To evaluate the accuracy of opportunistic measurements of volumetric bone mineral density (vBMD) in intraoperative multi-detector CT (MDCT) scans, using preoperative MDCT as the reference. This retrospective, single-center study included 105 patients (mean age: 73 ± 12.6 years, 53 women) who underwent spine surgery for various indications. All patients had preoperative MDCT with/without intravenous contrast and unenhanced intraoperative scans. VBMD of thoracolumbar vertebrae was automatically extracted using a convolutional neural network (CNN)-based framework with asynchronous calibration and contrast-phase correction. Vertebrae affected by artifacts, fractures, or severe degenerations were excluded. Root-mean-square errors (RMSEs) for associations between pair-wise vertebrae from preoperative and intraoperative vBMD values were calculated in linear regression models. Mean bias and 95

Deep learning models have achieved remarkable success in segmenting brain white matter lesions in multiple sclerosis (MS), becoming integral to both research and clinical workflows. While brain lesions have gained significant attention in MS research, the involvement of spinal cord lesions in MS is relatively understudied. This is largely owing to the variability in spinal cord magnetic resonance imaging (MRI) acquisition protocols, high individual anatomical differences, the complex morphology and size of spinal cord lesions, and lastly, the scarcity of labeled datasets required to develop robust segmentation tools. As a result, automatic segmentation of spinal cord MS lesions remains a significant challenge. Although some segmentation tools exist for spinal cord lesions, most have been developed using sagittal T2-weighted (T2w) sequences primarily focusing on cervical spines. With the growing importance of spinal cord imaging in MS, axial T2w scans are becoming increasingly relevant due to their superior sensitivity in detecting lesions compared to sagittal acquisition protocols. However, most existing segmentation methods struggle to effectively generalize to axial sequences due to differences in image characteristics caused by the highly anisotropic spinal cord scans. To address these challenges, we developed a robust, open-source lesion segmentation tool tailored specifically for axial T2w scans covering the whole spinal cord. We investigated key factors influencing lesion segmentation, including the impact of stitching together individually acquired spinal regions, straightening the spinal cord, and comparing the effectiveness of 2D and 3D convolutional neural networks (CNNs). Drawing on these insights, we trained a multi-center model using an extensive dataset of 582 MS patients, resulting in a dataset comprising an entirety of 2,167 scans. We empirically evaluated the model's segmentation performance across various spinal segments for lesions with varying sizes. Our model significantly outperforms the current state-of-the-art methods, providing consistent segmentation across cervical, thoracic, and lumbar regions. To support the broader research community, we integrate our model into the widely-used Spinal Cord Toolbox (v7.0 and above), making it accessible via the command sct_deepseg lesion_ms_axial_t2 -i .

Introduction The role of low-pathogenic bacteria cultured from removed spinal implants is unclear and the efficacy of perioperative single-dose antibiotics against such bacteria remains underexplored. Research question This study aims to investigate whether pedicle screw loosening is associated with pathogens and if the choice of perioperative antibiotics can prevent these bacteria. Methods A retrospective analysis was conducted on 93 patients with implants removed between 01/01/2018 and 03/31/2020. Patients with both loosened and non-loosened implants were included. The latter group was subdivided into cases where implants were exchanged due to adjacent segment degeneration (ASD) and those with elective implant removal after fracture healing. Bacterial cultures from removed implants were analyzed for resistance profiles against the prophylactic single-shot antibiotics administered during implantation. Patients with acute infection, spondylodiscitis, deep wound infection, empyema, and carbon/polyetheretherketone spinal implants were excluded. Results Bacterial isolates were detected in both loosened (41%) and non-loosened (27%) implants (p=0.23). The most frequently cultivated bacteria were Cutibacterium acnes and Staphylococcus epidermidis. Sensitivity to the administered antibiotics was 75%. While Cutibacterium acnes was entirely sensitive, Staphylococcus epidermidis was completely resistant. Patients with loosened implants without bacteria had a significantly lower bone mineral density (BMD) than patients with implants removed due to ASD. However, patients with loosened implants and positive bacterial cultures had comparable BMD to ASD patients. Conclusions The high rate of sensitive Cutibacterium acnes and resistant Staphylococcus epidermidis on removed spinal implants suggests a need to revisit current antimicrobial prophylaxis. Further research is required to determine the clinical significance of low-virulence bacteria, especially on non-loosened implants.

ObjectivesTo quantitatively investigate the age- and sex-related longitudinal changes in trabecular volumetric bone mineral density (vBMD) and vertebral body volume at the thoracolumbar spine in adults.MethodsWe retrospectively included 168 adults (mean age 58.7 ± 9.8 years, 51 women) who received ≥7 MDCT scans over a period of ≥6.5 years (mean follow-up 9.0 ± 2.1 years) for clinical reasons. Level-wise vBMD and vertebral body volume were extracted from 22720 thoracolumbar vertebrae using a convolutional neural network (CNN)-based framework with asynchronous calibration and correction of the contrast media phase. Human readers conducted semiquantitative assessment of fracture status and bony degenerations.ResultsIn the 40-60 years age group, women had a significantly higher trabecular vBMD than men at all thoracolumbar levels (p<0.05 to p<0.001). Conversely, men, on average, had larger vertebrae with lower vBMD. This sex difference in vBMD did not persist in the 60-80 years age group. While the lumbar (T12-L5) vBMD slopes in women only showed a non-significant trend of accelerated decline with age, vertebrae T1-11 displayed a distinct pattern, with women demonstrating a significantly accelerated decline compared to men (p<0.01 to p<0.0001). Between baseline and last follow-up examinations, the vertebral body volume slightly increased in women (T1-12: 1.1 ± 1.0 cm3; L1-5: 1.0 ± 1.4 cm3) and men (T1-12: 1.2 ± 1.3 cm3; L1-5: 1.5 ± 1.6 cm3). After excluding vertebrae with bony degenerations, the residual increase was only small in women (T1-12: 0.6 ± 0.6 cm3; L1-5: 0.7 ± 0.7 cm3) and men (T1-12: 0.7 ± 0.6 cm3; L1-5: 1.2 ± 0.8 cm3). In non-degenerated vertebrae, the mean change in volume was <5% of the respective vertebral body volumes.ConclusionSex differences in thoracolumbar vBMD were apparent before menopause, and disappeared after menopause, likely attributable to an accelerated and more profound vBMD decline in women at the thoracic spine. In patients without advanced spine degeneration, the overall volumetric changes in the vertebral body appeared subtle.

Abstract Background To investigate the reproducibility of automated volumetric bone mineral density (vBMD) measurements from routine thoracoabdominal computed tomography (CT) assessed with segmentations by a convolutional neural network and automated correction of contrast phases, on diverse scanners, with scanner-specific asynchronous or scanner-agnostic calibrations. Methods We obtained 679 observations from 278 CT scans in 121 patients (77 males, 63.6%) studied from 04/2019 to 06/2020. Observations consisted of two vBMD measurements from Δdifferent reconstruction kernels (n = 169), Δcontrast phases (n = 133), scan Δsessions (n = 123), Δscanners (n = 63), or Δall of the aforementioned (n = 20), and observations lacking scanner-specific calibration (n = 171). Precision was assessed using root-mean-square error (RMSE) and root-mean-square coefficient of variation (RMSCV). Cross-measurement agreement was assessed using Bland-Altman plots; outliers within 95% confidence interval of the limits of agreement were reviewed. Results Repeated measurements from Δdifferent reconstruction kernels were highly precise (RMSE 3.0 mg/cm3; RMSCV 1.3%), even for consecutive scans with different Δcontrast phases (RMSCV 2.9%). Measurements from different Δscan sessions or Δscanners showed decreased precision (RMSCV 4.7% and 4.9%, respectively). Plot-review identified 12 outliers from different scan Δsessions, with signs of hydropic decompensation. Observations with Δall differences showed decreased precision compared to those lacking scanner-specific calibration (RMSCV 5.9 and 3.7, respectively). Conclusion Automatic vBMD assessment from routine CT is precise across varying setups, when calibrated appropriately. Low precision was found in patients with signs of new or worsening hydropic decompensation, what should be considered an exclusion criterion for both opportunistic and dedicated quantitative CT. Relevance statement Automated CT-based vBMD measurements are precise in various scenarios, including cross-session and cross-scanner settings, and may therefore facilitate opportunistic screening for osteoporosis and surveillance of BMD in patients undergoing routine clinical CT scans. Key Points Artificial intelligence-based tools facilitate BMD measurements in routine clinical CT datasets. Automated BMD measurements are highly reproducible in various settings. Reliable, automated opportunistic osteoporosis diagnostics allow for large-scale application. Graphical Abstract

Spine MRIs are usually acquired in highly anisotropic 2D axial or sagittal slices. Vertebra structures are not fully resolved in these images, and multi-image superresolution by aligning scans to pair them is difficult due to partial volume effects and inter-vertebral movement during acquisition. Hence, we propose an unpaired inpainting superresolution algorithm that extrapolates the missing spine structures. We generate synthetic training pairs by multiple degradation functions that model the data shift and acquisition errors between sagittal slices and sagittal views of axial images. Our method employs modeling of the k-space point spread function and the interslice gap. Further, we imitate different MR acquisition challenges like histogram shifts, bias fields, interlace movement artifacts, Gaussian noise, and blur. This enables the training of diffusion-based superresolution models on scaling factors larger than 6$\times$ without real paired data. The low z-resolution in axial images prevents existing approaches from separating individual vertebrae instances. By applying this superresolution model to the z-dimension, we can generate images that allow a pre-trained segmentation model to distinguish between vertebrae and enable automatic segmentation and processing of axial images. We experimentally benchmark our method and show that diffusion-based superresolution outperforms state-of-the-art super-resolution models.

Background: Evidence from randomized controlled trials has shown a benefit for endovascular treatment in basilar artery occlusion. We aimed to show the effect of the recanalization result on outcome and determine the role of underlying basilar stenosis in a real-world setting. Methods: A retrospective, single-center study of patients who received endovascular treatment for basilar artery occlusion from March 2008 to June 2022 was conducted. Clinical and outcome characteristics were gathered. Multivariate logistic regression analysis was performed to predict poor outcomes (post-treatment mRS 5 or 6). MRS shift analysis was performed. Results: This study includes 210 patients (mean age, 71.4 years +/− 13.3 [standard deviation]; 124 men). The variables age (OR, 1.04; 95% CI: 1.01–1.08; p = 0.014), underlying basilar stenosis (OR: 4.86; 95% CI: 2.15–10.98; p < 0.001), admission NHISS (OR: 1.09; 95% CI: 1.04–1.13; p < 0.001), and TICI (OR: 1.89; 95% CI: 1.09–3.25; p = 0.022) independently predicted a poor outcome. Patients with occlusions due to underlying stenosis had significantly worse recanalization rates. Median post-treatment mRS in all patients with embolic occlusion was 4; IQR, 2–5 (only patients with embolic occlusion: mTICI 0-2a, median: 5 [IQR, 4–5.5]; mTICI 2b, median: 4 [IQR, 2.5–6]; mTICI 3, median: 3 [IQR, 1–5]; p = 0.037). Median post-treatment mRS in all patients with occlusions due to underlying stenosis was 5; IQR, 4–6 (only patients with embolic occlusions: mTICI 0-2a, median: 6 [IQR, 4.5–6]; mTICI 2b, median: 6 [IQR, 4.25–6]; mTICI 3, median: 5 [IQR, 3.5–5.25]; p = 0.059). Conclusions: Successful recanalization is essential for preventing poor outcomes in basilar artery occlusion. Underlying basilar stenosis diminishes the effect of successful recanalization on the overall outcome.

Oligodendrocytes continue to differentiate from their precursor cells even in adulthood, a process that can be modulated by neuronal activity and experience. Previous work has indicated that conditional ablation of oligodendrogenesis in adult mice leads to learning and memory deficits in a range of behavioral tasks. The current study replicated and re-evaluated evidence for a role of oligodendrogenesis in motor learning, using a complex running wheel task. Further, we found that ablating oligodendrogenesis alters brain microstructure (ex vivo MRI) and brain activity (in vivo EEG) independent of experience with the task. This suggests a role for adult oligodendrocyte formation in the maintenance of brain function and indicates that task-independent changes due to oligodendrogenesis ablation need to be considered when interpreting learning and memory deficits in this model.

This study aimed to validate the correlation between volumetric bone mineral density in the cervicothoracic and lumbar spine using measurements from opportunistic CT scans. The bone density assessment proved feasible, allowing us to propose optimal cut-off values for diagnosing osteoporosis and predicting vertebral fractures in the cervical and thoracic spine. To investigate the performance of cervicothoracic volumetric bone mineral density (vBMD), obtained through opportunistic quantitative computed tomography (QCT), in discriminating patients with/without osteoporosis and with/without vertebral fractures (VFs), using lumbar vBMD as the reference. Three hundred twenty-five patients (65.3 ± 19.2 years, 140 women) with routine non-contrast or contrast-enhanced multi-detector CT (MDCT) scans were included. Trabecular vBMD was automatically extracted from each vertebra using a convolutional neural network (CNN)–based framework (SpineQ software v1.0) with asynchronous calibration and contrast phase correction. The correlations of vBMD between each vertebra spanning C2–T12 and the averaged lumbar spine (L1–L3, or L4 and L5) vBMD values were analyzed, considering fracture status and degeneration. Vertebra-specific linear regression equations were used to approximate lumbar vBMD at the cervicothoracic spine. Cervicothoracic vBMD correlated well with lumbar vBMD (r = 0.79), with significant improvement after excluding degenerated vertebrae (p < 0.05; r = 0.89), except for C7-T3 and T9. Cervical (AUC = 0.94) and thoracic vBMD (AUC = 0.97) showed strong discriminatory ability for osteoporosis (vBMD < 80 mg/cm3). Excluding degenerated vertebrae at the cervical spine increased the AUC to 0.97. Cervical and thoracic vBMD (AUC = 0.74, AUC = 0.72) were comparable to lumbar vBMD (AUC = 0.72) in differentiating patients with and without prevalent VFs. Trabecular vBMD < 190 mg/cm3 for the cervical spine and < 100 mg/cm3 for the thoracic spine were potential indicators of osteoporosis, similar to < 80 mg/cm3 at the lumbar spine. Cervicothoracic vBMD may allow for determination of osteoporosis and prediction of VFs.

Arterial spin labeling (ASL) is a contrast agent-free magnetic resonance imaging (MRI) technique to measure cerebral blood flow (CBF). We sought to investigate effects of CBF within the infarct on outcome and risk of hemorrhagic transformation (HT). In 111 patients (median age: 74 years, 50 men) who had undergone mechanical thrombectomy (MT) for ischemic stroke of the anterior circulation (median interval: 4 days between MT and MRI), post-stroke %CBF difference from pseudo-continuous ASL was calculated within the diffusion-weighted imaging (DWI)-positive infarct territory following lesion segmentation in relationship to the unaffected contralateral side. Functional independence was defined as a modified Rankin Scale (mRS) of 0-2 at 90 days post-stroke. %CBF difference, pre-stroke mRS, and infarct volume were independently associated with functional independence in a multivariate regression model. %CBF difference was comparable between patients with and without HT. A subcohort of 10 patients with decreased infarct-CBF despite expanded Treatment in Cerebral Infarction (eTICI) 2c or 3 recanalization was identified (likely related to the no-reflow phenomenon). Outcome was significantly worse in this group compared to the remaining cohort. In conclusion, ASL-derived %CBF difference from the DWI-positive infarct territory independently predicted functional independence, but %CBF difference was not significantly associated with an increased risk of HT.

The atypical protein tyrosine phosphatase enzyme, dual-specificity phosphate 15 (DUSP15) is thought to be activated by myelin regulatory factor (MyRF) and to have a role in oligodendrocyte differentiation. Here, we assess whether Dusp15 is reduced in the hippocampus of mice with conditional knock-out of Myrf in oligodendrocyte precursor cells. Using quantitative polymerase chain reaction (qPCR) we found that Dusp15 expression was indeed lower in these mice. Alterations in myelin have been associated with Alzheimer's disease (AD), autism spectrum disorder (ASD) and attention deficit/hyperactivity disorder (ADHD). Symptoms of these disorders can include impairments of object recognition and attention. We, therefore tested the mice in the object recognition task (ORT) and 5-choice serial reaction time task (5CSRTT). However, we did not find behavioural impairments indicating that attentional abilities and object recognition are not impacted by reduced oligodendrogenesis and hippocampal Dusp15 expression. Gaining insight into the role of newly formed oligodendrocytes and Dusp15 expression is helpful for the development of well targeted treatments for myelin dysregulation.

Background: To investigate reproducibility of texture features and volumetric bone mineral density (vBMD) extracted from trabecular bone in the thoracolumbar spine in routine clinical multi-detector computed tomography (MDCT) data in a single scanner environment. Methods: Patients who underwent two routine clinical thoraco-abdominal MDCT exams at a single scanner with a time interval of 6 to 26 months (n=203, 131 males; time interval mean, 13 months; median, 12 months) were included in this observational study. Exclusion criteria were metabolic and hematological disorders, bone metastases, use of bone-active medications, and history of osteoporotic vertebral fractures (VFs) or prior diagnosis of osteoporosis. A convolutional neural network (CNN)-based framework was used for automated spine labeling and segmentation (T5-L5), asynchronous Hounsfield unit (HU)-to-BMD calibration, and correction for the intravenous contrast medium phase. Vertebral vBMD and six texture features [varianceglobal, entropy, short-run emphasis (SRE), long-run emphasis (LRE), run-length non-uniformity (RLN), and run percentage (RP)] were extracted for mid-(T5-T8) and lower thoracic (T9-T12), and lumbar vertebrae (L1-L5), respectively. Relative annual changes were calculated in texture features and vBMD for each vertebral level and sorted by sex, and changes were checked for statistical significance (P<0.05) using paired t-tests. Root mean square coefficient of variation (RMSCV) and root mean square error (RMSE) were calculated as measures of variability. Results: SRE, LRE, RLN, and RP exhibited substantial reproducibility with RMSCV-values below 2%, for both sexes and at all spine levels, while vBMD was less reproducible (RMSCV =11.9-16.2%). Entropy showed highest variability (RMSCV =4.34-7.69%) due to statistically significant increases [range, mean +/- standard deviation: (4.40 +/- 5.78)% to (8.36 +/- 8.66)%, P<0.001]. RMSCV of varianceglobal ranged from 1.60% to 3.03%. Conclusions: Opportunistic assessment of texture features in a single scanner environment using the presented CNN-based framework yields substantial reproducibility, outperforming vBMD reproducibility. Lowest scan-rescan variability was found for higher-order texture features. Further studies are warranted to determine, whether microarchitectural changes to the trabecular bone may be assessed through texture features.

ObjectiveEndovascular thrombectomy is a long-established therapy for acute basilar artery occlusion (aBAO). Unlike for anterior circulation stroke, cost-effectiveness of endovascular treatment has not been evaluated and is urgently needed to calculate expected health benefits and financial rewards. The aim of this study was therefore to simulate patient-level costs, analyze the economic potential of endovascular thrombectomy in patients with acute basilar artery occlusion (aBAO), and identify major determinants of cost-effectiveness. MethodsA Markov model was developed to compare outcome and cost parameters between patients treated by endovascular thrombectomy and patients treated by best medical care, based on four recent prospective clinical trials (ATTENTION, BAOCHE, BASICS, and BEST). Treatment outcomes were derived from the most recent literature. Uncertainty was addressed by deterministic and probabilistic sensitivity analyses. Willingness to pay per QALY thresholds were set at 1x gross domestic product per capita, as recommended by the World Health Organization. ResultsEndovascular treatment of acute aBAO stroke yielded an incremental gain of 1.71 quality-adjusted life-years per procedure with an incremental cost-effectiveness ratio of $7,596 per QALY. This was substantially lower than the Willingness to pay of $63,593 per QALY. Lifetime costs were most sensitive to costs of the endovascular procedure. ConclusionEndovascular treatment is cost-effective in patients with aBAO stroke.

ObjectivesOpportunistic quantitative computed tomography (oQCT) derived from non-dedicated routine CT has demonstrated high accuracy in diagnosing osteoporosis and predicting incident vertebral fractures (VFs). We aimed to investigate the cost-effectiveness of oQCT screening compared to dual-energy X-ray absorptiometry (DXA) as the standard of care for osteoporosis screening.MethodsThree screening strategies (“no osteoporosis screening”, “oQCT screening”, and “DXA screening”) after routine CT were simulated in a state-transition model for hypothetical cohorts of 1,000 patients (women and men aged 65 years) over a follow-up period of 5 years (base case). The primary outcomes were the cumulative costs and the quality-adjusted life years (QALYs) estimated from a U.S. health care perspective for the year 2022. Cost-effectiveness was assessed based on a willingness-to-pay (WTP) threshold of $70,249 per QALY. The secondary outcome was the number of prevented VFs. Deterministic and probabilistic sensitivity analyses were conducted to test the models’ robustness.ResultsCompared to DXA screening, oQCT screening increased QALYs in both sexes (additional 2.40 per 1,000 women and 1.44 per 1,000 men) and resulted in total costs of $3,199,016 and $950,359 vs. $3,262,934 and $933,077 for women and men, respectively. As a secondary outcome, oQCT screening prevented 2.6 and 2.0 additional VFs per 1,000 women and men, respectively. In the probabilistic sensitivity analysis, oQCT screening remained cost-effective in 88.3% (women) and 90.0% (men) of iterations.ConclusionoQCT screening is a cost-effective ancillary approach for osteoporosis screening and has the potential to prevent a substantial number of VFs if considered in daily clinical practice.

ObjectivesTo investigate vertebral osteoporotic fracture (VF) prediction by automatically extracted trabecular volumetric bone mineral density (vBMD) from routine CT, and to compare the model with fracture prevalence-based prediction models.MethodsThis single-center retrospective study included patients who underwent two thoraco-abdominal CT scans during clinical routine with an average inter-scan interval of 21.7 ± 13.1 months (range 5–52 months). Automatic spine segmentation and vBMD extraction was performed by a convolutional neural network framework (anduin.bonescreen.de). Mean vBMD was calculated for levels T5-8, T9-12, and L1-5. VFs were identified by an expert in spine imaging. Odds ratios (ORs) for prevalent and incident VFs were calculated for vBMD (per standard deviation decrease) at each level, for baseline VF prevalence (yes/no), and for baseline VF count (n) using logistic regression models, adjusted for age and sex. Models were compared using Akaike’s and Bayesian information criteria (AIC & BIC).Results420 patients (mean age, 63 years ± 9, 276 males) were included in this study. 40 (25 female) had prevalent and 24 (13 female) had incident VFs. Individuals with lower vBMD at any spine level had higher odds for VFs (L1-5, prevalent VF: OR,95%-CI,p: 2.2, 1.4–3.5,p=0.001; incident VF: 3.5, 1.8–6.9,p<0.001). In contrast, VF status (2.15, 0.72–6.43,p=0.170) and count (1.38, 0.89–2.12,p=0.147) performed worse in incident VF prediction. Information criteria revealed best fit for vBMD-based models (AIC vBMD=165.2; VF status=181.0; count=180.7).ConclusionsVF prediction based on automatically extracted vBMD from routine clinical MDCT outperforms prediction models based on VF status and count. These findings underline the importance of opportunistic quantitative osteoporosis screening in clinical routine MDCT data.

To determine the correlation between cervicothoracic and lumbar volumetric bone mineral density (vBMD) in an average cohort of adults and to identify specific diagnostic thresholds for the cervicothoracic spine on the individual subject level. In this HIPPA–compliant study, we retrospectively included 260 patients (59.7 ± 18.3 years, 105 women), who received a contrast-enhanced or non-contrast-enhanced CT scan. vBMD was extracted using an automated pipeline ( https://anduin.bonescreen.de ). The association of vBMD between each vertebra spanning C2–T12 and the averaged values at the lumbar spine (L1–L3) was analyzed before and after semiquantitative assessment of fracture status and degeneration, and respective vertebra-specific cut-off values for osteoporosis were calculated using linear regression. In both women and men, trabecular vBMD decreased with age in the cervical, thoracic, and lumbar regions. vBMD values of cervicothoracic vertebrae showed strong correlations with lumbar vertebrae (L1–L3), with a median Pearson value of r = 0.87 (range: rC2 = 0.76 to rT12 = 0.96). The correlation coefficients were significantly lower (p < 0.0001) without excluding fractured and degenerated vertebrae, median r = 0.82 (range: rC2 = 0.69 to rT12 = 0.93). Respective cut-off values for osteoporosis peaked at C4 (209.2 mg/ml) and decreased to 83.8 mg/ml at T12. Our data show a high correlation between clinically used mean L1–L3 values and vBMD values elsewhere in the spine, independent of age. The proposed cut-off values for the cervicothoracic spine therefore may allow the determination of low bone mass even in clinical cases where only parts of the spine are imaged.

OBJECTIVES:To determine the accuracy of an artificial neural network (ANN) for fully automated detection of the presence and phase of iodinated contrast agent in routine abdominal multidetector computed tomography (MDCT) scans and evaluate the effect of contrast correction for osteoporosis screening. METHODS:This HIPPA-compliant study retrospectively included 579 MDCT scans in 193 patients (62.4 ± 14.6 years, 48 women). Three different ANN models (2D DenseNet with random slice selection, 2D DenseNet with anatomy-guided slice selection, 3D DenseNet) were trained in 462 MDCT scans of 154 patients (threefold cross-validation), who underwent triphasic CT. All ANN models were tested in 117 unseen triphasic scans of 39 patients, as well as in a public MDCT dataset containing 311 patients. In the triphasic test scans, trabecular volumetric bone mineral density (BMD) was calculated using a fully automated pipeline. Root-mean-square errors (RMSE) of BMD measurements with and without correction for contrast application were calculated in comparison to nonenhanced (NE) scans. RESULTS:The 2D DenseNet with anatomy-guided slice selection outperformed the competing models and achieved an F1 score of 0.98 and an accuracy of 98.3% in the test set (public dataset: F1 score 0.93; accuracy 94.2%). Application of contrast agent resulted in significant BMD biases (all p < .001; portal-venous (PV): RMSE 18.7 mg/ml, mean difference 17.5 mg/ml; arterial (AR): RMSE 6.92 mg/ml, mean difference 5.68 mg/ml). After the fully automated correction, this bias was no longer significant (p > .05; PV: RMSE 9.45 mg/ml, mean difference 1.28 mg/ml; AR: RMSE 3.98 mg/ml, mean difference 0.94 mg/ml). CONCLUSION:Automatic detection of the contrast phase in multicenter CT data was achieved with high accuracy, minimizing the contrast-induced error in BMD measurements. KEY POINTS:• A 2D DenseNet with anatomy-guided slice selection achieved an F1 score of 0.98 and an accuracy of 98.3% in the test set. In a public dataset, an F1 score of 0.93 and an accuracy of 94.2% were obtained. • Automated adjustment for contrast injection improved the accuracy of lumbar bone mineral density measurements (RMSE 18.7 mg/ml vs. 9.45 mg/ml respectively, in the portal-venous phase). • An artificial neural network can reliably reveal the presence and phase of iodinated contrast agent in multidetector CT scans ( https://github.com/ferchonavarro/anatomy_guided_contrast_c ). This allows minimizing the contrast-induced error in opportunistic bone mineral density measurements.

ABSTRACT Opportunistic osteoporosis screening in nondedicated routine computed tomography (CT) is of increasing importance. The purpose of this study was to compare lumbar volumetric bone mineral density (vBMD) assessed by a convolutional neural network (CNN)-based framework in routine CT to vBMD from dedicated quantitative CT (QCT), and to evaluate the ability of vBMD and surrogate measurements of Hounsfield units (HU) to distinguish between patients with and without osteoporotic vertebral fractures (VFs). A total of 144 patients (median age: 70.7 years, 93 females) with clinical routine CT (eight different CT scanners, 120 kVp or 140 kVp, with and without intravenous contrast medium) and dedicated QCT acquired within ≤30 days were included. Vertebral measurements included (i) vBMD from the CNN-based approach including automated vertebral body labeling, segmentation, and correction of the contrast media phase for routine CT data (vBMD_OPP), (ii) vBMD from dedicated QCT (vBMD_QCT), and (iii) noncalibrated HU from vertebral bodies of routine CT data as previously proposed for immanent opportunistic osteoporosis screening based on CT attenuation. The intraclass correlation coefficient (ICC) for vBMD_QCT versus vBMD_OPP indicated better agreement (ICC = 0.913) than the ICC for vBMD_QCT versus noncalibrated HU (ICC = 0.704). Bland-Altman analysis showed data points from 137 patients (95.1%) within the limits of agreement (LOA) of −23.2 to 25.0 mg/cm3 for vBMD_QCT versus vBMD_OPP. Osteoporosis (vBMD <80 mg/cm3) was detected in 89 patients (vBMD_QCT) and 88 patients (vBMD_OPP), whereas no patient crossed the diagnostic thresholds from normal vBMD to osteoporosis or vice versa. In a subcohort of 88 patients (thoracolumbar spine covered by imaging for VF reading), 69 patients showed one or more prevalent VFs, and the performance for discrimination between patients with and without VFs was best for vBMD_OPP (area under the curve [AUC] = 0.862; 95% confidence interval [CI], 0.771–0.953). In conclusion, automated opportunistic osteoporosis screening in routine CT of various scanner setups is feasible and may demonstrate high diagnostic accuracy for prevalent VFs. © 2022 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).