Abdominal Radiology最新文献

The field of radiology education is undergoing a paradigm shift due to technological advancements and the increasing complexity of medical imaging. Traditional didactic teaching methods are progressively being supplemented or replaced by innovative pedagogical approaches that enhance engagement, competency, and clinical preparedness. This review examines the evolution of radiology education, highlighting novel teaching methodologies such as simulation-based training, artificial intelligence assisted learning, virtual and augmented reality, flipped classrooms, and case-based learning. Furthermore, this manuscript discusses the challenges of integrating these methodologies into radiology curricula and explores potential future directions in radiology education.

Background

Intrahepatic cholestasis of pregnancy (ICP) is a liver disorder that occurs during pregnancy, typically in the third trimester, and is associated with adverse fetal outcomes. Despite being well-recognized, the exact pathogenesis of ICP remains incompletely understood, with impaired hepatobiliary function hypothesized to play a significant role in its development.

Purpose

The study aims to evaluate the relationship between altered gallbladder motility and hepatobiliary dysfunction in pregnancies complicated by ICP.

Materials and methods

This was a prospective hospital-based study involving 90 participants: 30 non-pregnant controls and 60 pregnant women (30 with intrahepatic cholestasis of pregnancy (ICP), and 30 pregnant controls). The cholestasis group was further subdivided into an icteric group (elevated serum bilirubin levels) and an anicteric group (abnormal liver function tests but normal bilirubin levels). Gallbladder volume (GBV) was initially measured after an overnight fast (12–14 h). Following a standardized meal (75 g of bread and butter or 100 g of groundnut), GBV and ejection fraction were re-measured at multiple postprandial time points (15, 30, 45, 60, 75, and 90 min).

Results

The study enrolled 60 pregnant women in their third trimester (gestational weeks 28–40) and 30 non-pregnant controls. Among the pregnant women, 30 were diagnosed with ICP, which was further divided into an icteric group (n = 6) and an anicteric group (n = 24). The ICP group demonstrated consistently higher GB volumes and lower ejection fraction (EF) compared to non-pregnant and pregnant controls, with statistically significant differences across all time intervals. Within the ICP patients, the icteric group consistently exhibited higher GB volumes and lower EF values compared to anicteric group. Receiver operating characteristic (ROC) curve analysis revealed that the best post-prandial 15-min GB ejection fraction cutoff for predicting ICP was 53% and postprandial 15-min GB volume cutoff for predicting ICP was 9.5 ml,

Conclusion

Impaired gallbladder motility, characterized by increased GB volume and decreased EF, is a key feature of obstetric cholestasis. Gallbladder ultrasound measurements, particularly postprandial GB volume and EF, can serve as useful diagnostic tools for distinguishing between cholestasis of pregnancy and healthy pregnancies, with high sensitivity and specificity for identifying ICP in late pregnancy.

The kidney plays a crucial role in regulating the acid-base balance of the body. The development of kidney diseases is often accompanied by changes in the renal PH, protein, glucose, urea, and other metabolites. Chemical exchange saturation transfer (CEST) imaging is a new magnetic resonance imaging (MRI) technique based on the chemical exchange between solute and water protons. The potential of CEST imaging in renal metabolic imaging stems from its ability to provide image contrast based on specific molecular compositions at relevant millimolar physiological concentrations. In this article, we briefly introduce the fundamental principles of CEST imaging and its applications in various renal disorders, including acute kidney injury, chronic kidney disease, renal tumors, and renal allograft rejection. Finally, we discuss the challenges and prospects of CEST imaging in the field of kidney diseases.

Purpose

To evaluate the frequency and independent risk factors associated with transient severe motion artifact (TSM) during the arterial phase of gadoxetic acid–enhanced liver MRI in a Southeast Asian population.

Materials and methods

This retrospective study included 836 consecutive studies who underwent gadoxetic acid–enhanced MRI between October 2022 and October 2024 at a tertiary academic hospital. Two abdominal radiologists reviewed arterial phase images and graded motion artifacts using a validated 5-point scale; grades 4 and 5 were classified as TSM. Clinical and laboratory data were extracted from the electronic medical record. Univariable and multivariable logistic regression analyses were performed to identify predictors of TSM. Subgroup analyses were conducted based on age and body mass index (BMI).

Results

TSM artifacts were observed in 43 of 836 studies (5.14%). In multivariable analysis, older age (> 65 years) (adjusted OR 4.59; 95% CI 1.268–16.710; p = 0.021) and lower serum albumin (adjusted OR 0.33; 95% CI 0.143–0.761; p = 0.009) were independent predictors of TSM. Subgroup analyses demonstrated higher TSM incidence in patients aged ≥ 65 years (7.07% vs. 3.41%; p = 0.019) and in those with BMI > 30 kg/m² (11.11% vs. 4.73%; p = 0.052). Other variables, including sex, comorbidities, liver disease etiology, and fluid overload, were not significantly associated with TSM.

Conclusion

TSM during gadoxetic acid–enhanced liver MRI occurs in approximately 5% of patients and is independently associated with older age and lower serum albumin. Awareness of these risk factors can guide protocol optimization and personalized imaging strategies to improve arterial phase image quality.

Purpose

To explore the value of ultrasound (US) and contrast-enhanced ultrasound (CEUS) in differentiating mass-forming Pancreatitis (MFP) and pancreatic ductal adenocarcinoma (PDAC).

Methods

This retrospective study analyzed clinical and imaging data from 281 patients who underwent pancreatic CEUS between January 2018 and December 2023. Patients were randomly divided into training (n = 196) and validation (n = 85) sets. Logistic regression analyses were conducted to identify independent predictive imaging features for differentiating PDAC from MFP in the training set. Based on the identified predictors, two nomogram models were constructed: the US model and the US + CEUS model. The diagnostic performance of both models was assessed via the area under the receiver operating characteristic curve (AUC), calibration plots, Hosmer-Lemeshow test, and decision-curve analysis (DCA).

Results

Multivariate logistic regression analysis based on these factors identified taller-than-wide shape (P = 0.002, OR = 0.12), calcification (P = 0.003, OR = 13.76), and washout pattern (P = 0.002, OR = 0.13) as independent predictive factors for distinguishing PDAC from MFP. Compared to the US model, the US + CEUS model demonstrated better performance with AUC values 0.930 (95% CI: 0.895–0.965) in the training set and 0.914 (95% CI: 0.853–0.976) in the validation set. Calibration curve plots and the Hosmer-Lemeshow test (P > 0.05) confirmed that the model has good calibration, and DAC showed significant clinical benefit.

Conclusion

The nomogram model constructed using taller-than-wide shape, calcification, and washout pattern demonstrated excellent discriminative ability, accuracy, and clinical utility in differentiating PDAC from MFP.

Purpose

Investigate spleen diameter (d) and volume (v) relationship in patients with hematologic malignancies (HM) by determining volumetric thresholds that best correlate to established diameter thresholds for assessing response to treatment. Exploratorily, interrogate the impact of volumetric measurements in response categories and as a predictor of response.

Methods

Secondary analysis of prospectively collected clinical trial data of 382 patients with HM. Spleen diameters were computed following Lugano criteria and volumes using deep learning segmentation. d and v relationship was estimated using power regression model, volumetric thresholds ((:{v}_{threshold})) for treatment response estimated; threshold search to determine percentual change ((:{v}_{%}),) and minimum volumetric increase ((:{v}_{increase})) that maximize agreement with Lugano criteria performed. Spleen diameter and volume predictive performance for clinical response investigated using random forest model.

Results

(:v=2.24times:{d}^{2.14}) describes the relationship between spleen diameter and volume. (:{v}_{threshold}) for splenomegaly was 546 cm³. (:{v}_{threshold}), (:{v}_{%}), and (:{v}_{increase}) for assessing response resulting in highest agreement with Lugano criteria were 570 cm³, 73%, and 170 cm³, respectively. Predictive performance for response between diameter and volume were not significantly different (P=0.78).

Conclusion

This study provides empirical spleen volume threshold and percentual changes that best correlate with diameter thresholds, i.e., Lugano criteria, for assessment of response to treatment in patients with HM. In our dataset use of spleen volumetric thresholds versus diameter thresholds resulted in similar response assessment categories and did not signal differences in predictive values for response.

Rectal cancer (RC) is one of the most common malignant tumors of the digestive system and has an alarmingly high incidence and mortality rate globally. Compared to conventional imaging examinations, radiomics can extract quantitative features that reflect tumor heterogeneity and mine data from medical images. In this review, we discuss the potential value of multimodal MRI-based radiomics in the diagnosis and treatment of RC, with a special emphasis on the role of peritumoral tissue characteristics in clinical decision-making. Existing studies have shown that a radiomics model integrating intratumoral and peritumoral characteristics has good application prospects in RC staging evaluation, efficacy prediction, metastasis monitoring, recurrence early warning, and prognosis judgment. At the same time, this paper also objectively analyzes the existing methodological limitations in this field, including insufficient data standardization, inadequate model validation, limited sample size and poor reproducibility of results. By combining existing evidence, this review aimed to enhance the attention of clinicians and radiologists on the characteristics of peritumoral tissues and promote the translational application of radiomics technology in the individualized treatment of RC.

Graphical abstract

Objectives

To train and evaluate the performance of a machine learning triaging tool that identifies MRI negative for clinically significant prostate cancer and to compare this against non-MRI models.

Methods

2895 MRIs were collected from two sources (1630 internal, 1265 public) in this retrospective study. Risk models compared were: Prostate Cancer Prevention Trial Risk Calculator 2.0, Prostate Biopsy Collaborative Group Calculator, PSA density, U-Net segmentation, and U-Net combined with clinical parameters. The reference standard was histopathology or negative follow-up. Performance metrics were calculated by simulating a triaging workflow compared to radiologist interpreting all exams on a test set of 465 patients. Sensitivity and specificity differences were assessed using the McNemar test. Differences in PPV and NPV were assessed using the Leisenring, Alonzo and Pepe generalized score statistic. Equivalence test p-values were adjusted within each measure using Benjamini–Hochberg correction.

Results

Triaging using U-Net with clinical parameters reduced radiologist workload by 12.5% with sensitivity decrease from 93 to 90% (p = 0.023) and specificity increase from 39 to 47% (p < 0.001). This simulated workload reduction was greater than triaging with risk calculators (3.2% and 1.3%, p < 0.001), and comparable to PSA density (8.4%, p = 0.071) and U-Net alone (11.6%, p = 0.762). Both U-Net triaging strategies increased PPV (+ 2.8% p = 0.005 clinical, + 2.2% p = 0.020 nonclinical), unlike non-U-Net strategies (p > 0.05). NPV remained equivalent for all scenarios (p > 0.05). Clinically-informed U-Net triaging correctly ruled out 20 (13.4%) radiologist false positives (12 PI-RADS = 3, 8 PI-RADS = 4). Of the eight (3.6%) false negatives, two were misclassified by the radiologist. No misclassified case was interpreted as PI-RADS 5.

Conclusions

Prostate MRI triaging using machine learning could reduce radiologist workload by 12.5% with a 3% sensitivity decrease and 8% specificity increase, outperforming triaging using non-imaging-based risk models. Further prospective validation is required.

Objective

The diagnosis of liver fibrosis is usually based on histopathological examination of liver puncture specimens. Although liver puncture is accurate, it has invasive risks and high economic costs, which are difficult for some patients to accept. Therefore, this study uses deep learning technology to build a liver fibrosis diagnosis model to achieve non-invasive staging of liver fibrosis, avoid complications, and reduce costs.

Methods

This study uses ultrasound examination to obtain pure liver parenchyma image section data. With the consent of the patient, combined with the results of percutaneous liver puncture biopsy, the degree of liver fibrosis indicated by ultrasound examination data is judged. The concept of Fibrosis Contrast Layer (FCL) is creatively introduced in our experimental method, which can help our model more keenly capture the significant differences in the characteristics of liver fibrosis of various grades. Finally, through label fusion (LF), the characteristics of liver specimens of the same fibrosis stage are abstracted and fused to improve the accuracy and stability of the diagnostic model.

Results

Experimental evaluation demonstrated that our model achieved an accuracy of 85.6%, outperforming baseline models such as ResNet (81.9%), InceptionNet (80.9%), and VGG (80.8%). Even under a small-sample condition (30% data), the model maintained an accuracy of 84.8%, significantly outperforming traditional deep-learning models exhibiting sharp performance declines.

Conclusion

The training results show that in the whole sample data set and 30% small sample data set training environments, the FCLLF model’s test performance results are better than those of traditional deep learning models such as VGG, ResNet, and InceptionNet. The performance of the FCLLF model is more stable, especially in the small sample data set environment. Our proposed FCLLF model effectively improves the accuracy and stability of liver fibrosis staging using non-invasive ultrasound imaging.

Purpose

High-quality imaging is critical for accurate prostate cancer assessment using MRI. This study describes a quality improvement (QI) initiative for prostate MRI and evaluates the impact of image quality on diagnostic performance.

Methods

In this prospective study, 1328 patients underwent prostate MRI between March 2023 and March 2024. The QI initiative focused on patient preparation, protocol standardization, technologist education, and quality control. Image quality was prospectively scored using the Prostate Imaging Quality (PI-QUAL) system. Diagnostic performance was assessed by comparing the positive predictive value (PPV) of PI-RADS ≥ 3 lesions before and after the intervention and between exams with optimal (PI-QUAL ≥ 4) and suboptimal quality. Fisher’s exact test was used for comparisons. Clinically significant prostate cancer (csPCa) was defined as Gleason Grade Group ≥ 2.

Results

The proportion of PI-QUAL ≥ 4 exams increased from 67 to 84% after the intervention (p < 0.001). The PPV for any prostate cancer improved from 65 to 78% (p = 0.03), though the increase for csPCa (from 42 to 46%) was not statistically significant (p = 0.5). No significant difference in PPV was observed between optimal and suboptimal exams for any cancer (76% vs. 77%, p = 0.9) or csPCa (45% vs. 48%, p = 0.7).

Conclusion

The QI initiative significantly improved image quality and overall cancer detection rates. However, the association between image quality and csPCa detection was not statistically significant, possibly due to greater improvements in T2-weighted images vs. diffusion weighted images and high reader expertise.