Diagnostic and interventional radiology最新文献

Purpose: After the introduction of the Ovarian-Adnexal Reporting and Data System (O-RADS) for magnetic resonance imaging (MRI), several studies with diverse characteristics have been published to assess its diagnostic performance. This systematic review and meta-analysis aimed to assess the diagnostic performance of O-RADS MRI scoring for adnexal masses, accounting for the risk of selection bias.

Methods: The PubMed, Scopus, Web of Science, and Cochrane databases were searched for eligible studies. Borderline or malignant lesions were considered malignant. All O-RADS MRI scores ≥4 were considered positive. The quality of the studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. The pooled sensitivity, specificity, and likelihood ratio (LR) values were calculated, considering the risk of selection bias.

Results: Fifteen eligible studies were found, and five of them had a high risk of selection bias. Between-study heterogeneity was low-to-moderate for sensitivity but substantial for specificity (I2 values were 35.5% and 64.7%, respectively). The pooled sensitivity was significantly lower in the studies with a low risk of bias compared with those with a high risk of bias (93.0% and 97.5%, respectively; P = 0.043), whereas the pooled specificity was not different (90.4% for the overall population). The negative and positive LRs were 0.08 [95% confidence interval (CI) 0.05–0.11] and 10.0 (95% CI 7.7–12.9), respectively, for the studies with low risk of bias and 0.03 (95% CI 0.01–0.10) and 10.3 (95% CI 3.8–28.3), respectively, for those with high risk of bias.

Conclusion: The overall diagnostic performance of the O-RADS system is very high, particularly for ruling out borderline/malignant lesions, but with a moderate ruling-in potential. Studies with a high risk of selection bias lead to an overestimation of sensitivity.

Clinical significance: The O-RADS system demonstrates considerable diagnostic performance, particularly in ruling out borderline or malignant lesions, and should routinely be used in practice. The high between-study heterogeneity observed for specificity suggests the need for improvement in the consistent characterization of the benign lesions to reduce false positive rates.

Bevacizumab is a common strategy for the treatment of recurrent glioblastoma. Survival status is a crucial issue for patients with recurrent glioblastoma, and the apparent diffusion coefficient (ADC) values of the lower Gaussian curve have been reported to have the potential to predict prognosis in recurrent glioblastoma. In the present study, we aimed to clarify the survival prediction of ADC values in patients with recurrent glioblastoma receiving bevacizumab treatment through a systematic review and meta-analysis of randomized clinical trials, comparing ADC values higher than the cut-off values with those lower than the cut-off values to determine which type of ADC values can be associated with significant survival benefits. Different survival indicators were analyzed, including overall survival (OS) and progression-free survival (PFS). Ten studies with a total of 782 patients with recurrent glioblastoma were included. The focused outcomes were OS and PFS. Our results showed that ADC values lower than the cut-off values were associated with significant benefits for OS status compared with ADC values higher than the cut-off values. Similar significant benefits were observed for PFS. The meta-analysis results suggest that ADC values lower than the cut-off values might be associated with significant benefits for OS and PFS when compared with ADC values higher than the cut-off values. However, bias in relation to the different stages of recurrent glioblastoma and different types, doses, and regimens of bevacizumab should not be ignored.

Purpose: This study aimed to explore the relationship between the time from percutaneous vertebral augmentation (PVA) until subsequent fracture and the risk of new symptomatic fractures (NSFs) in untreated vertebrae at different distances from "augmented vertebrae".

Methods: Patients who underwent PVA for the treatment of osteoporotic vertebral compression fractures at the West China Hospital of Sichuan University from May 2014 to April 2019 were retrospectively recruited. Vertebrae not treated during PVA were stratified based on their distance from the nearest augmented vertebra and the time elapsed since PVA. Survival curves were plotted to compare the risk of NSFs in untreated vertebrae at different distances from augmented vertebrae. The Cox proportional hazards model was used to identify risk factors of NSFs in untreated vertebrae.

Results: total, 162 patients with 228 NSFs (2.760 vertebrae) were analyzed. More than half of the NSFs (56.6%) occurred within the first year after PVA. Rates and hazard ratios (HRs) of NSFs were higher in vertebrae located one segment away from the augmented vertebrae (21.0%, HR: 3.99, P < 0.001), two segments away (10.6%, HR: 1.97, P = 0.003), or three segments away (10.5%, HR: 2.26, P < 0.001) than in vertebrae located five or more segments away (3.81%, HR: 1.00). Similar results were observed regardless of whether the untreated vertebrae were located in the thoracolumbar junction. In addition to distance, other risk factors of NSFs were the thoracolumbar location of untreated vertebrae, the number of augmented vertebrae, and percutaneous vertebroplasty.

Conclusion: The risk of NSFs is greater for untreated vertebrae located closer to augmented vertebrae than for untreated vertebrae further away. This distance dependence occurs mainly within the three segments closest to the augmented vertebra. The risk of NSFs decreases with time after augmentation, and it is also related to the number of augmented vertebrae, the type of augmentation, and whether the untreated vertebrae are thoracolumbar or not.

Purpose: In this case-control study, we aimed to evaluate how muscle volume affects pelvic floor function by analyzing the levator ani muscle (LAM) using volumetric segmentation in addition to standard magnetic resonance (MR) defecography assessments.

Methods: We enrolled 85 patients with varying degrees of pelvic floor dysfunction (PFD) and 85 age- and gender-matched controls in this retrospective study. All patients had MR defecography images, while all controls had pelvic MR images obtained for other reasons. Group comparisons were performed using independent samples t-tests and Mann-Whitney U tests. The receiver operating curve (ROC) was constructed to establish a cut-off value for a normal LAM volume. Interrater reliability was assessed by calculating the intraclass correlation coefficient. A P value of less than 0.05 was considered statistically significant.

Results: Volumetric measurements revealed that the control group had higher LAM volumes, and the ROC curve analysis indicated a cut-off value of 38934.3 mm3 with a sensitivity of 0.812 and specificity of 0.8 for PFD assessment using LAM volumetric measurement. Gender did not significantly affect LAM volume in the control group.

Conclusion: Alongside the useful structural and functional information acquired from MR defecography images, volumetric analysis, and three-dimensional reconstructions of LAM may help to improve the accuracy of the diagnosis.

Purpose: Non-invasive methods for predicting pathological complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT) can provide distinct leverage in the management of patients with locally advanced rectal cancer (LARC). This study aimed to investigate whether including the golden-angle radial sparse parallel (GRASP) dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) perfusion parameter (K_trans), in addition to tumor regression grading (TRG) and apparent diffusion coefficient (ADC) values, can improve the predictive ability for pCR.

Methods: Patients with LARC who underwent nCRT and subsequent surgery were included. The imaging parameters were compared between patients with and without pCR. Receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive ability of these parameters for pCR.

Results: A total of 111 patients were included in the study. A pCR was obtained in 32 patients (28.8%). MRI-based TRG (mrTRG) showed a negative correlation with pCR (r = -0.61, P < 0.001), and the average ADC value showed a positive correlation with pCR (r = 0.62, P < 0.001). Before nCRT, K_trans in the pCR group was significantly higher than in the non-pCR group (1.30 ± 0.24 vs. 0.88 ± 0.34, P < 0.001), but no difference was identified after nCRT. Following ROC curve analysis, the area under the curve (AUC) of mrTRG (level 1-2), average ADC value, and K_trans value for predicting pCR were 0.738 [95% confidence interval (CI): 0.65-0.82], 0.78 (95% CI: 0.69-0.86), and 0.84 (95% CI: 0.77-0.92), respectively. The model combining the three parameters had significantly higher predictive ability for pCR (AUC: 0.94, 95% CI: 0.88-0.98).

Conclusion: The use of a combination of the GRASP DCE-MRI K_trans with mrTRG and ADC can lead to a better pCR predictive performance.

Purpose: To examine the diagnostic performance for the longitudinal extent of extrahepatic bile duct (EHD) cancer on computed tomography (CT) after biliary drainage (BD) and investigate the appropriate timing of magnetic resonance imaging (MRI) acquisition.

Methods: This retrospective study included patients who underwent curative-intent surgery for EHD cancer and CT pre- and post-BD between November 2005 and June 2021. The biliary segment-wise longitudinal tumor extent was evaluated according to the 2019 Korean Society of Abdominal Radiology consensus recommendations, with pre-BD CT, post-BD CT, and both pre- and post-BD CT. The performance for tumor detectability was compared using generalized estimating equation (GEE) method. When preoperative MRI was performed, patients were divided into two subgroups according to the timing of MRI with respect to BD, and the performance of MRI obtained pre- and post-BD was compared.

Results: In 105 patients (mean age: 67 ± 8 years; 74 men and 31 women), the performance for tumor detectability was superior using both CT scans compared with using post-BD CT alone (reader 1: sensitivity, 72.6% vs. 64.6%, P < 0.001; specificity, 96.9% vs. 94.8%, P = 0.063; reader 2: sensitivity, 77.2% vs. 72.9%, P = 0.126; specificity, 97.5% vs. 94.2%, P = 0.003), and it was comparable with using pre-BD CT alone. In biliary segments with a catheter, higher sensitivity and specificity were observed using both CT scans than using post-BD CT (reader 1: sensitivity, 74.4% vs. 67.5%, P = 0.006; specificity, 92.4% vs. 88.0%, P = 0.068; reader 2: sensitivity, 80.5% vs. 74.4%, P = 0.013; specificity, 94.3% vs. 88.0%, P = 0.016). Post-BD MRI (n = 30) exhibited a comparable performance to pre-BD MRI (n = 55) (reader 1: sensitivity, 77.9% vs. 75.0%, P = 0.605; specificity, 97.2% vs. 94.9%, P = 0.256; reader 2: sensitivity, 73.2% vs. 72.6%, P = 0.926; specificity, 98.4% vs. 94.9%, P = 0.068).

Conclusion: Pre-BD CT provided better diagnostic performance in the preoperative evaluation of EHD cancer. The longitudinal tumor extent could be accurately assessed with post-BD MRI, which was similar to pre-BD MRI.

Clinical significance: The acquisition of pre-BD CT could be beneficial for the preoperative evaluation of EHD cancer when BD is planned. Post-BD MRI would not be significantly affected by BD in terms of the diagnostic performance of the longitudinal tumor extent.

Purpose: The present study compares the diagnostic performance of unenhanced computed tomography (CT) radiomics-based machine learning (ML) classifiers and a radiologist in cystic renal masses (CRMs).

Methods: Patients with pathologically diagnosed CRMs from two hospitals were enrolled in the study. Unenhanced CT radiomic features were extracted for ML modeling in the training set (Guangzhou; 162 CRMs, 85 malignant). Total tumor segmentation was performed by two radiologists. Features with intraclass correlation coefficients of >0.75 were screened using univariate analysis, least absolute shrinkage and selection operator, and bidirectional elimination to construct random forest (RF), decision tree (DT), and k-nearest neighbor (KNN) models. External validation was performed in the Zhuhai set (45 CRMs, 30 malignant). All images were assessed by a radiologist. The ML models were evaluated using calibration curves, decision curves, and receiver operating characteristic (ROC) curves.

Results: Of the 207 patients (102 women; 59.1 ± 11.5 years), 92 (41 women; 58.0 ± 13.7 years) had benign CRMs, and 115 (61 women; 59.8 ± 11.4 years) had malignant CRMs. The accuracy, sensitivity, and specificity of the radiologist's diagnoses were 85.5%, 84.2%, and 91.1%, respectively [area under the (ROC) curve (AUC), 0.87]. The ML classifiers showed similar sensitivity (94.2%-100%), specificity (94.7%-100%), and accuracy (94.3%-100%) in the training set. In the validation set, KNN showed better sensitivity, accuracy, and AUC than DT and RF but weaker specificity. Calibration and decision curves showed excellent and good results in the training and validation set, respectively.

Conclusion: Unenhanced CT radiomics-based ML classifiers, especially KNN, may aid in screening CRMs.

Purpose: To evaluate the efficacy of interventional radiological (IR) procedures in iatrogenic urinary tract injury and investigate the factors affecting the outcome.

Methods: Fifty-eight patients (21 male) with a mean age of 50.3 ± 15.8 years referred for iatrogenic urinary tract injury were enrolled in this study. Technical success was defined as (i) successful placement of a nephrostomy catheter within the renal pelvis and/or (ii) successful antegrade ureteral stent placement (double J stent) between the renal pelvis and bladder lumen. Complete resolution was defined as maintained ureteral patency without an external drain and ureteral stent. The factors that may affect complete resolution [ureteral avulsion, ureterovaginal fistula (UVF), history of malignancy/radiotherapy, and time to IR management] were also investigated. The receiver operating characteristic analysis was performed to estimate the cut-off time point for the IR management timing affecting complete resolution.

Results: The technical success rate for nephrostomy and ureteral stent placement was 100% (n = 58/58) and 78% (n = 28/36), respectively. In 14 patients, non-dilated pelvicalyceal systems were evident. In 18 patients, no further intervention after percutaneous nephrostomy was performed due to (i) poor performance status (n = 6) and (ii) reconstruction surgery upon clinicians' and/or patients' request (n = 12). Reconstruction surgery was required in 11 of the remaining 40 patients due to failure of percutaneous treatment (n = 11/40, 27.5%). In six of the patients, ureteral stents could not be removed due to the development of benign ureteral strictures (n = 6/40, 15%). Our complete resolution rate was 57.5% (n = 23/40). Age, gender, type of surgery (endoscopic or open), side and location of the injury did not statistically affect the complete resolution rate. The presence of ureteral avulsion, history of malignancy and radiotherapy individually or in combination significantly affected the complete resolution rate negatively. The presence of UVF also had a negative effect on the complete resolution rate; however, it did not reach statistical significance. Delayed intervention was also a significant factor related to lower complete resolution. The optimal cut-off point of the time interval for favorable clinical outcome was found to be 0–19^th day following the surgery.

Conclusion: IR procedures are safe and effective in the management of iatrogenic urinary tract injuries. Antegrade ureteral stenting should be performed as soon as possible to establish ureteral integrity without the development of stricture.

Although artificial intelligence (AI) methods hold promise for medical imaging-based prediction tasks, their integration into medical practice may present a double-edged sword due to bias (i.e., systematic errors). AI algorithms have the potential to mitigate cognitive biases in human interpretation, but extensive research has highlighted the tendency of AI systems to internalize biases within their model. This fact, whether intentional or not, may ultimately lead to unintentional consequences in the clinical setting, potentially compromising patient outcomes. This concern is particularly important in medical imaging, where AI has been more progressively and widely embraced than any other medical field. A comprehensive understanding of bias at each stage of the AI pipeline is therefore essential to contribute to developing AI solutions that are not only less biased but also widely applicable. This international collaborative review effort aims to increase awareness within the medical imaging community about the importance of proactively identifying and addressing AI bias to prevent its negative consequences from being realized later. The authors began with the fundamentals of bias by explaining its different definitions and delineating various potential sources. Strategies for detecting and identifying bias were then outlined, followed by a review of techniques for its avoidance and mitigation. Moreover, ethical dimensions, challenges encountered, and prospects were discussed.

Teleconferencing can facilitate a multidisciplinary approach to teaching radiology to medical students. This study aimed to determine whether an online learning approach enables students to appreciate the interrelated roles of radiology and other specialties during the management of different medical cases. Turkish medical students attended five 60-90-minute online lectures delivered by radiologists and other specialists from the United States and Canada through Zoom meetings between November 2020 and January 2021. Student ambassadors from their respective Turkish medical schools recruited their classmates with guidance from the course director. Students took a pretest and posttest to assess the knowledge imparted from each session and a final course survey to assess their confidence in radiology and the value of the course. A paired t-test was used to assess pretest and posttest score differences. A 4-point Likert-type scale was used to assess confidence rating differences before and after attending the course sessions. A total of 1,458 Turkish medical students registered for the course. An average of 437 completed both pre- and posttests when accounting for all five sessions. Posttest scores were significantly higher than pretest scores for each session (P < 0.001). A total of 546 medical students completed the final course survey evaluation. Students' rating of their confidence in their radiology knowledge increased after taking the course (P < 0.001). Students who took our course gained an appreciation for the interrelated roles of different specialties in approaching medical diagnoses and interpreting radiological findings. These students also reported an increased confidence in radiology topics and rated the course highly relevant and insightful. Overall, our findings indicated that multidisciplinary online education can be feasibly implemented for medical students by video teleconferencing.