Automatic localization of anatomical landmarks in head cine fluoroscopy images via deep learning.

BACKGROUND: Fluoroscopy guided interventions (FGIs) pose a risk of prolonged radiation exposure; personalized patient dosimetry is necessary to improve patient safety during these procedures. However, current FGIs systems do not capture the precise exposure regions of the patient, making it challenging to perform patient-procedure-specific dosimetry. Thus, there is a pressing need to develop approaches to extract and use this information to enable personalized radiation dosimetry for interventional procedures. PURPOSE: To propose a deep learning (DL) approach for the automatic localization of 3D anatomical landmarks on randomly collimated and magnified 2D head fluoroscopy images. MATERIALS AND METHODS: The model was developed with datasets comprising 800 000 pseudo 2D synthetic images (mixture of vessel-enhanced and non-enhancement), each with 55 annotated anatomical landmarks (two are landmarks for eye lenses), generated from 135 retrospectively collected head computed tomography (CT) volumetric data. Before training, dynamic random cropping was performed to mimic the varied field-size collimation in FGI procedures. Gaussian-distributed additive noise was applied to each individual image to enhance the robustness of the DL model in handling image degradation that may occur during clinical image acquisition in a clinical environment. The model was trained with 629 370 synthetic images for approximately 275 000 iterations and evaluated against a synthetic image test set and a clinical fluoroscopy test set. RESULTS: The model shows good performance in estimating in- and out-of-image landmark positions and shows feasibility to instantiate the skull shape. The model successfully detected 96.4% and 92.5% 2D and 3D landmarks, respectively, within a 10 mm error on synthetic test images. It demonstrated an average of 3.6 ± 2.3 mm mean radial error and successfully detected 96.8% 2D landmarks within 10 mm error on clinical fluoroscopy images. CONCLUSION: Our deep-learning model successfully localizes anatomical landmarks and estimates the gross shape of skull structures from collimated 2D projection views. This method may help identify the exposure region required for patient-specific organ dosimetry in FGIs procedures.

Marrow Fat-Cortical Bone Relationship in ß-Thalassemia: A Study Using MRI.

BACKGROUND: Growing evidence suggests that marrow adipocytes play an active role in the regulation of bone metabolism and hematopoiesis. However, research on the relationship between bone and fat in the context of hematological diseases, particularly ß-thalassemia, remains limited. PURPOSE: To investigate the relationship between marrow fat and cortical bone thickness in ß-thalassemia and to identify key determinants influencing these variables. STUDY TYPE: Prospective. SUBJECTS: Thirty-five subjects in four subject groups of increasing disease severity: 6 healthy control (25.0 ± 5.3 years, 2 male), 4 ß-thalassemia minor, 13 intermedia, and 12 major (29.1 ± 6.4 years, 15 male). FIELD STRENGTH/SEQUENCE: 3.0 T, 3D fast low angle shot sequence and T1-weighted turbo spin echo. ASSESSMENT: Analyses on proton density fat fraction (PDFF) and R2* values in femur subregions (femoral head, greater trochanter, intertrochanteric, diaphysis, distal) and cortical thickness (CBI) of the subjects' left femur. Clinical data such as age, sex, body mass index (BMI), and disease severity were also included. STATISTICAL TESTS: One-way analysis of variance (ANOVA), mixed ANOVA, Pearson correlation and multiple regression. P-values <0.05 were considered significant. RESULTS: Bone marrow PDFF significantly varied between the femur subregions, F(2.89,89.63) = 44.185 and disease severity, F(1,3) = 12.357. A significant interaction between subject groups and femur subregions on bone marrow PDFF was observed, F(8.67,89.63) = 3.723. Notably, a moderate positive correlation was observed between PDFF and CBI (r = 0.33-0.45). Multiple regression models for both PDFF (R2 = 0.476, F(13,151) = 10.547) and CBI (R2 = 0.477, F(13,151) = 10.580) were significant. Significant predictors for PDFF were disease severity (ßTMi = 0.36, ßTMa = 0.17), CBI (ß = 0.24), R2* (ß = -0.32), and height (ß = -0.29) while for CBI, the significant determinants were sex (ß = -0.27), BMI (ß = 0.55), disease severity (ßTMi = 2.15), and PDFF (ß = 0.25). DATA CONCLUSION: This study revealed a positive correlation between bone marrow fat fraction and cortical bone thickness in ß-thalassemia with varying disease severity, potentially indicating a complex interplay between bone health and marrow composition. EVIDENCE LEVEL: 2 TECHNICAL EFFICACY: Stage 3.

Opportunistic Extraction of Quantitative CT Biomarkers: Turning the Incidental Into Prognostic Information.

The past two decades have seen a significant increase in the use of CT, with a corresponding rise in the mean population radiation dose. This rise in CT use has caused improved diagnostic certainty in conditions that were not previously routinely evaluated using CT, such as headaches, back pain, and chest pain. Unused data, unrelated to the primary diagnosis, embedded within these scans have the potential to provide organ-specific measurements that can be used to prognosticate or risk-profile patients for a wide variety of conditions. The recent increased availability of computing power, expertise and software for automated segmentation and measurements, assisted by artificial intelligence, provides a conducive environment for the deployment of these analyses into routine use. Data gathering from CT has the potential to add value to examinations and help offset the public perception of harm from radiation exposure. We review the potential for the collection of these data and propose the incorporation of this strategy into routine clinical practice.

Generation of fluoroscopy-alike radiographs as alternative datasets for deep learning in interventional radiology.

In fluoroscopy-guided interventions (FGIs), obtaining large quantities of labelled data for deep learning (DL) can be difficult. Synthetic labelled data can serve as an alternative, generated via pseudo 2D projections of CT volumetric data. However, contrasted vessels have low visibility in simple 2D projections of contrasted CT data. To overcome this, we propose an alternative method to generate fluoroscopy-like radiographs from contrasted head CT Angiography (CTA) volumetric data. The technique involves segmentation of brain tissue, bone, and contrasted vessels from CTA volumetric data, followed by an algorithm to adjust HU values, and finally, a standard ray-based projection is applied to generate the 2D image. The resulting synthetic images were compared to clinical fluoroscopy images for perceptual similarity and subject contrast measurements. Good perceptual similarity was demonstrated on vessel-enhanced synthetic images as compared to the clinical fluoroscopic images. Statistical tests of equivalence show that enhanced synthetic and clinical images have statistically equivalent mean subject contrast within 25% bounds. Furthermore, validation experiments confirmed that the proposed method for generating synthetic images improved the performance of DL models in certain regression tasks, such as localizing anatomical landmarks in clinical fluoroscopy images. Through enhanced pseudo 2D projection of CTA volume data, synthetic images with similar features to real clinical fluoroscopic images can be generated. The use of synthetic images as an alternative source for DL datasets represents a potential solution to the application of DL in FGIs procedures.

Evaluation of Therapeutic Efficacy and Imaging Capabilities of 153Sm2O3-Loaded Polystyrene Microspheres for Intra-Tumoural Radionuclide Therapy of Liver Cancer Using Sprague-Dawley Rat Model.

Introduction: Neutron-activated samarium-153-oxide-loaded polystyrene ([153Sm]Sm2O3-PS) microspheres has been developed in previous study as a potential theranostic agent for hepatic radioembolization. In this study, the therapeutic efficacy and diagnostic imaging capabilities of the formulation was assessed using liver cancer Sprague-Dawley (SD) rat model. Methods: Twelve male SD rats (150-200 g) that implanted with N1-S1 hepatoma cell line orthotopically were divided into two groups (study versus control) to monitor the tumour growth along 60 days of treatment. The study group received an intra-tumoural injection of approximately 37 MBq of [153Sm]Sm2O3-PS microspheres, while control group received an intra-tumoural injection of 0.1 mL of saline solution. A clinical single photon emission computed tomography/computed tomography (SPECT/CT) system was used to scan the rats at Day 5 post-injection to investigate the diagnostic imaging capabilities of the microspheres. All rats were monitored for change in tumour volume using a portable ultrasound system throughout the study period. Histopathological examination (HPE) was performed after the rats were euthanized at Day 60. Results: At Day 60, no tumour was observed on the ultrasound images of all rats in the study group. In contrast, the tumour volumes in the control group were 24-fold larger compared to baseline. Statistically significant difference was observed in tumour volumes between the study and control groups (p < 0.05). The SPECT/CT images clearly displayed the location of [153Sm]Sm2O3-PS in the liver tumour of all rats at Day 5 post-injection. Additionally, the [153Sm]Sm2O3-PS microspheres was visible on the CT images and this has added to the benefits of 153Sm as a CT contrast agent. The HPE results showed that the [153Sm]Sm2O3-PS microspheres remained concentrated at the injection site with no tumour cells observed in the study group. Conclusions: Neutron-activated [153Sm]Sm2O3-PS microspheres demonstrated excellent therapeutic and diagnostic imaging capabilities for theranostic treatment of liver cancer in a SD rat model. Further studies with different animal and tumour models are planned to validate this finding.

Operational nuclear research reactors in the Asia-Pacific with potential for medical radionuclide production.

Personalised cancer treatment is of growing importance and can be achieved via targeted radionuclide therapy. Radionuclides with theranostic properties are proving to be clinically effective and are widely used because diagnostic imaging and therapy can be accomplished using a single formulation that avoids additional procedures and unnecessary radiation burden to the patient. For diagnostic imaging, single photon emission computed tomography (SPECT) or positron emission tomography (PET) is used to obtain functional information noninvasively by detecting the gamma (γ) rays emitted from the radionuclide. For therapeutics, high linear energy transfer (LET) radiations such as alpha (α), beta (ß - ) or Auger electrons are used to kill cancerous cells in close proximity, whereas sparing the normal tissues surrounding the malignant tumour cells. One of the most important factors that lead to the sustainable development of nuclear medicine is the availability of functional radiopharmaceuticals. Nuclear research reactors play a vital role in the production of medical radionuclides for incorporation into clinical radiopharmaceuticals. The disruption of medical radionuclide supplies in recent years has highlighted the importance of ongoing research reactor operation. This article reviews the current status of operational nuclear research reactors in the Asia-Pacific region that have the potential for medical radionuclide production. It also discusses the different types of nuclear research reactors, their operating power, and the effects of thermal neutron flux in producing desirable radionuclides with high specific activity for clinical applications.

Ultrasound shear wave elastography for the evaluation of renal pathological changes in adult patients.

OBJECTIVE: Many studies have conflicting findings in using shear wave elastography (SWE) to assess renal fibrosis. This study reviews the use of SWE to evaluate pathological changes in native kidneys and renal allografts. It also tries to elucidate the confounding factors and care taken to ensure the results are consistent and reliable. METHODS: The review was carried out according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. Literature search was conducted in Pubmed, Web of Science and Scopus database up to 23 October 2021. To evaluate risk and bias applicability, the Cochrane risk-of bias tool and GRADE was used. The review was registered under PROSPERO CRD42021265303. RESULTS: A total of 2921 articles were identified. 104 full texts were examined and 26 studies included in systematic review. 11 studies performed on native kidneys and 15 studies on transplanted kidney. A wide range of impact factors was found that affect the accuracy of SWE of renal fibrosis in adult patients. CONCLUSIONS: Compared to point SWE, two-dimensional SWE with elastogram could enable better selection of the region of interest in kidneys, leading to reproducible results. Tracking waves were attenuated as the depth from skin to region of interest increased, therefore, SWE is not recommended for overweight or obese patients. Variable transducer forces might also affect SWE reproducibility, thus, training of operators to ensure consistent operator-dependent transducer forces may be helpful. ADVANCES IN KNOWLEDGE: This review provides a holistic insight on the efficiency of using SWE in evaluating pathological changes in native and transplanted kidneys, thereby contributing to the knowledge of its utilisation in clinical practice.

Impact of preoperative magnetic resonance imaging on surgery and eligibility for intraoperative radiotherapy in early breast cancer.

We looked at the usefulness of magnetic resonance imaging (MRI) in decision-making and surgical management of patients selected for intraoperative radiotherapy (IORT). We also compared lesion size measurements in different modalities (ultrasound (US), mammogram (MMG), MRI) against pathological size as the gold standard. 63 patients eligible for IORT based on clinical and imaging criteria over a 34-month period were enrolled. All had MMG and US, while 42 had additional preoperative MRI for locoregional preoperative staging. Imaging findings and pathological size concordances were analysed across the three modalities. MRI changed the surgical management of 5 patients (11.9%) whereby breast-conserving surgery (BCS) and IORT was cancelled due to detection of satellite lesion, tumor size exceeding 30mm and detection of axillary nodal metastases. Ten of 42 patients (23.8%) who underwent preoperative MRI were subjected to additional external beam radiotherapy (EBRT); 7 due to lymphovascular invasion (LVI), 2 due to involved margins, and 1 due to axillary lymph node metastatic carcinoma detected in the surgical specimen. Five of 21 (23.8%) patients without prior MRI were subjected to additional EBRT post-surgery; 3 had LVI and 2 had involved margins. The rest underwent BCS and IORT as planned. MRI and MMG show better imaging-pathological size correlation. Significant increase in the mean 'waiting time' were seen in the MRI group (34.1 days) compared to the conventional imaging group (24.4 days). MRI is a useful adjunct to conventional imaging and impacts decision making in IORT. It is also the best imaging modality to determine the actual tumour size.

Development of neutron-activated samarium-153-loaded polystyrene microspheres as a potential theranostic agent for hepatic radioembolization.

PURPOSE: Hepatic radioembolization is an effective minimally invasive treatment for primary and metastatic liver cancers. Yttrium-90 [90Y]-labelled resin or glass beads are typically used as the radioembolic agent for this treatment; however, these are not readily available in many countries. In this study, novel samarium-153 oxide-loaded polystyrene ([153Sm]Sm2O3-PS) microspheres were developed as a potential alternative to 90Y microspheres for hepatic radioembolization. METHODS: The [152Sm]Sm2O3-PS microspheres were synthesized using solid-in-oil-in-water solvent evaporation. The microspheres underwent neutron activation using a 1 MW open-pool research reactor to produce radioactive [153Sm]Sm2O3-PS microspheres via 152Sm(n,γ)153Sm reaction. Physicochemical characterization, gamma spectroscopy and in-vitro radionuclide retention efficiency were carried out to evaluate the properties and stability of the microspheres before and after neutron activation. RESULTS: The [153Sm]Sm2O3-PS microspheres achieved specific activity of 5.04 ± 0.52 GBq·g-1 after a 6 h neutron activation. Scanning electron microscopy and particle size analysis showed that the microspheres remained spherical with an average diameter of ~33 µm before and after neutron activation. No long half-life radionuclide and elemental impurities were found in the samples. The radionuclide retention efficiencies of the [153Sm]Sm2O3-PS microspheres at 550 h were 99.64 ± 0.07 and 98.76 ± 1.10% when tested in saline solution and human blood plasma, respectively. CONCLUSIONS: A neutron-activated [153Sm]Sm2O3-PS microsphere formulation was successfully developed for potential application as a theranostic agent for liver radioembolization. The microspheres achieved suitable physical properties for radioembolization and demonstrated high radionuclide retention efficiency in saline solution and human blood plasma.