Background/Objectives: To evaluate the potential complementary value of the 3D multi fast field echo (M-FFE) sequence for knee bone assessment at 1.5T MRI, focusing on the visualization of osseous abnormalities and its potential to complement a standard knee MRI protocol. Methods: A standard knee MRI protocol supplemented with a 3D M-FFE sequence was performed at 1.5T MRI in 49 outpatients. Two observers independently evaluated image quality, visualization of six anatomical knee structures, and osseous abnormalities using a 5-point Likert scale. Intra- and interrater reliability was assessed using intraclass correlation coefficients and Cohen’s kappa. Eleven examinations were reassessed for intra-rater reliability. Diagnostic confidence with the 3D M-FFE sequence was evaluated. Results: Image quality was consistently high, with sharpness and artifact scores of 4–5 in >95% of examinations, while contrast showed greater variability, and no examinations scored below 3. Osseous abnormalities were identified in 18/49 patients by each observer, with agreement in 12 cases. A total of 28 and 25 abnormalities were identified by the observers, most commonly osteochondral lesions and bone islands. Abnormalities were assigned visualization scores of 4–5. Inter-rater agreement for osseous abnormalities ranged from fair to substantial (kappa = 0.40–0.76), while intra-rater agreement varied from slight to almost perfect depending on anatomical location. Both observers reported increased perceived diagnostic confidence in osseous abnormalities with the addition of the 3D M-FFE sequence. Conclusions: The 3D M-FFE sequence demonstrated high image quality, robust anatomical visualization, and potential complementary value for osseous assessment at 1.5T MRI. Further studies using an appropriate reference standard are needed to establish diagnostic accuracy.
Objectives: To compare fracture detection rates and observer confidence between a high-frequency hockey-stick transducer and a conventional linear transducer for detecting non-displaced dentoalveolar fractures in a paired ex vivo model. Materials and Methods: This paired ex vivo study included 88 experimentally created non-displaced dentoalveolar fracture sites in fresh ovine specimens. Two oral and maxillofacial radiologists independently evaluated all sites with both a 12L3 conventional linear transducer (10–15 MHz) and an 18H5 hockey-stick transducer (10–18 MHz), with transducer order randomized. Cone-beam computed tomography (CBCT) confirmed fracture status. Within-observer paired fracture-detection outcomes were compared using the exact McNemar test, and a cluster-adjusted generalized estimating equation (GEE) analysis accounted for repeated assessments and clustering within ovine heads. Observer confidence was compared using the Wilcoxon signed-rank test. Results: Observer 1 detected 76/88 fracture sites (86.4%) with the hockey-stick transducer and 64/88 (72.7%) with the conventional linear transducer (p = 0.004). Observer 2 detected 74/88 sites (84.1%) and 62/88 (70.5%), respectively (p < 0.001). In the cluster-adjusted GEE analysis, the hockey-stick transducer was associated with higher odds of fracture detection (OR, 2.29; 95% CI, 1.84–2.86; p < 0.001). For both observers, confidence scores were higher with the hockey-stick transducer (median [IQR], 4 [3–4] vs. 3 [2–3]; both p < 0.001), with Cohen’s dz values of 0.849 and 0.791, respectively. In a separate pre-study calibration set, interobserver agreement was 95% (κ = 0.828; 95% CI, 0.503–1.000). Conclusions: In this ex vivo model, the hockey-stick transducer achieved consistently higher fracture-detection rates across both observers and greater observer confidence than the conventional linear transducer. The observed differences may reflect the combined technical and ergonomic characteristics of the two transducers rather than probe geometry alone. Prospective clinical studies including both fractured and non-fractured sites are required before conclusions regarding diagnostic accuracy or clinical superiority can be established.
Background/Objectives: Unilateral sinonasal lesions may be difficult to characterize on non-contrast computed tomography (CT) because inflammatory polyps, benign neoplasms, and malignant tumors can appear as similar unilateral soft-tissue masses. Non-contrast CT is widely available, but the diagnostic performance of relative CT number across inflammatory, benign neoplastic, and malignant sinonasal lesions within a single cohort remains insufficiently characterized. This study aimed to provide additional evidence regarding the diagnostic performance of relative CT number on non-contrast CT by comparing its values among sinonasal polyp (SP), inverted papilloma (IP), and squamous cell carcinoma (SCC), and by evaluating predefined binary diagnostic comparisons. Methods: This retrospective single-center study included 92 patients with histopathologically confirmed unilateral or unilateral-dominant sinonasal lesions: 49 SP, 24 IP, and 19 SCC. Preoperative non-contrast paranasal sinus CT images were reviewed. Relative CT number was calculated as the ratio of mean lesion attenuation to mean pontine attenuation. Group differences were analyzed using non-parametric tests, and diagnostic performance was assessed with receiver operating characteristic (ROC) curve analysis. Age-adjusted binary logistic regression analyses were also performed for the predefined binary comparisons. Sensitivity analyses were additionally adjusted for lesion ROI area. Results: Relative CT number differed significantly among the three groups (overall p < 0.001). ROC analysis showed an AUC of 0.815 (95% CI, 0.704–0.904) for SP versus IP, 0.667 (95% CI, 0.498–0.826) for IP versus SCC, 0.869 (95% CI, 0.792–0.932) for SP versus IP+SCC, and 0.937 (95% CI, 0.866–0.986) for SP versus SCC. After age adjustment, relative CT number was associated with the positive diagnostic category in all four comparisons; however, the IP–SCC association was no longer significant after ROI-area adjustment (p = 0.073). The highest diagnostic performance was observed for distinguishing SP from SCC. Conclusions: These preliminary findings suggest that relative CT number may serve as an adjunctive quantitative CT parameter for differentiating unilateral inflammatory sinonasal lesions from neoplastic lesions, with the strongest diagnostic performance observed for SP versus SCC. However, its standalone discrimination between IP and SCC was limited, and the age-adjusted association remains hypothesis-generating.
Objectives: This single-blinded ex vivo diagnostic accuracy study evaluated high-frequency ultrasonography for detecting nondisplaced dentoalveolar fractures and assessed the effect of simulated post-traumatic soft-tissue edema. Methods: Eighteen fresh sheep heads provided 108 anatomically distinct sites (54 fractured and 54 non-fractured controls). Two blinded oral and maxillofacial radiologists independently assessed each site, yielding 216 observer-level evaluations. Cone-beam tomography (CBCT) served as the reference standard. Edema was simulated by submucosal injection of 0.5, 1.0, or 1.5 mL sterile saline. Results: Simulated edema was associated with significantly lower odds of correct ultrasonographic classification in the GEE analysis accounting for within-site clustering (OR = 0.217; 95% CI, 0.096–0.491; p < 0.001). Accuracy decreased from 87.0% to 59.3%, sensitivity from 96.3% to 77.8%, and specificity from 77.8% to 40.7%. False-negative and false-positive rates increased from 3.7% to 22.2% and from 22.2% to 59.3%, respectively. Observer confidence also decreased significantly (3.53 ± 0.90 vs. 2.61 ± 0.86; p < 0.001). No statistically significant differences in diagnostic accuracy were detected among the three edema-volume groups. Conclusions: Simulated soft-tissue edema substantially impairs ultrasonographic detection of nondisplaced dentoalveolar fractures. These experimental findings suggest that negative ultrasonographic findings obtained in the presence of soft-tissue swelling should be interpreted cautiously, with complementary imaging considered when clinically indicated.
Purpose: To investigate whether radiomic characteristics of visually unremarkable basal ganglia on preprocedural non-contrast computed tomography (CT) are associated with subsequent post-thrombectomy hyperdense transformation (HT) in patients with acute ischemic stroke. Materials and Methods: This retrospective single-center study included a derivation cohort of 84 patients with preserved Alberta Stroke Programme Early CT Scores (ASPECTS ≥ 8), no visually appreciable basal ganglia abnormality, anterior circulation large-vessel occlusion, and complete reperfusion after mechanical thrombectomy between January 2022 and July 2023. An independent temporal validation cohort comprised 31 consecutive eligible patients treated between August and November 2023. The ipsilateral basal ganglia were manually segmented on preprocedural non-contrast CT, and radiomic features were extracted using LIFEx. Candidate features underwent interobserver reproducibility assessment and correlation filtering before least absolute shrinkage and selection operator logistic regression with 10-fold cross-validation. The final model was applied without modification to the temporal validation cohort. Nested cross-validation and learning-curve analysis were additionally performed to assess model-development optimism and stability. Results: The final model retained two texture features, Gray-Level Run Length Matrix Short Runs Emphasis and Gray-Level Size Zone Matrix Grey Level Variance. In the derivation cohort, the model achieved an area under the receiver operating characteristic curve (AUC) of 0.760 (95% CI, 0.652–0.855), with 48.8% sensitivity, 95.3% specificity, and 72.6% accuracy. When applied without modification to the independent temporal validation cohort, the AUC was 0.677 (95% CI, 0.475–0.858), with 33.3% sensitivity, 93.8% specificity, and 64.5% accuracy. Nested cross-validation yielded a mean AUC of 0.673 ± 0.037, with variability in feature selection across resampling iterations. Learning-curve analysis showed an early plateau, with little improvement in cross-validated discrimination as the training sample increased within the evaluated range. Conclusions: Radiomic analysis of visually unremarkable basal ganglia on preprocedural non-contrast CT identified an exploratory two-feature imaging signature associated with subsequent post-thrombectomy HT. The model retained modest discriminatory ability in temporal validation, although internal resampling demonstrated model-development optimism and limited feature-selection stability. These findings should be considered hypothesis-generating and warrant further evaluation using larger datasets, alternative modeling approaches, and independent multicenter validation.
