Abdominal Radiology最新文献

Purpose

To development of a preoperative CT-based radiomics model for predicting muscle invasion in patients with upper tract urothelial carcinoma below T3 stage.

Methods

163 consecutive patients who underwent radical nephroureterectomy for stage pT1–2 UTUC were retrospectively enrolled two medical centers (116 patients in training data and 47 patients in external validation data). Lesion segmentation, extraction and selection of radiomic features on pre-surgical CT urography, development and validation of predictive models were performed. Risk stratification of UTUC was evaluated. The diagnostic performance of the radiomics model and risk stratification was analyzed. Reference standard was histopathological analysis.

Results

Among 163 patients (mean age, 52 years ± 9 [standard deviation], 97 men), 61.5% had pT2 grade tumors. 1165 features with intraclass coefficients > 0.75 were retained for least absolute shrinkage and selection operator (LASSO) regression. Nine radiomic features with non-zero coefficients on LASSO regression were selected from the training dataset and used for constructing the radiomics model. Good discrimination capability of the predictive model was observed, as AUCs were 0.859 (95% CI, 0.782–0.917) in the training dataset and 0.821 (95% CI, 0.682–0.918) in the validation dataset, respectively. Based on judgement by the model, When the tumor length diameter > 3 cm, combining ureteroscopy biopsy would improve sensitivity and NPV to 0.86 (95% CI, 0.776–0.922), 0.81 (95% CI, 0.714–0.903).

Conclusion

The preoperative radiomics model showed promising diagnostic performance in predicting UTUC muscle invasion. This could help patients receive more accurate risk classification, especially help patients avoiding radical nephroureterectomy.

Graphical abstract

To development of a preoperative CT-based radiomics model for predicting muscle invasion in patients with upper tract urothelial carcinoma below T3 stage, 163 consecutive patients who underwent radical nephroureterectomy were retrospectively enrolled two medical centers. Nine radiomic features with non-zero coefficients on LASSO regression were selected. Good discrimination capability of the predictive model was observed, as AUCs were 0.859 (95% CI, 0.782-0.917) in the training dataset and 0.821 (95% CI, 0.682-0.918) in the validation dataset, respectively.

Purpose

High-grade clear cell renal cell carcinoma (ccRCC) is linked to lower survival rates and more aggressive disease progression. This study aims to assess the diagnostic performance of MRI-derived radiomics as a non-invasive approach for pre-operative differentiation of high-grade from low-grade ccRCC.

Methods

A systematic search was conducted across PubMed, Scopus, and Embase. Quality assessment was performed using QUADAS-2 and METRICS. Pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and area under the curve (AUC) were estimated using a bivariate model. Separate meta-analyses were conducted for radiomics models and combined models, where the latter integrated clinical and radiological features with radiomics. Subgroup analysis was performed to identify potential sources of heterogeneity. Sensitivity analysis was conducted to identify potential outliers.

Results

A total of 15 studies comprising 2,265 patients were included, with seven and six studies contributing to the meta-analysis of radiomics and combined models, respectively. The pooled estimates of the radiomics model were as follows: sensitivity, 0.78; specificity, 0.84; PLR, 4.17; NLR, 0.28; DOR, 17.34; and AUC, 0.84. For the combined model, the pooled sensitivity, specificity, PLR, NLR, DOR, and AUC were 0.87, 0.81, 3.78, 0.21, 28.57, and 0.90, respectively. Radiomics models trained on smaller cohorts exhibited a significantly higher pooled specificity and PLR than those trained on larger cohorts. Also, radiomics models based on single-user segmentation demonstrated a significantly higher pooled specificity compared to multi-user segmentation.

Conclusion

Radiomics has demonstrated potential as a non-invasive tool for grading ccRCC, with combined models achieving superior performance.

Graphical Abstract

Purpose

The aim of this study was to develop and validate CT venous phase image-based radiomics to predict BRAF gene mutation status in preoperative colorectal cancer patients.

Methods

In this study, 301 patients with pathologically confirmed colorectal cancer were retrospectively enrolled, comprising 225 from Centre I (73 mutant and 152 wild-type) and 76 from Centre II (36 mutant and 40 wild-type). The Centre I cohort was randomly divided into a training set (n = 158) and an internal validation set (n = 67) in a 7:3 ratio, while Centre II served as an independent external validation set (n = 76). The whole tumor region of interest was segmented, and radiomics characteristics were extracted. To explore whether tumor expansion could improve the performance of the study objectives, the tumor contour was extended by 3 mm in this study. Finally, a t-test, Pearson correlation, and LASSO regression were used to screen out features strongly associated with BRAF mutations. Based on these features, six classifiers—Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), K-Nearest Neighbors (KNN), and Extreme Gradient Boosting (XGBoost)—were constructed. The model performance and clinical utility were evaluated using receiver operating characteristic (ROC) curves, decision curve analysis, accuracy, sensitivity, and specificity.

Results

Gender was an independent predictor of BRAF mutations. The unexpanded RF model, constructed using 11 imaging histologic features, demonstrated the best predictive performance. For the training cohort, it achieved an AUC of 0.814 (95% CI 0.732–0.895), an accuracy of 0.810, and a sensitivity of 0.620. For the internal validation cohort, it achieved an AUC of 0.798 (95% CI 0.690–0.907), an accuracy of 0.761, and a sensitivity of 0.609. For the external validation cohort, it achieved an AUC of 0.737 (95% CI 0.616–0.847), an accuracy of 0.658, and a sensitivity of 0.667.

Conclusions

A machine learning model based on CT radiomics can effectively predict BRAF mutations in patients with colorectal cancer. The unexpanded RF model demonstrated optimal predictive performance.

Objectives

Adequate bowel preparation and tagging are critical in optimizing CTC performance. Iohexol has a higher safety profile than other available tagging agents. This study aims to determine if iohexol serves as an adequate fluid and stool tagging agent in conjunction with minimally cathartic bowel preparation.

Methods

In this prospective observational study, 50 participants ingested 50 mL of oral iohexol for tagging and 10 oz magnesium citrate for bowel preparation prior to CTC. Written informed consent was obtained. CTC was performed in all participants in at least two of the standard four positions (supine, prone, right decubitus, and left decubitus). Two board-certified abdominal radiologists independently scored the 6 colonic segments of participants who underwent successful CTC. The amount of residual fluid and solid stool, attenuation of tagged fluid, and efficacy of fluid and stool tagging were recorded in each segment. Statistical analyses were performed with R-4.4.0.

Results

47 participants (mean age 66.39 ± 8.65 years; 39 female) underwent successful CTC. Of 1252 total colonic segments, 14.8% had no residual fluid and 59.5% had < 25% residual fluid. 73.6% of segments with residual fluid demonstrated good tagging. The mean fluid tagging efficacy ratio for all segments was 0.737 (95% CI: 0.700–0.775) with mean attenuation of 467 HU. Fluid tagging efficacy decreased from the cecum (0.934) to rectum (0.493). 92.8% of segments had no residual solid stool. Of the 7.2% of segments containing solid stool, 4.7% of segments had submerged stool ≤ 5 mm, 0.8% had 1–3 pieces of retained stool between 6 and 9 mm, and 1.8% had > 3 pieces 6–9 mm or single pieces > 1 cm.

Conclusion

Low-volume (50 mL) iohexol is an effective fluid and fecal tagging agent for CTC with a minimally cathartic bowel preparation. This provides an easy option to label residual material and cleanse the bowel for patients undergoing CTC.

Graphical abstract

Bilateral adnexal lesions involve structures such as the ovaries, fallopian tubes, and surrounding tissues, arising from diverse etiologies, including inflammatory, infectious, neoplastic, and functional causes. Their variable presentation poses a diagnostic challenge in clinical practice, necessitating a multidisciplinary approach for accurate assessment and management. The American College of Radiology (ACR) introduced the Ovarian-Adnexal Reporting and Data System (O-RADS) as a standardized lexicon and risk stratification tool for evaluating adnexal lesions via ultrasound (US) and magnetic resonance imaging (MRI). While MRI is the most accurate modality for assessing indeterminate adnexal masses, bilateral lesions frequently present diagnostic dilemmas, particularly when they exhibit divergent O-RADS classifications or arise from different etiologies. The O-RADS system does not provide specific guidelines for bilateral lesions, requiring independent classification of each lesion, with management dictated by the highest assigned category. Certain pathologies demonstrate a propensity for bilateral involvement, underscoring the importance of recognizing their imaging characteristics and differential diagnoses. Integrating this knowledge into diagnostic reports enhances clinical decision-making and optimizes patient outcomes.

Graphical Abstract

Objective

To evaluate the possibility of performing renal vessel reconstruction on non-enhanced CT images using deep learning models.

Materials and methods

177 patients’ CT scans in the non-enhanced phase, arterial phase and venous phase were chosen. These data were randomly divided into the training set (n = 120), validation set (n = 20) and test set (n = 37). In training set and validation set, a radiologist marked out the right renal arteries and veins on non-enhanced CT phase images using contrast phases as references. Trained deep learning models were tested and evaluated on the test set. A radiologist performed renal vessel reconstruction on the test set without the contrast phase reference, and the results were used for comparison. Reconstruction using the arterial phase and venous phase was used as the gold standard.

Results

Without the contrast phase reference, both radiologist and model could accurately identify artery and vein main trunk. The accuracy was 91.9% vs. 97.3% (model vs. radiologist) in artery and 91.9% vs. 100% in vein, the difference was insignificant. The model had difficulty identify accessory arteries, the accuracy was significantly lower than radiologist (44.4% vs. 77.8%, p = 0.044). The model also had lower accuracy in accessory veins, but the difference was insignificant (64.3% vs. 85.7%, p = 0.094).

Conclusion

Deep learning models could accurately recognize the right renal artery and vein main trunk, and accuracy was comparable to that of radiologists. Although the current model still had difficulty recognizing small accessory vessels, further training and model optimization would solve these problems.

Purpose

Diagnosis of peritoneal invasion, lymph node metastasis, and hepatic metastasis is crucial in the decision-making process of ovarian tumor treatment. This study aimed to test the feasibility of low-dose abdominopelvic CT with an artificial intelligence iterative reconstruction (AIIR) for diagnosing peritoneal invasion, lymph node metastasis, and hepatic metastasis in pre-operative imaging of ovarian tumor.

Methods

This study prospectively enrolled 88 patients with pathology-confirmed ovarian tumors, where routine-dose CT at portal venous phase (120 kVp/ref. 200 mAs) with hybrid iterative reconstruction (HIR) was followed by a low-dose scan (120 kVp/ref. 40 mAs) with AIIR. The performance of diagnosing peritoneal invasion and lymph node metastasis was assessed using receiver operating characteristic (ROC) analysis with pathological results serving as the reference. The hepatic parenchymal metastases were diagnosed and signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were measured. The perihepatic structures were also scored on the clarity of porta hepatis, gallbladder fossa and intersegmental fissure.

Results

The effective dose of low-dose CT was 79.8% lower than that of routine-dose scan (2.64 ± 0.46 vs. 13.04 ± 2.25 mSv, p < 0.001). The low-dose AIIR showed similar area under the ROC curve (AUC) with routine-dose HIR for diagnosing both peritoneal invasion (0.961 vs. 0.960, p = 0.734) and lymph node metastasis (0.711 vs. 0.715, p = 0.355). The 10 hepatic parenchymal metastases were all accurately diagnosed on the two image sets. The low-dose AIIR exhibited higher SNR and CNR for hepatic parenchymal metastases and superior clarity for perihepatic structures.

Conclusion

In low-dose pre-operative CT of ovarian tumor, AIIR delivers similar diagnostic accuracy for peritoneal invasion, lymph node metastasis, and hepatic metastasis, as compared to routine-dose abdominopelvic CT. It is feasible and diagnostically safe to apply up to 80% dose reduction in CT imaging of ovarian tumor by using AIIR.

Graphical abstract

Purpose

To evaluate the clinicoradiological significance of intrahepatic periportal hyperintensity (PHI) detected by gadoxetate-enhanced hepatobiliary phase (HBP) MRI and T2-weighted imaging (T2WI), and to assess its potential as a noninvasive imaging biomarker for clinical stratification of liver injury in patients with cirrhosis.

Methods

This retrospective study included 37 cirrhotic patients with intrahepatic diffuse PHI on HBP imaging, who underwent gadoxetate-enhanced MRI between October 2019 and November 2023. PHI patterns were classified into two groups based on the spatial concordance between periportal enhancement areas on HBP and periportal hyperintense areas on T2WI. The matching group (Type A, n = 21) demonstrated complete spatial overlap between the two sequences. The mismatching group, comprised Type B (n = 11), in which PHI on HBP was immediately outside of that on T2WI, and Type C (n = 5), in which PHI was present on HBP but absent on T2WI. Clinical etiologies and liver biochemical markers (ALT, AST, GGT, TBil, DBil, ALP, Alb, TP) were compared across PHI subtypes.

Results

Type A PHI was predominantly associated with acute liver injury (e.g., acute viral hepatitis flares, drug-induced liver injury, autoimmune hepatitis), characterized by a strong ALT-AST correlation (r = 0.95, P < 0.001) and significantly elevated levels of ALT, AST, GGT, TBil, and DBil (all P < 0.001). In contrast, Types B and C PHI were primarily linked to chronic fibrotic conditions (e.g., HBV/HCV-related cirrhosis, primary biliary cholangitis, and primary sclerosing cholangitis), showing a strong TBil-DBil correlation (r = 0.95, P < 0.001) and moderately elevated ALP and Alb levels (P = 0.027 and P = 0.017, respectively). Receiver operating characteristic (ROC) analysis identified DBil > 37.5 μmol/L as the optimal threshold for differentiating Type A from Types B/C PHI (AUC = 0.922; sensitivity = 86.7%, specificity = 100%). Notably, HBP-doughnut nodules without arterial-phase hyperenhancement (APHE) were exclusively observed in the mismatching group (Type B: 4/11; Type C: 3/5), further supporting their association with chronic fibrotic changes.

Conclusion

PHI phenotyping based on HBP-T2WI spatial concordance enables accurate, noninvasive differentiation between acute inflammatory and chronic fibrotic liver injury in cirrhotic patients. When integrated with the DBil threshold, this imaging-based approach provides as a robust biomarker for clinical stratification of liver injury and may facilitate individualized diagnosis and therapeutic decision-making in chronic liver disease.

Focal liver lesions (FLLs) are common and are often first identified on abdominal ultrasound examinations. Although CT and MRI were historically required to noninvasively characterize many FLLs, introduction of microbubble contrast agents produced a groundbreaking change as contrast enhanced ultrasound (CEUS) showed vascularity to the capillary level for the first time. CEUS shows specific arterial phase enhancement patterns in benign lesions and accurately differentiates malignant lesions based on the timing and intensity of washout. Parametric time of arrival and microvascular imaging techniques can demonstrate vascularity in FLLs with significantly improved sensitivity compared with conventional Doppler techniques. Shear-wave elastography and quantitative ultrasound are generally used to evaluate diffuse liver disease but show promise in evaluation of FLLs.

Graphical abstract

Body packing, a method used to traffic illicit drugs, primarily involves the gastrointestinal tract as a concealment route. Commonly trafficked substances include cocaine, heroin, marijuana, methamphetamine, and cannabis, often sealed in handmade latex packets characterized by specific imaging signs. Prompt diagnosis is crucial for initiating appropriate treatment, recognizing complications, and ensuring proper medico-legal handling. Abdominal radiographs are the preferred initial imaging modality due to their low cost and widespread availability, though their sensitivity varies depending on packet size, location, and interpreter expertise. Abdominopelvic non-contrast CT is the gold standard for detecting gastrointestinal packages, offering high sensitivity and specificity. Low-dose CT protocols are recommended to minimize radiation exposure without compromising diagnostic accuracy, particularly for follow-up or in cases without complications. Contrast-enhanced CT is reserved for assessing suspected complications such as bowel obstruction or perforation. This pictorial review highlights key imaging findings correlated with clinical features, aiming to facilitate accurate recognition, timely intervention, and prevention of complications in suspected cases.

Graphical abstract