ACTA Radiologica (Test)

Acta Radiologica, Ahead of Print. <br/>BackgroundOpportunistic screening of osteoporosis using computed tomography (CT) should be optimized according to CT protocols. However, to date, no study has evaluated the different contrast phases of CT urography in osteoporosis screening.PurposeTo compare the diagnostic performance of different contrast phases of CT urography in opportunistic screening of osteoporosis.Material and MethodsIn this single-center retrospective study, 210 consecutive patients (181 men, 29 women; mean age = 72.0 ± 9.5 years) who underwent both CT urography and dual-energy X-ray absorptiometry (DXA) within 3 months in July to December 2020 were included. The attenuation value was measured at the center of the L1 vertebral body in the axial plane of the precontrast, corticomedullary, and excretory phases. The mean and standard deviation (SD) of attenuation values were calculated. The area under the receiver operating characteristic curve (AUC) in differentiating osteoporosis versus osteopenia/normal was measured in each phase, and non-parametric comparisons between precontrast and the other phases were performed.ResultsAccording to DXA results, 90, 90, and 30 patients were classified into the normal, osteopenia, and osteoporosis groups, respectively. The mean ± SD of attenuation values in the precontrast, corticomedullary, and excretory phases were 114.9 ± 47.8, 132.8 ± 49.3, and 126.2 ± 47.2 HU, respectively. A significant difference was observed between AUCs measured in the precontrast (0.804, 95% confidence interval [CI] = 0.717–0.890) and corticomedullary phases (0.760, 95% CI = 0.661–0.860) (P = 0.003) and those between the precontrast and excretory phases (0.774, 95% CI = 0.678–0.869) (P = 0.005).ConclusionThe precontrast phase outperformed the other phases in the opportunistic screening of osteoporosis using CT urography.

Acta Radiologica, Ahead of Print. <br/>BackgroundAccurate preoperative non-invasive assessment of HER2 expression in breast cancer is crucial for personalized treatment and prognostic stratification.PurposeTo evaluate the effectiveness of radiomics models based on multi-parametric magnetic resonance imaging (MRI) in distinguishing HER2 expression status in invasive breast cancer.Material and MethodsWe conducted a retrospective analysis of baseline MRI scans and clinical data from 400 patients with breast cancer between January 2018 and December 2019. Two-dimensional regions of interest were manually segmented on the maximum tumor images obtained from turbo inversion recovery magnitude (TIRM), dynamic contrast-enhanced magnetic resonance imaging phase 2 (DCE2), dynamic contrast-enhanced magnetic resonance imaging phase 4 (DCE4), diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC) sequences using ITK-SNAP software. Features were extracted and screened for dimensionality reduction. Logistic regression models were developed to predict HER2 expression status.ResultsIn distinguishing HER2-overexpression from non-HER2-overexpression, the DCE2 model outperformed other single-parameter models, with areas under the curve (AUCs) of 0.91 (training) and 0.88 (test). Combination models with DCE features showed significantly improved performance (P ≤ 0.001). The multiparameter model achieved the highest AUCs of 0.93 (training) and 0.91 (test). In distinguishing HER2-low from HER2-zero, the TIRM model performed best among single-parameter models, with AUCs of 0.80 (training) and 0.72 (test). The multiparameter model further enhanced prediction, yielding an AUC of 0.83 (test).ConclusionRadiomics models based on multi-parametric MRI features demonstrated strong clinical utility in assessing HER2 expression status in invasive breast cancer, particularly in identifying HER2-overexpression and HER2-low expression subtypes.

Acta Radiologica, Ahead of Print. <br/>BackgroundKnees with anterior cruciate ligament (ACL) injury have distinct knee bone morphology. However, the correlation between ACL injury and morphology of the proximal tibiofibular joint has not been investigated.PurposeTo compare proximal tibiofibular joint morphology in patients with injured and intact ACLs to aid in predicting and preventing ACL injuries.Material and MethodsA total of 50 patients with ACL injury and 50 individuals without structural knee damage (revealed by magnetic resonance imaging [MRI]) were included in this study. In the anteroposterior knee radiographs, the length between the proximal end of the fibular head and the medial and lateral tibia joint line were measured. In the axial knee MRI scans with the most proximal part of the fibular head, the distance calculation was performed between the most anterior point of the tibia plateau (MATP) and the most anterior point of fibular head (MAFH).ResultsThe mean length from the proximal end of the fibular head and the medial and lateral tibial joint lines was 8.7 ± 2.8 mm and 13.2 ± 2.6 mm in the ACL-injured group, and 6.5 ± 2.8 mm and 11.1 ± 2.4 mm in the control group. The length for both medial and lateral sides was significantly greater in the ACL-injured group. The length between the MATP and MAFH was significantly larger in the ACL-injured group than the control group.ConclusionA large distance between the proximal end of the fibular head and the tibial joint line and knees with posterior fibular head placement would be another anatomic risk factor of ACL injury.

Acta Radiologica, Ahead of Print. <br/>BackgroundNumerous studies have examined the prognosis of patients with relapsing-remitting multiple sclerosis (RRMS).PurposeTo seek a relationship between the gray matter/white matter (GM/WM) volume ratio, clinical features, neurofilament light chain (NfL), and glial fibrillary acidic protein (GFAP) in RRMS.Material and MethodsA total of 61 patients aged 18–70 years with RRMS and 20 controls were included. Neurological examinations were noted at the first attack and last visit. Blood samples were taken in remission.ResultsPatients’ WM volume was lower and the GM/WM volume ratio was larger than in the controls (P < 0.001). Patients’ WM correlated with GM volume (P < 0.001, r = 0.608). WM correlated with GM/WM volume ratio (P < 0.001, r = −0.632). For NfL and GFAP, no difference was found between patients and controls. Patients’ NfL correlated with GFAP (P < 0.001, r = 0.452). Age and disease duration were correlated with Gfap (P < 0.001, r = 0.434; P < 0.002, r = 0.381). No correlation was found between NfL/GFAP levels and all volumetric measurements of patients. Higher Expanded Disability Status Scale (EDSS), lower GM and WM volumes, and greater GM/WM volume ratio were detected in patients with the first pyramidal findings (P < 0.001).ConclusionPatients’ NfL and GFAP levels were lower than in controls and did not correlate with all volumes. GFAP is more informative, particularly in patients of advanced age and those with longstanding diseases. Higher EDSS and decreased GM and WM volumes were seen in patients with pyramidal symptoms at the first attack.

Acta Radiologica, Ahead of Print. <br/>Rater agreement is commonly assessed in radiologic studies concerning image quality. There are several methods of measuring rater agreement. To choose the appropriate method, the researcher needs to consider the scale of the outcome variable and the design of the study. This article provides a brief overview of available methods, focusing on the most practical and commonly used, including intraclass correlation, the Svensson method, variants of the kappa statistic, the agreement coefficient by Gwet (AC1/AC2), and Krippendorff's alpha. Additional methods that are not primarily intended for rater agreement analysis but are applied in some cases are also discussed.

Acta Radiologica, Ahead of Print. <br/>BackgroundIdentification and staging of rectal cancer are mainly based on the difference in signal intensity (SI) between the tumor and normal structures of the intestinal wall on T2-weighted imaging. However, differentiating stage pT0–T1 from pT2 rectal tumors is difficult using routine magnetic resonance imaging (MRI) sequences.PurposeTo summarize and explore whether MRI findings from routine imaging can help differentiate pT0–T1 from pT2 rectal tumors.Material and MethodsA total of 110 patients with pT0–T2 rectal cancer underwent preoperative pelvic MRI examinations and tumor resection without preoperative chemoradiotherapy. MRI findings of rectal lesions (including tumor location, shape, longitudinal length, maximum cross-section, submucosal high intensity [SHI], extramural fibrotic scarring, wall shrinkage, lesion-to-wall signal intensity ratio, and presence of lymph node with short axis over 3 mm) and clinical characteristics were analyzed by univariate and multivariate analyses to screen the independent factors associated with pathological results.ResultsOf all the lesions, 32 tumors were proved to be pT0–T1 and 78 tumors were pT2. Univariate and multivariate logistic regression analyses revealed that tumor shape (odds ratio [OR] = 24.607, P < 0.001), SHI (OR = 6.129, P = 0.002), and extramural fibrotic scarring (OR = 0.110, P = 0.007) were independent factors distinguishing pT0–T1 tumors from pT2 tumors. If the rectal lesion has a cauliflower-like shape with SHI and no extramural fibrotic scarring, it is more likely to be a pT0–T1 tumor.ConclusionThe imaging features obtained from the routine MRI sequence showed potential value for differentiating pT0–T1 from pT2 rectal tumors.

Acta Radiologica, Ahead of Print. <br/>BackgroundWhile the “swallow tail” sign observed in the substantia nigra (SN) on susceptibility map-weighted imaging (SMWI) serves as an effective marker for differentiating patients with Parkinson's disease (PD) from healthy individuals, its visual assessment proves inadequate in differentiating PD from atypical Parkinson syndromes (APS).PurposeTo employ radiomic features extracted from multi-echo SMWI of the SN to distinguish between PD and APS.Material and MethodsSMWI data were acquired from 63 PD patients, 38 APS patients, and 89 healthy controls. The participants were randomly assigned to either training or test groups in a 7:3 proportion. Utilizing the PyRadiomics software, a set of radiomic features were extracted from SN for analysis. Features underwent standardization via the maximum-minimum method, with 166 statistically significant features identified through independent t-tests. To minimize the risk of overfitting, the least absolute shrinkage and selection operator (LASSO) algorithm was implemented to identify and select the five most significant features from the radiomic dataset. Five distinct machine-learning classifiers were developed to distinguish between PD, APS, and healthy controls. The SHapley Additive Explanations was employed to gain insights into and visualize the relative importance of each feature within these models.ResultsMorphological, first-order, texture, and wavelet transform features of the SN emerged as the most crucial determinants. The light gradient-boosting machine model demonstrated superior performance in distinguishing between PD, APS, and healthy controls.ConclusionRadiomic features of the SN derived from SMWI show promise in differentiating PD from APS, potentially enhancing diagnostic accuracy in clinical settings.

Acta Radiologica, Ahead of Print. <br/>BackgroundRenal cell carcinomas (RCCs) adjacent to critical structures can be safely treated using percutaneous cryoablation with adjunctive techniques. However, there are only a few reports describing the factors affecting local tumor control after percutaneous cryoablation for such RCCs.PurposeTo evaluate the factors affecting local tumor control in the percutaneous cryoablation of RCCs adjacent to critical structures.Material and MethodsA total of 403 tumors from 328 patients were retrospectively reviewed. Patients with histologically proven RCCs adjacent to critical structures with at least 1 year of follow-up after treatment were included. Hereditary RCCs were excluded. Of the 403 tumors, 40 met the criteria. Owing to the tumor location, all procedures were performed using adjunctive techniques. Baseline, perioperative, and follow-up data were collected and factors affecting local tumor progression were evaluated. Comparisons between groups with and without local tumor progression were performed using Fisher's exact test or Wilcoxon's rank sum test.ResultsThe median follow-up was 1289.5 days. Local tumor progression was observed in 7/40 (17.5%) patients. Univariate analysis revealed that significant factors affecting local tumor progression were situated adjacent to the ureter (P = 0.005), requiring adjunctive techniques other than hydro- and/or pneumo-dissection (P = 0.005), as well as the distance between tumors and critical structures (P = 0.04). The ice-ball margin tended to be smaller in the group with local tumor progression but this was not significant (P = 0.07).ConclusionRenal tumors adjacent to the ureter may result in local tumor progression after cryoablation compared with tumors adjacent to other organs.

Acta Radiologica, Ahead of Print. <br/>BackgroundEmerging evidence suggests multiparametric prostate MRI's role in evaluating prostate cancer (PCa) aggressiveness.PurposeTo evaluate the utility of quantitative sequences in prostate MRI for assessing peripheral zone PCa aggressiveness, focusing on added value of relaxation maps to ADC.Material and MethodsIn total, 64 men undergoing preoperative bi-parametric prostate MRI with synthetic sequences were retrospectively enrolled. Radiologists delineated lesions on ADC maps and synthetic MRI, matched with digitized whole-mount specimens. PCa was graded as low, intermediate, or high based on Gleason grade. Statistical analyses assessed differences in T1, T2, PD, and ADC values across grades. For significant variables, AUC curves were generated. A logistic regression model evaluated the added diagnostic value of relaxation maps to ADC.ResultsOf the 95 PCa foci, 11 were low-grade, 45 intermediate-grade, and 39 high-grade. T1 (P = 0.4) and PD (P = 0.3) showed no significant differences, while T2 and ADC differed significantly (P < 0.001). Median T2 values were 98 ms (low-grade), 81 ms (intermediate), and 73 ms (high-grade), and median ADC values were 1006 mm²/s, 836 mm²/s, and 721 mm²/s, respectively. ADC yielded AUCs of 0.888 (95% CI=0.794–0.963; P < 0.0001) for low- versus intermediate-to-high-grade PCa, and 0.825 (95% CI=0.724–0.901; P < 0.0001) for low-to-intermediate versus high-grade. T2 AUCs were 0.770 (95% CI=0.605–0.908; P = 0.021) and 0.755 (95% CI=0.648–0.864; P = 0.0051), respectively. A logistic regression model combining T2 and ADC showed no significant improvement over ADC alone in differentiating low- versus intermediate-to-high-grade and low-to-intermediate versus high-grade PCa (P > 0.05).ConclusionT1 and PD maps from synthetic MRI showed limited utility in grading PCa. Although T2 values were effective, ADC values were superior in assessing PCa aggressiveness. The addition of T2 to ADC did not significantly improve diagnostic performance.

Acta Radiologica, Ahead of Print. <br/>BackgroundDigital breast tomosynthesis (DBT) has several limitations with respect to microcalcification depiction. The quality of DBT images may be improved by modifying technical parameters; however, the amount of raw data and postprocessing computational time is unacceptably huge. Therefore, once detected it still seems necessary to take additional spot images to characterize microcalcifications. Additional imaging increases cumulative dose. Radiation risks and protection are raising issues nowadays.PurposeTo compare image quality of magnification tomosynthesis (t-spot) and synthetic spot magnification (s-spot) to conventional spot images in characterization of microcalcifications.Material and MethodsThis prospective single institute reading study includes 45 patients with suspected microcalcifications initially found on mammography and referred to vacuum-assisted biopsy. Conventional spot and DBT were performed. Acquired DBT data were used to reconstruct t- and s-spots. All images were reviewed by four breast radiologists in two sessions. The morphology (BI-RADS) and extent of microcalcifications were recorded. Clinical performance was evaluated using area under the receiver operating characteristic curves (AUC) for BI-RADS and Bland–Altman plots for the maximum extent of microcalcifications.ResultsAll four readers preferred t-spot to s-spot. The overall AUC for t-spot was 0.67 and for spot 0.69 with a difference lower than a non-inferiority margin (delta = 0.012, 95% confidence interval [CI] = −0.067–0.09, P = 0.772). Three readers underestimate the extent of microcalcifications on t-spots for both benign and malignant cases (P < 0.05). The discordance becomes more evident with the increasing extent of the lesion.ConclusionThe performance of conventional and reconstructed spots was similar but reader-dependent and inconclusive.