3D Slicer as an image computing platform for the Quantitative Imaging Network

Sox2-overexpressing neural stem cells alleviate ventricular enlargement and neurological dysfunction in posthemorrhagic hydrocephalus

JOURNAL/nrgr/04.03/01300535-202602000-00045/figure1/v/2025-05-05T160104Z/r/image-tiff Neural stem cells (NSCs) have the potential for self-renewal and multidirectional differentiation, and their transplantation has achieved good efficacy in a variety of diseases. However, only 1%-10% of transplanted NSCs survive in the ischemic and hypoxic microenvironment of posthemorrhagic hydrocephalus. Sox2 is an important factor for NSCs to maintain proliferation. Therefore, Sox2-overexpressing NSCs (NSCSox2) may be more successful in improving neurological dysfunction after posthemorrhagic hydrocephalus. In this study, human NSCSox2 was transplanted into a posthemorrhagic hydrocephalus mouse model, and retinoic acid was administered to further promote NSC differentiation. The results showed that NSCSox2 attenuated the ventricular enlargement caused by posthemorrhagic hydrocephalus and improved neurological function. NSCSox2 also promoted nerve regeneration, inhibited neuroinflammation and promoted M2 polarization (anti-inflammatory phenotype), thereby reducing cerebrospinal fluid secretion in choroid plexus. These findings suggest that NSCSox2 rescued ventricular enlargement and neurological dysfunction induced by posthemorrhagic hydrocephalus through neural regeneration and modulation of inflammation.

Association of pectoralis muscle area on computed tomography with airflow limitation severity and respiratory outcomes in COPD: A population-based prospective cohort study

BACKGROUND

Previous studies have shown that patients with chronic obstructive pulmonary disease (COPD) of severe or very severe airflow limitation have a reduced pectoralis muscle area (PMA), which is associated with mortality. However, whether patients with COPD of mild or moderate airflow limitation also have a reduced PMA remains unclear. Additionally, limited evidence is available regarding the associations between PMA and respiratory symptoms, lung function, computed tomography (CT) imaging, lung function decline, and exacerbations. Therefore, we conducted this study to evaluate the presence of PMA reduction in COPD and to clarify its associations with the referred variables.

METHODS

This study was based on the subjects enrolled from July 2019 to December 2020 in the Early Chronic Obstructive Pulmonary Disease (ECOPD) study. Data including questionnaire, lung function, and CT imaging were collected. The PMA was quantified on full-inspiratory CT at the aortic arch level using predefined -50 and 90 Hounsfield unit attenuation ranges. Multivariate linear regression analyses were performed to assess the association between the PMA and airflow limitation severity, respiratory symptoms, lung function, emphysema, air trapping, and the annual decline in lung function. Cox proportional hazards analysis and Poisson regression analysis were used to evaluate the PMA and exacerbations after adjustment.

RESULTS

We included 1352 subjects at baseline (667 with normal spirometry, 685 with spirometry-defined COPD). The PMA was monotonically lower with progressive airflow limitation severity of COPD after adjusting for confounders (vs. normal spirometry; Global Initiative for Chronic Obstructive Lung Disease [GOLD] 1: β=-1.27, P=0.028; GOLD 2: β=-2.29, P<0.001; GOLD 3: β=-4.88, P<0.001; GOLD 4: β=-6.47, P=0.014). The PMA was negatively associated with the modified British Medical Research Council dyspnea scale (β=-0.005, P=0.026), COPD Assessment Test score (β=-0.06, P=0.001), emphysema (β=-0.07, P<0.001), and air trapping (β=-0.24, P<0.001) after adjustment. The PMA was positively associated with lung function (all P<0.05). Similar associations were discovered for the pectoralis major muscle area and pectoralis minor muscle area. After the 1-year follow-up, the PMA was associated with the annual decline in the post-bronchodilator forced expiratory volume in 1 s percent of predicted value (β=0.022, P=0.002) but not with the annual rate of exacerbations or the time to first exacerbation.

CONCLUSION

Patients with mild or moderate airflow limitation exhibit a reduced PMA. The PMA is associated with airflow limitation severity, respiratory symptoms, lung function, emphysema, and air trapping, suggesting that PMA measurement can assist with COPD assessment.

Assessment of renal insufficiency in patients with renal artery stenosis by multiparametric magnetic resonance imaging

OBJECTIVES

To evaluate the function of kidneys with renal artery stenosis using multiparametric magnetic resonance imaging, assess the diagnostic efficacy of multiparametric magnetic resonance imaging for single kidney dysfunction.

MATERIALS AND METHODS

Renal multiparametric magnetic resonance imaging was performed on 62 patients with RAS using the Philips Ingenia CX 3.0 T MRI machine. The scanning sequences included arterial spin labeling, phase contrast MRI, diffusion weighted imaging, T1 mapping, and blood oxygen level-dependent MRI. All patients underwent radionuclide renal dynamic imaging, and the glomerular filtration rate (GFR) was calculated to determine renal function. Individual kidneys from renal artery stenosis patients were classified into normal (GFR ≥ 30) and reduced (GFR < 30) groups and the ability of the uni- and multi-variate logistic regression model to predict the group was determined.

RESULTS

MR parameters demonstrated considerable diagnostic efficacy for single kidney dysfunction, with AUC range of 0.597- 0.864. The strongest predictor was mean renal artery blood flow. The sensitivity and specificity were 0.93 and 0.69AUC was 0.864. The strongest predictors of the renal microstructure were cortical apparent diffusion coeffecient and T1 value, with ROC AUCs of 0.756 and 0.741, sensitivities of 0.875 and 0.689, and specificities of 0.537 and 0.731. Multiparametric MRI combined with the values of cortical renal blood flow and cortical T1 exhibited the highest diagnostic efficacy, with an AUC of 0.92, and sensitivity of 0.919, and specificity of 0.743.

CONCLUSION

Multiparametric magnetic resonance imaging can effectively detect the single renal dysfunction of kidneys with renal artery stenosis, which holds promise for the diagnosis and prognosis of patients with renal artery stenosis.

Automatic deep learning segmentation of mandibular periodontal bone topography on cone-beam computed tomography images

OBJECTIVES

This study evaluated the performance of a multi-stage Segmentation Residual Network (SegResNet)-based deep learning (DL) model for the automatic segmentation of cone-beam computed tomography (CBCT) images of patients with stage III and IV periodontitis.

METHODS

Seventy pre-processed CBCT scans from patients undergoing periodontal rehabilitation were used for training and validation. The model was tested on 10 CBCT scans independent from the training dataset by comparing results with semi-automatic (SA) segmentations. Segmentation accuracy was assessed using the Dice similarity coefficient (DSC), Intersection over Union (IoU), and Hausdorff distance 95th percentile (HD95). Linear periodontal measurements were performed on four tooth surfaces to assess the validity of the DL segmentation in the periodontal region.

RESULTS

The DL model achieved a mean DSC of 0.9650 ± 0.0097, with an IoU of 0.9340 ± 0.0180 and HD95 of 0.4820 mm ± 0.1269 mm, showing strong agreement with SA segmentation. Linear measurements revealed high statistical correlations between the mesial, distal, and lingual surfaces, with intraclass correlation coefficients (ICC) of 0.9442 (p < 0.0001), 0.9232 (p < 0.0001), and 0.9598(p < 0.0001), respectively, while buccal measurements revealed lower consistency, with an ICC of 0.7481 (p < 0.0001). The DL method reduced the segmentation time by 47 times compared to the SA method.

CONCLUSIONS

Acquired 3D models may enable precise treatment planning in cases where conventional diagnostic modalities are insufficient. However, the robustness of the model must be increased to improve its general reliability and consistency at the buccal aspect of the periodontal region.

CLINICAL SIGNIFICANCE

This study presents a DL model for the CBCT-based segmentation of periodontal defects, demonstrating high accuracy and a 47-fold time reduction compared to SA methods, thus improving the feasibility of 3D diagnostics for advanced periodontitis.

Evaluation of paranasal sinus volume and pathway relationships in domestic pigs: A sex-based analysis using computerized tomography and three-dimensional modeling

This study comprehensively analyzed the paranasal sinuses of male and female pigs, focusing on volumetric measurements, surface areas, and interconnectivity. Fourteen pig heads (Sus scrofa domesticus, 8 females and 6 males) were examined using CT imaging. The frontal (FS), maxillary (MS), ethmoidal (ES), lacrimal (LS), and sphenoid (SS) sinuses were identified. FS was divided into three compartments: medial rostral frontal sinus (MRFS), lateral rostral frontal sinus (LRFS), and caudal frontal sinus (CFS). Significant sex-related differences were observed in total sinus volume and surface area. In females, LS and ES exhibited asymmetry in volume and surface area between both sides. In males, LS showed no volumetric asymmetry but had significant surface area differences. LRFS and MS showed significant sex-based differences in volume and surface area, while SS exhibited surface area asymmetry in females but no volumetric differences. This study's main limitation was the absence of magnetic resonance imaging. These findings provide a detailed anatomical and morphological understanding of pig paranasal sinuses, offering valuable insights for biomedical research, clinical applications, and laboratory studies. The study highlights the utility of CT imaging in revealing subtle anatomical variations and its potential for advancing research and surgical planning.

Printing fidelity assessment and micro-mechanical characterization of FDM-printed PLA/HA composite for maxillofacial and oral applications

The aim of this work is to assess the printing fidelity, explore the mechanical properties and optimize process parameters of FDM-printed PLA/HA composite samples for maxillofacial and oral applications. Pure PLA and PLA/HA composites with 20 % and 35 % HA content by weight were evaluated. The effects of printing temperature and HA loading on printing fidelity and mechanical properties, as well as the FDM printability of screws and patient-specific membranes, were assessed. Square samples with holes and beam samples, designed with geometrical features comparable to fixation plates and guided bone regeneration (GBR) membranes, were FDM-printed and analyzed for dimensional accuracy and mechanical performance. The results show that holes are geometrical features difficult to print with high accuracy and a printing temperature of 200 °C provides better accuracy and mechanical properties compared to 210 °C. Higher HA loading reduces printability fidelity and increases flexural elastic modulus while decreasing maximum flexural strength and strain. Prototypes of patient-specific GBR membranes and fixation screws were successfully printed using PLA/HA 20, demonstrating the feasibility of producing custom medical devices with FDM technology. Roughness analysis on GBR membranes in PLA/HA 20 revealed no significant differences between the external and internal surfaces or between different printing configurations. Moreover, the FDM printing process does not affect the homogeneous distribution of HA particles within the PLA matrix in PLA/HA 20 composite. The results suggest that a printing parameters optimization procedure is fundamental for achieving the best performance of PLA/HA composites in terms of printing fidelity and mechanical properties. PLA/HA 20 shows promise as a biodegradable alternative to non-biodegradable materials such as titanium, which is commonly used for maxillofacial and oral applications.

Comparative voxel-based dosimetry with pre-treatment Tc-99m SPECT/CT and post-treatment Y-90 PET/MRI for radioembolization with Y-90 microspheres

PURPOSE

The aim of this study was to investigate the differences between voxel-based dosimetry and the mean absorbed doses calculated with pre-treatment Tc-99 m-MAA SPECT/CT and post-treatment Y-90 PET/MR images. We also sought to present a detailed comparison of dose-volume histograms (DVHs) calculated from pre- and post-treatment imaging.

METHODS

A total of 47 treatments and 41 patients were included in the analysis as six of the treatments were retreatments of the six patients. Multicompartment and voxel-based dosimetry were performed with pre-treatment Tc-99m-MAA SPECT/CT and Y-90 PET/MRI. Correlation coefficients between the two imaging methods for the mean absorbed dose of tumor, whole liver normal tissue, perfused normal tissue, T/N ratio and tumor D10, D50 and D90 values were calculated. Additionally, differences between these values were also evaluated with Bland-Altman plots.

RESULTS

Pre-treatment Tc-99m-MAA SPECT/CT accurately predicted the dose values for healthy liver parenchyma calculated with Y-90 PET/MRI but showed lower accuracy in predicting T/N ratio and tumor doses. There were significant variations in tumor-absorbed doses for both glass and resin microspheres. Additionally, D90 values were higher when calculated with SPECT/CT than with PET/MRI, whereas D10 values were higher in PET/MRI compared to SPECT/CT.

CONCLUSION

The findings in our study suggest that Tc-99m-MAA SPECT/CT had higher accuracy in predicting the dose to the healthy liver parenchyma compared to the tumor, maintaining its importance in treatment planning.

Survival prediction with radiomics for patients with IDH mutated lower-grade glioma

PURPOSE

Adult patients with diffuse lower-grade gliomas (dLGG) show heterogeneous survival outcomes, complicating postoperative treatment planning. Treating all patients early increases the risk of long-term side effects, while delayed treatment may lead to impaired survival. Refinement of prognostic models could optimize timing of treatment. Conventional radiological features are prognostic in dLGG, but MRI could carry more prognostic information. This study aimed to investigate MRI-based radiomics survival models and compare them with clinical models.

METHODS

Two clinical survival models were created: a preoperative model (tumor volume) and a full clinical model (tumor volume, extent of resection, tumor subtype). Radiomics features were extracted from preoperative MRI. The dataset was divided into training set and unseen test set (70:30). Model performance was evaluated on test set with Uno's concordance index (c-index). Risk groups were created by the best performing model's predictions.

RESULTS

207 patients with mutated IDH (mIDH) dLGG were included. The preoperative clinical, full clinical and radiomics models showed c-indexes of 0.70, 0.71 and 0.75 respectively on test set for overall survival. The radiomics model included four features of tumor diameter and tumor heterogeneity. The combined full clinical and radiomics model showed best performance with c-index = 0.79. The survival difference between high- and low-risk patients according to the combined model was both statistically significant and clinically relevant.

CONCLUSION

Radiomics can capture quantitative prognostic information in patients with dLGG. Combined models show promise of synergetic effects and should be studied further in astrocytoma and oligodendroglioma patients separately for optimal modelling of individual risks.

Evolving hemodynamics in Fontan circulation: age-driven insights in patients with interrupted inferior vena cava and hepatoazygos shunt

Patients with interrupted inferior vena cava (I-IVC) and azygos continuation who undergo Fontan completion via hepatoazygos shunting exhibit unique hemodynamic challenges. This study evaluates age-related shifts in systemic venous return dominance, hepatic flow distribution (HFD), power loss (PL), and flow disturbances using patient-specific computational fluid dynamics (CFD). Data analysis from 95 patients with I-IVC showed a nonlinear shift in upper-to-lower body systemic flow dominance with ratios of 2, 1, and 0.5 (correlating to ages ∼3, ∼10, and ∼20, respectively). CFD simulations for 17 selected patients revealed a trend of increasing HFD toward the right pulmonary artery, with median splits of 45%-49%, 48%-52%, and 40%-60% for the respective flow ratios. Power loss increased significantly with lower-body flow dominance. Median values for absolute PL were 4.75 mW (ratio 2), 16.5 mW (ratio 1), and 33.7 mW (ratio 0.5). Indexed PL showed a similar trend, rising from 0.04 mW/m2 to 0.11 mW/m2 across the flow ratios. Vorticity and viscous dissipation rates, key metrics of flow disturbances, also increased with lower-body flow dominance, showing strong correlations with PL (R = 0.58-0.76). Kruskal-Wallis-based statistical analysis identified significant statistical differences in absolute PL (P = 0.0045) and flow disturbances (P < 0.001), emphasizing the impact of age-related flow dynamics on Fontan efficiency. Our findings emphasize the need for targeted interventions in patients with I-IVC with azygos continuation to mitigate evolving hemodynamic inefficiencies and optimize Fontan outcomes during critical growth periods.NEW & NOTEWORTHY Evaluate how age-driven changes in patients with interrupted inferior vena cava impact Fontan efficiency. Using patient-specific computational fluid dynamics, our study reveals nonlinear flow dynamics, increasing power loss, and evolving hepatic flow distribution, emphasizing the need for tailored interventions to optimize outcomes.

Spatio-temporal dynamics of sound-induced vestibular processing: insights from stereo-EEG recordings

Numerous functions rely on the activation of the vestibular system, resulting in widespread activation of cortical brain regions. However, although the topographical organization of vestibular processing is relatively well understood, the temporal dynamics of this information processing remain insufficiently explored. In this study, we conducted an in-depth analysis of intracerebral recordings from 107 patients (123 implanted hemispheres) to investigate the cortical response to acoustic and sound-induced vestibular stimuli (SVS), thus unveiling the spatiotemporal dynamics of vestibular processing. Our findings revealed the existence of distinct early components (phasic peak, 20-40 ms) localized in Heschl's area, planum temporale, retroinsula, posterior insular cortex, PFcm, parietal operculum, and structures above the Sylvian fissure. Moreover, we identified later, tonic components (peaking at 50-80 ms) characterized by an extended duration, returning to baseline between 200 and 300 ms. Remarkably, these latter components exclusively involved the perisylvian cortices. The findings demonstrated that the early stages of human otolithic vestibular information processing involve both parallel and hierarchical pathways distributed across the perisylvian and peri‑Rolandic regions, rather than being restricted to a single primary cortical area. Furthermore, two distinct streams reminiscent of the dorsal/ventral dichotomy with specific spatio-temporal characteristics were identified. Collectively, our study uncovers a complex and interconnected cortical network that underlies vestibular processing, shedding light on the temporal dynamics of this essential sensory system. These findings pave the way for a deeper understanding of the functional organization of the vestibular system and its implications for sensory perception and motor control.

Can MRI radiomics predict neck metastasis at initial diagnosis in patients with squamous cell carcinoma of the tongue?

OBJECTIVES

This study aimed to predict the risk of neck metastasis in patients with tongue squamous cell carcinoma by performing radiomics analysis of preoperative magnetic resonance (MR) images.

METHODS

In total, 143 patients with primary tongue cancer were enrolled and divided into training and validation groups. The presence of neck metastases was assessed after at least 6-month of follow-up. Using fat-suppressed T2-weighted images, two observers manually set the volume of interest at the tumor site using the 3D Slicer software and extracted 107 image features. The analysis was based on significant differences between patient groups in the presence or absence of neck metastases by the Mann-Whitney U test and good agreement with intra- and inter-observer intra-class correlation coefficients exceeding 0.9. In addition, two characteristics that were determined to be very useful for diagnosing neck metastases were selected by receiver operating characteristic analysis and evaluated for goodness of fit in the validation data.

RESULTS

Neck metastases were identified in 57 of 143 patients and divided into 121 training and 22 validation datasets. Using the combined criteria of Major Axis Length of 3D-shape features and Joint Entropy of the gray-level co-occurrence matrix, neck metastases were identified in 80.7% of the cases; the validation data predicted neck metastases in 80% of the cases.

CONCLUSIONS

MR-imaging texture analysis of the primary tumor helps predict neck metastasis in patients with tongue cancer. The proposed criteria are simple yet useful for identifying groups of patients who may require neck dissection.

Alterations in Sphenoid Anatomy in Craniosynostosis: Implications for Fronto-Orbital Advancement

ObjectiveFronto-orbital advancement involves removal of the fronto-orbital bandeau. Visualization of the saw blade is lost as it passes through the fronto-orbital-sphenoid junction (FOSJ), placing the temporal lobe at risk of injury. We aim to provide a 3D analysis of the space surrounding this osteotomy to differentiate various types of craniosynostoses.DesignRetrospective cohortSettingInstitutional.PatientsThirty patients with isolated unicoronal synostosis, nonsyndromic bicoronal synostosis, metopic synostosis, Apert syndrome, Crouzon syndrome, and Muenke syndrome.InterventionsCT scans conducted between 2 months to 2 years of age were 3D reconstructed to compare craniometrics against normal controls.Main Outcome Measure(s)Craniometrics.ResultsThe mean bone thickness of the FOSJ at the level of the supraorbital rim was significantly small for the Apert, unicoronal and bicoronal groups. The mean vertical height of the middle cranial fossa from the lesser sphenoid wing was significantly greater in the unicoronal group. The mean vertical height of the tip of the temporal lobe from the lateral sphenoid ridge was greater in the unicoronal, isolated bicoronal, and Apert groups. The mean corneal protrusion beyond the lateral orbital rim was significantly greater in the Apert and unicoronal groups. The mean horizontal depth of the orbit was smallest in the Apert group. The mean vertical distance between the dacryon and the foramen cecum, and the mean volume of temporal lobe beneath the sphenoid shelf were the largest in the Apert group.ConclusionsPatients with Apert syndrome have the most unfavorable morphology of the anterior and middle cranial fossae.

Associations of tongue and hyoid position, tongue volume, and pharyngeal airway dimensions with various dentoskeletal growth patterns

OBJECTIVE

This study investigated the association between tongue and hyoid position, tongue volume, and pharyngeal airway dimensions with craniofacial growth patterns in the sagittal, vertical, and transverse planes.

METHODS

Cone beam computed tomography was used to assess 185 non-growing subjects (mean age, 28.7 ± 9.5 years). Multivariate linear regression analyses evaluated relationships between tongue and airway variables, and cephalometric/dental arch measurements.

RESULTS

Class III skeletal patterns-reflected by lower ANB and higher APDI-were significantly correlated with anteriorly positioned hyoids (ANB: β = 0.249; APDI: β = -0.291), and lower tongue positions at the tongue tip (ANB: β = -0.231; APDI: β = 0.166) and in the posterior area (ANB: β = -0.186; APDI: β = 0.196), and greater tongue volume (APDI: β = 0.174). Hyperdivergent vertical patterns-indicated by a lower ODI-were significantly correlated with a lower tongue tip position (β = -0.311) and posterior tongue position (β = -0.230). Regarding transverse dimensions, tongue volume showed positive correlations with upper intermolar width (β = 0.349), lower intercanine width (β = 0.130), lower intermolar width (β = 0.311), and a negative correlation with upper intercanine width (β = -0.299).

CONCLUSIONS

Sagittal and vertical craniofacial patterns are interrelated and show associations with tongue and hyoid position, as well as tongue volume. Transverse dental arch dimensions are correlated not only with tongue position and volume but also with pharyngeal airway volume.

AI-Based screening for thoracic aortic aneurysms in routine breast MRI

Prognosis for thoracic aortic aneurysms is significantly worse for women than men, with a higher mortality rate observed among female patients. The increasing use of magnetic resonance breast imaging (MRI) offers a unique opportunity for simultaneous detection of both breast cancer and thoracic aortic aneurysms. We retrospectively validate a fully-automated artificial neural network (ANN) pipeline on 5057 breast MRI examinations from public (Duke University Hospital/EA1141 trial) and in-house (Erlangen University Hospital) data. The ANN, benchmarked against 3D-ground-truth segmentations, clinical reports, and a multireader panel, demonstrates high technical robustness (dice/clDice 0.88-0.91/0.97-0.99) across different vendors and field strengths. The ANN improves aneurysm detection rates by 3.5-fold compared with routine clinical readings, highlighting its potential to improve early diagnosis and patient outcomes. Notably, a higher odds ratio (OR = 2.29, CI: [0.55,9.61]) for thoracic aortic aneurysms is observed in women with breast cancer or breast cancer history, suggesting potential further benefits from integrated simultaneous assessment for cancer and aortic aneurysms.

Automated Segmentation of Thoracic Aortic Lumen and Vessel Wall on 3D Bright- and Black-Blood MRI using nnU-Net

BACKGROUND

Magnetic resonance angiography (MRA) is an important tool for aortic assessment in several cardiovascular diseases. Assessment of MRA images relies on manual segmentation; a time-intensive process that is subject to operator variability. We aimed to optimize and validate two deep-learning models for automatic segmentation of the aortic lumen and vessel wall in high-resolution ECG-triggered free-breathing respiratory motion-corrected 3D bright- and black-blood MRA images.

METHODS

Manual segmentation, serving as the ground truth, was performed on 25 bright-blood and 15 black-blood 3D MRA image sets acquired with the iT2PrepIR-BOOST sequence (1.5T) in thoracic aortopathy patients. The training was performed with nnU-Net for bright-blood (lumen) and black-blood image sets (lumen and vessel wall). Training consisted of a 70:20:10% training: validation: testing split. Inference was run on datasets (single vendor) from different centres (UK, Spain, and Australia), sequences (iT2PrepIR-BOOST, T2 prepared CMRA, and TWIST MRA), acquired resolutions (from 0.9 mm3 to 3 mm3), and field strengths (0.55T, 1.5T, and 3T). Predictive measurements comprised Dice Similarity Coefficient (DSC), and Intersection over Union (IoU). Postprocessing (3D slicer) included centreline extraction, diameter measurement, and curved planar reformatting (CPR).

RESULTS

The optimal configuration was the 3D U-Net. Bright blood segmentation at 1.5T on iT2PrepIR-BOOST datasets (1.3 and 1.8 mm3) and 3D CMRA datasets (0.9 mm3) resulted in DSC ≥ 0.96 and IoU ≥ 0.92. For bright-blood segmentation on 3D CMRA at 0.55T, the nnUNet achieved DSC and IoU scores of 0.93 and 0.88 at 1.5 mm³, and 0.68 and 0.52 at 3.0 mm³, respectively. DSC and IoU scores of 0.89 and 0.82 were obtained for CMRA image sets (1 mm3) at 1.5T (Barcelona dataset). DSC and IoU score of the BRnnUNet model were 0.90 and 0.82 respectively for the contrast-enhanced dataset (TWIST MRA). Lumen segmentation on black blood 1.5T iT2PrepIR-BOOST image sets achieved DSC ≥ 0.95 and IoU ≥ 0.90, and vessel wall segmentation resulted in DSC ≥ 0.80 and IoU ≥ 0.67. Automated centreline tracking, diameter measurement and CPR were successfully implemented in all subjects.

CONCLUSION

Automated aortic lumen and wall segmentation on 3D bright- and black-blood image sets demonstrated excellent agreement with ground truth. This technique demonstrates a fast and comprehensive assessment of aortic morphology with great potential for future clinical application in various cardiovascular diseases.

Patient-specific three-dimensional-printed models for canine adrenalectomy: a report of three cases

Case history: Three dogs with adrenal masses scheduled for adrenalectomy were prospectively enrolled into a study to investigate the effectiveness of a 1:1-scale, three-dimensional (3D) printed model of neoplastic adrenal glands to aid surgical planning and provide intra-operative assistance during adrenalectomy in dogs.Case 1 presented with anorexia, lethargy and a distended abdomen; Case 2 with loss of appetite, behavioural changes, and vocalisation; and Case 3 with mild inappetence during the previous 15 days.Clinical and imaging findings: On physical examination, mild abdominal pain was noted in all cases. Case 1 was consistently mildly hypertensive over repeated measurements. All cases had mild or moderate elevations in the activities of alanine aminotransferase and aspartate aminotransferase, and the concentration of C-reactive protein. Cases 1 and 2 also had mild leucocytosis. Abdominal CT revealed a left-sided adrenal tumour with caval invasion in Case 1, and right-sided adrenal tumours without caval invasion in Cases 2 and 3. 3D-printed models were created from the CT scan. Different colours were assigned to anatomical structures for better visualisation. Measurements of six anatomical landmarks were made on CT images and on the 3D-printed model. The median absolute difference in measurements taken from the model and the CT scan was 0.75 (min 0, max 3.2) mm.Treatment and outcome: All dogs underwent surgical removal of the adrenal tumour via sterno-pubic celiotomy. Placing the 3D model on the operating table in the same orientation as the patient allowed for precise pre-planning of the dissection depth. Printed without the fat, and fibrous and capsular tissues that typically cover the retroperitoneal space, the model helped the surgeon to visualise vascular structures that were still covered by connective tissue in the patient. Subjectively, the use of 3D models improved surgical planning and execution by enhancing the understanding of anatomical relationships and enabling the accurate identification of surgical landmarks.No major intra-operative complications were reported. Post-operative outcomes were favourable, with no significant complications observed.Clinical relevance: The use of 3D-printed models in adrenal surgeries for dogs may enhance the surgeon's spatial awareness and intra-operative confidence. We recommend that these models are used in conjunction with CT imaging for effective pre-operative planning. Further research with larger sample sizes and a control group would allow a fuller exploration of the benefits of 3D-printed models in veterinary surgical practices.

Optimization of the size and location of the FOVs for CBCT capture of impacted lower third molars

BACKGROUND

Cone beam computed tomography (CBCT) is an established diagnostic tool for impacted wisdom teeth (third molars (3 M)) in proximity of the mandibular nerve canal. This study aims to define the minimum field-of-view (FOV) size and its localization to reduce radiation exposure. As reference, the chin rest of the CBCT device was used.

METHODS

Three-dimensional CBCT data sets were used to analyze the bilateral positions and dimensions of the wisdom teeth. A total of 215 wisdom teeth from a study population with a mean age of 21 years, including data from 82 male and 58 female patients, were mapped. By transformation into a common coordinate space using the device's chin rest as a joint denominator, the optimal size and location for uni- and bilateral capture of the wisdom teeth were determined, for both best-case and worst-case scenarios with regard to patient positioning.

RESULTS

The minimal FOVs for the lower 3 M capture were H 23.5 mm × R 35.4 mm in the best-case scenario assuming optimal patient positioning and H 35.4 mm × R 36.6 mm in the worst-case scenario with rotational deviation along the transversal axis. For the upper 3 M, the minimal FOVs were H 29.9 mm × R 29.2 mm in the best-case scenario and H 38.6 mm × R 35.6 mm in the worst-case scenario. Unilateral capture of both the upper and lower 3 M required FOV dimensions of H 51.7 mm × R 39.8 mm and H 44.8 mm × R 36.8 mm, respectively. For bilateral capture of all four 3 M, the best-case FOV was H 44.8 mm × R 84.8 mm and the worst-case FOV was H 51.7 mm × R 85.6 mm.

DISCUSSION

This research provides indication-specific FOVs for uni- and bilateral imaging of the upper and lower 3 M. Taking into account optimal clinical practices for CBCT imaging, this study aims to propose clinically feasible FOV dimensions while meeting the technical specifications of commonly used CBCT devices. Clinical application of the results may help reduce radiation exposure of patients receiving CBCT imaging of the wisdom teeth. Transfer of the present results to other CBCT devices requires further research.

TRIAL REGISTRATION

The study is registered in the German Trial Register with the number DRKS00026149, 2024/02/21.

An accuracy assessment of SlicerAutoscoperM - software for tracking skeletal structures in multi-plane videoradiography datasets

Tracking skeletal joint kinematics in vivo with biplane videoradiography (BVR) can rigorously address a range of important questions in musculoskeletal research. Here we report on SlicerAutoscoperM (SAM), an upgrade of the markerless tracking software (Autoscoper) and its migration into the established computing environment of 3DSlicer and addition of a comprehensive pre-processing module to provide a standardized workflow. We present the accuracy and agreement in tracking four skeletal joints by four research groups. Accuracy was assessed by comparing marker-generated and SAM kinematics for bones of the foot, knee, shoulder, and wrist. Bland-Altman analyses quantified bias (mean error) and limits of agreement (LOA). Tracking accuracy was robust for all joints. In the foot, mean error (bias) was less than 0.5° (1.8°) and 0.8 mm (3.1 mm). In the knee, mean error was less than 1.0° (1.5°) and 0.4 mm (0.8 mm). In the shoulder, mean translational error for both the humerus and scapula was less than 0.2 mm (0.7 mm). Rotational error was highest in Roll and Pitch for the humerus, 1.9° (4.8°) and 1.7° (4.6°), respectively, and Yaw was 0.3° (2.1°). The scapula rotational bias was less than 0.2° (0.7°). In the wrist, the error was less than 0.05° (1.2°) and 0.5 mm (1.2 mm). Our data demonstrate that SAM is an accurate image-based skeletal motion tracking tool. With broad adoption, SAM will promote collaboration, simplify the harmonization of methods between study sites for large multi-center research studies, lower the entry bar for early-stage investigators, and facilitate translations toward clinical use.

Association Between the Amount of Carotid Perivascular Adipose Tissue and Prior Ischemic Stroke: An MR Imaging Study

BACKGROUND

This study aimed to explore the association between the amount of carotid artery perivascular adipose tissue (PVAT) quantified by magnetic resonance imaging and prior cerebral infarction.

METHODS

A total of 139 patients (mean age, 64.4±8.2 years; 112 men) with moderate-to-severe atherosclerotic stenosis referred to carotid endarterectomy were included and underwent multicontrast magnetic resonance vessel wall and brain imaging. The amount of carotid artery PVAT with vulnerable plaque components on magnetic resonance images of each patient was quantitatively analyzed, and the measurements included the average PVAT area, PVAT area index, and PVAT volume index. The amount measurements of PVAT at slices with vulnerable plaque between patients with and without prior cerebral infarction were compared. Logistic regression analyses were conducted to determine the association between the amount measurements of PVAT and prior cerebral infarction.

RESULTS

Patients with prior cerebral infarction showed significantly higher PVAT area, PVAT area index, and PVAT volume index compared with those without (all P<0.01). The carotid PVAT area (odds ratio [OR], 1.015 [95% CI, 1.003-1.028]; P=0.018), PVAT area index (OR, 2.051 [95% CI, 1.084-3.880]; P=0.027), and PVAT volume index (OR, 2.864 [95% CI, 1.343-6.108]; P=0.006) on the index side were significantly associated with prior cerebral infarction in univariate logistic regression. After adjusting for clinical confounding factors and plaque features, the associations between carotid PVAT area (OR, 1.028 [95% CI, 1.008-1.048]; P=0.006), PVAT area index (OR, 3.587 [95% CI, 1.451-8.870]; P=0.006), and PVAT volume index (OR, 6.053 [95% CI, 2.048-17.889]; P=0.001) and prior cerebral infarction remained statistically significant.

CONCLUSIONS

The amount of PVAT in carotid artery with vulnerable plaques is independently associated with prior cerebral infarction and may, therefore, be related to the occurrence of ischemic stroke.

Adaptive reinforcement learning is causally supported by anterior cingulate cortex and striatum

Reinforcement learning can benefit from adaptive strategies that adjust exploration-exploitation levels, leverage working memory, or guide attention toward relevant information. We tested how the anterior cingulate cortex (ACC) and the striatum support these processes during learning of feature-based attention at varying feature uncertainty and motivational saliency. Brief, gaze-contingent electrical stimulation affected adaptive reinforcement learning in ACC and the striatum at high feature uncertainty, but in opposite ways. ACC stimulation impaired learning, while striatum stimulation improved learning. Modeling showed that ACC stimulation impaired optimizing exploration and use of prediction errors to reduce uncertainty, while striatum stimulation improved the updating of value expectations. These findings were consistent with neuronal selectivity. In ACC, neurons tracked error history and fired more strongly during more uncertain choices, while in the striatum, neurons fired more strongly during more certain, higher-value choices. These results show that the ACC and the striatum optimize the guidance of exploration toward reward-relevant objects during periods of uncertainty.

Two new cichlid species of the genus Labrochromis from rocky reefs of Lake Victoria, Tanzania (Perciformes, Cichlidae)

Lake Victoria is home to a unique and taxonomically understudied species flock of endemic haplochromine cichlid fishes, with many morphologically specialized trophic groups and many different species in each of them. One of several mollusk-eating trophic groups are the pharyngeal snail-crushers of the genus Labrochromis Regan, 1920. Currently, six species from Lake Victoria have been described in this genus, none of which occupies rocky shores and reefs. Rocky shores and reefs of Lake Victoria, however harbor rich assemblages of habitat-specialized cichlids and these include snail-crushers. Here two new species of Labrochromis are described from this habitat in the Southeastern part of Lake Victoria. These species are distinct in their ecology, morphology, and male nuptial coloration from all previously described Labrochromis species, and they are distinct from each other in oral dentition, morphology, and stripe pattern. These species are named Labrochromismawe sp. nov. and Labrochromismawepili sp. nov. Both are currently only known from the Mwanza and Speke Gulf regions of the lake in Tanzania.

Fast automated creation of digital twins for virtual mechanical testing of ovine fractured tibiae

Virtual mechanical testing with image-based digital twins enables subject-specific insights about the mechanical progression of bone fracture healing directly from imaging data. However, this technique is currently limited by the need for commercial software packages that require manual input to create finite element (FE) models from computed tomography (CT) scans. The purpose of this study was to develop automated image analysis algorithms that can create subject-specific models from CT scans without a human in the loop. Two competing techniques were developed and tested on an imaging dataset consisting of 26 intact and 44 osteotomized ovine tibiae. In both techniques, the raw image was cropped to an efficient bounding box, downsampled, segmented by an element-formation threshold, and cleaned up for efficient FE analysis using voxel-based meshes. The key difference between contour-free (CFT) and snake-reliant (SRT) techniques was threshold- and contour-based segmentation of images, respectively, before bounding box detection. The contours were detected using a snake that balanced desired aspects of the contours through energy minimization. Virtual torsion tests were performed and the results were validated by comparison to ground-truth experimental data. The CFT and SRT models produced nearly identical predictions of virtual torsional rigidity and both methods reliably replicated the physical tests. Models generated by SRT were faster to solve, but model preparation and solution combined was faster by CFT. Automatic digital twin creation by CFT is therefore recommended except where other downstream analyses require systematic spatial data sampling of the bone, which is only achieved by SRT.

Single-Cell Transcriptomics Revealed White Matter Repair Following Subarachnoid Hemorrhage

Existing research indicates the potential for white matter injury repair during the subacute phase following subarachnoid hemorrhage (SAH). However, elucidating the role of brain cell subpopulations in the acute and subacute phases of SAH pathogenesis remains challenging due to the cellular heterogeneity of the central nervous system. In this study, single-cell RNA sequencing was conducted on SAH model mice to delineate distinct cell populations. Gene Set Enrichment Analysis was performed to identify involved pathways, and cellular interactions were explored using the CellChat package in R software. Validation of the findings involved a comprehensive approach, including magnetic resonance imaging, immunofluorescence double staining, and Western blot analyses. This study identified ten major brain clusters with cell type-specific gene expression patterns. Notably, we observed infiltration and clonal expansion of reparative microglia in white matter-enriched regions during the subacute stage after SAH. Additionally, microglia-associated pleiotrophin (PTN) was identified as having a role in mediating the regulation of oligodendrocyte precursor cells (OPCs) in SAH model mice, implicating the activation of the mTOR signaling pathway. These findings emphasize the vital role of microglia-OPC interactions might occur via the PTN pathway, potentially contributing to white matter repair during the subacute phase after SAH. Our analysis revealed precise transcriptional changes in the acute and subacute phases after SAH, offering insights into the mechanism of SAH and for the development of drugs that target-specific cell subtypes.

MRI Radiomics Analysis in the Diagnostic Differentiation of Malignant Soft Tissue Myxoid Sarcomas From Benign Soft Tissue Musculoskeletal Myxomas

BACKGROUND

Differentiation of benign myxomas and malignant myxoid sarcomas can be difficult with an overlapping spectrum of morphologic MR findings.

PURPOSE

To assess the diagnostic utility of MRI radiomics in the differentiation of musculoskeletal myxomas and myxoid sarcomas.

STUDY TYPE

Retrospective.

POPULATION

A total of 523 patients were included; histologically proven myxomas (N = 201) and myxoid sarcomas (N = 322), randomly divided (70:30) into training:test subsets.

SEQUENCE/FIELD STRENGTH

T1-weighted (T1W), T2-weighted fat-suppressed (fluid-sensitive), and T1-weighted post-contrast (T1W + C) sequences at 1.0 T, 1.5 T, or 3.0 T.

ASSESSMENT

Seven semantic (qualitative) tumor features were assessed in each case. Manual 3D tumor segmentations performed with radiomics features extracted from T1W, fluid-sensitive, and T1W + C acquisitions. Models were constructed based on radiomic features from individual sequences and from their combination, both with and without the addition of qualitative tumor features.

STATISTICAL TESTS

Intraclass correlation evaluated in 60 cases segmented by three readers. Features with intraclass correlation <0.7 excluded from further analysis. Boruta feature selection and Random Forest modeling performed using the training-dataset, with resultant models used to assess class discrimination (myxoma vs. myxoid sarcoma) in the test dataset. Radiomics score defined as probability class = myxoma. Logistic regression modeling employed to estimate performance of the radiomics score. Area under the receiver operating characteristic curve (AUC) was used to assess diagnostic performance, and DeLong's test to assess performance between constructed models. A P-value <0.05 was considered significant.

RESULTS

Four qualitative semantic features showed significant predictive power in class discrimination. Radiomic models demonstrated excellent differentiation of myxomas from myxoid sarcomas: AUC of 0.9271 (T1W), 0.9049 (fluid-sensitive), and 0.9179 (T1W + C). Incorporation of multiparametric data or semantic features did not significantly improve model performance (P ≥ 0.08) compared to radiomic models derived from any individual MRI sequence alone.

DATA CONCLUSION

MRI radiomics appears to be accurate in the differentiation of myxomas from myxoid sarcomas. Classification performance did not improve when incorporating qualitative features or multiparametric imaging data.

PLAIN LANGUAGE SUMMARY

Accurately distinguishing between benign soft tissue myxomas and malignant myxoid sarcomas is essential for guiding appropriate management but remains challenging with conventional MRI interpretation. This study utilized radiomics, a method that extracts quantitative mathematically derived features from images, to develop predictive models based on routine MRI examination. Analyzing over 500 cases, MRI radiomics demonstrated excellent diagnostic accuracy in differentiating between benign myxomas and malignant myxoid sarcomas, highlighting the potential of the technique, as a powerful non-invasive tool that could complement current diagnostic approaches, and enhance clinical decision-making in patients with soft tissue myxoid tumors of the musculoskeletal system.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

MR Elastography for Classification of Focal Liver Lesions Using Viscoelastic Parameters: A Pilot Study Based on Intrinsic and Extrinsic Activations

BACKGROUND

Intrinsic activation MR elastography (iMRE) uses cardiovascular pulsations to assess tissue viscoelastic properties. Applying it to focal liver lesions extends its capabilities.

PURPOSE

To assess the viscoelastic parameters of focal liver lesions measured by iMRE and compare its diagnostic performance with extrinsic MRE (eMRE) for differentiating malignant and benign lesions.

STUDY TYPE

Prospective.

POPULATION

A total of 55 participants underwent MRI with research MRE sequences; 32 participants with 17 malignant and 15 benign lesions underwent both iMRE and eMRE. FIELD STRENGTH/SEQUENCE: iMRE at ~1 Hz heart rate used a 3 T scanner with a modified four-dimensional (4D)-quantitative flow gradient-echo phase contrast and low-velocity encoding cardiac-triggered technique. eMRE employed a gradient-echo sequence at 30, 40, and 60 Hz.

ASSESSMENT

Liver displacements were measured using 4D-phase contrast and reconstructed via a nonlinear inversion algorithm to determine shear stiffness (SS) and damping ratio (DR). iMRE parameters were normalized to the corresponding values from the spleen. Lesions were manually segmented, and image quality was reviewed.

STATISTICAL TESTS

Kruskal-Wallis, Mann-Whitney, Dunn's test, and areas under receiver operating characteristic curves (AUC) were assessed.

RESULTS

SS was significantly higher in malignant than benign lesions with iMRE at 1 Hz (3.69 ± 1.31 vs. 1.63 ± 0.45) and eMRE at 30 Hz (3.76 ± 1.12 vs. 2.60 ± 1.26 kPa), 40 Hz (3.76 ± 1.12 vs. 2.60 ± 1.26 kPa), and 60 Hz (7.32 ± 2.87 vs. 2.48 ± 1.12 kPa). DR was also significantly higher in malignant than benign lesions at 40 Hz (0.36 ± 0.11 vs. 0.21 ± 0.01) and 60 Hz (0.89 ± 0.86 vs. 0.22 ± 0.09). The AUC were 0.86 for iMRE SS, 0.87-0.98 for eMRE SS, 0.47 for iMRE DR, and 0.62-0.86 for eMRE DR.

DATA CONCLUSION

Cardiac-activated iMRE can characterize liver lesions and differentiate malignant from benign lesions through normalized SS maps.

LEVEL OF EVIDENCE

2 TECHNICAL EFFICACY: Stage 2.

Differentiating Glioma Recurrence and Pseudoprogression by APTw CEST MRI

OBJECTIVES

Recurrent glioma is highly treatment resistant due to its metabolic, cellular, and molecular heterogeneity and invasiveness. Tumor monitoring by conventional MRI has shortcomings to assess these key glioma characteristics. Recent studies introduced chemical exchange saturation transfer for metabolic imaging in oncology and assessed its diagnostic value for newly diagnosed glioma. This prospective study investigates amide proton transfer-weighted (APTw) MRI at 3 T as an imaging biomarker to elucidate the molecular heterogeneity and invasion patterns of recurrent glioma in comparison to pseudoprogression (PsPD).

MATERIALS AND METHODS

We performed a monocenter, prospective trial and screened 371 glioma patients who received tumor monitoring between August 2021 and March 2024 at our institution. The study included IDH wildtype astrocytoma and IDH mutant astrocytoma and oligodendroglioma, graded according to the WHO 2021 classification. Patients had received clinical standard of care treatment including surgical resection and radiochemotherapy prior to study inclusion. Patients were monitored by 3 monthly MRI follow-up imaging, and response assessment was performed according to the RANO criteria. Within this cohort, we identified 30 patients who presented with recurrent glioma and 12 patients with PsPD. In addition to standard anatomical sequences (FLAIR and T1-w Gd-enhanced sequences), MRI included APTw imaging. After sequence co-registration, semiautomated segmentation was performed of the FLAIR lesion, CE lesion, resection cavity, and the contralateral normal-appearing white matter, and APTw signals were quantified in these regions of interest.

RESULTS

APTw values were highest in solid, Gd-enhancing tumor parts as compared with the nonenhancing FLAIR lesion (APTw: 1.99% vs 1.36%, P = 0.001), whereas there were no detectable APTw alterations in the normal-appearing white matter (APTw: 0.005%, P < 0.001 compared with FLAIR). Patients with progressive disease had higher APTw levels compared with patients with PsPD (APTw: 1.99% vs 1.26%, P = 0.008). Chemical exchange saturation transfer identified heterogeneity within the FLAIR lesion that was not detectable by conventional sequences. There were also focal APTw signal peaks within contrast enhancing lesions as putative metabolic hotspots within recurrent glioma. The resection cavity developed an APTw increase at recurrence that was not detectable prior to recurrence nor in patients with PsPD (APTw before recurrence: 0.6% vs 2.68% at recurrence, P = 0.03).

CONCLUSIONS

Our study shows that APTw imaging can differentiate PD and PsPD. We identify previously undetectable imaging patterns during glioma recurrence, which include alterations within resection cavity associated with disease progression. Our work highlights the clinical potential of APTw imaging for glioma monitoring and further establishes it as an imaging biomarker in neuro-oncology.

Seizure Burden and Clinical Risk Factors in Glioma-Related Epilepsy: Insights From MRI Voxel-Based Lesion-Symptom Mapping

BACKGROUND

Epilepsy is the most common preoperative symptom in patients with supratentorial gliomas. Identifying tumor locations and clinical factors associated with preoperative epilepsy is important for understanding seizure risk.

PURPOSE

To investigate the key brain areas and risk factors associated with preoperative seizures in glioma patients.

STUDY TYPE

Retrospective.

POPULATION

A total of 735 patients with primary diffuse supratentorial gliomas (372 low grade; 363 high grade) with preoperative MRI and pathology data.

FIELD STRENGTH/SEQUENCE

Axial T2-weighted fast spin-echo sequence at 3.0 T.

ASSESSMENT

Seizure burden was defined as the number of preoperative seizures within 6 months. Tumor and high-signal edema areas on T2 images were considered involved regions. A voxel-based lesion-symptom mapping analysis was used to identify voxels associated with seizure burden. The involvement of peak voxels (those most associated with seizure burden) and clinical factors were assessed as risk factors for preoperative seizure.

STATISTICAL TESTS

Univariable and multivariable binary and ordinal logistic regression analyses and chi-square tests were performed, with results reported as odds ratios (ORs) and 95% confidence intervals. A P-value <0.05 was considered significant.

RESULTS

A total of 448 patients experienced preoperative seizures. Significant seizure burden-related voxels were located in the right hippocampus and left insular cortex (based on 1000 permutation tests), with significant differences observed in both low- and high-grade tumors. Tumor involvement in the peak voxel region was an independent risk factor for an increased burden of preoperative seizures (OR = 6.98). Additionally, multivariable binary logistic regression results indicated that 1p/19q codeletion (OR = 1.51), intermediate tumor volume (24.299-97.066 cm3), and involvement of the peak voxel (OR = 6.06) were independent risk factors for preoperative glioma-related epilepsy.

CONCLUSION

Voxel areas identified through voxel-based lesion-symptom mapping analysis, along with clinical factors, show associations with clinical seizure burden, offering insights for assessing seizure burden for glioma patients.

LEVEL OF EVIDENCE

4 TECHNICAL EFFICACY: Stage 1.

Whole Brain 3D T1 Mapping in Multiple Sclerosis Using Standard Clinical Images Compared to MP2RAGE and MR Fingerprinting

Quantitative T1 and T2 mapping is a useful tool to assess properties of healthy and diseased tissues. However, clinical diagnostic imaging remains dominated by relaxation-weighted imaging without direct collection of relaxation maps. Dedicated research sequences such as MR fingerprinting can save time and improve resolution over classical gold standard quantitative MRI (qMRI) methods, although they are not widely adopted in clinical studies. We investigate the use of clinical sequences in conjunction with prior knowledge provided by machine learning to elucidate T1 maps of brain in routine imaging studies without the need for specialized sequences. A classification learner was trained on T1w (magnetization prepared rapid gradient echo [MPRAGE]) and T2w (fluid-attenuated inversion recovery [FLAIR]) data (2.6 million voxels) from multiple sclerosis (MS) patients at 3T, compared to gold standard inversion recovery fast spin echo T1 maps in five healthy subjects, and tested on eight MS patients. In the MS patient test, the results of the machine learner-produced T1 maps were compared to MP2RAGE and MR fingerprinting T1 maps in seven tissue regions of the brain: cortical grey matter, white matter, cerebrospinal fluid, caudate, putamen and globus pallidus. Additionally, T1s in lesion-segmented tissue was compared using the three different methods. The machine learner (ML) method had excellent agreement with MP2RAGE, with all average tissue deviations less than 3.2%, with T1 lesion variation of 0.1%-5.3% across the eight patients. The machine learning method provides a valuable and accurate estimation of T1 values in the human brain while using data from standard clinical sequences and allowing retrospective reconstruction from past studies without the need for new quantitative techniques.

Microsurgical Anatomy of the Common Tendinous Ring and Its Surgical Implications

BACKGROUND AND OBJECTIVES

The common tendinous ring (CTR), also known as the common annular tendon or annulus of Zinn, is a critical anatomic structure located at the convergence of the orbital apex, superior orbital fissure (SOF), optic canal, and the anterior aspect of the lateral sellar compartment. It plays a vital role in both neurosurgical and neuro-ophthalmological interventions. The aim of this study was to delineate the complex 3-dimensional (3D) topography of the CTR and explore its implications for surgical procedures.

METHODS

Ten formalin-fixed skull base specimens from adult Chinese cadavers were meticulously dissected to investigate the morphology of the CTR, focusing particularly on its relationship with the 4 extraocular rectus tendons, the optic strut, the SOF, and the optic canal. Additional skull base specimens were subjected to 3D surface scanning, computed tomography, and histopathological examinations to deepen our understanding of the CTR's structural complexities.

RESULTS

The CTR establishes a spatial, 3D tendinous assembly, encompassing 4 rectus tendons, 2 tendinous connections, and a singular common tendinous root. These components interlink to form a distinctive dual-ring configuration, featuring the optic foramen and the oculomotor foramen. The posterior part of the superior rectus tendon demarcates the common boundary between these 2 foramina. The oculomotor foramen itself serves as the central sector of the SOF. Precise incisions of the medial and lateral tendinous connections and fusions are essential for safely opening the CTR.

CONCLUSION

The structural composition, interconnections, and dual-ring configuration of the CTR are crucial for precise and safe surgery of orbital apex and adjacent regions.

FedSynthCT-Brain: A federated learning framework for multi-institutional brain MRI-to-CT synthesis

The generation of Synthetic Computed Tomography (sCT) images has become a pivotal methodology in modern clinical practice, particularly in the context of Radiotherapy (RT) treatment planning. The use of sCT enables the calculation of doses, pushing towards Magnetic Resonance Imaging (MRI) guided radiotherapy treatments. Moreover, with the introduction of MRI-Positron Emission Tomography (PET) hybrid scanners, the derivation of sCT from MRI can improve the attenuation correction of PET images. Deep learning methods for MRI-to-sCT have shown promising results, but their reliance on single-centre training dataset limits generalisation capabilities to diverse clinical settings. Moreover, creating centralised multi-centre datasets may pose privacy concerns. To address the aforementioned issues, we introduced FedSynthCT-Brain, an approach based on the Federated Learning (FL) paradigm for MRI-to-sCT in brain imaging. This is among the first applications of FL for MRI-to-sCT, employing a cross-silo horizontal FL approach that allows multiple centres to collaboratively train a U-Net-based deep learning model. We validated our method using real multicentre data from four European and American centres, simulating heterogeneous scanner types and acquisition modalities, and tested its performance on an independent dataset from a centre outside the federation. In the case of the unseen centre, the federated model achieved a median Mean Absolute Error (MAE) of 102.0 HU across 23 patients, with an interquartile range of 96.7-110.5 HU. The median (interquartile range) for the Structural Similarity Index (SSIM) and the Peak Signal to Noise Ratio (PNSR) were 0.89 (0.86-0.89) and 26.58 (25.52-27.42), respectively. The analysis of the results showed acceptable performances of the federated approach, thus highlighting the potential of FL to enhance MRI-to-sCT to improve generalisability and advancing safe and equitable clinical applications while fostering collaboration and preserving data privacy.

Serial reconstruction of Hominini manual phalanges

Serial morphology is the study of repeating, sequentially arranged structures in organisms, focusing on their development, organization, and evolutionary significance. Manual digits in primates, exhibiting proximodistal and radioulnar homology, can be analyzed using a serial morphological approach. This method offers a potential tool for reconstructing serial elements of hominin fossil record. Therefore, this study aims to analyze serial homology in proximal and intermediate phalanges of extant and extinct Hominini species to validate a novel methodology for reconstructing missing bones within the hand. For this purpose, we designed a template (27 true landmarks and 128 curve semilandmarks) for proximal and intermediate phalanges of digits II-V, applied to Homo sapiens (n = 125), Homo neanderthalensis (n = 9), Homo naledi (n = 9), Australopithecus sediba (n = 2), and Pan troglodytes (n = 122). Missing data were estimated using bilateral symmetry or geometric estimation methods. We used principal component analysis to quantify and examine morphological variability for each phalanx. Each serially reconstructed phalanx was validated by comparing the estimated to the original morphology using generalized Procrustes analysis and Procrustes distances, principal component analysis, and the Mann-Whitney U test. The results highlight both similarities and differences in serial homology between Pan and hominins, reflecting a shared developmental 'blueprint' alongside interspecific morphological variations influenced by genetic and functional factors. Finally, serial reconstruction with homologous elements is possible in the proximal and intermediate manual phalanges of Hominini, being more accurate with a proximal-proximal or intermediate-intermediate disposition, offering potential for reconstructing missing fossil hominin manual phalanges.

Radiomics of lung ventilation/perfusion tomographic imaging in pulmonary embolism diagnosis

PURPOSE

The aim of this study was to develop a machine learning model (named V/P-mics) to identify pulmonary embolism based on lung ventilation/perfusion single-photon emission tomography (V/P-SPECT) images.

METHODS

We retrospectively collected the data of 260 patients from one hospital who underwent V/P-SPECT. Patients were randomly assigned to training and testing groups in a 7:3 ratio. We created an internal further validation group using data of an additional 35 patients from the same hospital, and an external further validation group using data of 30 patients from another hospital. We constructed 35 models and selected one for further optimization. The generalizability of V/P-mics was proven by comparing the area under the curve (AUC) of the testing group, internal and external further validation groups. The diagnostic accuracy and efficiency of V/P-mics was compared with that of nuclear physicians.

RESULTS

V/P-mics showed excellent generalizability, with no statistical difference in AUC among the testing, internal further validation, and external further validation groups (0.938 vs. 0.923 vs. 0.990, all P values > 0.05). The AUC of V/P-mics was close to that of the senior physician (0.923 vs. 0.975, P = 0.332), but significantly higher than the junior physician (0.923 vs. 0.725, P = 0.050). Furthermore, V/P-mics significantly shortened the diagnosis time as compared to the junior physician (100 ± 16 s vs. 240 ± 37 s, P = 0.001).

CONCLUSION

The V/P-mics had good discrimination and generalizability and significantly shortened the diagnosis time for patients with pulmonary embolism. Of note, the model showed excellent interpretability.

Automated Cone Beam Computed Tomography Segmentation of Multiple Impacted Teeth With or Without Association to Rare Diseases: Evaluation of Four Deep Learning-Based Methods

OBJECTIVE

To assess the accuracy of three commercially available and one open-source deep learning (DL) solutions for automatic tooth segmentation in cone beam computed tomography (CBCT) images of patients with multiple dental impactions.

MATERIALS AND METHODS

Twenty patients (20 CBCT scans) were selected from a retrospective cohort of individuals with multiple dental impactions. For each CBCT scan, one reference segmentation and four DL segmentations of the maxillary and mandibular teeth were obtained. Reference segmentations were generated by experts using a semi-automatic process. DL segmentations were automatically generated according to the manufacturer's instructions. Quantitative and qualitative evaluations of each DL segmentation were performed by comparing it with expert-generated segmentation. The quantitative metrics used were Dice similarity coefficient (DSC) and the normalized surface distance (NSD).

RESULTS

The patients had an average of 12 retained teeth, with 12 of them diagnosed with a rare disease. DSC values ranged from 88.5% ± 3.2% to 95.6% ± 1.2%, and NSD values ranged from 95.3% ± 2.7% to 97.4% ± 6.5%. The number of completely unsegmented teeth ranged from 1 (0.1%) to 41 (6.0%). Two solutions (Diagnocat and DentalSegmentator) outperformed the others across all tested parameters.

CONCLUSION

All the tested methods showed a mean NSD of approximately 95%, proving their overall efficiency for tooth segmentation. The accuracy of the methods varied among the four tested solutions owing to the presence of impacted teeth in our CBCT scans. DL solutions are evolving rapidly, and their future performance cannot be predicted based on our results.

Relationship between magnetization transfer ratio and axial compressive strain in tibiofemoral articular cartilage

PURPOSE

The objective of this study was to determine the relationship between magnetization transfer ratio (MTR) and strain in ex vivo bovine cartilage using 9.4T magnetic resonance imaging (MRI) and to compare this relationship to the relationship between T2 and strain.

METHODS

A previously designed custom electropneumatic loading device was used to compress together osteochondral blocks of bovine femoral and tibial cartilage and bone within a 9.4T Bruker MRI scanner. Stepwise loads were applied to compress cartilage to targets of 10%, 20% and 30% strain. Images were acquired for unloaded cartilage and after each load had been held constant for 20 min to minimize creep effects. A custom algorithm was used to quantify T2 and MTR (average, by depth, and column wise) in the region of contact, and to calculate axial cartilage strain. Repeated measures correlation was used to investigate potential correlations between MTR and strain, and T2 and strain.

RESULTS

Mean MTR vs strain correlations were rrm = -0.90 (CI = -0.96 to -0.75) in the tibia and rrm = -0.70 (CI = -0.87 to -0.35) in the femur. Mean T2 vs strain correlations were rrm = -0.84 (CI = -0.94 to -0.59) in the tibia and rrm = -0.11 (CI = -0.57 to 0.41) in the femur. Column-wise analyses produced negligible or weak correlations (-0.07 to -0.34 for T2 and MTR).

CONCLUSIONS

MTR increases with strain in the region of contact, and MTR is more strongly correlated to strain than T2. MTR shows promise as a surrogate measure of strain, especially when averaged over the region of contact.

Altered vertebral biomechanical properties in prostate cancer patients following androgen deprivation therapy

Androgen deprivation therapy (ADT) for localised and metastatic prostate cancer (PCa) is known to improve survival in patients but has been associated with negative long-term impacts on the skeleton, including decreased bone mineral density (BMD) and increased fracture risk. Generally, dual-enery X-ray absorptiometry (DXA) measurements of areal BMD (aBMD) of vertebrae are used clinically to assess bone health. However, a prediction of vertebral bone strength requires information that aBMD cannot provide, such as geometry and volumetric BMD (vBMD). This study aims to investigate the effect of ADT on the densitometric (aBMD, trabecular vBMD, integral vBMD) and mechanical integrity (failure load and failure strength) of vertebrae, using a combination of DXA, quantitative computed tomography (QCT) and finite element (FE) modelling. For the FE analyses, 3D models were reconstructed from QCT images of 26 ADT treated patients, and their matched controls, collected as part of the ANTELOPE clinical trial. The ADT treated group experienced significantly decreased trabecular and integral vBMD (trabecular vBMD: -18 %, p < 0.001, integral vBMD: -11 %, p < 0.001) compared to control patients that showed no significant temporal changes (trabecular vBMD p = 0.037, integral vBMD p = 0.56). A similar trend was seen in the ADT treated group for the failure load and failure strength, where a decrease of 14 % was observed (p < 0.001). When comparing the proficiency in predicting the mechanical properties from densitometric properties, the integral vBMD performed best in the pooled data (r = 0.86-0.87, p < 0.001) closely followed by trabecular vBMD (r = 0.73-0.75, p < 0.001) with aBMD having a much weaker predictive ability (r = 0.19-0.21, p < 0.01). In conclusion, ADT significantly reduced both the densitometric properties and the mechanical strength of vertebrae. A stronger relationship between both trabecular vBMD and integral vBMD with the mechanical properties than the aBMD was observed, suggesting that such clinical measurements could improve predictions of fracture risk in prostate cancer patients treated with ADT.

Minimally invasive balloon reduction of acute Hill-Sachs lesions and impression fractures after shoulder dislocation: an experimental approach

BACKGROUND

The shoulder is the joint most often affected by dislocations. It is known that bony defects of the glenoid and/or humerus can lead to recurrent dislocations even after arthroscopic shoulder stabilization. To prevent recurrent instability, it appears reasonable to reduce fresh and larger Hill-Sachs lesions (off-track lesions). In spine surgery, balloon kyphoplasty is widely used for fracture reduction. The aim of this study was to determine whether the balloon dilatation procedure is transferable to the humeral head and under which conditions it can be successful.

METHODS

In this biomechanical proof-of-concept study, a preformed "impression stamp" was used to create standardized Hill-Sachs impressions (Calandra grade III lesion) on 13 human humeral specimens. A kyphoplasty balloon was then placed intraosseously and close to the defect using a special monolateral trocar and the defect was reduced by balloon dilatation. Computed tomography imaging was performed on the native specimens, after defect placement and after reduction. Both the defect size and the extent of reduction were measured radiologically, as well as the bone density. Also, the pressure and volume necessary for reduction were measured.

RESULTS

Thirteen proximal humerus specimens were examined. The mean age of the 7 body donors (3 female) was 85 years (standard deviation [SD] ±8.1). The average bone density was 49.0 mg/ml (SD ±26.9) and therefore defined as osteoporotic. The average defect depth (difference to native, measured from a best-fit circle) was 4.1 mm (SD ±1.4), corresponding to an average defect volume of 1203 mm³ (SD ±949). After balloon dilatation, the average achieved reduction (difference to the defect) was 4.3 mm (SD ±1.9) or 1223 mm³ (SD ±1059). This results in an average difference from native to reduced of -0.2 mm (SD ±0.5) or 20 mm³ (SD ±212) and represents no significant difference compared to baseline values (P = .90 or P = .75, respective), but a slight over-reduction in the volumetric measurements. In all humeral specimens, a reduction of at least 2.6 mm or 221 mm³ compared to the defect situation was achieved. An average volume of 3.4 ml (SD ±1.1) and pressure of 271 psi (SD ±81) were required for reduction.

CONCLUSION

The present results demonstrate that minimally invasive balloon reposition of large acute Hill-Sachs lesions appears methodologically feasible. The required pressure for reduction seems to be higher compared to previous studies. Further studies are necessary for the development of this technique and its reproducible translation into clinical practice.

Does Covariation Between Cranial and Pelvic Shapes Alleviate the Obstetric Dilemma? Insights From a Brazilian Sample

Cranial and pelvic bones could have evolved in response to each other during human evolutionary history due to the increasingly tight fit between the baby's head and the mother's pelvis during delivery. A recently identified covariation pattern between these sets of bones and stature has shown important evidence of such an evolutionary trade-off, alleviating the chances of obstructed labor. Here, we tested the validity of this covariation pattern in a different sample, from a population with known high rates of C-section. 98 computed tomographies were used to perform statistical covariation tests (two-block partial least squares and ANOVA Procrustes) between pelvic and cranial shape, as well as other anthropometric variables, like stature, body mass, and BMI. Additionally, measurements were taken from cranial and pelvic bones for classic morphometric analyses. The results have shown an important sexual dimorphism in pelvic bones' shape but no correlation between them and stature or cranial size or shape. In terms of size, the sexual dimorphism on the true pelvis was also noticeable. The fact that the results obtained from this sample do not corroborate previous findings suggests the absence of this pattern in some populations. It also suggests that the current ideal rates of C-sections proposed by the World Health Organization might not be considering the existing diversity among human populations that may account for variable levels of difficulties in birth.

Impact of bulk density assignment of bone on MRI-based abdominal region radiotherapy planning for MR-linac workflow

PURPOSE

The purpose of the present study was to evaluate the impact of bone relative electron density (rED) assignment on radiotherapy planning for the abdominal region.

METHODS

Twenty patients who received abdominal radiotherapy using MR-Linac and underwent magnetic resonance imaging (MRI) and computed tomography (CT) simulation were analyzed. The reference plan (RP) was established using both CT and MR image sets (RP_CT and RP_MRI). The RP_MRI utilized the bulk density method. The recalculated RPs derived from various rED assignment methods were evaluated for comparison on both datasets. The RPs were recalculated by excluding rED assignment for bones (scenario A). Based on the International Commission on Radiation Units and Measurements report, lung contours were assigned rED of 0.258, and body contours were assigned 1.000 (scenario B) and 1.019 (scenario C). Dose volume histogram (DVH) differences between the three recalculated scenarios and RPs were evaluated. D95, D99, and D1cc were evaluated for target volumes, including gross tumor volume, internal target volume, and planning target volume. DVH parameters, including D1cc for each abdominal organ at risk (OAR) and the mean dose to the liver and kidneys, were evaluated. Three-dimensional local gamma analysis was conducted to assess dose distribution differences between the three recalculated scenarios and RPs.

RESULTS

In all scenarios of the CT- and MRI-based validation, the average gamma pass rates (2%/2 mm) were higher than 95%. In the CT-based validation, all target DVHs across the 20 patients showed that none exceeded 2% error in scenario A, whereas 2% and 14% exceeded the threshold in scenarios B and C, respectively. For OARs in CT and MRI-based validation, absolute maximum dose differences when compared with those of the RP were 0.19 Gy and 0.22 Gy, respectively, in scenario A.

CONCLUSION

Excluding bone rED considerations in abdominal treatment planning may not yield notable clinical differences.

Developing a Reproducible Radiomics Model for Diagnosis of Active Crohn's Disease on CT Enterography Across Annotation Variations and Acquisition Differences

To systematically identify radiomics features on CT enterography (CTE) scans which can accurately diagnose active Crohn's disease across multiple sources of variation. Retrospective study of CTE scans curated between 2013 and 2015, comprising 164 subjects (65 male, 99 female; all patients were over the age of 18) with endoscopic confirmation for the presence or absence of active Crohn's disease. All patients had three distinct sets of scans available (full and reduced dose, where the latter had been reconstructed via two different methods), acquired on a single scanner at a single institution. Radiomics descriptors from annotated terminal ileum regions were individually and systematically evaluated for resilience to different imaging variations (changes in dose/reconstruction, batch effects, and simulated annotation differences) via multiple reproducibility measures. Multiple radiomics models (by accounting for each source of variation) were evaluated in terms of classifier area under the ROC curve (AUC) for identifying patients with active Crohn's disease, across separate discovery and hold-out validation cohorts. Radiomics descriptors selected based on resiliency to multiple sources of imaging variation yielded the highest overall classification performance in the discovery cohort (AUC = 0.79 ± 0.04) which also best generalized in hold-out validation (AUC = 0.81). Performance was maintained across multiple doses and reconstructions while also being significantly better (p < 0.001) than non-resilient descriptors or descriptors only resilient to a single source of variation. Radiomics features can accurately diagnose active Crohn's disease on CTE scans across multiple sources of imaging variation via systematic analysis of reproducibility measures. Clinical utility and translatability of radiomics features for diagnosis and characterization of Crohn's disease on CTE scans will be contingent on their reproducibility across multiple types and sources of imaging variation.

Reinforcement learning for safe autonomous two-device navigation of cerebral vessels in mechanical thrombectomy

PURPOSE

Autonomous systems in mechanical thrombectomy (MT) hold promise for reducing procedure times, minimizing radiation exposure, and enhancing patient safety. However, current reinforcement learning (RL) methods only reach the carotid arteries, are not generalizable to other patient vasculatures, and do not consider safety. We propose a safe dual-device RL algorithm that can navigate beyond the carotid arteries to cerebral vessels.

METHODS

We used the Simulation Open Framework Architecture to represent the intricacies of cerebral vessels, and a modified Soft Actor-Critic RL algorithm to learn, for the first time, the navigation of micro-catheters and micro-guidewires. We incorporate patient safety metrics into our reward function by integrating guidewire tip forces. Inverse RL is used with demonstrator data on 12 patient-specific vascular cases.

RESULTS

Our simulation demonstrates successful autonomous navigation within unseen cerebral vessels, achieving a 96% success rate, 7.0 s procedure time, and 0.24 N mean forces, well below the proposed 1.5 N vessel rupture threshold.

CONCLUSION

To the best of our knowledge, our proposed autonomous system for MT two-device navigation reaches cerebral vessels, considers safety, and is generalizable to unseen patient-specific cases for the first time. We envisage future work will extend the validation to vasculatures of different complexity and on in vitro models. While our contributions pave the way toward deploying agents in clinical settings, safety and trustworthiness will be crucial elements to consider when proposing new methodology.

Frontal Bone Resorption after Frontofacial Monobloc Advancement in FGFR -Related Craniosynostoses: Predictive Factors

BACKGROUND

Frontofacial monobloc advancement (FFMBA) with internal distraction is a key procedure in managing FGFR -related craniosynostoses. Resorption of the frontal bone flap can occur months to years postoperatively. This study aimed to identify clinical factors associated with the extent of frontal bone resorption in patients with Crouzon or Pfeiffer syndrome.

METHODS

A retrospective single-center study was conducted on children with Crouzon or Pfeiffer syndrome who underwent FFMBA between May of 2008 and October of 2021. Computed tomography scans were analyzed at 4 points in time: early postoperatively (before distraction), at the end of distraction, 2 years after surgery, and, optionally, 5 years after surgery. Variables examined included demographic data, genetic mutation, prior craniofacial procedures, presence of tracheostomy or ventriculoperitoneal shunt, indication for FFMBA, perioperative parameters (osteosynthesis type, advancement, or surgery duration), and postoperative parameters (cerebrospinal fluid leak, infection, distraction protocol, or retrofrontal dead space).

RESULTS

A total of 63 patients were included (mean age, 2.91 ± 1.61 years). At 2 years postoperatively, the mean frontal bone resorption was 4.83 ± 5.35 cm² (range, 0 to 27.3 cm²). Significant factors associated with increased resorption included use of absorbable sutures compared to steel wires (9.83 versus 3.99 cm²; P = 0.04) and the presence of a retrofrontal dead space at 2 years (13.12 versus 3.63 cm²; P < 0.001).

CONCLUSIONS

Frontal bone resorption following FFMBA is significantly associated with the use of absorbable sutures and persistent retrofrontal dead space. Rigid osteosynthesis should be favored, and distractor activation should be performed with follow-up of brain expansion.

CLINICAL QUESTION/LEVEL OF EVIDENCE

Risk, III.

Eliminating the second CT scan of dual-tracer total-body PET/CT via deep learning-based image synthesis and registration

PURPOSE

This study aims to develop and validate a deep learning framework designed to eliminate the second CT scan of dual-tracer total-body PET/CT imaging.

METHODS

We retrospectively included three cohorts of 247 patients who underwent dual-tracer total-body PET/CT imaging on two separate days (time interval:1-11 days). Out of these, 167 underwent [68Ga]Ga-DOTATATE/[18F]FDG, 50 underwent [68Ga]Ga-PSMA-11/[18F]FDG, and 30 underwent [68Ga]Ga-FAPI-04/[18F]FDG. A deep learning framework was developed that integrates a registration generative adversarial network (RegGAN) with non-rigid registration techniques. This approach allows for the transformation of attenuation-correction CT (ACCT) images from the first scan into pseudo-ACCT images for the second scan, which are then used for attenuation and scatter correction (ASC) of the second tracer PET images. Additionally, the derived registration transform facilitates dual-tracer image fusion and analysis. The deep learning-based ASC PET images were evaluated using quantitative metrics, including mean absolute error (MAE), peak signal-to-noise ratio (PSNR), and structural similarity index measure (SSIM) across the whole body and specific regions. Furthermore, the quantitative accuracy of PET images was assessed by calculating standardized uptake value (SUV) bias in normal organs and lesions.

RESULTS

The MAE for whole-body pseudo-ACCT images ranged from 97.64 to 112.59 HU across four tracers. The deep learning-based ASC PET images demonstrated high similarity to the ground-truth PET images. The MAE of SUV for whole-body PET images was 0.06 for [68Ga]Ga-DOTATATE, 0.08 for [68Ga]Ga-PSMA-11, 0.06 for [68Ga]Ga-FAPI-04, and 0.05 for [18F]FDG, respectively. Additionally, the median absolute percent deviation of SUV was less than 2.6% for all normal organs, while the mean absolute percent deviation of SUV was less than 3.6% for lesions across four tracers.

CONCLUSION

The proposed deep learning framework, combining RegGAN and non-rigid registration, shows promise in reducing CT radiation dose for dual-tracer total-body PET/CT imaging, with successful validation across multiple tracers.

Quantitative and Qualitative Symmetry Analysis of Open Reduction and Fixation of Zygomatic Complex Fractures

Zygomatic complex (ZMC) fractures are among the most common craniofacial injuries, impacting both function and esthetics. This study evaluates the effectiveness of open reduction and internal fixation (ORIF) in restoring facial symmetry following ZMC fractures. Sixteen patients with unilateral ZMC fractures underwent a retrospective analysis comparing preoperative and postoperative computed tomography (CT) scans to a control group of ten individuals without facial fractures. Quantitative metrics, including root mean square distance (RMSD) and heatmap analysis, were used alongside a qualitative zygoma fracture scale to assess outcomes. Postoperative results showed significant improvements in facial symmetry, with RMSD values approaching those of the control group. Heatmap analysis revealed that 50% of patients achieved deviations within 2-4 mm and 31% within 2 mm, highlighting the effectiveness of ORIF. More complex fractures exhibited higher residual asymmetry, emphasizing the influence of fracture severity on surgical outcomes. The zygoma fracture scale correlated with heatmap results, supporting its value as a complementary assessment tool. These findings demonstrate ORIF's capability to restore symmetry while identifying areas for improvement in managing complex fractures. The study underscores the need for enhanced imaging and standardized evaluation methods to optimize surgical precision and outcomes in craniofacial trauma care.

Surface-based versus voxel-based finite element head models: comparative analyses of strain responses

Finite element (FE) models of the human head are important injury assessment tools but developing a high-quality, hexahedral-meshed FE head model without compromising geometric accuracy is a challenging task. Important brain features, such as the cortical folds and ventricles, were captured only in a handful of FE head models that were primarily developed from two meshing techniques, i.e., surface-based meshing with conforming elements to capture the interfacial boundaries and voxel-based meshing by converting the segmented voxels into elements with and without mesh smoothing. Despite these advancements, little knowledge existed of how similar the strain responses were between surface- and voxel-based FE head models. This study uniquely addressed this gap by presenting three anatomically detailed models - a surface-based model with conforming meshes to capture the cortical folds-subarachnoid cerebrospinal fluid and brain-ventricle interfaces, and two voxel-based models (with and without mesh smoothing) - derived from the same imaging dataset. All numerical settings in the three models were exactly the same, except for the meshes. These three models were employed to simulate head impacts. The results showed that, when calculating commonly used injury metrics, including the percentile strains below the maximum (e.g., 99 percentile strain) and the volume of brain element with the strain over certain thresholds, the responses of the three models were virtually identical. Different strain patterns existed between the surface- and the voxel-based models at the interfacial boundary (e.g., sulci and gyri in the cortex, regions adjacent to the falx and tentorium) with strain differences exceeding 0.1, but remarkable similarities were noted at the non-interfacial region. The mesh smoothing procedure marginally reduced the strain discrepancies between the voxel- and surface-based model. This study yielded new quantitative insights into the general similarity in the strain responses between the surface- and voxel-based FE head models and underscored that caution should be exercised when using the strain at the interface to predict injury.

A radiomics approach to distinguish Progressive Supranuclear Palsy Richardson's syndrome from other phenotypes starting from MR images

BACKGROUND AND OBJECTIVE

Progressive Supranuclear Palsy (PSP) is an uncommon neurodegenerative disorder with different clinical onset, including Richardson's syndrome (PSP-RS) and other variant phenotypes (vPSP). Recognising the clinical progression of different phenotypes would enhance the accuracy of detection and treatment of PSP. The study goal was to identify radiomic biomarkers for distinguishing PSP phenotypes extracted from T1-weighted magnetic resonance images (MRI).

METHODS

Forty PSP patients (20 PSP-RS and 20 vPSP) took part in the present work. Radiomic features were collected from 21 regions of interest (ROIs) mainly from frontal cortex, supratentorial white matter, basal nuclei, brainstem, cerebellum, 3rd and 4th ventricles. After features selection, three tree-based machine learning (ML) classifiers were implemented to classify PSP phenotypes.

RESULTS

10 out of 21 ROIs performed best about sensitivity, specificity, accuracy and area under the receiver operating characteristic curve (AUCROC). Particularly, features extracted from the pons region obtained the best accuracy (0.92) and AUCROC (0.83) values while by using the other 10 ROIs, evaluation metrics range from 0.67 to 0.83. Eight features of the Gray Level Dependence Matrix were recurrently extracted for the 10 ROIs. Furthermore, by combining these ROIs, the results exceeded 0.83 in phenotypes classification and the selected areas were brain stem, pons, occipital white matter, precentral gyrus and thalamus regions.

CONCLUSIONS

Based on the achieved results, our proposed approach could represent a promising tool for distinguishing PSP-RS from vPSP.

Staged cingulotomy and capsulotomy for trigeminal neuropathic pain: A case report

Central lesioning techniques, such as cingulotomy and capsulotomy, have demonstrated efficacy in managing chronic pain, particularly in complex cases. These procedures are particularly useful for patients who are non-responsive to conventional treatments. The current study presents, to the best of our knowledge, the first case of a patient received combined stereotactic radiofrequency (RF) cingulotomy and GammaKnife capsulotomy (GKC), and who achieved good pain control after. The patient initially presented with severe, continuous left-sided facial pain unresponsive to medication and subsequently underwent a bilateral stereotactic RF cingulotomy. Pain levels were monitored using the Numeric Rating Scale. Patient satisfaction, as well as neurological and cognitive functions, were evaluated over a 2-year period. At 1-month post-cingulotomy, the patient's pain reduced from the initial score of 9/10 to 1/10, and the patient remained pain-free for the subsequent 3 months. Upon recurrence, the conducted GKC resulted in a more sustained reduction of pain to 2-3/10 at 1-year post-procedure. The patient reported high satisfaction with no observed neurological or cognitive deficits. Magnetic resonance imaging scans confirmed the presence of capsulotomy lesions in addition to the existing cingulotomy lesions. The combined cingulotomy and capsulotomy approach provided significant and sustained pain relief in the patient, indicating the potential of these techniques as advanced treatment options for refractory trigeminal neuropathic pain. This case is the first documented instance of the combined use of cingulotomy and capsulotomy, suggesting their viability for patients requiring higher-tier pain management strategies.

Robotic Secondary Cytoreduction with 3D Reconstruction for Isolated Recurrent Ovarian Cancer: A Stepwise Approach to Splenectomy

BACKGROUND

Isolated parenchymal splenic relapse is a rare condition experienced by patients presenting with recurrent ovarian cancer (ROC).1-3 In such cases, complete secondary cytoreductive surgery (SCS) followed by chemotherapy offers significant overall survival benefits for platinum-sensitive ROC patients. Randomized trials such as DESKTOP-3 and SOC-1 described splenectomy during SCS in 6-15% of patients.4,5 Robotic-assisted surgery (RAS) and advanced three-dimensional (3D) imaging reconstruction can be integrated when choosing minimally invasive surgery (MIS) to reduce the risks associated with splenic procedures.6-8 METHODOLOGY: The case of a 70-year-old BRCA2-mutated patient with single-site splenic ROC is presented. Computed tomography (CT) scan was reviewed by an expert radiologist during a multidisciplinary tumor board. The images were manually segmented using 3D Slicer software to obtain the final 3D reconstruction.9 Using the da Vinci Xi™ robot (Intuitive Surgical, Sunnyvale, CA, USA), a medial-to-lateral spleen dissection was performed. Several key surgical steps were followed to avoid tumor manipulation and subsequently minimize potential neoplastic spread.

RESULTS

Intraoperative ultrasound confirmed lesion localization, and advanced robotic instruments facilitated precise hilum control, reducing the risk of bleeding and pancreatic tail injury. Robotic SCS with complete tumor resection was achieved. No intraoperative or postoperative complications were reported. Final histology confirmed the diagnosis of high-grade ROC.

CONCLUSION

RAS represents a viable option for SCS in ROC patients with isolated parenchymal localization. The integration of 3D reconstruction with RAS allows for a tailored approach in complex cases. A high-level of expertise and appropriate identification of candidates for MIS-SCS are required to achieve the best outcomes for ROC patients.

Evaluation of dosimetric and spatial accuracy of a virtual cone technique for radiosurgery using linac-integrated CBCT-based polymer gel dosimetry

PURPOSE

This study evaluates the dosimetric and geometric precision of a virtual cone technique using CBCT-based polymer gel dosimetry, enabling radiation delivery, and imaging readout within an identical spatial coordinate system.

METHODS

We created a C# script for a virtual cone technique that generates a treatment plan with 10 gantry arcs at 0°, 36°, 72°, 288°, and 324° couch angles, with 2 arcs per couch angle using 45° and 135° collimator angles. Two verification plans using Eclipse v15.6 (AcurosXB) were created with 20 Gy at the maximum dose for: (1) a cylindrical gel, with an additional calibration region; (2) a 3D printed anthropomorphic skull phantom with a gel insert. The 50% isodose (10 Gy) width through the central axis of the axial and sagittal planes (SPs) were measured for the gel experiment. The distance between the centers-of-masses of the 10 Gy isodose region of the plan and the gel (skull phantom) were calculated for an end-to-end spatial accuracy test.

RESULTS

The maximum point dose measured with gel was within 1% of the plan, though the gel measured 50% isodose widths of 5.56± $\; \pm \;$ 0.02 mm, 5.65 ± $ \pm \;$ 0.04 mm, 4.23 ± $ \pm \;$ 0.01 mm for axial (anterior-posterior), axial (left-right), sagittal (superior-inferior) respectively, which were slightly narrower than Eclipse (1.29 mm maximum difference in the SP due to CBCT slice thickness). The center-of-mass distance was 0.66 mm for the gel experiment, and 0.94 mm for complete end-to-end testing with the anthropomorphic phantom, including CBCT setup (kV-MV isocenter uncertainty).

CONCLUSION

The 50% isodose width of the gel measurement was 5.15 mm (mean), which was tighter than our Eclipse v15.6 beam model. The end-to-end spatial accuracy test, only achievable with gel dosimetry using CBCT readout, resulted in sub-millimeter accuracy. This study demonstrates the value of gel dosimetry in verifying the dosimetric and spatial accuracy of this high precision, stereotactic technique.

Femoral bone growth predictions based on personalized multi-scale simulations: validation and sensitivity analysis of a mechanobiological model

Musculoskeletal function is pivotal to long-term health. However, various patient groups develop torsional deformities, leading to clinical, functional problems. Understanding the interplay between movement pattern, bone loading and growth is crucial for improving the functional mobility of these patients and preserving long-term health. Multi-scale simulations in combination with a mechanobiological bone growth model have been used to estimate bone loads and predict femoral growth trends based on cross-sectional data. The lack of longitudinal data in the previous studies hindered refinements of the mechanobiological model and validation of subject-specific growth predictions, thereby limiting clinical applications. This study aimed to validate the growth predictions using magnetic resonance images and motion capture data-collected longitudinally-from ten growing children. Additionally, a sensitivity analysis was conducted to refine model parameters. A linear regression model based on physical activity information, anthropometric data and predictions from the refined mechanobiological model explained 70% of femoral anteversion development. Notably, the direction of femoral development was accurately predicted in 18 out of 20 femurs, suggesting that growth predictions could help to revolutionize treatment strategies for torsional deformities.

An augmented reality overlay for navigated prostatectomy using fiducial-free 2D-3D registration

PURPOSE

Markerless navigation in minimally invasive surgery is still an unsolved challenge. Many proposed navigation systems for minimally invasive surgeries rely on stereoscopic images, while in clinical practice oftentimes monocular endoscopes are used. Combined with the lack of automatic video-based navigation systems for prostatectomies, this paper explores methods to tackle both research gaps at the same time for robot-assisted prostatectomies.

METHODS

In order to realize a semi-automatic augmented reality overlay for navigated prostatectomy, the camera pose w.r.t. the prostate needs to be estimated. We developed a method where visual cues are drawn on top of the organ after an initial manual alignment, simultaneously creating matching landmarks on the 2D and 3D data. Starting from this key frame, the cues are then tracked in the endoscopic video. Both PnPRansac and differentiable rendering are then explored to perform 2D-3D registration for each frame.

RESULTS

We performed experiments on synthetic and in vivo data. On synthetic data differentiable rendering can achieve a median target registration error of 6.11 mm. Both PnPRansac and differentiable rendering are feasible methods for 2D-3D registration.

CONCLUSION

We demonstrated a video-based markerless augmented reality overlay for navigated prostatectomy, using visual cues as an anchor.