is inserted into Meckel's cave or not. The absence of an objective evaluation metric has become apparent. To investigate the relationship between the morphology of Meckel's Cave and the balloon used in percutaneous microcompression for trigeminal neuralgia and establish objective criteria for assessing balloon shape in percutaneous microcompression procedures. This retrospective study included 58 consecutive patients with primary trigeminal neuralgia. Data included demographic, clinical outcomes, and morphological features of Meckel's cave and the balloon obtained from MRI and Dyna-CT imaging. MRI of Meckel's cave and Dyna-CT of intraoperative balloon were modeled, and the morphological characteristics and correlation were analyzed. The reconstructed balloon presented a fuller morphology expanding outward and upward on the basis of Meckel's cave. The projected area of balloon was strongly positively correlated with the projected area of Meckel's cave. The Pearson correlation coefficients were 0.812 (P<0.001) for axial view, 0.898 (P<0.001) for sagittal view and 0.813 (P<0.001) for coronal view. Similarity analysis showed that the sagittal projection image of Meckel's cave and that of the balloon had good similarity. This study reveals that the balloon in percutaneous microcompression essentially represents an expanded morphology of Meckel's cave, extending outward and upward. There is a strong positive correlation between the volume and projected area of the balloon and that of Meckel's cave. Notably, the sagittal projection image of Meckel's cave serves as a reliable predictor of the intraoperative balloon shape. This method has a certain generalizability and can help providing objective criteria for judging balloon shape during percutaneous microcompression procedures.

Radiologic-Pathologic Correlation of Cardiac Tumors: Updated 2021 WHO Tumor Classification.

Cardiac tumors, although rare, carry high morbidity and mortality rates. They are commonly first identified either at echocardiography or incidentally at thoracoabdominal CT performed for noncardiac indications. Multimodality imaging often helps to determine the cause of these masses. Cardiac tumors comprise a distinct category in the World Health Organization (WHO) classification of tumors. The updated 2021 WHO classification of tumors of the heart incorporates new entities and reclassifies others. In the new classification system, papillary fibroelastoma is recognized as the most common primary cardiac neoplasm. Pseudotumors including thrombi and anatomic variants (eg, crista terminalis, accessory papillary muscles, or coumadin ridge) are the most common intracardiac masses identified at imaging. Cardiac metastases are substantially more common than primary cardiac tumors. Although echocardiography is usually the first examination, cardiac MRI is the modality of choice for the identification and characterization of cardiac masses. Cardiac CT serves as an alternative in patients who cannot tolerate MRI. PET performed with CT or MRI enables metabolic characterization of malignant cardiac masses. Imaging individualized to a particular tumor type and location is crucial for treatment planning. Tumor terminology changes as our understanding of tumor biology and behavior evolves. Familiarity with the updated classification system is important as a guide to radiologic investigation and medical or surgical management. ©RSNA, 2024 Supplemental material is available for this article.

Accurate object localization facilitates automatic esophagus segmentation in deep learning.

BACKGROUND: Currently, automatic esophagus segmentation remains a challenging task due to its small size, low contrast, and large shape variation. We aimed to improve the performance of esophagus segmentation in deep learning by applying a strategy that involves locating the object first and then performing the segmentation task. METHODS: A total of 100 cases with thoracic computed tomography scans from two publicly available datasets were used in this study. A modified CenterNet, an object location network, was employed to locate the center of the esophagus for each slice. Subsequently, the 3D U-net and 2D U-net_coarse models were trained to segment the esophagus based on the predicted object center. A 2D U-net_fine model was trained based on the updated object center according to the 3D U-net model. The dice similarity coefficient and the 95% Hausdorff distance were used as quantitative evaluation indexes for the delineation performance. The characteristics of the automatically delineated esophageal contours by the 2D U-net and 3D U-net models were summarized. Additionally, the impact of the accuracy of object localization on the delineation performance was analyzed. Finally, the delineation performance in different segments of the esophagus was also summarized. RESULTS: The mean dice coefficient of the 3D U-net, 2D U-net_coarse, and 2D U-net_fine models were 0.77, 0.81, and 0.82, respectively. The 95% Hausdorff distance for the above models was 6.55, 3.57, and 3.76, respectively. Compared with the 2D U-net, the 3D U-net has a lower incidence of delineating wrong objects and a higher incidence of missing objects. After using the fine object center, the average dice coefficient was improved by 5.5% in the cases with a dice coefficient less than 0.75, while that value was only 0.3% in the cases with a dice coefficient greater than 0.75. The dice coefficients were lower for the esophagus between the orifice of the inferior and the pulmonary bifurcation compared with the other regions. CONCLUSION: The 3D U-net model tended to delineate fewer incorrect objects but also miss more objects. Two-stage strategy with accurate object location could enhance the robustness of the segmentation model and significantly improve the esophageal delineation performance, especially for cases with poor delineation results.

Updated 2022 ACC/AHA Guideline Improves Concordance Between TTE and CT in Monitoring Marfan Snydrome and Related Disorders, but Relevant Measurement Differences Remain Frequent.

Background: Patients diagnosed with Marfan syndrome or a related syndrome require frequent aorta monitoring using imaging techniques like transthoracic echocardiography (TTE) and computed tomography (CT). Accurate aortic measurement is crucial, as even slight enlargement (>2 mm) often necessitates surgical intervention. The 2022 ACC/AHA guideline for Aortic Disease Diagnosis and Management includes updated imaging recommendations. We aimed to compare these with the 2010 guideline. Methods: This retrospective study involved 137 patients with Marfan syndrome or a related disorder, undergoing TTE and ECG-triggered CT. Aortic diameter measurements were taken based on the old 2010 guideline (TTE: inner edge to inner edge, CT: external diameter) and the new 2022 guideline (TTE: leading edge to leading edge, CT: internal diameter). Bland-Altman plots compared measurement differences. Results: Using the 2022 guideline significantly reduced differences outside the clinical agreement limit from 49% to 26% for the aortic sinus and from 41% to 29% for the ascending aorta. Mean differences were -0.30 mm for the aortic sinus and +1.12 mm for the ascending aorta using the 2022 guideline, compared to -2.66 mm and +1.21 mm using the 2010 guideline. Conclusion: This study demonstrates for the first time that the 2022 ACC/AHA guideline improves concordance between ECG-triggered CT and TTE measurements in Marfan syndrome patients, crucial for preventing life-threatening aortic complications. However, the frequency of differences >2 mm remains high. Clinical Relevance/Application: Accurate aortic diameter measurement is vital for patients at risk of fatal aortic complications. While the 2022 guideline enhances concordance between imaging modalities, frequent differences >2 mm persist, potentially impacting decisions on aortic repair. The risk of repeat radiation exposure from ECG-triggered CT, considered the 'gold standard', continues to be justified.

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.

Influence of deep learning image reconstruction algorithm for reducing radiation dose and image noise compared to iterative reconstruction and filtered back projection for head and chest computed tomography examinations: a systematic review.

Background: The most recent advances in Computed Tomography (CT) image reconstruction technology are Deep learning image reconstruction (DLIR) algorithms. Due to drawbacks in Iterative reconstruction (IR) techniques such as negative image texture and nonlinear spatial resolutions, DLIRs are gradually replacing them. However, the potential use of DLIR in Head and Chest CT has to be examined further. Hence, the purpose of the study is to review the influence of DLIR on Radiation dose (RD), Image noise (IN), and outcomes of the studies compared with IR and FBP in Head and Chest CT examinations. Methods: We performed a detailed search in PubMed, Scopus, Web of Science, Cochrane Library, and Embase to find the articles reported using DLIR for Head and Chest CT examinations between 2017 to 2023. Data were retrieved from the short-listed studies using Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines. Results: Out of 196 articles searched, 15 articles were included. A total of 1292 sample size was included. 14 articles were rated as high and 1 article as moderate quality. All studies compared DLIR to IR techniques. 5 studies compared DLIR with IR and FBP. The review showed that DLIR improved IQ, and reduced RD and IN for CT Head and Chest examinations. Conclusions: DLIR algorithm have demonstrated a noted enhancement in IQ with reduced IN for CT Head and Chest examinations at lower dose compared with IR and FBP. DLIR showed potential for enhancing patient care by reducing radiation risks and increasing diagnostic accuracy.

The application of deep learning in abdominal trauma diagnosis by CT imaging.

BACKGROUND: Abdominal computed tomography (CT) scan is a crucial imaging modality for creating cross-sectional images of the abdominal area, particularly in cases of abdominal trauma, which is commonly encountered in traumatic injuries. However, interpreting CT images is a challenge, especially in emergency. Therefore, we developed a novel deep learning algorithm-based detection method for the initial screening of abdominal internal organ injuries. METHODS: We utilized a dataset provided by the Kaggle competition, comprising 3,147 patients, of which 855 were diagnosed with abdominal trauma, accounting for 27.16% of the total patient population. Following image data pre-processing, we employed a 2D semantic segmentation model to segment the images and constructed a 2.5D classification model to assess the probability of injury for each organ. Subsequently, we evaluated the algorithm's performance using 5k-fold cross-validation. RESULTS: With particularly noteworthy performance in detecting renal injury on abdominal CT scans, we achieved an acceptable accuracy of 0.932 (with a positive predictive value (PPV) of 0.888, negative predictive value (NPV) of 0.943, sensitivity of 0.887, and specificity of 0.944). Furthermore, the accuracy for liver injury detection was 0.873 (with PPV of 0.789, NPV of 0.895, sensitivity of 0.789, and specificity of 0.895), while for spleen injury, it was 0.771 (with PPV of 0.630, NPV of 0.814, sensitivity of 0.626, and specificity of 0.816). CONCLUSIONS: The deep learning model demonstrated the capability to identify multiple organ injuries simultaneously on CT scans and holds potential for application in preliminary screening and adjunctive diagnosis of trauma cases beyond abdominal injuries.