Potential of GPT-4 for Detecting Errors in Radiology Reports: Implications for Reporting Accuracy.

Background Errors in radiology reports may occur because of resident-to-attending discrepancies, speech recognition inaccuracies, and large workload. Large language models, such as GPT-4 (ChatGPT; OpenAI), may assist in generating reports. Purpose To assess effectiveness of GPT-4 in identifying common errors in radiology reports, focusing on performance, time, and cost-efficiency. Materials and Methods In this retrospective study, 200 radiology reports (radiography and cross-sectional imaging [CT and MRI]) were compiled between June 2023 and December 2023 at one institution. There were 150 errors from five common error categories (omission, insertion, spelling, side confusion, and other) intentionally inserted into 100 of the reports and used as the reference standard. Six radiologists (two senior radiologists, two attending physicians, and two residents) and GPT-4 were tasked with detecting these errors. Overall error detection performance, error detection in the five error categories, and reading time were assessed using Wald χ2 tests and paired-sample t tests. Results GPT-4 (detection rate, 82.7%;124 of 150; 95% CI: 75.8, 87.9) matched the average detection performance of radiologists independent of their experience (senior radiologists, 89.3% [134 of 150; 95% CI: 83.4, 93.3]; attending physicians, 80.0% [120 of 150; 95% CI: 72.9, 85.6]; residents, 80.0% [120 of 150; 95% CI: 72.9, 85.6]; P value range, .522-.99). One senior radiologist outperformed GPT-4 (detection rate, 94.7%; 142 of 150; 95% CI: 89.8, 97.3; P = .006). GPT-4 required less processing time per radiology report than the fastest human reader in the study (mean reading time, 3.5 seconds ± 0.5 [SD] vs 25.1 seconds ± 20.1, respectively; P < .001; Cohen d = -1.08). The use of GPT-4 resulted in lower mean correction cost per report than the most cost-efficient radiologist ($0.03 ± 0.01 vs $0.42 ± 0.41; P < .001; Cohen d = -1.12). Conclusion The radiology report error detection rate of GPT-4 was comparable with that of radiologists, potentially reducing work hours and cost. © RSNA, 2024 See also the editorial by Forman in this issue.

Highly compressed SENSE accelerated relaxation-enhanced angiography without contrast and triggering (REACT) for fast non-contrast enhanced magnetic resonance angiography of the neck: Clinical evaluation in patients with acute ischemic stroke at 3 tesla.

BACKGROUND AND PURPOSE: Long acquisition times limit the feasibility of established non-contrast-enhanced MRA (non-CE-MRA) techniques. The purpose of this study was to evaluate a highly accelerated flow-independent sequence (Relaxation-Enhanced Angiography without Contrast and Triggering [REACT]) for imaging of the extracranial arteries in acute ischemic stroke (AIS). MATERIALS AND METHODS: Compressed SENSE (CS) accelerated (factor 7) 3D isotropic REACT (fixed scan time: 01:22 min, reconstructed voxel size 0.625 × 0.625 × 0.75 mm3) and CE-MRA (CS factor 6, scan time: 1:08 min, reconstructed voxel size 0.5 mm3) were acquired in 76 AIS patients (69.4 ±â€¯14.3 years, 33 females) at 3 Tesla. Two radiologists assessed scans for the presence of internal carotid artery (ICA) stenosis and stated their diagnostic confidence using a 5-point scale (5 = excellent). Vessel quality of cervical arteries as well as the impact of artifacts and image noise were scored on 5-point scales (5 = excellent/none). Apparent signal- and contrast-to-noise ratios (aSNR/aCNR) were measured for the common carotid artery (CCA) and ICA (C1-segment). RESULTS: REACT provided a sensitivity of 88.5% and specificity of 100% for clinically relevant (≥50%) ICA stenosis with substantial concordance to CE-MRA regarding stenosis grading (Cohen's kappa 0.778) and similar diagnostic confidence (REACT: mean 4.5 ±â€¯0.4 vs. CE-MRA: 4.5 ±â€¯0.6; P = 0.674). Presence of artifacts (3.6 ±â€¯0.5 vs. 3.5 ±â€¯0.7; P = 0.985) and vessel quality (all segments: 3.6 ±â€¯0.7 vs. 3.8 ±â€¯0.7; P = 0.004) were comparable between both techniques with REACT showing higher scores at the CCA (4.3 ±â€¯0.6 vs. 3.8 ±â€¯0.9; P < 0.001) and CE-MRA at V2- (3.3 ±â€¯0.5 vs. 3.9 ±â€¯0.8; P < 0.001) and V3-segments (3.3 ±â€¯0.5 vs. 4.0 ±â€¯0.8; P < 0.001). For all vessels, REACT showed a lower impact of image noise (3.8 ±â€¯0.6 vs. 3.6 ±â€¯0.7; P = 0.024) while yielding higher aSNR (52.5 ±â€¯15.1 vs. 37.9 ±â€¯12.5; P < 0.001) and aCNR (49.4 ±â€¯15.0 vs. 34.7 ±â€¯12.3; P < 0.001) for all vessels combined. CONCLUSIONS: In patients with acute ischemic stroke, highly accelerated REACT provides an accurate detection of ICA stenosis with vessel quality and scan time comparable to CE-MRA.

The worldwide COVID-19 pandemic caused a decline in sonographic examinations - is this a continuing trend?

PURPOSE: Due to the increasing number of COVID-19 infections since spring 2020 the patient care workflow underwent changes in Germany. To minimize face-to-face exposure and reduce infection risk, non-time-critical elective medical procedures were postponed. Since ultrasound examinations include non-time-critical elective examinations and often can be substituted by other imaging modalities not requiring direct patient contact, the number of examinations has declined significantly. The aim of this study is to quantify the baseline number of ultrasound examinations in the years before, during, and in the early post-pandemic period of the COVID-19 pandemic (since January 2015 to September 2023), and to measure the number of examinations at different German university hospitals. MATERIALS AND METHODS: The number of examinations was assessed based on a web-based database at all participating clinics at the indicated time points. RESULTS: N = 288 562 sonographic examinations from four sites were included in the present investigation. From January 2020 to June 2020, a significantly lower number of examinations of n = 591.21 vs. 698.43 (p = 0.01) per month and included center was performed. Also, excluding the initial pandemic period until June 2020, significantly fewer ultrasound examinations were performed compared to pre-pandemic years 648.1 vs. 698.4 (p < 0.05), per month and included center, while here differences between the individual centers were observed. In the late phase of the pandemic (n = 681.96) and in the post-pandemic phase (as defined by the WHO criteria from May 2023; n = 739.95), the number of sonographic examinations returned to pre-pandemic levels. CONCLUSION: The decline in the number of sonographic examinations caused by the COVID-19 pandemic was initially largely intentional and can be illustrated quantitatively. After an initial abrupt decline in sonographic examinations, the pre-pandemic levels could not be reached for a long time, which could be due to restructuring of patient care and follow-up treatment. In the post-pandemic phase, the pre-pandemic level has been achieved again. The reasons for a prolonged reduction in ultrasound examinations are discussed in this article. KEY POINTS: · During the pandemic, significantly fewer ultrasound examinations were performed in the included centers.. · The number of examinations could not be reach the pre-pandemic level for a long time, which could be due to restructuring of patient care and follow-up treatment.. · Identifying causes for sonographic exam reduction is crucial in pandemic preparedness to uphold healthcare quality and continuity for all patients.. · The prolonged decline in sonographic examinations during the pandemic does not represent a lasting trend, as evidenced by the return to pre-pandemic levels..

Standardization of a CT Protocol for Imaging Patients with Suspected COVID-19-A RACOON Project.

CT protocols that diagnose COVID-19 vary in regard to the associated radiation exposure and the desired image quality (IQ). This study aims to evaluate CT protocols of hospitals participating in the RACOON (Radiological Cooperative Network) project, consolidating CT protocols to provide recommendations and strategies for future pandemics. In this retrospective study, CT acquisitions of COVID-19 patients scanned between March 2020 and October 2020 (RACOON phase 1) were included, and all non-contrast protocols were evaluated. For this purpose, CT protocol parameters, IQ ratings, radiation exposure (CTDIvol), and central patient diameters were sampled. Eventually, the data from 14 sites and 534 CT acquisitions were analyzed. IQ was rated good for 81% of the evaluated examinations. Motion, beam-hardening artefacts, or image noise were reasons for a suboptimal IQ. The tube potential ranged between 80 and 140 kVp, with the majority between 100 and 120 kVp. CTDIvol was 3.7 ± 3.4 mGy. Most healthcare facilities included did not have a specific non-contrast CT protocol. Furthermore, CT protocols for chest imaging varied in their settings and radiation exposure. In future, it will be necessary to make recommendations regarding the required IQ and protocol parameters for the majority of CT scanners to enable comparable IQ as well as radiation exposure for different sites but identical diagnostic questions.

Chronic Hyponatremia and Brain Structure and Function Before and After Treatment.

RATIONALE & OBJECTIVE: Hyponatremia is the most common electrolyte disorder and is associated with significant morbidity and mortality. This study investigated neurocognitive impairment, brain volume, and alterations in magnetic resonance imaging (MRI)-based measures of cerebral function in patients before and after treatment for hyponatremia. STUDY DESIGN: Prospective cohort study. SETTING & PARTICIPANTS: Patients with presumed chronic hyponatremia without signs of hypo- or hypervolemia treated in the emergency department of a German tertiary-care hospital. EXPOSURE: Hyponatremia (ie, plasma sodium concentration [Na+]<125mmol/L) before and after treatment leading to [Na+]>130mmol/L. OUTCOMES: Standardized neuropsychological testing (Mini-Mental State Examination, DemTect, Trail Making Test A/B, Beck Depression Inventory, Timed Up and Go) and resting-state MRI were performed before and after treatment of hyponatremia to assess total brain and white and gray matter volumes as well as neuronal activity and its synchronization. ANALYTICAL APPROACH: Changes in outcomes after treatment for hyponatremia assessed using bootstrapped confidence intervals and Cohen d statistic. Associations between parameters were assessed using correlation analyses. RESULTS: During a 3.7-year period, 26 patients were enrolled. Complete data were available for 21 patients. Mean [Na+]s were 118.4mmol/L before treatment and 135.5mmol/L after treatment. Most measures of cognition improved significantly. Comparison of MRI studies showed a decrease in brain tissue volumes, neuronal activity, and synchronization across all gray matter after normalization of [Na+]. Volume effects were particularly prominent in the hippocampus. During hyponatremia, synchronization of neuronal activity was negatively correlated with [Na+] (r=-0.836; 95% CI, -0.979 to-0.446) and cognitive function (Mini-Mental State Examination, r=-0.523; 95% CI, -0.805 to-0.069; DemTect, r=-0.744; 95% CI, -0.951 to-0.385; and Trail Making Test A, r=0.692; 95% CI, 0.255-0.922). LIMITATIONS: Small sample size, insufficient quality of several MRI scans as a result of motion artifact. CONCLUSIONS: Resolution of hyponatremia was associated with improved cognition and reductions in brain volumes and neuronal activity. Impaired cognition during hyponatremia is closely linked to increased neuronal activity rather than to tissue volumes. Furthermore, the hippocampus appears to be particularly susceptible to hyponatremia, exhibiting pronounced changes in tissue volume. PLAIN-LANGUAGE SUMMARY: Hyponatremia is a common clinical problem, and patients often present with neurologic symptoms that are at least partially reversible. This study used neuropsychological testing and magnetic resonance imaging to examine patients during and after correction of hyponatremia. Treatment led to an improvement in patients' cognition as well as a decrease in their brain volumes, spontaneous neuronal activity, and synchronized neuronal activity between remote brain regions. Volume effects were particularly prominent in the hippocampus, an area of the brain that is important for the modulation of memory. During hyponatremia, patients with the lowest sodium concentrations had the highest levels of synchronized neuronal activity and the poorest cognitive test results.

An inducible Cd79b mutation confers ibrutinib sensitivity in mouse models of Myd88-driven diffuse large B-cell lymphoma.

ABSTRACT: Diffuse large B-cell lymphoma (DLBCL) is the most common aggressive lymphoma and constitutes a highly heterogenous disease. Recent comprehensive genomic profiling revealed the identity of numerous molecularly defined DLBCL subtypes, including a cluster which is characterized by recurrent aberrations in MYD88, CD79B, and BCL2, as well as various lesions promoting a block in plasma cell differentiation, including PRDM1, TBL1XR1, and SPIB. Here, we generated a series of autochthonous mouse models to mimic this DLBCL cluster and specifically focused on the impact of Cd79b mutations in this setting. We show that canonical Cd79b immunoreceptor tyrosine-based activation motif (ITAM) mutations do not accelerate Myd88- and BCL2-driven lymphomagenesis. Cd79b-mutant murine DLBCL were enriched for IgM surface expression, reminiscent of their human counterparts. Moreover, Cd79b-mutant lymphomas displayed a robust formation of cytoplasmic signaling complexes involving MYD88, CD79B, MALT1, and BTK. These complexes were disrupted upon pharmacological BTK inhibition. The BTK inhibitor-mediated disruption of these signaling complexes translated into a selective ibrutinib sensitivity of lymphomas harboring combined Cd79b and Myd88 mutations. Altogether, this in-depth cross-species comparison provides a framework for the development of molecularly targeted therapeutic intervention strategies in DLBCL.

Compressed SENSE accelerated 3D single-breath-hold late gadolinium enhancement cardiovascular magnetic resonance with isotropic resolution: clinical evaluation.

Aim: The purpose of this study was to investigate the clinical application of Compressed SENSE accelerated single-breath-hold LGE with 3D isotropic resolution compared to conventional LGE imaging acquired in multiple breath-holds. Material & Methods: This was a retrospective, single-center study including 105 examinations of 101 patients (48.2 ± 16.8 years, 47 females). All patients underwent conventional breath-hold and 3D single-breath-hold (0.96 × 0.96 × 1.1 mm3 reconstructed voxel size, Compressed SENSE factor 6.5) LGE sequences at 1.5 T in clinical routine for the evaluation of ischemic or non-ischemic cardiomyopathies. Two radiologists independently evaluated the left ventricle (LV) for the presence of hyperenhancing lesions in each sequence, including localization and transmural extent, while assessing their scar edge sharpness (SES). Confidence of LGE assessment, image quality (IQ), and artifacts were also rated. The impact of LV ejection fraction (LVEF), heart rate, body mass index (BMI), and gender as possible confounders on IQ, artifacts, and confidence of LGE assessment was evaluated employing ordinal logistic regression analysis. Results: Using 3D single-breath-hold LGE readers detected more hyperenhancing lesions compared to conventional breath-hold LGE (n = 246 vs. n = 216 of 1,785 analyzed segments, 13.8% vs. 12.1%; p < 0.0001), pronounced at subendocardial, midmyocardial, and subepicardial localizations and for 1%-50% of transmural extent. SES was rated superior in 3D single-breath-hold LGE (4.1 ± 0.8 vs. 3.3 ± 0.8; p < 0.001). 3D single-breath-hold LGE yielded more artifacts (3.8 ± 1.0 vs. 4.0 ± 3.8; p = 0.002) whereas IQ (4.1 ± 1.0 vs. 4.2 ± 0.9; p = 0.122) and confidence of LGE assessment (4.3 ± 0.9 vs. 4.3 ± 0.8; p = 0.374) were comparable between both techniques. Female gender negatively influenced artifacts in 3D single-breath-hold LGE (p = 0.0028) while increased heart rate led to decreased IQ in conventional breath-hold LGE (p = 0.0029). Conclusions: In clinical routine, Compressed SENSE accelerated 3D single-breath-hold LGE yields image quality and confidence of LGE assessment comparable to conventional breath-hold LGE while providing improved delineation of smaller LGE lesions with superior scar edge sharpness. Given the fast acquisition of 3D single-breath-hold LGE, the technique holds potential to drastically reduce the examination time of CMR.

Reconstruction of shoulder MRI using deep learning and compressed sensing: a validation study on healthy volunteers.

BACKGROUND: To investigate the potential of combining compressed sensing (CS) and deep learning (DL) for accelerated two-dimensional (2D) and three-dimensional (3D) magnetic resonance imaging (MRI) of the shoulder. METHODS: Twenty healthy volunteers were examined using at 3-T scanner with a fat-saturated, coronal, 2D proton density-weighted sequence with four acceleration levels (2.3, 4, 6, and 8) and a 3D sequence with three acceleration levels (8, 10, and 13), all accelerated with CS and reconstructed using the conventional algorithm and a new DL-based algorithm (CS-AI). Subjective image quality was evaluated by two blinded readers using 6 criteria on a 5-point Likert scale (overall impression, artifacts, and delineation of the subscapularis tendon, bone, acromioclavicular joint, and glenoid labrum). Objective image quality was measured by calculating signal-to-noise-ratio, contrast-to-noise-ratio, and a structural similarity index measure. All reconstructions were compared to the clinical standard (CS 2D acceleration factor 2.3; CS 3D acceleration factor 8). Additionally, subjective and objective image quality were compared between CS and CS-AI with the same acceleration levels. RESULTS: Both 2D and 3D sequences reconstructed with CS-AI achieved on average significantly better subjective and objective image quality compared to sequences reconstructed with CS with the same acceleration factor (p ≤ 0.011). Comparing CS-AI to the reference sequences showed that 4-fold acceleration for 2D sequences and 13-fold acceleration for 3D sequences without significant loss of quality (p ≥ 0.058). CONCLUSIONS: For MRI of the shoulder at 3 T, a DL-based algorithm allowed additional acceleration of acquisition times compared to the conventional approach. RELEVANCE STATEMENT: The combination of deep-learning and compressed sensing hold the potential for further scan time reduction in 2D and 3D imaging of the shoulder while providing overall better objective and subjective image quality compared to the conventional approach. TRIAL REGISTRATION: DRKS00024156. KEY POINTS: • Combination of compressed sensing and deep learning improved image quality and allows for significant acceleration of shoulder MRI. • Deep learning-based algorithm achieved better subjective and objective image quality than conventional compressed sensing. • For shoulder MRI at 3 T, 40% faster image acquisition for 2D sequences and 38% faster image acquisition for 3D sequences may be possible.

Somatic rearrangements causing oncogenic ectodomain deletions of FGFR1 in squamous cell lung cancer.

The discovery of frequent 8p11-p12 amplifications in squamous cell lung cancer (SQLC) has fueled hopes that FGFR1, located inside this amplicon, might be a therapeutic target. In a clinical trial, only 11% of patients with 8p11 amplification (detected by FISH) responded to FGFR kinase inhibitor treatment. To understand the mechanism of FGFR1 dependency, we performed deep genomic characterization of 52 SQLCs with 8p11-p12 amplification, including 10 tumors obtained from patients who had been treated with FGFR inhibitors. We discovered somatically altered variants of FGFR1 with deletion of exons 1-8 that resulted from intragenic tail-to-tail rearrangements. These ectodomain-deficient FGFR1 variants (ΔEC-FGFR1) were expressed in the affected tumors and were tumorigenic in both in vitro and in vivo models of lung cancer. Mechanistically, breakage-fusion-bridges were the source of 8p11-p12 amplification, resulting from frequent head-to-head and tail-to-tail rearrangements. Generally, tail-to-tail rearrangements within or in close proximity upstream of FGFR1 were associated with FGFR1 dependency. Thus, the genomic events shaping the architecture of the 8p11-p12 amplicon provide a mechanistic explanation for the emergence of FGFR1-driven SQLC. Specifically, we believe that FGFR1 ectodomain-deficient and FGFR1-centered amplifications caused by tail-to-tail rearrangements are a novel somatic genomic event that might be predictive of therapeutically relevant FGFR1 dependency.

Automated localization and segmentation of cervical lymph nodes on contrast-enhanced CT using a 3D foveal fully convolutional neural network.

BACKGROUND: In the management of cancer patients, determination of TNM status is essential for treatment decision-making and therefore closely linked to clinical outcome and survival. Here, we developed a tool for automatic three-dimensional (3D) localization and segmentation of cervical lymph nodes (LNs) on contrast-enhanced computed tomography (CECT) examinations. METHODS: In this IRB-approved retrospective single-center study, 187 CECT examinations of the head and neck region from patients with various primary diseases were collected from our local database, and 3656 LNs (19.5 ± 14.9 LNs/CECT, mean ± standard deviation) with a short-axis diameter (SAD) ≥ 5 mm were segmented manually by expert physicians. With these data, we trained an independent fully convolutional neural network based on 3D foveal patches. Testing was performed on 30 independent CECTs with 925 segmented LNs with an SAD ≥ 5 mm. RESULTS: In total, 4,581 LNs were segmented in 217 CECTs. The model achieved an average localization rate (LR), i.e., percentage of localized LNs/CECT, of 78.0% in the validation dataset. In the test dataset, average LR was 81.1% with a mean Dice coefficient of 0.71. For enlarged LNs with a SAD ≥ 10 mm, LR was 96.2%. In the test dataset, the false-positive rate was 2.4 LNs/CECT. CONCLUSIONS: Our trained AI model demonstrated a good overall performance in the consistent automatic localization and 3D segmentation of physiological and metastatic cervical LNs with a SAD ≥ 5 mm on CECTs. This could aid clinical localization and automatic 3D segmentation, which can benefit clinical care and radiomics research. RELEVANCE STATEMENT: Our AI model is a time-saving tool for 3D segmentation of cervical lymph nodes on contrast-enhanced CT scans and serves as a solid base for N staging in clinical practice and further radiomics research. KEY POINTS: • Determination of N status in TNM staging is essential for therapy planning in oncology. • Segmenting cervical lymph nodes manually is highly time-consuming in clinical practice. • Our model provides a robust, automated 3D segmentation of cervical lymph nodes. • It achieves a high accuracy for localization especially of enlarged lymph nodes. • These segmentations should assist clinical care and radiomics research.

Transfer-Learning Deep Radiomics and Hand-Crafted Radiomics for Classifying Lymph Nodes from Contrast-Enhanced Computed Tomography in Lung Cancer.

OBJECTIVES: Positron emission tomography (PET) is currently considered the non-invasive reference standard for lymph node (N-)staging in lung cancer. However, not all patients can undergo this diagnostic procedure due to high costs, limited availability, and additional radiation exposure. The purpose of this study was to predict the PET result from traditional contrast-enhanced computed tomography (CT) and to test different feature extraction strategies. METHODS: In this study, 100 lung cancer patients underwent a contrast-enhanced 18F-fluorodeoxyglucose (FDG) PET/CT scan between August 2012 and December 2019. We trained machine learning models to predict FDG uptake in the subsequent PET scan. Model inputs were composed of (i) traditional "hand-crafted" radiomics features from the segmented lymph nodes, (ii) deep features derived from a pretrained EfficientNet-CNN, and (iii) a hybrid approach combining (i) and (ii). RESULTS: In total, 2734 lymph nodes [555 (20.3%) PET-positive] from 100 patients [49% female; mean age 65, SD: 14] with lung cancer (60% adenocarcinoma, 21% plate epithelial carcinoma, 8% small-cell lung cancer) were included in this study. The area under the receiver operating characteristic curve (AUC) ranged from 0.79 to 0.87, and the scaled Brier score (SBS) ranged from 16 to 36%. The random forest model (iii) yielded the best results [AUC 0.871 (0.865-0.878), SBS 35.8 (34.2-37.2)] and had significantly higher model performance than both approaches alone (AUC: p < 0.001, z = 8.8 and z = 22.4; SBS: p < 0.001, z = 11.4 and z = 26.6, against (i) and (ii), respectively). CONCLUSION: Both traditional radiomics features and transfer-learning deep radiomics features provide relevant and complementary information for non-invasive N-staging in lung cancer.

Physiological iodine uptake of the spine's bone marrow in dual-energy computed tomography - using artificial intelligence to define reference values based on 678 CT examinations of 189 individuals.

Purpose: The bone marrow's iodine uptake in dual-energy CT (DECT) is elevated in malignant disease. We aimed to investigate the physiological range of bone marrow iodine uptake after intravenous contrast application, and examine its dependence on vBMD, iodine blood pool, patient age, and sex. Method: Retrospective analysis of oncological patients without evidence of metastatic disease. DECT examinations were performed on a spectral detector CT scanner in portal venous contrast phase. The thoracic and lumbar spine were segmented by a pre-trained neural network, obtaining volumetric iodine concentration data [mg/ml]. vBMD was assessed using a phantomless, CE-certified software [mg/cm3]. The iodine blood pool was estimated by ROI-based measurements in the great abdominal vessels. A multivariate regression model was fit with the dependent variable "median bone marrow iodine uptake". Standardized regression coefficients (ß) were calculated to assess the impact of each covariate. Results: 678 consecutive DECT exams of 189 individuals (93 female, age 61.4 ± 16.0 years) were evaluated. AI-based segmentation provided volumetric data of 97.9% of the included vertebrae (n=11,286). The 95th percentile of bone marrow iodine uptake, as a surrogate for the upper margin of the physiological distribution, ranged between 4.7-6.4 mg/ml. vBMD (p <0.001, mean ß=0.50) and portal vein iodine blood pool (p <0.001, mean ß=0.43) mediated the strongest impact. Based thereon, adjusted reference values were calculated. Conclusion: The bone marrow iodine uptake demonstrates a distinct profile depending on vBMD, iodine blood pool, patient age, and sex. This study is the first to provide the adjusted reference values.