64-slice CT Research Articles

Prior Attention Network for Multi-Lesion Segmentation in Medical Images.

The accurate segmentation of multiple types of lesions from adjacent tissues in medical images is significant in clinical practice. Convolutional neural networks (CNNs) based on the coarse-to-fine strategy have been widely used in this field. However, multi-lesion segmentation remains to be challenging due to the uncertainty in size, contrast, and high interclass similarity of tissues. In addition, the commonly adopted cascaded strategy is rather demanding in terms of hardware, which limits the potential of clinical deployment. To address the problems above, we propose a novel Prior Attention Network (PANet) that follows the coarse-to-fine strategy to perform multi-lesion segmentation in medical images. The proposed network achieves the two steps of segmentation in a single network by inserting a lesion-related spatial attention mechanism in the network. Further, we also propose the intermediate supervision strategy for generating lesion-related attention to acquire the regions of interest (ROIs), which accelerates the convergence and obviously improves the segmentation performance. We have investigated the proposed segmentation framework in two applications: 2D segmentation of multiple lung infections in lung CT slices and 3D segmentation of multiple lesions in brain MRIs. Experimental results show that in both 2D and 3D segmentation tasks our proposed network achieves better performance with less computational cost compared with cascaded networks. The proposed network can be regarded as a universal solution to multi-lesion segmentation in both 2D and 3D tasks. The source code is available at https://github.com/hsiangyuzhao/PANet.

The proximity of the middle hepatic vein to the hepatic hilus: a retrospective radiological study.

The middle hepatic vein (MHV) is an important landmark in anatomical hemihepatectomy. The proximity between the MHV and the hilar plate was suspected to be associated with tumor exposure during left hemihepatectomy for advanced perihilar cholangiocarcinoma and is reported to facilitate a dorsal approach to the MHV during laparoscopic hemihepatectomy. However, the precise distance between these locations is unknown. To investigate the "accurate and normal" distance between the MHV and the hilar plate, the present study focused on patients who presented without perihilar tumor. One hundred and sixty-eight consecutive patients who underwent pancreatoduodenectomy were included. Retrospective radiological measurement was performed using preoperative multi-detector row CT. The optimized CT slices perpendicular to the MHV were made using the multiplanar reconstruction technique. The shortest distance between the MHV and the hilar plate was measured on the left and right sides on the perpendicular slices. The diameters of the left and right hepatic ducts were also measured. The distance was 9.0mm (1.9-20.0mm) on the left side and 11.3mm (2.3-21.8) on the right side (p < 0.001). The distance on the left side was < 10mm in 60% of patients (n = 100). Only one-third of patients (n = 55) had a distance of ≥ 10mm on both sides. As the hepatic ducts became more dilated, the distance from the MHV to the hilar plate became shorter. The MHV was located in close proximity to the hepatic hilus, especially on the left side.

The value of bronchocele attenuation in pulmonary computed tomography in assessment of allergic bronchopulmonary aspergillosis in the background of cystic fibrosis: A cross-sectional study.

There is no specific test in the definitive diagnostic approach to Allergic bronchopulmonary aspergillosis (ABPA) especially in the background of cystic fibrosis, but comprehensive and simultaneous clinical, radiological and serological examination will be the basis of ABPA diagnosis. The increasing in attenuation of bronchoceles in imaging has recently been proposed as a valuable diagnostic criterion. The present study aimed to assess bronchocele attenuation in pulmonary CT scan of patients with complicated cystic fibrosis for diagnosis of ABPA. This cross-sectional study was performed on 74 consecutive patients aged 3-18 years suffering cystic fibrosis presented with exacerbation of pulmonary symptoms and were suspected of having ABPA. All were examined by 16 Slice CT Scan and the density of bronchoceles above 5mm in diameter were measured in Hounsfield unit. The total serum IgE titer, skin prick test for aspergillus and anti-aspergillus IgG and IgE level were obtained for all subjects and both cutoff values of IgE level (>500 IU/mL and >1000 IU/mL) were considered as the criteria for ABPA diagnosis. Considering IgE level of greater than 500 IU/mL and 1000 IU/mL as the diagnostic criteria, 24.3% and 10.8% had evidence of ABPA, respectively. Considering the two pointed diagnostic IgE ranges and based on the analysis of the area under the ROC curve, bronchocele attenuation could effectively predict the presence of ABPA with the best cutoff values of 37.25 (with a sensitivity of 70.6% and a specificity of 66.7%) and 40.00 (with a sensitivity of 85.7% and a specificity of 65.1%), respectively. The presence of bronchocele and an increase in its attenuation on CT scan will be diagnostic for the occurrence of ABPA.

Diagnostic accuracy of lower-dose cardiac CT in evaluating young infants with non-coronary complex congenital heart disease on a 64-slice multidetector CT scanner.

Redoable precise and non-invasive diagnostic imaging modality with the least radiation dose is essential for infants with congenital heart disease (CHD) . To investigate the accuracy and estimate the radiation dose of our cardiac computed tomography (CCT) protocol. A total of 82 infants with CHD underwent non-ECG-gated CCT without contrast timing scanning techniques and were retrospectively studied. The image quality and radiation dose were estimated. The radiation dose was compared statistically to virtual retrospective ECG-gated and prospective ECG-triggering scanning modes. The diagnostic accuracy was assessed assuming the surgical results as the diagnostic gold standard. Most exams showed a high quality with low radiation doses compared to previous studies. The mean effective dose (ED) was 0.39 ± 1.2, significantly lower than that of the virtual retrospective ECG-gated and prospective ECG-triggering scanning and lower than in previous studies. Our CCT protocol has achieved a diagnostic accuracy of 99.52% with a sensitivity of 94.83% and specificity of 99.91%. Non-ECG-gated CCT without contrast timing techniques can detect the non-coronary cardiovascular defects of CHD in infants with an ultralow radiation dose and a high diagnostic accuracy.

Retrospective study of Ct brain imaging findings indicated for infants at a tertiary care hospital

Computed tomography in the diagnosis of pathologies in children is becoming increasingly popular.&#x0D; The aim: to study findings of referrals for CT scans (Computerised Tomography) of the brain in children in the department of Radiology in a developing environment.&#x0D; Materials and methods: retrospective imaging observational study was done in a 1000 bedded tertiary care hospital in South India in the year 2022 from April to August for 5 months under 16 slice CT scan. Cases are referred from the department of Pediatrics to the department of Radiology as part of the routine clinical evaluation and treatment protocol.&#x0D; Results: Our sample includes 100 infants; after exclusion criteria total of 60 infants' brain CTs were taken into account, and analysed their imaging from radiology department records. Of 60 cases, 18 (30 %) showed HIE Pattern, and 42 (70 %) had normal plain CT Brain findings. However, due to clinical signs and symptoms, they are evaluated under сontrast CT imaging (after checking creatinine levels), showing 20 cases (33.3 %) are standard, 10 cases (16.6 %) show meningitis, and 2 cases (3.3 %) show SOL. Out of 18 cases of HIE, 10 cases (16.6 %) are under less than 6 months and they undergone neuro sonogram showing 4 cases (6.6 %) normal NSG, 2 cases (3.3 %) showing grade IV HIE, another 2 cases (3.3 %) showing grade II, III HIE Findings.&#x0D; Conclusions: Our study concluded that plain CT brain showed normal in the majority of the cases which came to the department of Radiology after admission to the hospital. Those cases with strong clinical history are evaluated with contrast CT and evaluated the findings. After contrast imaging, most cases showing their infective, obstructive, and other causes of illness are interpreted. However, its role is minimal for seizures in which CT/CECT shows normal study. Under 6 months, NSG is a suitable method for evaluating and screening the infant's brain.

An Automatic Segmentation Method for Lung Tumor Based on Improved Region Growing Algorithm.

In medical image processing, accurate segmentation of lung tumors is very important. Computer-aided accurate segmentation can effectively assist doctors in surgery planning and treatment decisions. Although the accurate segmentation results of lung tumors can provide a reliable basis for clinical treatment, the key to obtaining accurate segmentation results is how to improve the segmentation performance of the algorithm. We propose an automatic segmentation method for lung tumors based on an improved region growing algorithm, which uses the prior information on lung tumors to achieve an automatic selection of the initial seed point. The proposed method includes a seed point expansion mechanism and an automatic threshold update mechanism and takes the combination of multiple segmentation results as the final segmentation result. In the lung image database consortium (LIDC-IDRI) dataset, we designed 10 experiments to test the proposed method and compare it with 4 popular segmentation methods. The experimental results show that the average dice coefficient obtained by the proposed method is 0.936 ± 0.027, and the average Jaccard distance is 0.114 ± 0.049. The average dice coefficient obtained by the proposed method is 0.107, 0.053, 0.040, and 0.156, higher than that of the other four methods, respectively. This study proves that the proposed method can automatically segment lung tumors in CT slices and has suitable segmentation performance.

Traceability and quality assessment of Douglas fir (Pseudotsuga menziesii (Mirb.) Franco) logs: the TreeTrace_Douglas database

Key messageThe TreeTrace_Douglas database includes images and measurements at several stages of the processing of Douglas fir logs, from sawmill logyard to machine grading and destructive testing of boards, and is suitable for research on quality assessment and traceability. A total of 52 long logs, 156 short logs, 208 wood discs, and 346 boards were analyzed. The image data includes RGB images of log ends and board ends, RGB images and CT slices of strips, and a set of images of the boards (RGB, laser, and X-rays) obtained with an industrial board grading machine. The measurements include wood density, growth ring widths, pith and board location in the logs, heartwood and sapwood areas, mechanical properties of each board obtained by vibratory and static testing, and visual grading of the boards. Dataset is available at https://doi.org/10.15454/YUNEGL and associated metadata are available at https://metadata-afs.nancy.inra.fr/geonetwork/srv/fre/catalog.search#/metadata/d9eef6e4-f195-41f4-b6c2-2ab46adc637e.

The use of dynamic CT imaging for tracking mandibular movements in a phantom

Purpose. The objective of this study was to analyse the possibilities of using 4D CT scanning for the tracking of patients’ mandibles. Methods. A clinical 256-slice Revolution CT was used in obtaining 4D CT scans without table movement, with a novel mandibular phantom, mounted on a programmable six degrees-of-freedom Stewart Platform in motion. The phantom was used to simulate mandibular motions which are combinations of rotations with translations (depression, elevation, protrusion, retrusion and laterotrusion). The phantom was scanned five times during identical motion patterns with a dynamic CT acquisition protocol. An image processing workflow consisting of a pairwise rigid registration and semi-automatic segmentation was developed to extract kinematic parameters (cardan angles and point-of-interest displacements) from the dynamic sequences. Reproducibility was investigated by the 95% confidence interval and the absorbed organ dose to organs of interest in the primary beam were also estimated and compared to those of a standard CT scan of the brain Results. The maximum average 95% confidence interval for the displacement across all time points for the five repetitions was 0.61 mm (Y axis). In terms of rotations, the maximum average 95% confidence interval across all time points for the five repetitions was 1.39° (X axis). The effective dose for the dynamic scan was found to be 1.3 mSv, for a CTDIvol of 63.95 mGy and a DLP of 1023.14 mGycm. The absorbed organ doses were similar to organ doses during a clinical head CT scan. Conclusions. A framework is proposed to use 4D CT scanning as a possible methodology to evaluate the motion of the temporomandibular joint. The scanning protocol allows to visualise the motion by applying a semi-automated segmentation and registration. A graphical representation of all displacements in the three spatial dimensions can depict multiple points-of-interest at once during the same acquisition. A novel type of phantom was also introduced which simulates mandibular movement with six degrees-of-freedom (three translations and three rotations).

Estimation of effective and organ dose from chest CT

The aim of this study was to estimate radiation doses from high resolution chest computed tomography and pulmonary angiography examinations. This retrospective study was conducted between March 2021 and May 2022 for adult patients underwent CT examination in Royal Care Hospital - Sudan. A total of 180 patients were examined with 64-slice CT system (Aquilion 64, Toshiba Medical System Corporation). Data was collected using a validated survey, including scanner information, patient demographic data, scan protocol, and dose indices (CT dose index volume (CTDIvol) and dose length product (DLP)). Effective and organ doses were calculated by VirtualDose software based on ICRP 103.The result of the study revealed that CTDIvol for high resolution chest computed tomography and pulmonary angiography were 15.41 mGy and 14.61 mGy respectively. DLP were 476.5 mGy/cm and 520 mGy/cm, respectively. The calculated effective doses (ED) were 8.95 mSv and 8.57 mSv, respectively. The highest organ dose was recorded in the heart (20.09 mGy) and the lowest at the thyroid (3.41 mGy).The reported measurements provide a variety of data that CT departments can use for efforts aimed at quality improvement. Dose calculation software facilitates efficient documentation and comparison of local radiation dose data with national DRLs and helps to implement optimization if DRLs are exceeded the limits.

A deep learning approach for automated diagnosis of pulmonary embolism on computed tomographic pulmonary angiography

BackgroundComputed tomographic pulmonary angiography (CTPA) is the diagnostic standard for confirming pulmonary embolism (PE). Since PE is a life-threatening condition, early diagnosis and treatment are critical to avoid PE-associated morbidity and mortality. However, PE remains subject to misdiagnosis.MethodsWe retrospectively identified 251 CTPAs performed at a tertiary care hospital between January 2018 to January 2021. The scans were classified as positive (n = 55) and negative (n = 196) for PE based on the annotations made by board-certified radiologists. A fully anonymized CT slice served as input for the detection of PE by the 2D segmentation model comprising U-Net architecture with Xception encoder. The diagnostic performance of the model was calculated at both the scan and the slice levels.ResultsThe model correctly identified 44 out of 55 scans as positive for PE and 146 out of 196 scans as negative for PE with a sensitivity of 0.80 [95% CI 0.68, 0.89], a specificity of 0.74 [95% CI 0.68, 0.80], and an accuracy of 0.76 [95% CI 0.70, 0.81]. On slice level, 4817 out of 5183 slices were marked as positive for the presence of emboli with a specificity of 0.89 [95% CI 0.88, 0.89], a sensitivity of 0.93 [95% CI 0.92, 0.94], and an accuracy of 0.89 [95% CI 0.887, 0.890]. The model also achieved an AUROC of 0.85 [0.78, 0.90] and 0.94 [0.936, 0.941] at scan level and slice level, respectively for the detection of PE.ConclusionThe development of an AI model and its use for the identification of pulmonary embolism will support healthcare workers by reducing the rate of missed findings and minimizing the time required to screen the scans.

Abstract 15175: Differentiation Between Intramural Hematoma and Thrombosed False Lumen in Aortic Dissection Using Latest Spectral Computed Tomography With Dual-Layer Detector

Introduction: Differentiating intramural hematoma (IH) from a thrombosed false lumen (FL) in aortic dissection (AD) is difficult by CT since both are regarded as AD without a tear. IH may be caused by vasa vasorum rupture, and therefore differentiation should be by histology. Spectral CT with a dual layer detector can generate CT images at any keV after ordinary acquisition using virtual monoenergetic settings (VMS). Hypothesis: Using spectral CT (range 40-200 keV), and a spectral distribution of CT values for aortic true lumen (TL) and FL, we can differentiate IH from thrombosed FL in AD in early and late-phase enhanced CT values. Methods: We performed enhanced 128-slice spectral CT 7500 (Philips) scans with a dual-layer detector in 8 patients (5 males, 71±14 years), with suspected IH or thrombosed FL in AD in the ascending aorta (Group 1), and 6 patients (4 males, 67±12 years) with thrombosed FL in AD in the descending aorta (Group 2). VMS images at various keV values were created from spectral-based images. We measured TL, FL or IH lumens in axial images at each keV in early and late phases using contrast. Results: On spectral profile analysis, in Group 1, lower keV values were associated with almost no and slightly increased mean CT values for FL or IH, in early and late phases using contrast. In Group 2, lower keV values were associated with slightly and moderately increased mean CT values for FL, in early and late phases using contrast. The degree of increments in CT values in lower keV images in late phase with contrast was significantly smaller for FL or IH in ascending aortas (n=8) than for FL in descending aortas (n=6) (P&lt;0.05). Conclusions: On CT, TL, FL, and IH lumens, in early and late phases using contrast are unique. IH may be caused by vasa vasorum rupture and thrombosed FL in AD spreading contrast in late phase. This may cause a difference in the degree of increments in CT values in lower keV images in late phase using contrast for FL or IH in the ascending aorta and for FL in the descending aorta.

Cochlear promontory anatomy relevant to development of subendosteal and transpromontory electrodes using 192-section ultra-high resolution temporal bone CT imaging.

To characterize normative adult ranges for cochlear promontory thickness relevant to the development of subendosteal and transpromontory electrodes to rehabilitate various neurotologic disorders. Adults (≥18 years). In vivo radiologic assessment using a 192-slice CT scanner (Force-192; Siemens Healthcare) with ultrahigh-resolution scan mode combined and iterative reconstruction. Cochlear promontory thickness. Among 48 included patients (96 ears), the mean (SD) age was 56 (18) years (range 25-94) and included 25 (52%) women. Of that 12 patients (25%) had osteopenia (n=6) or osteoporosis (n=6). The mean (SD) body mass index was 28 (5)kg/m2. The mean (SD) promontory thickness for the 96 temporal bones under study was 1.22 (0.24)mm (range 0.55-1.85). There was not a statistically significant association between age and promontory thickness (correlation coefficient .08; p=.44). Promontory thickness was significantly greater for men than women (mean 1.28 vs. 1.17 mm; p=.03) and increased with increasing body mass index (correlation coefficient .30; p=.004). Last, promontory thickness was significantly less for patients with osteopenia or osteoporosis compared with those without these conditions (mean 1.09 vs. 1.27 mm; p=.002). Cochlear promontory thickness can vary by almost 1.5 mm across patients and is significantly associated with patient sex, body mass index, and comorbid osteopenia/osteoporosis. Subendosteal and transpromontory electrode placement techniques must account for this degree of variability. IV.

Abstract 11285: Synthetic-Contrast-Enhanced Computed Tomography Creating by Machine Learning Without Contrast Media

Introduction: Contrast-enhanced computed tomography (CECT) is the gold standard for the diagnosis of aortic dissection (AD), at the same time, CECT has risks of nephropathy and allergic reactions caused by contrast media. Hypothesis: Synthetic-contrast-enhanced computed tomography (SECT) created from non-contrast-enhanced computed tomography (NECT) by machine learning without contrast agents could be an alternative to CECT in the diagnosis of AD. Methods: 50 patients with type A AD and 200 patients without AD who underwent NECT and CECT in pairs at Kobe university hospital were randomly selected and used for model training and testing. 3120 CT slices were used for training and 379 slices were used for testing. AD detection algorithms for NECT or CECT were developed by convolutional neural networks using NECT datasets or CECT datasets respectively. A SECT creating algorithm was developed by u-net using the pairs of CT datasets. To provide visual explanations of diagnosis, the gradient-weighted class activation mapping (Grad-CAM) method was adapted. Results: The provided sensitivities of AD detection algorithms for NECT or SECT were 90 % and 97 %, and the specificities were 100 % and 100 % respectively. When the developed AD detection algorithms for CECT were adapted to SECT, the provided sensitivity of 100 %, and sensitivity of 95 %. Most slices SECT of patients with AD demonstrated intimal flap and/or thrombosed false lumen as the almost same as CECT. The Grad-CAM visualization revealed that the AD detection algorithm using SECT images identified intimal flap or thrombosed false lumen in images more clearly than the AD detection algorithm using NECT images. Conclusions: The SECT creating algorithm improved diagnostic sensitivity for detecting AD using NECT datasets, which suggests the potential of the proposed method to support the clinical practice by reducing complications of contrast agents.

Casting-on: Technical study of a unique hollow-cast copper vessel dated to the sixth dynasty using X-ray computerized tomography

The article studied casting-on technique that was used to manufacture a hollow-cast vessel with a narrow neck. Casting-on technique, lost-wax addition, is casting a new part to attach with a pre-existing object whether cast or hammered. The case study is a small hollow-cast vessel dated back to the Sixth Dynasty [2465–2323 BCE]. It was excavated in Pepe-nefer mastaba at the Southern Upper Egypt, Edfu region. Non-invasive multi-detector computerized tomography (MDCT) and micro-analytical techniques (SEM-EDS, PXRD) were used to investigate the studied vessel. The metallographic microscope was used to examine the microstructure of both the pre-cast body and the casting-on part. SEM-EDS analysis was used to identify their elemental compositions. PXRD was used to analyze the brownish-red patina on the entire vessel’s outer surface. Also, PXRD was used to identify the chemical composition of the core material that inside the vessel. Axial, sagittal 2D CT slices and 3D CT reconstruction images revealed both the pre-cast body and the casting-on part were made by direct lost-wax casting technique. SEM-EDS analyses identified the chemical compositions of the pre-cast body and the casting-on part as copper-lead and copper-arsenic alloys respectively with a trace of iron in both. PXRD identified the cuprite with trace of copper as the chemical composition of the brownish-red patina that cover the entire vessel’s outer surface. The study proved that the casting-on technique was already practiced during the Sixth Dynasty as a metallurgical repair method for an imperfect pre-cast vessel. CT results helped to draw an ideational schematic representation for the casting process of both the two vessel’s parts. The study concluded that the studied vessel may be one of the early trails to manufacture the hollow-cast copper vessels during the Old kingdom.

Role of High-Resolution CT Thorax in Diagnosing Interstitial Lung Disease and Its Association With Smoking and Connective Tissue Disorder.

Introduction Interstitial lung diseases (ILDs)primarily affect the interstitium, an alveolar wall tissue between the capillary endothelium and the alveolar epithelium. The term 'interstitial,'however, is misleading since alveolar spaces, peripheral airways, and vessels can be involved in most of these disorders.They often require a multidisciplinary diagnosis i.e., an integration of clinical, radiological, and pathological findings. A chest radiograph is relatively insensitive because of nonspecific patterns. Generally, these disorders can progress to irreversible pulmonary fibrosis and are an important cause of morbidity and mortality. It is critical to make a prompt and accurate diagnosis of the underlying causes so that patients can be managed appropriately. ILD is subdivided into idiopathic interstitial pneumonia, of which idiopathic pulmonary fibrosis (IPF) is one subset, and diffuse parenchymal lung diseases, which may be secondary to a variety of occupational or environmental exposures or others. Theycan complicate multiple rheumatic or connective tissue diseases (CTDs). Apart from ILD, other forms of lung damage involving the pleura, vasculature, airways, and lymphatic tissuecan complicate CTDs. Aims Aims include studying the role of high-resolution computed tomography (HRCT) in diagnosing various ILDs based on morphologic patterns,evaluating the correlation between ILD and various connective tissue disordersand the prevalence of complications in such patients, andevaluating the association of smoking with various ILDs. Methods This is a retrospective study in which HRCT thorax was performed on a 128-slice PhilipsCT scanner machine on 50 patients from December 2020 to February 2022 in SVPHospital, Ahmedabad. No age orgender bias was followed. Result Out of 50 patients studied, 19 (38%) patients had the usual interstitial pneumonia (UIP) pattern and 12 (24%) had the nonspecific interstitial pneumonia (NSIP) pattern. These two were the most common among all ILD patterns. Other patterns found were hypersensitivity pneumonitis(5; 10%), respiratory bronchiolitis-related ILD (3;6%), and organizing pneumonia (2; 4%). Innine patients, the morphologic pattern was either subtle (3; 6%) or mixed (6; 12%), and the final diagnosis remained inconclusive; patients were advised clinical correlation and biopsy. Eleven (22%) patients had a history of smoking. Amongsmokers, the most common pattern was UIP while all patients with respiratory bronchiolitis (RB) ILD had a history of smoking. Fourteen (28%) patients showed a positive association with CTD. Amongthem, rheumatoid arthritis (RA) was the most commonCTD and the most common pattern among RApatients was UIP. Ten (20%) of patients developed pulmonary arterial hypertension, of which two patients who had connective tissue disorder developed pulmonary arterial hypertension at a young age(24 years). The rest of the patients who developed pulmonary arterial hypertension were above 45 years of age. Among these, two were smokers. Conclusion HRCT plays an important role in the diagnosis of ILD on the basis of various morphological patterns. CTD plays a significant role in the development of ILD. UIP is the most common ILD amongpatients with a smoking history and RA. NSIP Is the most common in patients with CTD other than RA. Pulmonary arterial hypertension (PAH) develops early in patients with CTD. There is a significant risk of the development of PAH in patients with chronicILD.

Reduced trabecular bone mineral density and thinner cortices in men with distal forearm fractures.

Men with distal forearm fractures have reduced bone density, an increased risk of osteoporosis and of further fractures. The aim of the study was to investigate the structural determinants of these observations using quantitative CT (qCT). Ninety six men with low-trauma distal forearm fracture and 101 age-matched healthy control subjects were recruited. All subjects underwent a quantitative CT on a standard 64-slice whole body CT scanner. These were analysed on Mindways QCT PRO™ Software to generate volumetric and geometric data at the lumbar spine, femoral neck and total hip, ultra-distal and distal 33% radius. Biochemical investigations, health questionnaires and measurements of bone turnover were made. Men with fracture had significantly lower total and trabecular vBMD at all sites. The greatest percentage reduction was at the ultra-distal radius (13.5% total and 11.7% trabecular vBMD). In the fracture group cortical vBMD was significantly higher in the femoral neck (p<0.001) and maintained at the ultra-distal radius compared with control subjects. However, cortical cross-sectional area (CSA) and thickness were significantly reduced at the femoral neck (p<0.001 and p=0.002 respectively) and forearm sites (CSA ultradistal radius p=0.001, cortical thickness p=0.002, CSA distal one third radius p=0.045 and cortical thickness p=0.005). Cross sectional moment of inertia (CSMI) and section moduli were significantly reduced at the femoral neck (CSMI1 p=0.002, CSMI2 p=0.012 and section moduli Z1 p<0.001, Z2 p=0.004) and the ultra-distal radius (CSMI1 p=0.008 and section moduli Z1 p=0.018, Z2 p=0.007). In stepwise logistic regression analysis distal forearm fracture showed the strongest association with a model comprising ultra-distal forearm trabecular vBMD (negative), procollagen type I N-terminal propeptide (PINP, positive) and sex hormone binding globulin (SHBG, negative). In conclusion, these observations explain the structural reasons for the increased fracture risk in men with distal forearm fractures.

The anatomical growth of the thoracic cage in adolescents with specific reference to axial growth comparing the right and left hemithorax.

This study has demonstrated the changing volume of both the anterior and posterior thorax in normal adolescents (without spinal or thoracic deformity), differentiating for both sex and age, to further understand how the thorax grows, along with the differences in growth between the anterior and posterior thorax. The thorax was measured on axial CT slices at every vertebral level from T3 to T12 in a series of scans previous taken for routine clinical care. Measurements taken were the anteroposterior thoracic distance and the area of the anterior and posterior rib prominences on either side of the thorax. Data was analyzed per vertebral level, differentiating for age and sex. There were 486 CT scans analyzed (257 males and 229 females) between the ages of 8 and 18 years. The analysis identified that for the anterior thorax, there are three phases of growth with an initial slow increase in volume, followed by a stabilization of little growth, followed by another phase of a more rapid increase in volume. For the posterior thorax, there was a gradual increase in area with increasing age. This study demonstrates that the shape of the thorax is age and sex dependent, with males having both a greater width and depth of thorax compared to females. Of particular note is the difference in patterns of growth between the anterior and posterior thorax. This information will add to the understanding of normal growth, which will aid in the management of conditions where that growth is disturbed.

Understanding ASPECTS of stroke: Inter-rater reliability between emergency medicine physician and radiologist in a rural setup.

The purpose is to determine the inter-rater reliability in grading ASPECTS score, between emergency medicine physician at first contact and radiologist among patients with acute ischemic stroke. We conducted a prospective analysis of 765 acute ischemic stroke cases referred to the Department of Radiodiagnosis in a rural-based hospital in South India, during January 2017 to October 2021. Non-contrast computed tomography (NCCT) scans of the brain were performed using GE Bright Speed Elite 128 Slice CT Scanner. ASPECTS score was calculated separately by an emergency medicine physician and radiologist. Inter-rater reliability for total and dichotomized ASPECTS (≥6 and <6) scores were assessed using statistical analysis (ICC and Cohen ĸ coefficients) on SPSS software (v17.0). Inter-rater agreement for total and dichotomized ASPECTS was substantial (ICC 0.79 and Cohen ĸ 0.68) between the emergency physician and the radiologist. Mean difference in ASPECTS between the two readers was only 0.15 with standard deviation of 1.58. No proportionality bias was detected. The Bland-Altman plot was constructed to demonstrate the distribution of ASPECTS differences between the two readers. Substantial inter-rater agreement was noted in grading ASPECTS between emergency medicine physician at first contact and radiologist, thereby confirming its robustness even in a rural setting.

Evaluating Accuracy, Completion Time and Usability of Everyday Touch Devices for Contouring

<h3>Purpose/Objective(s)</h3> Adaptive radiation therapy aims to account for patients' continuous anatomical changes and update the treatment plans on a daily basis. Despite the potential benefits, technological limitations hinder clinical adoption: daily re-evaluation of contours associated with adaptive planning is time-intensive and difficult to access, and often requires direct provider input which classically involves visualizing and editing contours on immobile desktop computers. This project evaluated the efficacy of everyday touch devices (i.e., tablet and phone) for contouring using a novel cross-platform interface. <h3>Materials/Methods</h3> Guided by user-centered and iterative design principles, we developed a cross-platform web-based contouring application which facilitates CT image navigation and target delineation. In a within-subject study design, we invited 8 radiation oncology residents to contour a liver using 4 different conditions: Desktop & Mouse, Tablet & Stylus, Tablet & Finger, and Phone & Finger. We compared the time spent contouring and assessed accuracy with Dice Similarity Coefficients (DSC). With each interface we administered a System Usability Scale (SUS) questionnaire which evaluates perceived ease of use. We incorporated repeated-measure ANOVA to assess differences between groups. <h3>Results</h3> Among the participants, using a Tablet & Finger and Phone & Finger produced the shortest time spent contouring per CT slice (<i>p</i> < .001), while Desktop & Mouse led to the longest time per slice (see Table). When comparing participant contours to a baseline gold standard, results were similar between all devices, though the Desktop & Mouse yielded a slightly lower contouring accuracy, and the most accurate was the Tablet & Stylus (<i>p</i> < .0001). Usability assessment after contouring found that participants scored Tablet & Stylus the highest, followed by Tablet & Finger, Desktop & Mouse, and Phone & Finger (<i>p</i> < .01). <h3>Conclusion</h3> Everyday touch devices (i.e., tablet and phone) not only met the standard contouring efficiency with a Desktop & Mouse, but also exceeded expectations in terms of both time spent and accuracy, especially contouring with a Tablet & Stylus which also received significantly better usability scores. Given the improved mobility with touch devices, these results suggest promising implications in addressing technological challenges associated with adaptive radiation therapy.

Comparisons of Clinical and Reference Standard Contours to AI Auto-Segmentation: An Evaluation of 5 Commercial Models in Head and Neck Organ at Risk Delineation

<h3>Purpose/Objective(s)</h3> Auto-segmentation of organs at risk (OARs) using artificial intelligence (AI) has the potential to improve the quality of radiation oncology contouring while also improving efficiency. OAR segmentation models can be used to standardize across centers and within cooperative group trials and as quality benchmarks. However, variation exists in the delineation and quality of OARs in commercially available products. We evaluated 5 commercially available products to assess whether models could outperform the clinical OAR contours relative to the reference standard. <h3>Materials/Methods</h3> We archived clinically used contours of 35 H&N patients; additionally, disease site experts contoured 42 anatomically correct structures on each scan using international cooperative group and departmental OAR consensus guidelines. The disease site expert contours were considered the reference standard, while the clinically treated contours were used as the benchmark values. We applied the 5 commercially available AI based auto-segmentation tools to the same 35 CT datasets, generating a total of 4395 structures. To compare these structures volumetric and overlap Dice Similarity Coefficient (DSC) were evaluated. Overlap DSC only compares CT slices where contours from both structure sets are present, reducing the impact on DSC values for differing contouring practices for tubular OARs. <h3>Results</h3> Models 1-5 generated a range of 16 to 27 structures. The least commonly delineated structures were the external auditory canals, mastoids, and nasal cavity. For structures contoured by all models, the most variation as measured by volumetric DSC was found in the spinal cord and oral cavity, whereas the least variation was found in the brain. Detailed evaluation of each structure across these platforms has been analyzed and is available for presentation as well as evaluation of modes of failure. Briefly, there were 33 complete contour misses out of 4395 inferences (0.75%). Selecting the 9 OARs delineated by all 5 models, the overall median volumetric DSC values were 0.84, 0.86, 0.80, 0.79, and 0.83, respectively. Median volumetric DSC for all structures in each model were 0.75, 0.74, 0.71, 0.69, and 0.61, with 2 of 5 models performing at or better than clinical contours when compared to the reference standard. <h3>Conclusion</h3> Across 5 commercially available AI segmentation models, 2 of 5 models performed at or above the clinical standard, with significant heterogeneity in the number and definition of OARs. No models performed to expert qualitative review, which was based on international cooperative group consensus guidelines. Available segmentation tools have significant room for improvement and more formal consensus standards of OARs are needed. Saliency models by voxel and second check quality systems can be developed by employing more than 1 model.