Computed Tomography Slices Research Articles

ULTRASONOGRAPHY VERSUS COMPUTED TOMOGRAPHY FOR DETECTION OF UROLITHIASIS

Urolithiasis, or urinary stone disease, is a common condition with potentially severe health implications if undetected or untreated. While computed tomography (CT) is the gold standard for detecting urolithiasis, ultrasonography (USG) is often preferred in clinical practice due to its accessibility, non-invasiveness, and lower cost. Evaluating the diagnostic accuracy of USG compared to CT can help guide its use in diagnosing urolithiasis. Objective: To determine the diagnostic accuracy of ultrasonography in detecting urolithiasis, using computed tomography as the gold standard. Methods: This cross-sectional study included 320 patients suspected of urolithiasis, aged 18-70, who were examined at the Department of Radiology in collaboration with the Department of Urology at Sir Ganga Ram Hospital, Lahore from January 2024 to July 2024. CT scans were conducted using the Toshiba Aquilion TSX-101A/4 slice CT Scanner, while ultrasound imaging was performed with the Toshiba Aplio 500, using a 3.5 MHz curved transducer. Data analysis focused on diagnostic accuracy, sensitivity, and specificity of USG compared to CT. Results: The mean age of the patients was 42.08±15.38 years, with a male predominance of 54.4% (174/320). The average body mass index (BMI) was 24.48±4.57 kg/m², and the mean stone size detected was 18.18±10.86 mm. Ultrasonography demonstrated a sensitivity of 95.8%, specificity of 100%, and overall diagnostic accuracy of 96.5% in detecting urolithiasis when compared to CT. Conclusion: Ultrasonography shows high sensitivity (95.8%) and excellent specificity (100%) in detecting urolithiasis, with an average detected stone size of 18.18±10.86 mm. These findings support the use of USG as a reliable, accessible diagnostic tool for urolithiasis, especially when CT is not readily available.

Neural network-based automated proptosis measurement using computed tomography images for patients with thyroid-associated orbitopathy

The purpose of this study was to evaluate the clinical feasibility and reliability of a neural network (NN)-based automated proptosis measurement system using computed tomography (CT) images. An automated proptosis measurement system was developed using the CT images of 200 eyes from 100 patients diagnosed with thyroid-associated orbitopathy. We compared the proptosis value obtained from the proposed automated system with the values obtained from the Hertel exophthalmometer and manual measurements of CT slices. The average measurement values were 17.77 ± 2.47 mm with the Hertel exophthalmometer, 18.87 ± 2.68 mm with the manual measurement of CT slices, and 19.30 ± 2.76 mm with the proposed automated system. There was no significant difference in the proptosis values measured using the manual and automated NN-based methods (p = 0.241). The values obtained from manual measurement and automated measurement using CT images showed excellent agreement with an intraclass correlation coefficient of 0.95. Based on the Bland-Altman plots, the 95% limits of agreement between manual CT and NN-based measurements were much smaller than those between Hertel exophthalmometer measurement and manual and NN-based measurements. In conclusion, automated NN-based measurements could provide a straightforward and efficient method for measuring proptosis using CT images.

CT image segmentation of foxtail millet seeds based on semantic segmentation model VGG16-UNet

Foxtail millet is an important minor cereal crop rich in nutrients. Due to the small size of its seeds, there is little information on the diversity of its seed structure among germplasms, limiting the identification of genes controlling seed development and germination. This paper utilized X-ray computed tomography (CT) scanning technology and deep learning models to reveal the microstructure of foxtail millet seeds, gaining insights into their internal features, distribution, and composition. A total of 100 foxtail millet varieties were scanned with X-ray computed tomography to obtain 3D reconstruction images and slices. Pre-processing steps were adopted to improve image segmentation accuracy, including noise reduction, rotation, contrast enhancement, and brightness enhancement. The experiment revealed that traditional OpenCV image processing methods failed to achieve precise segmentation, whereas deep learning models exhibited outstanding performance in segmenting seed CT slice images. We compared UNet, PSPNet, and DeepLabV3 models, selected different backbones and optimizers based on the dataset, and continuously adjusted learning rates and maximum training epochs to train the models. Results demonstrated that VGG16-UNet achieved an accuracy of 99.19% on the foxtail millet seed CT slice image dataset, outperforming PSPNet and DeepLabV3 models. Compared to ResNet-UNet, VGG16-UNet shows an improvement of approximately 3.18% in accuracy, demonstrating superior performance in accurately segmenting the inner glume, outer glume, embryo, and endosperm under various adhesion conditions. Accurate segmentation of foxtail millet CT images enables analysis of embryo size, endosperm size, and glume thickness, which impact germination, growth, and nutrition. This study fills a gap in small grain structure research, offering insights to optimize agriculture and molecular breeding for improved yield and quality.

Radiation exposure of patients undergoing multiple-slice CT scans in Romania

Abstract Ionizing radiation is largely used in medicine due to its role in the diagnosis and treatment of patients. In the past 50 years computed tomography (CT) has become indispensable for medical practice because it offers sub-millimeter cross-sectional imaging of any region of the body. New CT scanners use thinner slices, providing more accurate results, but also an increased radiation exposure with possible long-term harmful effects for the patients. The parameter used as an indicator of radiation dose in CT is the dose-length product (DLP). The study included 4507 patients who were CT scanned using a 16, 32, 64 or 128-slice CT machine. DLP was measured for head, sinuses, thorax, abdomen, pelvis, abdomen+pelvis, and trunk scans - regions for which the health authorities have established national diagnostic reference levels (DRL - the expected radiation dose received by a patient in CT). One-way ANOVA was used to analyse data considering the use of contrast agents, the gender, age, and weight of patients. Average DLP values were higher (up to 80%) than national DRL for all regions in 64 and 128-slice scans. The highest DLP values were registered in contrast CT for all scanners (p < 0.01); the maximum average (a threefold increase compared to DRL) was measured in the 32-slice sinuses contrast CT. Average DLP correlated with patients’ gender and weight (p < 0.01), but the influence was lower compared to the use of contrast agents. In conclusion, CT scans acquired using thinner slices lead to more accurate results and high-quality images, but the study suggests an increased radiation exposure. CT has an uncontestable contribution to a precise and rapid diagnosis, but more attention must be given to the stochastic effects risks associated with the radiation dose received during a CT scan. Are all scans necessary? Do all patients really need to receive high doses? The role of the practitioner is decisive, and the justifiable process is “a must” in all radiodiagnosis practices. Key messages • CT doses must be kept “as low as reasonably achievable” while maintaining diagnostic image quality. • Justification, dose limitation, and optimisation are fundamental principles of radioprotection.

Diagnosis of COVID-19 from computed tomography slices using flower pollination algorithm, k-nearest neighbor, and support vector machine classifiers

Coronavirus disease 19 (COVID-19), caused by the severe acute respiratory syndrome-coronavirus-2 virus, is commonly diagnosed through imaging techniques such as computed tomography (CT) scans, which reveal characteristic lung lesions. In this study, we propose a computer-aided diagnosis (CAD) system to assist in the early detection of COVID-19 from CT lung slices, leveraging advanced machine-learning algorithms for precise and efficient analysis. To achieve this, we developed a CAD system that diagnoses COVID-19 from CT lung slices. An adaptive Wiener filter was applied to remove noise from the CT images. The chest tissues were then segmented using an optimal thresholding method to extract regions of interest, which represent the COVID-19 lesions under investigation. The feature vectors were divided into training and testing with an 80/20 ratio. A wrapper-based flower pollination algorithm was employed alongside the k-nearest neighbor classifier to select the optimal feature set. These selected features were subsequently used to train a support vector machine (SVM) classifier. With feature selection, the SVM achieved an accuracy of 91.30% on a real-time dataset, outperforming seven other machine learning classifiers (radial basis function-SVM, k nearest neighbor, linear discriminant analysis, random forest, naïve Bayes, AdaBoost, extreme gradient boosting) and four deep learning classifiers (convolutional neural network, recurrent neural network, long short term memory, Bidirectional long short term memory). For the publicly available COVID-19 CT dataset, an accuracy of 88.18% was achieved. In conclusion, our COVID-19 CAD system improves diagnostic accuracy, with future work aimed at enhancing efficiency and expanding to covariant detection and severity assessment.

MaskAppendix: Backbone-Enriched Mask R-CNN Based on Grad-CAM for Automatic Appendix Segmentation.

A leading cause of emergency abdominal surgery, appendicitis is a common condition affecting millions of people worldwide. Automatic and accurate segmentation of the appendix from medical imaging is a challenging task, due to its small size, variability in shape, and proximity to other anatomical structures. In this study, we propose a backbone-enriched Mask R-CNN architecture (MaskAppendix) on the Detectron platform, enhanced with Gradient-weighted Class Activation Mapping (Grad-CAM), for precise appendix segmentation on computed tomography (CT) scans. In the proposed MaskAppendix deep learning model, ResNet101 network is used as the backbone. By integrating Grad-CAM into the MaskAppendix network, our model improves feature localization, allowing it to better capture subtle variations in appendix morphology. We conduct extensive experiments on a dataset of abdominal CT scans, demonstrating that our method achieves state-of-the-art performance in appendix segmentation, outperforming traditional segmentation techniques in terms of both accuracy and robustness. In the automatic segmentation of the appendix region in CT slices, a DSC score of 87.17% was achieved with the proposed approach, and the results obtained have the potential to improve clinical diagnostic accuracy. This framework provides an effective tool for aiding clinicians in the diagnosis of appendicitis and other related conditions, reducing the potential for diagnostic errors and enhancing clinical workflow efficiency.

A Study on the Breakthrough Pressure of Porous Rock Formations in Gas Storage Reservoirs

Breakthrough pressure of natural gas in rock is an important evaluation parameter in gas reservoir development. In this study, experimental measurements of porosity and permeability of the caprock core were carried out. A digital core model was obtained by body rendering of rock CT (Computed Tomography) slices using Avizo software. Then, the grid model was imported into Fluent software for a numerical simulation of natural gas breakthrough pressure in the core. The innovations of this paper include (1) an accurate calculation method of cap breakthrough pressure based on the VOF method; (2) the method performs a perceptual analysis of average pore radius, wettability, roughness height, and viscosity ratio.

Role of computed tomography enteroclysis in the evaluation of small bowel diseases

Background: Examining small bowel loops within the peritoneal cavity poses significant challenges because of their length and overlapping nature. Aims and Objectives: This study aimed to evaluate the utility of computed tomography enteroclysis (CTE) in diagnosing small intestinal diseases and to understand its clinical application. Materials and Methods: The study was conducted from November 2019 to October 2021 in 30 patients undergoing CTE treatment. The day before treatment, bowel preparation included a low-residue diet, lots of fluids, an overnight fast of 8–12 h, and laxatives. A 16-slice Aquilion Lightning computed tomography (CT) scanner was used in general hospitals, and a 128-slice GE Healthcare helical CT scanner was used in multispecialty hospitals. Results: Among the 30 patients, 20 (66.7%) were male, and 10 (33.3%) were female, predominantly aged between 30 and 60 years old. Positive findings were noted in 23 patients (76.7%), including intestinal tuberculosis, inflammatory bowel disease, small bowel obstruction, and neoplastic lesions. Various small intestinal pathologies, such as non-specific edema, intussusception, and ischemic bowel disease, were also identified. Conclusion: This raises the standard of current small bowel imaging by delivering suitable intestinal distension using the enteral volume challenge principle and helps identify and characterize small-bowel problems.

Radiomics and 256-slice-dual-energy CT in the automated diagnosis of mild acute pancreatitis: the innovation of formal methods and high-resolution CT.

Acute pancreatitis (AP) is a common disease, and several scores aim to assess its prognosis. Our study aims to automatically recognize mild AP from computed tomography (CT) images in patients with acute abdominal pain but uncertain diagnosis from clinical and serological data through Radiomic model based on formal methods (FMs). We retrospectively reviewed the CT scans acquired with Dual Source 256-slice CT scanner (Somatom Definition Flash; Siemens Healthineers, Erlangen, Germany) of 80 patients admitted to the radiology unit of Antonio Cardarelli hospital (Naples) with acute abdominal pain. Patients were divided into 2 groups: 40 underwent showed a healthy pancreatic gland, and 40 affected by four different grades (CTSI 0, 1, 2, 3) of mild pancreatitis at CT without clear clinical presentation or biochemical findings. Segmentation was manually performed. Radiologists identified 6 patients with a high expression of diseases (CTSI 3) to formulate a formal property (Rule) to detect AP in the testing set automatically. Once the rule was formulated, and Model Checker classified 70 patients into "healthy" or "unhealthy". The model achieved:accuracy81%, precision 78% and recall81%. Combining FMs results with radiologists agreement, and applying the mode in clinical practice, the global accuracy would have been 100%. Our model was reliable to automatically detect mild AP at primary diagnosis even in uncertain presentation and it will be tested prospectively in clinical practice.

Automatic detection of hidden defects and qualification of additively manufactured parts using X-ray computed tomography and computer vision

Additive manufacturing (AM) has the unique capability to produce parts with complex three-dimensional structures with features that are internal to the part and hidden from view. Such internal features are difficult to inspect and may have defects that affect the function of a part. X-ray computed tomography (CT) is one of the only methods for nondestructive inspection (NDI) of the interior of AM parts. This paper explores how machine learning (ML) methods can be used to analyze CT scans of AM parts to automatically identify the presence of defects in geometric features. We designed a nozzle part with internal three-dimensional (3D) channels and introduced five different synthetic defects whose presence could affect the nozzle performance. A resin-based AM process fabricated 155 nozzle parts that included both defect-free parts and parts with synthetic defects. CT scans were collected for each part and processed into 68,510 image cross sections. The extracted images were used to train an ML model based on the ResNet34 architecture. The model can automatically identify and classify defects in individual CT slice images with over 98% accuracy. High model accuracy is possible with training on as few as 30 parts. The research demonstrates the potential of ML methods to automatically identify hidden defects and qualify AM parts using X-ray CT scans.

Intelligent void identification of particle packing system of caved ore and rock

Analyzing the void structure of the packing system of caved ore and rock and establishing the causal relationship between void structure and macroscopic properties are critical research areas in the field of gravity flow. However, the current method of analyzing void structure primarily involves manual threshold segmentation, leading to low analysis efficiency and the need for improved accuracy. This study introduces a novel approach using a dataset of two-dimensional Computed Tomography (CT) slices of an irregular limestone particle packing system. The primary goal is to enhance existing algorithms by focusing on data and network aspects to address the challenges in identifying this type of image dataset. As a result, a model named Data Attention gate with Recurrent Residual convolutional neural network based on U-net (DA2RU-net) is developed and evaluated for its convergence, accuracy, and robustness. The findings indicate the following: (1) To identify voids in the packing system of caved ore and rock using CT slices, it is advisable to employ 1000 images with dimensions between 300 and 400 pixels. It is essential to ensure that both large and small voids are represented in the image dataset. (2) The DA2RU-net model demonstrates superior performance with an average Dice value of 0.9859 for identifying large voids and 0.9654 for identifying small voids, surpassing other iterations of the U-net model and traditional algorithms. (3) The DA2RU-net model shows robustness to variations in brightness levels.

Tomographic Assessment of Fusion Rate, Implant-Endplate Contact Area, Subsidence, and Alignment With Lumbar Personalized Interbody Implants at 1-Year Follow-Up.

Incongruity between irregularly shaped vertebral endplates and the uniform surfaces of stock interbody fusion cages has been identified as contributing to cage subsidence, pseudarthrosis, and unpredictable alignment. Advances in manufacturing techniques have driven the development of personalized interbody cages (PICs) that can match individual endplate morphology and provide the exact shape and size needed to fill the disc space and achieve the planned correction. This study used computed tomography (CT) imaging to evaluate the implant-endplate contact area, fusion, subsidence, and achievement of planned alignment correction in patients receiving PIC devices. This retrospective study included patients treated for adult spinal deformity at a single site and implanted with PIC devices at L4 to L5 or L5 to S1 for segmental stabilization and alignment correction, who received 1-year postoperative CT images as part of their standard of care. An evaluation using 3-dimensional thin-section scans was conducted. Implant-endplate contact and signs of fusion were assessed in each CT slice across both endplates. The degree of subsidence as well as measures of segmental and global lumbar alignment were also assessed. Fifteen patients were included in the study, with a mean age of 68.2 years. Follow-up ranged between 9 and 14 months. Twenty-six total lumbar levels were implanted; 20 with PIC devices via the anterior lumbar interbody fusion approach, 2 with stock cages via the anterior lumbar interbody fusion approach, and 4 with PIC devices via the transforaminal lumbar interbody fusion approach. CT analysis of PIC-implanted levels found an overall implant-endplate contact area ratio of 93.9%, a subsidence rate of 4.5%, a fusion rate of 100%, and satisfactory segmental and global lumbar correction compared with the preoperative plan. PIC implants can provide nearly complete contact with endplate surfaces regardless of the individual endplate morphology. Subsidence, fusion, and alignment assessments in this tomographic study illustrated results consistent with the benefits of a personalized interbody implant.

Automatic personal identification using a single CT image.

Computer vision (CV) mimics human vision, enabling computers to automatically compare radiological images from recent examinations with a large image database for unique identification, crucial in emergency scenarios involving unknown patients or deceased individuals. This study aims to extend a CV-based personal identification method from orthopantomograms (OPGs) to computed tomography (CT) examinations using single CT slices. The study analyzed 819 cranial computed tomography (CCT) examinations from 722 individuals, focusing on single CT slices from six anatomical regions to explore their potential for CV-based personal identification in 69 procedures. CV automatically identifies and describes interesting features in images, which can be recognized in a reference image and then designated as matching points. In this study, the number of matching points was used as an indicator for identification. Across six different regions, identification rates ranged from 41/69 (59%) to 69/69 (100%) across over 700 possible identities. Comparison of images from the same individual achieved higher matching points, averaging 6.32 ± 0.52% (100% represents the maximum possible matching points), while images of different individuals averaged 0.94 ± 0.15%. Reliable matching points are found in the teeth, maxilla, cervical spine, skull bones, and paranasal sinuses, with the maxillary sinuses and ethmoidal cells being particularly suitable for identification due to their abundant matching points. Unambiguous identification of individuals based on a single CT slice is achievable, with maxillary sinus CT slices showing the highest identification rates. However, metal artifacts, especially from dental prosthetics, and various head positions can hinder identification. Radiology possesses a multitude of reference images for a CV database, facilitating automated CV-based personal identification in emergency examinations or cases involving unknown deceased individuals. This enhances patient care and communication with relatives by granting access to medical history. Unknown individuals in radiology or forensics pose challenges, addressed through automatic CV-based identification methods. A single CT slice highlighting the maxillary sinuses is particularly effective for personal identification. Radiology plays a pivotal role in automated personal identification by leveraging its extensive image database.

Fabrication and Evaluation of Thyroid Shield from Silicone Rubber-Cooper and its Comparison to Tube Current Modulation in CT Examination

This study aimed to fabricate thyroid shields made from Silicone Rubber (SR)-Cooper (Cu) material, analyze the effectiveness of SR-Cu shields in reducing radiation dose, and compare them with tube current modulation (TCM) in computed tomography (CT) examination. Thyroid shields were made from SR-Cu with Cu percentages of 0, 5, 10, 15, and 20%. The thyroid shields were positioned over the neck of the anthropomorphic phantom. Scanning was performed using a GE 128-slice CT scanner with fixed tube current of 150 mA and tube current modulation (TCM). The elatiscity of thyroid shields was tested using an universal testing machine (UTM). The ability of thyroid shields for dose reduction was measured using a 10X6-3CT Radcal detector, and quality of the resulted images was characterized with metric of signal-to-noise ratio (SNR) at the anterior, posterior, and lateral areas of the neck area of anthropomorphic phantom. It is found that the elasticity of the thyroid shields increased from 0.09 to 0.12 N/mm2 for Cu percentages from 0 to 20%. The measured dose decreased as the percentage of Cu increased. 20% of SR-Cu was able to reduce the dose by 32.4% for the fix tube current. In comparison, the TCM technique reduced the dose by 44.5%. Therefore, dose reduction using the developed shields is lower than using TCM approach. It is also found that implementation of the thyroid shields did not reduce image quality significantly. It is found that there were no apparent artifacts in the images. The highest SNR was found in the image with 20% SR-Cu, which was 3.84. In comparison, the SNR using the TCM approach was 3.59. In conclusion SR-Cu shields were successfully developed and they reduced dose with relatively consistent of image quality.

Reproducibility of a single-volume dynamic CT myocardial blood flow measurement technique: validation in a swine model

BackgroundWe prospectively assessed the reproducibility of a novel low-dose single-volume dynamic computed tomography (CT) myocardial blood flow measurement technique.MethodsThirty-four pairs of measurements were made under rest and stress conditions in 13 swine (54.3 ± 12.3 kg). One or two acquisition pairs were acquired in each animal with a 10-min delay between each pair. Contrast (370 mgI/mL; 0.5 mL/kg) and a diluted contrast/saline chaser (0.5 mL/kg; 30:70 contrast/saline) were injected peripherally at 5 mL/s, followed by bolus tracking and acquisition of a single volume scan (100 kVp; 200 mA) with a 320-slice CT scanner. Bolus tracking and single volume scan data were used to derive perfusion in mL/min/g using a first-pass analysis model; the coronary perfusion territories of the left anterior descending (LAD), left circumflex (LCx), and right coronary artery (RCA) were automatically assigned using a previously validated minimum-cost path technique. The reproducibility of CT myocardial perfusion measurement within the LAD, LCx, RCA, and the whole myocardium was assessed via regression analysis. The average CT dose index (CTDI) of perfusion measurement was recorded.ResultsThe repeated first (Pmyo1) and second (Pmyo2) single-volume CT perfusion measurements were related by Pmyo2 = 1.01Pmyo1 − 0.03(ρ = 0.96; RMSE = 0.08 mL/min/g; RMSE = 0.07 mL/min/g) for the whole myocardium, and by Preg2 = 0.86Preg1 + 0.13(ρ = 0.87; RMSE = 0.31 mL/min/g; RMSE = 0.29 mL/min/g) for the LAD, LCx, and RCA perfusion territories. The average CTDI of the single-volume CT perfusion measurement was 10.5 mGy.ConclusionThe single-volume CT blood flow measurement technique provides reproducible low-dose myocardial perfusion measurement using only bolus tracking data and a single whole-heart volume scan.Relevance statementThe single-volume CT blood flow measurement technique is a noninvasive tool that reproducibly measures myocardial perfusion and provides coronary CT angiograms, allowing for simultaneous anatomic-physiologic assessment of myocardial ischemia.Key PointsA low-dose single-volume dynamic CT myocardial blood flow measurement technique is reproducible.Motion misregistration artifacts are eliminated using a single-volume CT perfusion technique.This technique enables combined anatomic-physiologic assessment of coronary artery disease.Graphical

Thoracic sarcopenia measured by Hounsfield unit average calculation predicts morbidity and mortality in coronary artery bypass grafting.

The aim of the study was to investigate the potential prognostic role of preoperative measurement of erector spinae myosteatosis with Hounsfield unit average calculation as a marker for sarcopenia and frailty in patients undergoing coronary bypass surgery. Preoperative computer tomography-derived measurements of 479 consecutive patients undergoing coronary bypass surgery between January 2017 and December 2019 were retrospectively performed. The erector spinae muscle at the level of the 12th vertebra was manually outlined bilaterally on the axial computer tomography slices and Hounsfield unit average calculation was performed. The lower quartile of muscle density values was defined as myosteatotic and thus sarcopenic. Sarcopenic (n = 121) versus non-sarcopenic patients (n = 358) were compared regarding postoperative morbidity and short- and long-term mortality. Results were adjusted for age, body mass index, atrial fibrillation and hypertension using inverse probability weighting. Sarcopenia was associated with higher 30-day mortality (4.1% vs 0.8%; P = 0.012), mid-term mortality after 1 year (9.3% vs 3.1%; P = 0.047) and 2 years (10.8% vs 4.2%; P = 0.047). Long-term mortality (5 years) was 20.8% for sarcopenic and 13.0% for non-sarcopenic patients but was not found to be significantly different (P = 0.089). Sarcopenia was associated with higher rates of reintubation (7.5% vs 1.1%; P < 0.001), sternal wound infections (7.5% vs 2.8%; P = 0.039) and acute kidney injury requiring haemodialysis (2.5% vs 0.4%; P = 0.021). In patients undergoing coronary bypass surgery, sarcopenia was associated with increased short-term mortality, mid-term mortality and morbidity. The measurement of erector spinae myosteatosis could be an easy and useful parameter in preoperative risk assessment.

IQAGPT: computed tomography image quality assessment with vision-language and ChatGPT models

Large language models (LLMs), such as ChatGPT, have demonstrated impressive capabilities in various tasks and attracted increasing interest as a natural language interface across many domains. Recently, large vision-language models (VLMs) that learn rich vision–language correlation from image–text pairs, like BLIP-2 and GPT-4, have been intensively investigated. However, despite these developments, the application of LLMs and VLMs in image quality assessment (IQA), particularly in medical imaging, remains unexplored. This is valuable for objective performance evaluation and potential supplement or even replacement of radiologists’ opinions. To this end, this study introduces IQAGPT, an innovative computed tomography (CT) IQA system that integrates image-quality captioning VLM with ChatGPT to generate quality scores and textual reports. First, a CT-IQA dataset comprising 1,000 CT slices with diverse quality levels is professionally annotated and compiled for training and evaluation. To better leverage the capabilities of LLMs, the annotated quality scores are converted into semantically rich text descriptions using a prompt template. Second, the image-quality captioning VLM is fine-tuned on the CT-IQA dataset to generate quality descriptions. The captioning model fuses image and text features through cross-modal attention. Third, based on the quality descriptions, users verbally request ChatGPT to rate image-quality scores or produce radiological quality reports. Results demonstrate the feasibility of assessing image quality using LLMs. The proposed IQAGPT outperformed GPT-4 and CLIP-IQA, as well as multitask classification and regression models that solely rely on images.

Can acetabular dysplasia be measured on axial CT? A measurement for trauma surgeons

BackgroundAcetabular dysplasia has been theorized as a risk factor instability amongst common acetabular fractures, such as posterior wall (PW) fractures. However, common radiographic measurements often cannot be acquired in trauma patients. We evaluated axial computed tomography (CT) scans to identify novel, easily-obtained measurements that correlate with acetabular dysplasia for use in surgical indications and planning. MethodsPatients with known acetabular dysplasia undergoing elective periacetabular osteotomy were selected. A different group of patients without pelvic ring or acetabular fractures from an institutional trauma registry were selected as a comparison group. Standard indices of dysplasia were collected, such as center-edge angle (CEA). Acetabular geometric measurements were taken at three axial levels: 0 – cranial CT slice at the dome; 1 – cranial CT where the dome is an incomplete circle; 2 – cranial CT with femoral head visible. Distances between levels were also calculated: Levels 0–1 (Dome Height; DH), Levels 1–2 (Head Height; HH), and Dome-Head Difference (DH – HH = DHD). ResultsDH, HH, and DHD were all significantly correlated with CEA, Tonnis angle, and Sharp's angle in dysplastic hips. All dysplastic hips had DH ≤ 2.5 mm and HH ≥ 1.25 mm. DHD ≤ 0 mm was most specific (93.6 % sensitive, 77.3 % specific) for predicting dysplasia. ConclusionDH ≤ 2.5 mm, HH ≥ 1.25 mm, and DHD ≤ 0 mm were independently associated with dysplasia on axial CT scans. These measurements may be quickly and easily used by trauma surgeons to assess a trauma-based axial CT scan for acetabular dysplasia.

Radiomic feature robustness in CT scans affected by fiducial marker induced streak artifacts for patients with hepatocellular carcinoma.

Stereotactic body radiation therapy for hepatocellular carcinoma necessitates the implantation of gold fiducial markers in the liver, resulting in artifacts on computed tomography (CT) scans, which affect radiomic feature values. This report aims to assess the effect of these artifacts on radiomic features and how removing CT slices affects radiomic features extracted from 3D regions of interest (ROI). First, the range variation in 38 tumor contours unaffected by artifacts was assessed after sequentially and randomly removing 25%, 50%, 75% of slices. Subsequently, the agreement of feature values before and after removing ROI slices containing artifacts from 186 patients' CT scans was assessed with Lin's concordance correlation coefficient (CCC) and a Wilcoxon signed-rank test. In artifact-free tumor volumes, at least 71% of features remain robust with up to 50% of slices removed, while 56% remains robust with up to 75% of slices removed. When comparing contours before and after removing slices containing artifacts, around a third of features in the tumor contour and surrounding area remain robust (CCC>0.9), compared to 44% in the healthy liver. Concerning the tumor, 13% (Gray Level Size Zone Matrix) to 61% (first order) of the features remain robust (CCC>0.9). Over 90% of features differ significantly as assessed by Wilcoxon signed-rank test, however. This study demonstrates that removing slices containing artifacts is a feasible solution for the CT fiducial problem in this patient population and provides insight into which features are affected.

Generative adversarial network-based reconstruction of healthy anatomy for anomaly detection in brain CT scans.

To help radiologists examine the growing number of computed tomography (CT) scans, automatic anomaly detection is an ongoing focus of medical imaging research. Radiologists must analyze a CT scan by searching for any deviation from normal healthy anatomy. We propose an approach to detecting abnormalities in axial 2D CT slice images of the brain. Although much research has been done on detecting abnormalities in magnetic resonance images of the brain, there is little work on CT scans, where abnormalities are more difficult to detect due to the low image contrast that must be represented by the model used. We use a generative adversarial network (GAN) to learn normal brain anatomy in the first step and compare two approaches to image reconstruction: training an encoder in the second step and using iterative optimization during inference. Then, we analyze the differences from the original scan to detect and localize anomalies in the brain. Our approach can reconstruct healthy anatomy with good image contrast for brain CT scans. We obtain median Dice scores of 0.71 on our hemorrhage test data and 0.43 on our test set with additional tumor images from publicly available data sources. We also compare our models to a state-of-the-art autoencoder and a diffusion model and obtain qualitatively more accurate reconstructions. Without defining anomalies during training, a GAN-based network was used to learn healthy anatomy for brain CT scans. Notably, our approach is not limited to the localization of hemorrhages and tumors and could thus be used to detect structural anatomical changes and other lesions.