A Preoperative MRI-Based Radiomics-Clinicopathological Classifier to Predict the Recurrence of Pituitary Macroadenoma Within 5 Years

Objective: To investigate the ability of a MRI-based radiomics-clinicopathological model to predict pituitary macroadenoma (PMA) recurrence within 5 years.Materials and Methods: We recruited 74 recurrent and 94 non-recurrent subjects, following first surgery with 5-year follow-up data. Univariate and multivariate analyses were conducted to identify independent clinicopathological risk factors. Two independent and blinded neuroradiologists used 3D-Slicer software to manually delineate whole tumors using preoperative axial contrast-enhanced T1WI (CE-T1WI) images. 3D-Slicer was then used to extract radiomics features from segmented tumors. Dimensionality reduction was carried out by the least absolute shrinkage and selection operator (LASSO). Two multilayer perceptron (MLP) models were established, including independent clinicopathological risk factors (Model 1) and a combination of screened radiomics features and independent clinicopathological markers (Model 2). The predictive performance of these models was evaluated by receiver operator characteristic (ROC) curve analysis.Results: In total, 1,130 features were identified, and 4 of these were selected by LASSO. In the test set, the area under the curve (AUC) of Model 2 was superior to Model 1 {0.783, [95% confidence interval (CI): 0.718—.860] vs. 0.739, (95% CI: 0.665–0.818)}. Model 2 also yielded the higher accuracy (0.808 vs. 0.692), sensitivity (0.826 vs. 0.652), and specificity (0.793 vs. 0.724) than Model 1.Conclusions: The integrated classifier was superior to a clinical classifier and may facilitate the prediction of individualized prognosis and therapy.

Factors Affecting Volume Reduction Velocity for Arteriovenous Malformations After Treatment With Dose-Stage Stereotactic Radiosurgery

Background and PurposeThe purpose of this study was to identify morphologic and dosimetric features associated with volume reduction velocity for arteriovenous malformation (AVM) after dose-stage stereotactic radiosurgery (DS-SRS).MethodsThirty patients with intracranial AVM were treated with DS fractionated SRS at Beijing Tiantan Hospital from 2011 to 2019. The AVM nidus was automatically segmented from DICOMRT files using the 3D Slicer software. The change in lesion volume was obtained from the decrease in the planning target volume (PTV) between the two treatment sessions. The volume reduction velocity was measured by the change in volume divided by the time interval between treatments. Fourteen morphologic features of AVM prior to treatment were extracted from the PTV using ‘Pyradiomics’ implemented in Python. Along with other dosimetric features, univariate and multivariate analyses were performed to explore predictors of the volume reduction velocity.ResultsAmong the 15 male (50.0%) and 15 female (50.0%) patients enrolled in this study, 17 patients (56.7%) initially presented with hemorrhage. The mean treatment interval between the initial and second SRS was 35.73 months. In multivariate analysis, the SurfaceVolumeRatio was the only independent factor associated with the volume reduction velocity (p=0.010, odds ratio=0.720, 95% confidence interval: 0.560–0.925). The area under the curve of this feature for predicting the volume reduction velocity after the initial treatment of DS-SRS was 0.83. (p=0.0018).ConclusionsThe morphologic features correlated well with the volume reduction velocity in patients with intracranial AVM who underwent DS-SRS treatment. The SurfaceVolumeRatio could predict the rate of volume reduction of AVMs after DS-SRS.

Quality Assurance Assessment of Diffusion-Weighted and T2-Weighted Magnetic Resonance Imaging Registration and Contour Propagation for Head and Neck Cancer Radiotherapy

Background/Purpose: Adequate image registration of anatomic and functional MRI scans is necessary for MR-guided head and neck cancer (HNC) adaptive radiotherapy planning. Despite the quantitative capabilities of diffusion-weighted imaging (DWI) MRI for treatment plan adaptation, geometric distortion remains a considerable limitation. Therefore, we systematically investigated various deformable image registration (DIR) algorithms to co-register DWI and T2-weighted (T2W) images. Materials/Methods: We compared post-acquisition registration algorithms from three software packages (ADMIRE, Velocity, and 3D Slicer) applied to T2W and DWI MRI images in twenty HNC patients. In addition, we investigated implicit rigid registration (no algorithm applied) as a control comparator. Ground truth segmentations of radiotherapy structures (tumor and organs at risk) were generated by a physician expert on both image sequences. Three additional experts provided segmentations for five cases for interobserver variability studies. For each registration approach, structures were propagated from T2W to DWI images. These propagated structures were then compared with ground truth DWI structures using the Dice similarity coefficient (DSC), false-negative DSC, false-positive DSC, surface DSC, 95% Hausdorff distance, and mean surface distance. Results: 19 left submandibular glands, 18 right submandibular glands, 20 left parotid glands, 20 right parotid glands, 20 spinal cords, 9 brainstems, and 12 tumors were delineated. ADMIRE, the atlas-based auto segmentation DIR algorithm, demonstrated improved performance over implicit rigid registrations for most comparison metrics and structures (Bonferroni-corrected p < 0.05), while Velocity and 3D Slicer algorithms did not. Moreover, the ADMIRE methods significantly improved performance in individual and pooled analysis compared to all other methods. Interobserver variability analysis revealed no significant difference between observers (p > 0.05). Conclusions: Certain deformable registration software packages, such as those provided by ADMIRE, may be favorable for registering T2W and DWI images. These results are important to ensure the appropriate selection of registration strategies for MR-guided radiotherapy.

Creating a training set for artificial intelligence from initial segmentations of airways

Airways segmentation is important for research about pulmonary disease but require a large amount of time by trained specialists. We used an openly available software to improve airways segmentations obtained from an artificial intelligence (AI) tool and retrained the tool to get a better performance. Fifteen initial airway segmentations from low-dose chest computed tomography scans were obtained with a 3D-Unet AI tool previously trained on Danish Lung Cancer Screening Trial and Erasmus-MC Sophia datasets. Segmentations were manually corrected in 3D Slicer. The corrected airway segmentations were used to retrain the 3D-Unet. Airway measurements were automatically obtained and included count, airway length and luminal diameter per generation from the segmentations. Correcting segmentations required 2–4 h per scan. Manually corrected segmentations had more branches (p < 0.001), longer airways (p < 0.001) and smaller luminal diameters (p = 0.004) than initial segmentations. Segmentations from retrained 3D-Unets trended towards more branches and longer airways compared to the initial segmentations. The largest changes were seen in airways from 6th generation onwards. Manual correction results in significantly improved segmentations and is potentially a useful and time-efficient method to improve the AI tool performance on a specific hospital or research dataset.

Three-dimensional semi-automated volumetric assessment of the pulp space of teeth following regenerative dental procedures

The volumetric change that occurs in the pulp space over time represents a critical measure when it comes to determining the secondary outcomes of regenerative endodontic procedures (REPs). However, to date, only a few studies have investigated the accuracy of the available domain-specialized medical imaging tools with regard to three-dimensional (3D) volumetric assessment. This study sought to compare the accuracy of two different artificial intelligence-based medical imaging programs namely OsiriX MD (v 9.0, Pixmeo SARL, Bernex Switzerland, https://www.osirix-viewer.com) and 3D Slicer (http://www.slicer.org), in terms of estimating the volume of the pulp space following a REP. An Invitro assessment was performed to check the reliability and sensitivity of the two medical imaging programs in use. For the subsequent clinical application, pre- and post-procedure cone beam computed tomography scans of 35 immature permanent teeth with necrotic pulp and periradicular pathosis that had been treated with a cell-homing concept-based REP were processed using the two biomedical DICOM software programs (OsiriX MD and 3D Slicer). The volumetric changes in the teeth’s pulp spaces were assessed using semi-automated techniques in both programs. The data were statistically analyzed using t-tests and paired t-tests (P = 0.05). The pulp space volumes measured using both programs revealed a statistically significant decrease in the pulp space volume following the REP (P < 0.05), with no significant difference being found between the two programs (P > 0.05). The mean decreases in the pulp space volumes measured using OsiriX MD and 3D Slicer were 25.06% ± 19.45% and 26.10% ± 18.90%, respectively. The open-source software (3D Slicer) was found to be as accurate as the commercially available software with regard to the volumetric assessment of the post-REP pulp space. This study was the first to demonstrate the step-by-step application of 3D Slicer, a user-friendly and easily accessible open-source multiplatform software program for the segmentation and volume estimation of the pulp spaces of teeth treated with REPs.

Abstract 1122‐000188: 3D Analysis of Atherosclerotic Plaque Enhancement and the Parent Vessel

Introduction : High resolution vessel wall imaging (HR‐VWI) is a promising tool in studying intracerebral atherosclerotic disease. The analysis of the interplay between the patterns of enhancement between the plaque and its parent vessel can generate further insights on the biology of these lesions. We have developed a 3D method of plaque and parent vessel analysis. Methods : Images from fifty‐five plaques were obtained using 7T HR‐VWI. T1 and T1+Gd sequences were performed. 3D reconstructions of the plaque and its parent vessel were generated with 3D Slicer. Using an in‐house code, probes were orthogonally extended from the lumen of the vessel into the vessel wall and the plaque. Signal intensity values were then normalized to the corpus callosum. 3D heat maps and histograms were generated from hundreds of data points. A detailed analysis of the morphology of the histograms was performed to determine the uptake of gadolinium (Gd) by the parent vessel and the plaque. Variations in the width of the histogram were measured with the standard deviation. Results : Forty‐one patients with 55 plaques (41 culprit and 15 non culprit) were included. There was no difference in enhancement in T1‐pre between culprit and non‐culprit plaques when compared to the parent vessel (width = 0.14 ± 0.05 and 0.14 ± 0.03, respectively; p = 0.91). On the T1+Gd culprit plaques were significantly more enhancing compared to the parent vessel (0.26 ± 0.10) than non‐culprit plaques (0.20 ± 0.06) (p = 0.02). The presence of an enhancing plaque creates a bimodal distribution that increases the width of the histogram curve (figure). Conclusions : Culprit plaques exhibit different patterns of enhancement relative to the parent vessel compared to non‐culprit plaques. Histogram analysis of the parent vessel and its plaques provides a new set of metrics that may be used as a biomarker of disease progression.

P.201 Disruption of the Frontal Aslant Tract is Associated with Transient Aphasia and not Agraphia: A Neurosurgical Case Report

Background: The frontal aslant tract (FAT) is a recently discovered white-matter tract connecting the medial superior frontal gyrus to the inferior frontal gyrus. There is increasing evidence for its importance in speech initiation and production. Despite this, there remains limited evidence demonstrating clinical outcomes when lesioning this tract. Methods: Medical records for the case were reviewed. Imaging was exported and tractography was performed using 3D Slicer. Results: A 58-year-old female presented with a focal seizure and imaging demonstrating a left frontal extra-axial lesion. She underwent a left frontal craniotomy for tumour debulking and biopsy. The final pathology was consistent with a diffuse large B-cell lymphoma. Postoperatively, she presented with expressive aphasia without agraphia. She was able to write out answers to questions she could not verbalize. We used tractography to provide evidence of postoperative disruption to her left FAT. At a 6-week clinical follow-up, her language deficits were clinically resolved. Conclusions: To our knowledge, this is the first reported case of aphasia without agraphia seen with disruption of the left FAT. Further elucidating clinical outcomes of disrupting the dominant FAT will aid in improved patient counselling, prognostication and neurosurgical planning. Further research is required to investigate the mechanisms underlying language recovery and handwriting.

Increase in Brain Volume After Aneurysmal Subarachnoid Hemorrhage Leads to Unfavorable Outcome: A Retrospective Study Quantified by CT Scan

Background and Purpose: Primary brain swelling occurs in aneurysmal subarachnoid hemorrhage (aSAH) patients. The absence of a dynamic quantitative method restricts further study of primary brain swelling. This study compared differences in the change rate of brain volume (CRBV) between patients with and without primary brain swelling in the early stage of aSAH. Moreover, the relationship between CRBV and clinical outcomes was evaluated.Methods: Patients hospitalized within 24 h after aSAH were included in this retrospective study. Utilizing a qualitative standard established before the study to recognize primary brain swelling through brain CT after aSAH, clinical outcomes after 3 months of SAH were evaluated with a modified Rankin scale (mRS). The brain volume (BV) of each patient was calculated with a semiautomatic threshold algorithm of 3D-slicer, and the change in brain volume (CIBV) was obtained by subtracting the two extreme values (CIBV = BVmax – BVmin). The CRBV was obtained by CIBV/BVmin × 100%. The CRBV values that predicted unfavorable prognoses were estimated.Results: In total, 130 subjects were enrolled in the study. The mean CRBV in the non-swelling group and swelling group were 4.37% (±4.77) and 11.87% (±6.84), respectively (p &lt; 0.05). CRBV was positively correlated with the length of hospital stay, blood in the ambient cistern, blood in the lateral ventricle, and lateral ventricular volume (Spearman ρ = 0.334; p &lt; 0.001; Pearson ρ = 0.269, p = 0.002; Pearson ρ = 0.278, p = 0.001; Pearson ρ = 0.233, p = 0.008, respectively). Analysis of variance showed significant differences in CIBV, CRBV, blood in the ambient cistern, blood in the lateral ventricle, and lateral ventricular volume among varying modified Fisher scale (mFisher), with higher admission mFisher scale, indicating larger values of these variables. After adjusting for risk factors, the model showed that for every 1% increase in the CRBV, the probability of poor clinical prognosis increased by a factor of 1.236 (95% CI = 1.056–1.446). In the stratified analysis, the odds of worse clinical outcomes increased with increases in the CRBV. Receiver operating characteristic curve analysis showed that HH grade, mFisher scale, and score of CRBV (SCRBV) had diagnostic performance for predicting unfavorable clinical outcomes.Conclusion: Primary brain swelling increases brain volume after aSAH. The CRBV quantified by 3D-Slicer can be used as a volumetric representation of the degree of brain swelling. A larger CRBV in the early stage of aSAH is associated with poor prognosis. The CRBV can be used as a neuroimaging biomarker of early brain injury after bleeding and may be an effective predictor of patients' clinical prognoses.