RPS 1210 - When MR substitutes CT and CT replaces MR

Author Block: M. L. Schnitzer, F. Herr, N. Hohmann, C. A. Dascalescu, B. Holzapfel, J. Ricke, N. Hesse, J. Arnholdt, P. Reidler; Munich/DE
Purpose: To evaluate a novel CT-like MRI sequence for ARCO classification, reader confidence, and image quality, and compare its diagnostic performance with standard MRI and radiography.
Methods or Background: In this retrospective study, 21 patients with suspected femoral head necrosis underwent MRI (T1, T2, CT-like) and radiography. Two experienced readers independently assessed ARCO stage (I–IV), reader confidence, and image quality using a binary scale (1 = high/good, 2 = low/poor). Mean scores were calculated. Interrater reliability was assessed using intraclass correlation coefficients (ICC); intermodality agreement with weighted Cohen’s Kappa. Paired comparisons used Wilcoxon signed-rank tests.
Results or Findings: CT-like MRI yielded higher ARCO scores (2.23 ± 0.78) than radiography (1.75 ± 1.01, p = 0.001), T1 (2.08 ± 0.73, p = 0.010), and T2 (2.08 ± 0.69, p = 0.023). Reader confidence was superior with CT-like MRI (1.11 ± 0.25) compared to T1 (1.53 ± 0.47), T2 (1.50 ± 0.46), and radiography (1.43 ± 0.37), all p < 0.01. Image quality ratings were highest for CT-like (1.11 ± 0.30), outperforming T1 (1.86 ± 0.51) and T2 (1.75 ± 0.51), p < 0.001. Agreement in ARCO staging was excellent with T1 (κ = 0.904) and T2 (κ = 0.869), and substantial with radiography (κ = 0.709). Interrater reliability was high for ARCO staging (ICC 0.82–0.93), lower for confidence (ICC < 0.40), and variable for image quality, with CT-like showing the best agreement (ICC = 0.84).
Conclusion: This CT-like MRI sequence improves diagnostic confidence and staging accuracy while avoiding ionizing radiation and reducing the need for additional imaging. It supports a more sustainable approach by minimizing exposure and streamlining diagnostics.
Limitations: Single-center design and small sample size may limit generalizability. The sequence requires specialized protocols not yet widely available.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the local ethics committee of (clinical trial number: DRKS00030822, date of registration: 12 January 2023).

Author Block: R. F. Valenzuela, E. Duran-Sierra, M. Antony, J. Espinoza, C. M. Costelloe, W. Murphy, D. Wells, B. Amini, S. Lo; Houston, TX/US
Purpose: Zero-Echo-Time (ZTE) and Ultrashort-Echo-Time (UTE) sequences are designed to characterize short-T2 materials. By utilizing ray-sum-rendering and inverse-logarithm-rescaling, these sequences can enhance bone contrast, resembling CT. ZTE publications often rely on perception analysis to validate its use as a CT-succedaneum. We aim to quantitatively compare Ozteo (GE-ZTE) MRI and CT using radiomic analysis, first on normal bone and subsequently on sclerotic and lucent lesions.
Methods or Background: Starting July 2025, Ozteo has become part of our institutional routine MSK-pelvic-MRI across nine GE magnets (two-3T and seven-1.5T). An ongoing prospective collection included 26 patients with bone-metastases screening scanned with ZTE+CT within six months. Manual-volume-of-interest (VOI) segmentations were created from healthy right hips, using the left hip in the presence of right hip pathology or hardware. An in-house Python-based sequence-neutral software extracted 107 radiomic features, comprising 14 shape, 18 first-order, and 75 texture features. Statistical analysis compared ZTE MRI-and CT-derived radiomic features using two-tailed non-parametric Wilcoxon signed-rank paired tests.
Results or Findings: Gender: 42% male/58% female. Age-average: 50 years (range: 24-76). 22 right and four left hip VOI were segmented. The average time between ZTE-MRI and CT was two months (range: 1-day/6-months). The ZTE-MRI vs. CT statistical analysis revealed that 16-out-of-107 (15%) radiomic features were significantly similar (P>=0.05), including: glszm_largeareaemphasis (P=0.0796), glszm_zonevariance (P=0.0796), and firstorder_maximum (P=0.0559). Maximum differences (9.3%, P=2.98x10-8) included firstorder_robustmeanabsolutedeviation, firstorder_interquartilerange, firstorder_meanabsolutedeviation, firstorder_variance, glszm_graylevelvariance and firstorder_range.
Conclusion: Detached from subjective-perceptual evaluations, radiomics analysis from CT versus ZTE-MRI provided objective-quantitative evidence of maximal-similarities (15% of radiomic features) and maximal-differences (9.3%), outlining the true range of features where ZTE may be considered comparable with conventional CT. Based on preliminary data, we hypothesize that the normal bone feature range may differ from sclerotic and lytic bone lesions.
Limitations: Small sample size.
Funding for this study: The John S. Dunn, Sr. Distinguished Chair in Diagnostic imaging and M.R Evelyn Hudson Foundation Endowed Professorship.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: F. Abel, R. Sutter, F. Schmaranzer; Zürich/CH
Purpose: The acetabular cup version in patients with total hip arthroplasty (THA) is a key parameter influencing hip stability and functional outcomes. Although CT is the reference standard for assessing cup orientation, MRI has emerged as the first-line cross-sectional imaging modality for evaluating postoperative complications. This study compared MRI- and CT-based measurements of acetabular cup version in terms of agreement, reproducibility, and reliability.
Methods or Background: Patients who underwent THA between 2015 and 2025 and received both postoperative CT and MRI were retrospectively reviewed. MRI was performed using optimized metal-artifact reduction sequences. Acetabular cup version was measured on axial high-bandwidth T1-weighted turbo-spin echo sequences and corresponding CT scans by two musculoskeletal radiologists. Inter- and intra-reader, as well as inter-modality agreement, were assessed using intraclass correlation coefficients (ICC). Bland-Altman plots evaluated systematic bias.
Results or Findings: Thirty patients were included (median age, 69.5 years; 14 women). Mean acetabular version ranged from 31.20°–31.22° (standard deviation [SD], 9.40–10.30) and from 30.05°–30.21° (SD, 8.90–9.32) for CT and MRI for both readers, respectively.
Inter-reader reproducibility was excellent for CT (ICC, 0.96) and good for MRI (ICC, 0.76). Intra-reader reliability was excellent for both modalities (ICC, CT: 0.99; MRI: 0.94). Bland-Altman analysis showed no statistically significant differences between CT and MRI measurements with a slight positive bias for MRI (mean pooled difference, 1.05°; p = 0.06; limits of agreement, –4.66° to 6.76°) and good to excellent inter-modality agreement for both readers (ICC, 0.86–0.90).
Conclusion: MRI provides reliable measurements of acetabular cup version in THA, demonstrating good-to-excellent agreement with CT. MRI can be used for estimating cup version within clinically acceptable limits.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the local ethics committee (Cantonal Ethics Committee Zurich, approval number: 2023-00969).

Author Block: S. Liu; Beijing/CN
Purpose: Multi-material decomposition (MMD), a key application of DECT, has shown potential in musculoskeletal research. This study aimed to compare the diagnostic performance of DECT-based MMD with standard CT and MRI in detecting rotator cuff disease.
Methods or Background: This study evaluated patients diagnosed with rotator cuff disease who underwent third-generation dual-source DECT and 3.0T MRI within a 2-week interval between December 2023 and November 2024. Shoulder arthroscopy served as the reference standard. Six readers independently assessed rotator cuff tears and determine the degree of supraspinatus tendon diseases using standard CT, DECT-based MMD and MRI. Area under the curve (AUC), sensitivity, specificity, positive/negative predictive values and accuracy were calculated for the diagnosis of rotator cuff disease. Friedman test was used to analyze the radiologists’ diagnostic confidence across the three image types.
Results or Findings: In total of 103 patients (mean age: 50.0 ± 15.6 years) underwent shoulder arthroscopy. MMD demonstrated a higher average AUC for diagnosing rotator cuff tears (88% vs. 65%, P < 0.001) and supraspinatus tendon disease (86% vs. 63%, P < 0.001) compared to standard CT. Its diagnostic performance for supraspinatus tendon disease (91% vs. 90%, P = 0.35) and full-thickness tears (95% vs. 93%, P = 0.11) was comparable to that of MRI.
Conclusion: DECT-based MMD demonstrated superior diagnostic performance and reliability for detecting rotator cuff diseases compared to standard CT, with accuracy comparable to that of MRI in detecting supraspinatus tendon tears. DECT-based MMD offers a promising approach for the opportunistic detection of rotator cuff diseases.
Limitations: First, as a retrospective single-center study, the generalizability of our findings are limited. Second, its ability to identify partial-thickness tears and subtle tendon degeneration—conditions more frequently encountered in clinical practice—remains limited.
Funding for this study: This work was supported by National Key Research and Development Program of China (No.2023YFC2410703)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This retrospective study was approved by the Institutional Review Board of Peking University Third Hospital (Approval Number: M2024188)

Author Block: C. Beltramini¹, F. Piacentino¹, F. Fontana¹, A. Coppola¹, D. Mesiano¹, C. Recaldini¹, A. M. Ierardi², G. Carrafiello², M. Venturini¹; ¹Varese/IT, ²Milan/IT
Purpose: Although MRI with fat-suppression sequences is the gold standard for diagnosis of bone marrow edema (BME), Dual-Layer Spectral CT (DL-SCT) with electron density (ED) provides a viable alternative, particularly in situations where an MRI is not accessible. Using MRI as the reference standard, this study analyzed how DL-SCT with ED reconstructions may be a valid alternative in the detection of BME.
Methods or Background: This retrospective study included 28 patients with a suspected diagnosis of BME via MRI conducted between March and September 2024. Patients underwent DL-SCT using ED reconstructions obtained through IntelliSpace software v. 12.1. Images were evaluated by two experienced radiologists and one young radiologist in a blinded way, giving a grade from 0 to 3 to classify BME (0 absence; 1 mild; 2 moderate; 3 severe). To reduce the recall bias effect, the order of image evaluations was set differently for each reader. p-Values were considered significant when <0.05. Fleiss' Kappa was used to assess inter-rater reliability: agreement was considered poor for k < 0; slight for k 0.01-0.20; fair for 0.21-0.40; moderate for 0.41-0.60; substantial for 0.61-0.80; and almost perfect for 0.81-1.00.
Results or Findings: All the readers detected the presence or absence of BME using DL-SCT. Inter-rater reliability for grade 0 resulted in 1 (p-value < 0.001); for grade 1: 0.21 (p-value < 0.001); for grade 2: 0.197 (p-value < 0.001); and for grade 3: 0.515 (p-value < 0.001).
Conclusion: ED reconstructions allowed the identification of BME presence or absence in all analyzed cases, thus suggesting DL-SCT as a potentially effective method for its detection.
Limitations: Heterogeneous sample that doesn't consider metabolic disorders different from osteoporosis or malignancy.
Funding for this study: None
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: N. Castro, J. Catala March; Barcelona/ES
Purpose: Dual-energy CBCT (DE-CBCT) represents a significant technological evolution in musculoskeletal imaging, offering superior tissue differentiation compared to single-energy CBCT (SE-CBCT) through material decomposition. While MRI remains the gold standard for soft tissue evaluation, DE-CBCT emerges as a promising alternative when MRI is contraindicated or unavailable.
This prospective study evaluated the diagnostic performance of DE-CBCT against SE-CBCT for bone structures and MRI for soft tissues, focusing on its clinical application in trauma and joint diseases.
Methods or Background: Thirty participants (45.57±17.93 years) with suspected or confirmed pathology and MRI inclusion criteria were enrolled. Each participant underwent DE-CBCT (80kVp; 140kVp), SE-CBCT (120kVp), and MRI (proton density and T1-weighted).
Two blinded, MSK-experienced radiologists independently assessed bone detail, soft tissue contrast, and diagnostic confidence using a 5-point Likert scale.
Results or Findings: Inter-reader agreement was higher for DE-CBCT than SE-CBCT (k=0.82 vs 0.71). For soft tissues, MRI showed superior agreement over DE-CBCT (k=0.94 vs 0.76).
No significant differences in bone diagnostic accuracy emerged between DE-CBCT and SE-CBCT, although DE-CBCT received higher qualitative scores. Differences with MRI were notable for soft tissues (p<0.05), confirming its superior visualisation. However, DE-CBCT affirmed accurate visualisation of articular cartilage and tendons. Although not statistically significant, DE-CBCT provided greater confidence in complex fractures, such as intra-articular wrist and ankle injuries, while reducing the need for additional MRI and speeding emergency decision-making.
Conclusion: DE-CBCT is an effective tool for recent trauma, especially when MRI is unavailable. It accurately detects fractures and soft tissue changes, such as bone oedema and joint effusions, offering a comprehensive post-traumatic assessment. Clinically, DE-CBCT should be integrated into emergency departments for rapid triage of complex trauma and prioritised over SE-CBCT for bone oedema or cartilage evaluation.
Limitations: Increment of radiation exposure compared to SE-CBCT
Funding for this study: None
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: E. Vilkelyte, N. F. Kämmerling, E. Tesselaar; Linköping/SE
Purpose: To evaluate the diagnostic utility of bone marrow edema (BME) maps generated using dual source photon-counting CT (PCCT), compared with ultra-high-resolution (UHR) standard reconstructions, in patients with suspected fractures in different .
Methods or Background: This study included 33 patients (37 examinations; mean age 59.5 ± 22.7 years, range 13–88) with suspected fractures at various anatomical sites including spine, hip, and wrist, examined on a clinical PCCT system using a dual-source acquisition mode. Two radiologists with 9, and 10 years of experience, and a second-year radiology resident, independently reviewed anonymised datasets in two sessions: first, UHR images only; second, the same cases (in randomized order) supplemented with BME maps, and assessed fracture presence and diagnostic confidence (5-point scale). Image quality and evaluability of BME maps, and inter-rater agreement were analysed.
Results or Findings: Agreement on fracture detection improved from moderate without BME maps (κ=0.514, p<0.001) to substantial when BME maps were available (κ=0.711, p<0.001). Overall diagnostic confidence increased significantly with BME maps (median 4→5, p=0.007). In 7% of cases, BME maps led to newly diagnosed fractures, while 4% of suspected fractures were excluded. The change in diagnosed fractures was more frequent for the radiology resident. Interobserver agreement on the presence of an edema signal in the BME maps was low (AC1=0.197, p=0.074), and image quality ratings for the BME maps differed significantly between readers (p<0.001). BME maps were considered non-evaluable due to artifacts in 4/37 (11%) of examinations.
Conclusion: BME maps derived from PCCT improve interobserver agreement and diagnostic confidence in fracture assessment. They may aid in detection and exclusion of subtle fractures beyond UHR images alone.
Limitations: Limited sample size; heterogeneous patient cohort; absence of independent diagnostic reference standard for fractures;
Funding for this study: N.a.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: M. Bozzolo, G. Gallarato, A. Balderi, E. Roberto, S. Amabile, D. Buccicardi; Cuneo/IT
Purpose: To explore the role of Z-effective maps as an additional tool, in combination with Water/Calcium maps, for detecting vertebral bone marrow edema (BME) in spinal fractures with dual energy ct (DECT), with MRI as reference standard.
Methods or Background: We analysed 261 vertebral bodies from 41 patients with acute spinal trauma who underwent DECT (GE Revolution CT ES 128, Rapid-Switching 80–140 kVp) and MRI. Qualitative analysis compared Z-effective and Water/Calcium maps with MRI, assessing BME presence and extent. Two independent readers evaluated all examinations blinded to MRI results. Quantitative analysis included HU, Z-effective, Water/Calcium, SNR, and spectral curves between edematous and healthy bone.
Results or Findings: Z-effective and Water/Calcium maps showed sensitivity 93.8–96.2%, specificity 98.9–99.5%, accuracy 97.3–98.5%, with inter-reader agreement (κ >0.96). Quantitative values were higher in edematous bone: HU (298 vs 53, p<0.001), Z-effective (10.36 vs 8.39, p<0.001), Water/Calcium (1071 vs 1002 mg/cm³, p<0.001), and SNR (14.4 vs 2.3, p<0.001). Z-effective strongly correlated with HU (r=0.92, p<0.001), while correlation with Water/Calcium was moderate (r=0.37, p=0.016), indicating complementary diagnostic value. Spectral curves confirmed higher attenuation in BME than healthy bone (p<0.001). ROC analysis showed excellent diagnostic performance (AUC Z-effective 0.97–0.98; Water/Calcium 0.96–0.97).
Conclusion: Z-effective maps show diagnostic accuracy comparable to Water/Calcium maps in detecting vertebral BME and provide complementary information. Their integration strengthens the role of DECT as a reliable tool for bone marrow edema assessment. In selected clinical and emergency settings, DECT may represent a practical alternative to MRI when not available or contraindicated.
Limitations: The time interval, although short, between DECT and MRI may have influenced the extent of edema.

Relatively small patient sample.

Variability of MRI scanners, potentially affecting quality and interpretation.
Funding for this study: This study received no external funding.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: C. Obermüller¹, U. Bach¹, F. Zecca², F. Heidt¹, M. Lohezic¹, R. Guggenberger³, E. Burian¹, J. Kroschke¹, F. Ensle¹; ¹Zürich/CH, ²L'Aquila/IT, ³Winterthur/CH
Purpose: To assess the impact of deep learning-based (DL) chemical shift-artifact correction (CSC) on bone assessment in zero echo time (ZTE) MRI of the lumbar spine, compared to non-DL and DL-reconstructions. ZTE is a bone-specific MRI sequence suffering from CS artifacts, particularly in the spine and at 3T.
Methods or Background: This retrospective study included 60 clinical patients with ZTE lumbar spine MRI at 3T. Raw k-space data was reconstructed offline with a prototype DLCSC algorithm (specifically designed for CS reduction), prototype DL algorithm without artifact correction and non-DL technique. Quantitative image sharpness was plotted with line profiles over vertebrae L2-4, analysed with repeated-measures ANOVA. Two radiologists assessed qualitative pathology-related criteria (pathology type, conspicuity, impact of artifacts, diagnostic confidence, n = 22) and diagnostic image quality (cortical and trabecular bone delineation, noise, artifacts, n = 38) using a 4-point scale. Ordinal data was analysed using the Friedman test and weighted Cohen's kappa.
Results or Findings: For quantitative sharpness, the mean score for DLCSC (0.388) was significantly higher than for non-DL (0.269; post-hoc p = .010) and DL (0.247; post-hoc p = .001). For the qualitative assessment, mean scores consistently improved from non-DL to DL to DLCSC, respectively: Diagnostic Image Quality (p < 0.001): means were 0.95, 1.63, and 2.37. Pathology Conspicuity (p < 0.001): means were 1.27, 1.82, and 2.18. Diagnostic Confidence (p < 0.001): means were 1.86, 2.32, and 2.59. Image Noise (p < 0.001): means were 1.18, 1.76, and 2.55. Inter-reader agreement for the qualitative assessment was moderate (k=0.45-0.6).
Conclusion: The DLCSC technique significantly enhanced bone assessment of the lumbar spine, compared to DL and non-DL reconstructions, including the evaluation of pathologies. This advancement extends the utility of ZTE for comprehensive spine imaging.
Limitations: Modest sample size
Funding for this study: No funding was received for this study (Basec-Nr 2021-02408).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study received Ethics Committee approval.