Radiology最新文献

Background Left ventricular mass (LVM) is an established marker of cardiovascular risk; however, long-term follow-up studies in individuals with low to intermediate risk are lacking. Purpose To assess the sex-specific association of LVM measured with cardiac MRI with cardiovascular outcomes in those with a less than 20% 10-year risk of cardiovascular disease (CVD). Materials and Methods A total of 1528 volunteers older than 40 years of age with no history of CVD, a 10-year risk of CVD of less than 20%, and a B-type natriuretic peptide level greater than their sex-specific median underwent cardiac MRI between June 2008 and February 2013 as part of the Tayside Screening for Cardiac Events, or TASCFORCE, prospective study. LVM was indexed to body surface area, and the LVM-to-volume ratio was calculated. Follow-up for cardiovascular events was performed using national electronic health records. Cox proportional hazard models and Kaplan-Meier curves were applied to assess the impact of LVM. Results A total of 1495 participants (mean age, 54.5 years ± 8.3 [SD]; 925 female, 570 male) completed cardiac MRI, with a median follow-up of 10 years (IQR, 3 years). In female participants, LVM was associated with age, blood pressure, smoking status, and cholesterol level, while in male participants, LVM was associated with age and blood pressure. In female participants, the LVM-to-volume ratio was associated with cardiovascular events (hazard ratio [HR], 2.3 [95% CI: 1.1, 4.9] for the highest quartile vs the lowest quartile), while the LVM was not. In male participants, the LVM was associated with cardiovascular events (HR, 3.2 [95% CI: 1.5,7.0] for the highest quartile vs the lowest quartile), while the LVM-to-volume ratio was not. Conclusion In those with low to intermediate risk without established CVD, different remodeling patterns predict cardiovascular events, with increased LVM predictive in male participants, while LVM-to-volume ratio is predictive in female participants. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Garot and Duhamel in this issue.

The management of hepatocellular carcinoma (HCC) is undergoing transformational changes due to the emergence of various novel immunotherapies and their combination with image-guided locoregional therapies. In this setting, immunotherapy is expected to become one of the standards of care in both neoadjuvant and adjuvant settings across all disease stages of HCC. Currently, more than 50 ongoing prospective clinical trials are investigating various end points for the combination of immunotherapy with both percutaneous and catheter-directed therapies. This review will outline essential tumor microenvironment mechanisms responsible for disease evolution and therapy resistance, discuss the rationale for combining locoregional therapy with immunotherapy, summarize ongoing clinical trials, and report on developing imaging end points and novel biomarkers that are relevant to both diagnostic and interventional radiologists participating in the management of HCC.

Background Interstitial lung abnormalities (ILAs) are incidental CT findings suggesting early interstitial lung disease. However ILA prevalence data are scarce in an unselected routine clinical setting. Purpose To evaluate the prevalence, underreporting rate, and potential clinical impact of ILAs recognizable on either abdominal CT scans or thoracoabdominal CT scans in a routine clinical setting of unselected patients. Materials and Methods Consecutive abdominal or thoracoabdominal CT scans from unselected inpatients and outpatients (age, ≥50 years; without any available prior chest CT and no clinical history of disease against the diagnosis of ILA) from a single-center tertiary hospital between January 2008 and December 2015 were retrospectively reviewed for the presence of ILAs and compared with the original clinical reports from the CT scans. Radiologic progression of ILA was evaluated by comparing consecutive CT points. Multivariable models adjusted for age, sex, race/ethnicity, oncologic disease, and cardiovascular disease were used to assess factors associated with odds of ILAs progression and all-cause and cause-specific mortality. Results Among 21 118 patients (median age, 72 years [IQR, 64-80 years]; 11 028 [52.2%] female patients), ILAs were observed in 362 (1.7%) patients, notably in 222 (1.0%) patients who had fibrotic features at CT. ILAs were recognized in 122 of 9415 (1.3%) and 240 of 11 703 (2.1%) of abdominal and thoracoabdominal CT scans, respectively. Of available original reports for 360 patients, 158 (43.9%) of all ILAs were not originally reported. Traction bronchiectasis index was the CT factor associated with higher odds of ILA progression (odds ratio, 3.47; 95% CI: 1.83, 6.58; P < .001). Fibrotic ILAs had a fourfold higher risk of respiratory-cause mortality (hazard ratio, 4.01; 95% CI: 2.02, 7.92; P < .001) compared with patients without ILAs. Conclusion The prevalence of ILAs was 1.7% in a large, unselected sample of patients who underwent either abdominal or thoracoabdominal CT for various clinical indications. Despite their prognostic significance, 43.9% of ILAs were unreported. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Hata in this issue.

Background Because of its ability to help assess the presence of subtle morphologic changes in bone and bone marrow edema, dual-energy CT (DECT) could be an alternative to MRI in the diagnosis of bone stress injury that includes a stress fracture (SF) and stress reaction (SR). Purpose To determine the diagnostic accuracy of DECT in identifying bone stress injury of the lower limb using MRI as the reference standard. Materials and Methods This prospective study, conducted between June 2021 and January 2024, included consecutive patients clinically suspected of having stress injury (SF or SR) of the lower limb (leg, ankle, or foot). Imaging diagnosis was based on the absence or presence of cortical or periosteal changes, bone marrow edema, or a fracture line. The diagnostic performance of four blinded independent readers was determined for the entire cohort and for the subset of participants without fracture lines. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were calculated. Interobserver agreement was evaluated with Kendall coefficient of concordance (W). Results A total of 184 patients (mean age, 58 years ± 17 [SD]; 97 male) were enrolled. At MRI, 107 of 184 participants (58%) had positive diagnoses, including 70 with SF and 37 with SR. The mean overall sensitivity and specificity were 91% (390 of 428; 95% CI: 0.85, 0.95) and 93% (287 of 308; 95% CI: 0.87, 0.97), respectively, with an AUC of 0.94 (95% CI: 0.91, 0.97). Among patients without fracture lines (n = 114), the mean overall sensitivity and specificity of DECT were 79% (117 of 148; 95% CI: 0.65, 0.88) and 93% (287 of 308; 95% CI: 0.87, 0.97), respectively, with an AUC of 0.87 (95% CI: 0.81, 0.94). The interobserver agreement was very good for diagnosis of SF and SR combined (Kendall W = 0.90) and SR alone (Kendall W = 0.84). Conclusion DECT helped to accurately diagnose bone stress injury of the lower limb by identifying fracture lines and osseous stress reactions. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Breighner in this issue.

Background CT plays an important role in the opportunistic identification of hepatic steatosis. CT performance for steatosis detection has been inconsistent across various studies, and no clear guidelines on optimum thresholds have been established. Purpose To conduct a systematic review and meta-analysis to assess CT diagnostic accuracy in hepatic steatosis detection and to determine reliable cutoffs for the commonly mentioned measures in the literature. Materials and Methods A systematic search of the PubMed, Embase, and Scopus databases (English-language studies published from September 1977 to January 2024) was performed. Studies evaluating the diagnostic accuracy of noncontrast CT (NCCT), contrast-enhanced (CECT), and dual-energy CT (DECT) for hepatic steatosis detection were included. Reference standards included biopsy, MRI proton density fat fraction (PDFF), or NCCT. In several CECT and DECT studies, NCCT was used as the reference standard, necessitating subgroup analysis. Statistical analysis included a random-effects meta-analysis, assessment of heterogeneity with use of the I² statistic, and meta-regression to explore potential sources of heterogeneity. When available, mean liver attenuation, liver-spleen attenuation difference, liver to spleen attenuation ratio, and the DECT-derived fat fraction for hepatic steatosis diagnosis were assessed. Results Forty-two studies (14 186 participants) were included. NCCT had a sensitivity and specificity of 72% and 88%, respectively, for steatosis (>5% fat at biopsy) detection and 82% and 94% for at least moderate steatosis (over 20%-33% fat at biopsy) detection. CECT had a sensitivity and specificity of 66% and 90% for steatosis detection and 68% and 93% for at least moderate steatosis detection. DECT had a sensitivity and specificity of 85% and 88% for steatosis detection. In the subgroup analysis, the sensitivity and specificity for detecting steatosis were 80% and 99% for CECT and 84% and 93% for DECT. There was heterogeneity among studies focusing on CECT and DECT. Liver attenuation less than 40-45 HU, liver-spleen attenuation difference less than -5 to 0 HU, and liver to spleen attenuation ratio less than 0.9-1 achieved high specificity for detection of at least moderate steatosis. Conclusion NCCT showed high performance for detection of at least moderate steatosis. © RSNA, 2024 Supplemental material is available for this article.

Background It is unclear whether artificial intelligence (AI) explanations help or hurt radiologists and other physicians in AI-assisted radiologic diagnostic decision-making. Purpose To test whether the type of AI explanation and the correctness and confidence level of AI advice impact physician diagnostic performance, perception of AI advice usefulness, and trust in AI advice for chest radiograph diagnosis. Materials and Methods A multicenter, prospective randomized study was conducted from April 2022 to September 2022. Two types of AI explanations prevalent in medical imaging-local (feature-based) explanations and global (prototype-based) explanations-were a between-participant factor, while AI correctness and confidence were within-participant factors. Radiologists (task experts) and internal or emergency medicine physicians (task nonexperts) received a chest radiograph to read; then, simulated AI advice was presented. Generalized linear mixed-effects models were used to analyze the effects of the experimental variables on diagnostic accuracy, efficiency, physician perception of AI usefulness, and "simple trust" (ie, speed of alignment with or divergence from AI advice); the control variables included knowledge of AI, demographic characteristics, and task expertise. Holm-Sidak corrections were used to adjust for multiple comparisons. Results Data from 220 physicians (median age, 30 years [IQR, 28-32.75 years]; 146 male participants) were analyzed. Compared with global AI explanations, local AI explanations yielded better physician diagnostic accuracy when the AI advice was correct (β = 0.86; P value adjusted for multiple comparisons [P_adj] < .001) and increased diagnostic efficiency overall by reducing the time spent considering AI advice (β = -0.19; P_adj = .01). While there were interaction effects of explanation type, AI confidence level, and physician task expertise on diagnostic accuracy (β = -1.05; P_adj = .04), there was no evidence that AI explanation type or AI confidence level significantly affected subjective measures (physician diagnostic confidence and perception of AI usefulness). Finally, radiologists and nonradiologists placed greater simple trust in local AI explanations than in global explanations, regardless of the correctness of the AI advice (β = 1.32; P_adj = .048). Conclusion The type of AI explanation impacted physician diagnostic performance and trust in AI, even when physicians themselves were not aware of such effects. © RSNA, 2024 Supplemental material is available for this article.

Background Deep learning (DL) algorithms have shown promising results in mammographic screening either compared to a single reader or, when deployed in conjunction with a human reader, compared with double reading. Purpose To externally validate the performance of three DL algorithms as mammographic screen readers in an independent UK data set. Materials and Methods Three commercial DL algorithms (DL-1, DL-2, and DL-3) were retrospectively investigated from January 2022 to June 2022 using consecutive full-field digital mammograms collected at two UK sites during 1 year (2017). Normal cases with 3-year follow-up and histopathologically proven cancer cases detected either at screening (that round or next) or within the 3-year interval were included. A preset specificity threshold equivalent to a single reader was applied. Performance was evaluated for stand-alone DL reading compared with single human reading, and for DL reading combined with human reading compared with double reading, using sensitivity and specificity as the primary metrics. P < .025 was considered to indicate statistical significance for noninferiority testing. Results A total of 26 722 cases (median patient age, 59.0 years [IQR, 54.0-63.0 years]) with mammograms acquired using machines from two vendors were included. Cases included 332 screen-detected, 174 interval, and 254 next-round cancers. Two of three stand-alone DL algorithms achieved noninferior sensitivity (DL-1: 64.8%, P < .001; DL-2: 56.7%, P = .03; DL-3: 58.9%, P < .001) compared with the single first reader (62.8%), and specificity was noninferior for DL-1 (92.8%; P < .001) and DL-2 (96.8%; P < .001) and superior for DL-3 (97.9%; P < .001) compared with the single first reader (96.5%). Combining the DL algorithms with human readers achieved noninferior sensitivity (67.0%, 65.6%, and 65.4% for DL-1, DL-2, and DL-3, respectively; P < .001 for all) compared with double reading (67.4%), and superior specificity (97.4%, 97.6%, and 97.6%; P < .001 for all) compared with double reading (97.1%). Conclusion Use of stand-alone DL algorithms in combination with a human reader could maintain screening accuracy while reducing workload. Published under a CC BY 4.0 license. Supplemental material is available for this article.

Background Studies suggest that readers experience perceptual adaptation when interpreting batched screening mammograms, which may serve as a mechanism for improved performance. Purpose To analyze clinical digital breast tomosynthesis (DBT) screening data to evaluate changes in reader performance during sequential batch reading. Materials and Methods This observational retrospective study used data from the radiology information system collected for screening DBT examinations performed from January 2018 to December 2019. The reference standard was established based on pathology results, if applicable, or imaging findings at 1-year follow-up. Examinations were aggregated into batches, defined as a sequence of cases for a given reader with differences in interexamination time of 10 minutes or less. Mixed-effect models were used to evaluate performance and timing across batch positions. Statistical adjustments accounted for potential confounders, including patient characteristics (age, breast density), reading factors (day of week, time of day), and between-reader heterogeneity. Results The dataset included 121 652 examinations (median patient age, 61 years [IQR, 53-69 years]), including 1081 cancers interpreted by 15 radiologists. Unadjusted false-positive rates decreased from an averaged 15.5% at the first within-batch examination to 10.5% after three sequentially read examinations (P < .001), without a significant change in sensitivity (82.6% vs 84.2%; P = .15). The interpretation time consistently decreased, achieving a substantial reduction in longer batches (average 2.8 to 2.2 minutes for noncancer examinations; P < .001). Adjustment for sampling bias and confounders (patient age, breast density, day of week, time of day; all P < .005) reduced the effect on false-positive rates (11.5% to 9.4% after three examinations; P < .001) and interpretation time (3.2 to 2.7 minutes for noncancer examinations; P < .001). Conclusion Radiologists showed improved recall rates and interpretation times without change in sensitivity over the course of batch reading of DBT screening examinations. © RSNA, 2024 See also the editorial by Iima and Satake in this issue.