Comparing 3D, 2.5D, and 2D Approaches to Brain Image Auto-Segmentation

Deep-learning methods for auto-segmenting brain images either segment one slice of the image (2D), five consecutive slices of the image (2.5D), or an entire volume of the image (3D). Whether one approach is superior for auto-segmenting brain images is not known. We compared these three approaches (3D, 2.5D, and 2D) across three auto-segmentation models (capsule networks, UNets, and nnUNets) to segment brain structures. We used 3430 brain MRIs, acquired in a multi-institutional study, to train and test our models. We used the following performance metrics: segmentation accuracy, performance with limited training data, required computational memory, and computational speed during training and deployment. The 3D, 2.5D, and 2D approaches respectively gave the highest to lowest Dice scores across all models. 3D models maintained higher Dice scores when the training set size was decreased from 3199 MRIs down to 60 MRIs. 3D models converged 20% to 40% faster during training and were 30% to 50% faster during deployment. However, 3D models require 20 times more computational memory compared to 2.5D or 2D models. This study showed that 3D models are more accurate, maintain better performance with limited training data, and are faster to train and deploy. However, 3D models require more computational memory compared to 2.5D or 2D models.

Machine Learning Models for Prediction of Posttreatment Recurrence in Early-Stage Hepatocellular Carcinoma Using Pretreatment Clinical and MRI Features: A Proof-of-Concept Study

BACKGROUND. Posttreatment recurrence is an unpredictable complication after liver transplant for hepatocellular carcinoma (HCC) that is associated with poor survival. Biomarkers are needed to estimate recurrence risk before organ allocation. OBJECTIVE. This proof-of-concept study evaluated the use of machine learning (ML) to predict recurrence from pretreatment laboratory, clinical, and MRI data in patients with early-stage HCC initially eligible for liver transplant. METHODS. This retrospective study included 120 patients (88 men, 32 women; median age, 60.0 years) with early-stage HCC diagnosed who were initially eligible for liver transplant and underwent treatment by transplant, resection, or thermal ablation between June 2005 and March 2018. Patients underwent pretreatment MRI and posttreatment imaging surveillance. Imaging features were extracted from postcontrast phases of pretreatment MRI examinations using a pretrained convolutional neural network. Pretreatment clinical characteristics (including laboratory data) and extracted imaging features were integrated to develop three ML models (clinical model, imaging model, combined model) for predicting recurrence within six time frames ranging from 1 through 6 years after treatment. Kaplan-Meier analysis with time to recurrence as the endpoint was used to assess the clinical relevance of model predictions. RESULTS. Tumor recurred in 44 of 120 (36.7%) patients during follow-up. The three models predicted recurrence with AUCs across the six time frames of 0.60-0.78 (clinical model), 0.71-0.85 (imaging model), and 0.62-0.86 (combined model). The mean AUC was higher for the imaging model than the clinical model (0.76 vs 0.68, respectively; p = .03), but the mean AUC was not significantly different between the clinical and combined models or between the imaging and combined models (p > .05). Kaplan-Meier curves were significantly different between patients predicted to be at low risk and those predicted to be at high risk by all three models for the 2-, 3-, 4-, 5-, and 6-year time frames (p < .05). CONCLUSION. The findings suggest that ML-based models can predict recurrence before therapy allocation in patients with early-stage HCC initially eligible for liver transplant. Adding MRI data as model input improved predictive performance over clinical parameters alone. The combined model did not surpass the imaging model's performance. CLINICAL IMPACT. ML-based models applied to currently underutilized imaging features may help design more reliable criteria for organ allocation and liver transplant eligibility.

DuDoSS: Deep-learning-based dual-domain sinogram synthesis from sparsely sampled projections of cardiac SPECT

PURPOSE

Myocardial perfusion imaging (MPI) using single-photon emission-computed tomography (SPECT) is widely applied for the diagnosis of cardiovascular diseases. In clinical practice, the long scanning procedures and acquisition time might induce patient anxiety and discomfort, motion artifacts, and misalignments between SPECT and computed tomography (CT). Reducing the number of projection angles provides a solution that results in a shorter scanning time. However, fewer projection angles might cause lower reconstruction accuracy, higher noise level, and reconstruction artifacts due to reduced angular sampling. We developed a deep-learning-based approach for high-quality SPECT image reconstruction using sparsely sampled projections.

METHODS

We proposed a novel deep-learning-based dual-domain sinogram synthesis (DuDoSS) method to recover full-view projections from sparsely sampled projections of cardiac SPECT. DuDoSS utilized the SPECT images predicted in the image domain as guidance to generate synthetic full-view projections in the sinogram domain. The synthetic projections were then reconstructed into non-attenuation-corrected and attenuation-corrected (AC) SPECT images for voxel-wise and segment-wise quantitative evaluations in terms of normalized mean square error (NMSE) and absolute percent error (APE). Previous deep-learning-based approaches, including direct sinogram generation (Direct Sino2Sino) and direct image prediction (Direct Img2Img), were tested in this study for comparison. The dataset used in this study included a total of 500 anonymized clinical stress-state MPI studies acquired on a GE NM/CT 850 scanner with 60 projection angles following the injection of 99m Tc-tetrofosmin.

RESULTS

Our proposed DuDoSS generated more consistent synthetic projections and SPECT images with the ground truth than other approaches. The average voxel-wise NMSE between the synthetic projections by DuDoSS and the ground-truth full-view projections was 2.08% ± 0.81%, as compared to 2.21% ± 0.86% (p < 0.001) by Direct Sino2Sino. The averaged voxel-wise NMSE between the AC SPECT images by DuDoSS and the ground-truth AC SPECT images was 1.63% ± 0.72%, as compared to 1.84% ± 0.79% (p < 0.001) by Direct Sino2Sino and 1.90% ± 0.66% (p < 0.001) by Direct Img2Img. The averaged segment-wise APE between the AC SPECT images by DuDoSS and the ground-truth AC SPECT images was 3.87% ± 3.23%, as compared to 3.95% ± 3.21% (p = 0.023) by Direct Img2Img and 4.46% ± 3.58% (p < 0.001) by Direct Sino2Sino.

CONCLUSIONS

Our proposed DuDoSS is feasible to generate accurate synthetic full-view projections from sparsely sampled projections for cardiac SPECT. The synthetic projections and reconstructed SPECT images generated from DuDoSS are more consistent with the ground-truth full-view projections and SPECT images than other approaches. DuDoSS can potentially enable fast data acquisition of cardiac SPECT.

Association of Baseline Frailty with Patient-Reported Outcomes in Systemic Lupus Erythematosus at 1 Year

The relationship of baseline frailty with subsequent patient-reported outcomes in systemic lupus erythematosus (SLE) remains unclear. We assessed these associations in a pilot prospective cohort study. Frailty based on the FRAIL scale and the Fried phenotype and patient-reported outcomes, namely Patient Reported Outcomes Measurement Information System computerized adaptive tests and Valued Life Activities disability, were measured at baseline and 1 year among women aged 18-70 years with SLE enrolled at a single center. Differences in Patient Reported Outcomes Measurement Information System computerized adaptive tests between frail and non-frail participants were evaluated using Wilcoxon rank sum tests, and the association of baseline frailty with self-report disability at 1 year was estimated using linear regression. Of 51 participants, 24% (FRAIL scale) and 16% (Fried phenotype) met criteria for frailty at baseline despite median age of 55.0 and 56.0 years, respectively. Women with (versus without) baseline frailty using either measure had worse 1-year Patient Reported Outcomes Measurement Information System computerized adaptive test scores across multiple domains and greater self-report disability. Baseline frailty was significantly associated with self-report disability at 1 year (FRAIL scale: parameter estimate 0.55, 95% confidence interval (CI) 0.21-0.89, p<0.01; Fried phenotype: parameter estimate 0.61, 95% CI 0.22-1.00, p<0.01), including only slight attenuation after adjustment for SLE cumulative organ damage (FRAIL scale: parameter estimate 0.45, 95% CI 0.09-0.81, p=0.02; Fried phenotype: parameter estimate 0.49, 95% CI 0.09-0.90, p=0.02). These preliminary findings support frailty as an independent risk factor for clinically relevant patient-reported outcomes, including disability onset, among women with SLE.

Conventional vs. Drug-Eluting Beads Transarterial Chemoembolization for Unresectable Hepatocellular Carcinoma-A Propensity Score Weighted Comparison of Efficacy and Safety

This study compared the efficacy and safety of conventional transarterial chemoembolization (cTACE) with drug-eluting beads (DEB)-TACE in patients with unresectable hepatocellular carcinoma (HCC). This retrospective analysis included 370 patients with HCC treated with cTACE (n = 248) or DEB-TACE (n = 122) (January 2000-July 2014). Overall survival (OS) was assessed using uni- and multivariate Cox proportional hazards models and Kaplan-Meier analysis. Additionally, baseline imaging was assessed, and clinical and laboratory toxicities were recorded. Propensity score weighting via a generalized boosted model was applied to account for group heterogeneity. There was no significant difference in OS between cTACE (20 months) and DEB-TACE patients (24.3 months, ratio 1.271, 95% confidence interval 0.876-1.69; p = 0.392). However, in patients with infiltrative disease, cTACE achieved longer OS (25.1 months) compared to DEB-TACE (9.2 months, ratio 0.366, 0.191-0.702; p = 0.003), whereas DEB-TACE proved more effective in nodular disease (39.4 months) than cTACE (18 months, ratio 0.458, 0.308-0681; p = 0.007). Adverse events occurred with similar frequency, except for abdominal pain, which was observed more frequently after DEB-TACE (101/116; 87.1%) than cTACE (119/157; 75.8%; p = 0.02). In conclusion, these findings suggest that tumor morphology and distribution should be used as parameters to inform decisions on the selection of embolic materials for TACE for a more personalized treatment planning in patients with unresectable HCC.

NIMG-20. INCORPORATION OF AI-BASED AUTOSEGMENTATION AND CLASSIFICATION INTO NEURORADIOLOGY WORKFLOW: PACS-BASED AI TO BUILD YALE GLIOMA DATASET

PURPOSE

Translation of AI algorithms into clinical practice is significantly limited by lack of large individual hospital-based datasets with expert annotations. Current methods for generation of annotated imaging data are significantly limited due to inefficient imaging data transfer, complicated annotation software, and time required for experts to generate ground truth information. We incorporated AI tools for auto-segmentation of gliomas into PACS that is used at our institution for reading clinical studies and developed a workflow for annotation of images and development of volumetric segmentations in neuroradiology clinical workflow. Material: 1990 patients from Yale Radiation Oncology Registry (2012-2019) were identified. Segmentations were performed using a UNETR algorithm trained on BRaTS 2021 and an internal dataset of manually segmented tumors. Segmentations were validated by a board-certified neuro-radiologist and natively embedded PyRadiomics in PACS was used for feature extraction.

RESULTS

In 7 Months (05/2021 - 08/2021, 03/2022 - 05/2022) segmentations and annotations were performed in 835 patients (322 female, 467 male, 46 unknown, mean age 53 yrs). Dataset includes 275 Grade 4 Gliomas (54 Grade 3, 100 Grade 2, 31 Grade 1, 375 unknown). Molecular subtypes include IDH (113 mutated, 498 wildtype, 2 Equivocal, 222 unknown), 1p/19q (87 deleted or co-deleted, 122 intact, 626 unknown), MGMT promotor (182 methylated, 95 partially methylated, 275 unmethylated, 283 unknown), EGFR (76 amplified, 177 not amplified, 582 unknown), ATRX (40 mutated, 157 retained, 638 unknown), Ki-67 (616 known, 219 unknown) and p53 (549 known, 286 unknown). Classification of gliomas between grade 3/4 and grade 1/2, yielded AUC of 0.85.

CONCLUSION

We developed a method for incorporation of volumetric segmentation, feature extraction, and classification that is easily incorporated into neuroradiology workflow. These tools allowed us to annotate over 100 gliomas per month, thus establishing a proof of concept for rapid development of annotated imaging database for AI applications.

NIMG-02. PACS-INTEGRATED AUTO-SEGMENTATION WORKFLOW FOR BRAIN METASTASES USING NNU-NET

PURPOSE

Monitoring metastatic disease to the brain is laborious and time-consuming, especially in the setting of multiple metastases and when performed manually. Response assessment in brain metastases based on maximal unidimensional diameter as per the RANO-BM guideline is commonly performed1, however, accurate volumetric lesion estimates can be crucial for clinical decision-making2 and enhance outcome prediction3. We propose a deep learning (DL)-based auto-segmentation approach with the potential for improvement of time-efficiency, reproducibility and robustness against inter-rater variability. Materials and

METHODS

We retrospectively retrieved 259 patients with a total number of 916 lesions from our institutional database from 2014 - 2021. Patients with prior history of local radiation therapy or surgery were excluded. Manually generated trainee segmentations were revised and adjusted by a board-certified radiologist and served as ground truth for evaluation of segmentation accuracy. Model performance was tested via dice-similarity-coefficient (DSC). Volumetric measurements were then obtained within our PACS-integrated workflow on Visage 7 (Visage Imaging, Inc., San Diego, CA) at the click of one button.

RESULTS

For model training and evaluation, a train-test split of 90:10 on patient-level was performed (n= 234:25 (Patients), n= 861:55 (Lesions). A DL-algorithm (nnUNet) was incrementally trained on 10 batches of 23 patients. The DSC of the U-Net gradually increased throughout the training process and heuristically reached a plateau of 0.85. The sensitivity of the algorithm was 83% with detection of 46 out of 55 lesions in the testing dataset. The lesions that were not detected by the algorithm were below 5 mm in size. The false positive rate was 8% (n=4/50).

CONCLUSION

Our study demonstrates the feasibility of PACS-based integration of automatized segmentation workflows of brain metastases. An incremental-training approach is recommended to adapt DL algorithms to specific hospital settings.

NIMG-07. APPLYING A GLIOMA-TRAINED DEEP LEARNING AUTO-SEGMENTATION TOOL ON BM PRE- AND POST-RADIOSURGERY

PURPOSE

Stereotactic radiosurgery (SRS) has become the mainstay to treat BM. Follow-up MRI provides important information on lesion treatment response and guides future therapy planning. Volumetric measurements of BM have shown promise over traditional uni- and two-dimensional measurements in more accurate and repeatable assessment. However, routine clinical use has yet to be achieved because the workflow is laborious. In previous work, we developed a PACS-integrated deep learning algorithm for automatic high- and low-grade glioma 3D segmentation. In this work, we applied this U-Net to segment BM on pre- and post-Gamma Knife (GK) MRI and evaluated the performance.

METHODS

10 pre- and post-GK studies were autosegmented in five randomly selected patients (melanoma n= 3, breast n= 2). The glioma trained algorithm segmented the “Whole Tumor” (tumor core+peritumoral edema on T2w-FLAIR) and “Tumor Core” (CE tumor core+necrosis on SPGR). The AI generated segmentation was then revised as needed by a board-certified neuroradiologist and the dice-similarity-coefficient (DSC) between the revised and automatic volumetric segmentations were calculated.

RESULTS

Four patients had multicentric (2-4 BM) lesions. The mean± SD DSC for Whole Tumor and Tumor Core were 0.92±0.06 and 0.46±0.30 for pretreatment, 0.84±0.09 and 0.41±0.25 for posttreatment BM, respectively. The tool detected lesions with a sensitivity of 45% (5/11) for pretreatment and 50% (3/6) for posttreatment lesions. Three pretreatment and all posttreatment lesions that were not detected by the autosegmentation tool showed a very faint hyperintense peritumoral edema in T2w-FLAIR.

CONCLUSION

Volumetric segmentation of edema on FLAIR using the glioma-trained segmentation algorithm on pre- and post-GK BM did not require major adjustment of segmentation if it detects the lesion. On the other hand, with low sensitivity of lesion detection and low DSC for enhancing component, dedicated training of the algorithm on annotated BM data will be needed.

Clinical implementation of artificial intelligence in neuroradiology with development of a novel workflow-efficient picture archiving and communication system-based automated brain tumor segmentation and radiomic feature extraction

Purpose

Personalized interpretation of medical images is critical for optimum patient care, but current tools available to physicians to perform quantitative analysis of patient’s medical images in real time are significantly limited. In this work, we describe a novel platform within PACS for volumetric analysis of images and thus development of large expert annotated datasets in parallel with radiologist performing the reading that are critically needed for development of clinically meaningful AI algorithms. Specifically, we implemented a deep learning-based algorithm for automated brain tumor segmentation and radiomics extraction, and embedded it into PACS to accelerate a supervised, end-to- end workflow for image annotation and radiomic feature extraction.

Materials and methods

An algorithm was trained to segment whole primary brain tumors on FLAIR images from multi-institutional glioma BraTS 2021 dataset. Algorithm was validated using internal dataset from Yale New Haven Health (YHHH) and compared (by Dice similarity coefficient [DSC]) to radiologist manual segmentation. A UNETR deep-learning was embedded into Visage 7 (Visage Imaging, Inc., San Diego, CA, United States) diagnostic workstation. The automatically segmented brain tumor was pliable for manual modification. PyRadiomics (Harvard Medical School, Boston, MA) was natively embedded into Visage 7 for feature extraction from the brain tumor segmentations.

Results

UNETR brain tumor segmentation took on average 4 s and the median DSC was 86%, which is similar to published literature but lower than the RSNA ASNR MICCAI BRATS challenge 2021. Finally, extraction of 106 radiomic features within PACS took on average 5.8 ± 0.01 s. The extracted radiomic features did not vary over time of extraction or whether they were extracted within PACS or outside of PACS. The ability to perform segmentation and feature extraction before radiologist opens the study was made available in the workflow. Opening the study in PACS, allows the radiologists to verify the segmentation and thus annotate the study.

Conclusion

Integration of image processing algorithms for tumor auto-segmentation and feature extraction into PACS allows curation of large datasets of annotated medical images and can accelerate translation of research into development of personalized medicine applications in the clinic. The ability to use familiar clinical tools to revise the AI segmentations and natively embedding the segmentation and radiomic feature extraction tools on the diagnostic workstation accelerates the process to generate ground-truth data.

Analysis of Tumor Burden as a Biomarker for Patient Survival with Neuroendocrine Tumor Liver Metastases Undergoing Intra-Arterial Therapies: A Single-Center Retrospective Analysis

PURPOSE

To assess the value of quantitative analysis of tumor burden on baseline MRI for prediction of survival in patients with neuroendocrine tumor liver metastases (NELM) undergoing intra-arterial therapies.

MATERIALS AND METHODS

This retrospective single-center analysis included 122 patients with NELM who received conventional (n = 74) or drug-eluting beads, (n = 20) chemoembolization and radioembolization (n = 28) from 2000 to 2014. Overall tumor diameter (1D) and area (2D) of up to 3 largest liver lesions were measured on baseline arterially contrast enhanced MR images. Three-dimensional quantitative analysis was performed using the qEASL tool (IntelliSpace Portal Version 8, Philips) to calculate enhancing tumor burden (the ratio between enhancing tumor volume and total liver volume). Based on Q-statistics, patients were stratified into low tumor burden (TB) or high TB.

RESULTS

The survival curves were significantly separated between low TB and high TB groups for 1D (p < 0.001), 2D (p < 0.001) and enhancing TB (p = 0.008) measurements, with, respectively, 2.7, 2.6 and 2.2 times longer median overall survival (MOS) in the low TB group (p < 0.001, p < 0.001 and p = 0.008). Multivariate analysis showed that 1D, 2D, and enhancing TB were independent prognostic factors for MOS, with respective hazard ratios of 0.4 (95%CI: 0.2-0.6, p < 0.001), 0.4 (95%CI: 0.3-0.7, p < 0.001) and 0.5 (95%CI: 0.3-0.8, p = 0.003).

CONCLUSION

The overall tumor diameter, overall tumor area, and enhancing tumor burden are strong prognostic factors of overall survival in patients with neuroendocrine tumor liver metastases undergoing intra-arterial therapies.

[Factors associated with death and attrition in HIV-infected children under initial antiretroviral therapy in Guangxi Zhuang Autonomous Region, 2004 - 2019]

Objective: To investigate death and attrition in HIV-infected children under initial antiretroviral therapy (ART) and associated factors in Guangxi Zhuang autonomous region. Methods: This retrospective cohort study was conducted in HIV-infected children under initial ART in Guangxi from 2004 to 2019, data from ART information system of National comprehensive AIDS prevention and treatment information system. Cox proportional hazards models were used to assess factors associated with the death and attrition. Results: In 943 HIV-infected children, the overall mortality and attrition rates were 1.00/100 person-years and 0.77/100 person-years, respectively. The mortality and attrition rates within the first year of ART were 3.90/100 person-years and 1.67/100 person-years, respectively. The cumulative survival rate during the first, second, fifth and tenth year after ART was 96.14%, 95.80%, 93.68% and 91.54%, respectively. Multivariate Cox proportional hazards models results showed that being female (aHR=2.00, 95%CI: 1.17-3.40), CD4⁺T lymphocytes (CD4) counts before ART <200 cells/μl (aHR=2.79, 95%CI: 1.54-5.06), weight-for-age Z score before ART <-2 (aHR=2.38, 95%CI: 1.32-4.26), hemoglobin before ART <80 g/L (aHR=2.47, 95%CI: 1.24-4.92), initial ART with LPV/r (aHR=5.05, 95%CI: 1.15-22.12) were significantly associated with death; being female (aHR=2.23, 95%CI: 1.22-4.07) and initial ART with LPV/r (aHR=2.02, 95%CI: 1.07-3.79) were significantly associated with attrition. Conclusions: The effect of ART in HIV-infected children in Guangxi was better, but the mortality and attrition rates were high within the first year of treatment. It is necessary to strengthen the training in medical staff and health education in HIV-infected children and their parents in order to improve the treatment effect.

COVID-19 Pandemic Impact on US Childhood Caries and Potential Mitigation

Non-Hispanic Black (NHB) and Hispanic and low-income US children have a higher prevalence of untreated caries than their higher-income and non-Hispanic White (NHW) counterparts. Due to the COVID-19 pandemic, many dental offices and school sealant programs closed beginning March 2020. We examine the effect of reduced access to restorative care and sealants on the oral health of children from low-income households overall and by race/ethnicity and how increased sealant delivery in September 2022 could mitigate these effects. We used Markov chain Monte Carlo simulation to model COVID-19's impact on first permanent molar (1M) caries incidence and loss in quality of life (disability-adjusted life years [DALYs]) due to time lived with 1M untreated caries. Our model followed a cohort of children aged 7 y in March 2020 until February 2024. Model inputs were primarily obtained from published studies and nationally representative data. Excess DALYs per 1,000 children attributable to reduced access to care during the pandemic were 1.48 overall and greater for Hispanic (2.07) and NHB (1.75) children than for NHW children (0.94). Excess incidence of 1M caries over 4 y was 2.28 percentage points overall and greater for Hispanic (2.63) and NHB (2.40) children than for NHW (1.96) children. Delivering sealants to 50% of eligible 1Ms in September 2022 would not completely mitigate COVID-19's health access impact: overall excess DALYs would decrease to 1.05, and absolute disparities in excess DALYs between NHW children and Hispanic and NHB children would remain but decrease by 0.38 and 0.33, respectively. Sealing 40% of eligible 1Ms, however, would bring overall 4-y caries incidence down to pre-COVID-19 levels and eliminate the differential effect of the pandemic on children from minority groups. The pandemic's negative impact on the oral health of children from low-income households and increased disparities could be partially mitigated with increased sealant delivery.

Holter electrocardiogram research trends and hotspots: bibliometrics and visual analysis

OBJECTIVE

With the help of metrology, we can identify research hotspots and development trends in dynamic electrocardiography, and thereby provide corresponding reference material to aid further theoretical research.

MATERIALS AND METHODS

All research data derived from the core collection of Web of Science, and all searches were completed on the same day (February 6, 2022). The obtained data were stored in plain text format and imported into CiteSpace for subsequent analysis. Citation analysis and visualization technology were used to draw a visual map of the research elements, using factors such as annual literature volume, country, journal, author, abstract, keywords, and citation.

RESULTS

After screening, 2,937 papers were obtained. Research on ambulatory electrocardiography is increasing worldwide every year. Using research hotspots, keyword-clustering time-zone maps, and high-frequency emerging words, the research in this field was roughly divided into two stages, with 2017 as the divider. The first stage primarily focuses on areas such as atrial fibrillation, stroke, autonomic nerve function, catheter ablation, and T-wave alternation. The second stage saw the focus shift to wearable devices, sudden cardiac death, obstructive sleep apnea, feature extraction, cryptogenic stroke, and similar topics.

CONCLUSIONS

With the development of various wearable technologies, the daily monitoring of healthy people engaged in sporting activities and the development of innovative analysis algorithms providing more accurate data may represent the hotspots and direction of future research.

PET Image Denoising using a Deep-Learning Method for Extremely Obese Patients

The image quality in clinical PET scan can be severely degraded due to high noise levels in extremely obese patients. Our work aimed to reduce the noise in clinical PET images of extremely obese subjects to the noise level of lean subject images, to ensure consistent imaging quality. The noise level was measured by normalized standard deviation (NSTD) derived from a liver region of interest. A deep learning-based noise reduction method with a fully 3D patch-based U-Net was used. Two U-Nets, U-Nets A and B, were trained on datasets with 40% and 10% count levels derived from 100 lean subjects, respectively. The clinical PET images of 10 extremely obese subjects were denoised using the two U-Nets. The results showed the noise levels of the images with 40% counts of lean subjects were consistent with those of the extremely obese subjects. U-Net A effectively reduced the noise in the images of the extremely obese patients while preserving the fine structures. The liver NSTD improved from 0.13±0.04 to 0.08±0.03 after noise reduction (p = 0.01). After denoising, the image noise level of extremely obese subjects was similar to that of lean subjects, in terms of liver NSTD (0.08±0.03 vs. 0.08±0.02, p = 0.74). In contrast, U-Net B over-smoothed the images of extremely obese patients, resulting in blurred fine structures. In a pilot reader study comparing extremely obese patients without and with U-Net A, the difference was not significant. In conclusion, the U-Net trained by datasets from lean subjects with matched count level can provide promising denoising performance for extremely obese subjects while maintaining image resolution, though further clinical evaluation is needed.

Response assessment methods for patients with hepatic metastasis from rare tumor primaries undergoing transarterial chemoembolization

PURPOSE

This study assessed the response to conventional transarterial chemoembolization (cTACE) in patients with liver metastases from rare tumor primaries using one-dimensional (1D) and three-dimensional (3D) quantitative response assessment methods, and investigate the relationship of lipiodol deposition in predicting response.

MATERIALS AND METHODS

This retrospective bicentric study included 16 patients with hepatic metastases from rare tumors treated with cTACE between 2002 and 2017. Multi-phasic MR imaging obtained before and after cTACE was used for assessment of response. Response evaluation criteria in solid tumors (RECIST) and modified-RECIST (mRECIST) were utilized for 1D response assessment, and volumetric RECIST (vRECIST) and enhancement-based quantitative European Association for Study of the Liver EASL (qEASL) were used for 3D response assessment. The same day post-cTACE CT scan was analyzed to quantify intratumoral lipiodol deposition (%).

RESULTS

The mean and standard deviation (SD) of diameter of treated lesions per targeted area was 7.5 ± 5.4 cm, and the mean and SD of number of metastases in each targeted area was 4.2 ± 4.6. cTACE was technically successful in all patients, without major complications. While RECIST and vRECIST methods did not allocate patients with partial response, mRECIST and qEASL identified patients with partial response. Intratumoral lipiodol deposition significantly predicted treatment response according qEASL (R² = 0.470, p < 0.01), while no association was shown between lipiodol deposition within treated tumor area and RECIST or mRECIST (p > 0.212).

CONCLUSION

3D quantitative volumetric response analysis can be used for stratification of response to cTACE in patients with hepatic metastases originating from rare primary tumors. Lipiodol deposition could potentially be used as an early surrogate to predict response to cTACE.

Dual-energy x-ray absorptiometry scanner mismatch in follow-up bone mineral density testing

Scanner mismatch occurs frequently with follow-up dual-energy x-ray absorptiometry (DXA) scans. Nearly one-in-five follow-up DXA scans were conducted on non-cross-calibrated scanners (scanner mismatch) and more than a quarter of patients who had a follow-up DXA scan had experienced scanner mismatch.

INTRODUCTION

Detecting significant changes in bone mineral density (BMD) with dual-energy x-ray absorptiometry (DXA) scanners relies on the least significant change (LSC). Results from two different DXA scanners can only be compared, albeit with decreased sensitivity for change, if the LSC between the two scanners has been directly determined through cross-calibration. Performing follow-up DXA scans on non-cross-calibrated scanners (scanner mismatch) has safety and economic implications. This study aims to determine the proportion of scanner mismatch occurring at a population level.

METHODS

All patients who completed at least two DXA scans between 1 April 2009 and 31 December 2018 in the province of Alberta, Canada, were identified using population-based health services databases. Scanner mismatch was defined as a follow-up DXA scan completed on a DXA scanner that differed from and was not cross-calibrated to the previous DXA scanner. Multivariate logistic regression models were used to assess predictive factors that may contribute to scanner mismatch.

RESULTS

A total of 264,866 patients with 470,641 follow-up DXA scans were identified. Scanner mismatch occurred in 18.9% of follow-up DXA scans; 28.7% of patients experienced at least one scanner mismatch. Longer duration between scans (OR 1.25, 95% CI 1.24-1.26) and major osteoporotic fracture history before index scan (OR 1.06, 95% CI 1.03-1.08) increased risk of scanner mismatch. Osteoporosis medication use before index scan (OR 0.89; 95% CI 0.88-0.91), recency of follow-up scans (OR 0.98, 95% CI 0.73-0.98), female sex (OR 0.97, 95% CI 0.94-1.00), and age at last scan (OR 0.99, 95% CI 0.99-1.00) were associated with lower risk of scanner mismatch.

CONCLUSION

Scanner mismatch is a common problem, occurring in one-in-five follow-up DXA scans and affecting more than a quarter of patients. Interventions to reduce this large proportion of scanner mismatch are necessary.

Adopting a systems view of disrupting crisis-driven food insecurity

Objectives

During the COVID crisis, the incidence of food insecurity worsened around the globe. We were reminded that: food insecurity existed before COVID, worsened during this crisis, and will unfortunately be a persistent phenomenon in the post-COVID world. It is evident that to counter this public health threat, systematic changes will need to happen. In this short communication, we introduce the notion of a systems-oriented framework that can guide appropriate actions for us to disrupt future food insecurity crises.

Study design

This short communication identifies preliminary observations based on relevant past studies that documented the impact of COVID-19 on food insecurity, and the researchers’ conceptualization of a framework on how we may address future crisis-driven food insecurity challenges.

Methods

Systems-oriented framework was conceptualized based on preliminary observations in studies that investigated food insecurity during the COVID-19 pandemic.

Results

This short communication explores the notion of a systems-oriented framework as a guide to future action to prevent crisis-driven food insecurity.

Conclusions

The systems-oriented framework emphasizes the importance of action across macro, meso, and micro levels, and synchronization to maximize synergies.