Journal of Computer Assisted Tomography最新文献

Objective: This study aims to evaluate the predictive value of multiparameter characteristics of coronary computed tomography angiography (CCTA) plaque and the ratio of coronary artery volume to myocardial mass (V/M) in guiding percutaneous coronary stent implantation (PCI) in patients diagnosed with unstable angina.

Methods: Patients who underwent CCTA and coronary angiography (CAG) within 2 months were retrospectively analyzed. According to CAG results, patients were divided into a medical therapy group (n=41) and a PCI revascularization group (n=37). The plaque characteristics and V/M were quantitatively evaluated. The parameters included minimum lumen area at stenosis (MLA), maximum area stenosis (MAS), maximum diameter stenosis (MDS), total plaque burden (TPB), plaque length, plaque volume, and each component volume within the plaque. Fractional flow reserve (FFR) and pericoronary fat attenuation index (FAI) were calculated based on CCTA. Artificial intelligence software was employed to compare the differences in each parameter between the 2 groups at both the vessel and plaque levels.

Results: The PCI group had higher MAS, MDS, TPB, FAI, noncalcified plaque volume and lipid plaque volume, and significantly lower V/M, MLA, and CT-derived fractional flow reserve (FFRCT). V/M, TPB, MLA, FFRCT, and FAI are important influencing factors of PCI. The combined model of MLA, FFRCT, and FAI had the largest area under the ROC curve (AUC=0.920), and had the best performance in predicting PCI.

Conclusions: The integration of AI-derived multiparameter features from one-stop CCTA significantly enhances the accuracy of predicting PCI in angina pectoris patients, evaluating at the plaque, vessel, and patient levels.

In the rapidly evolving landscape of medical education, artificial intelligence (AI) holds transformative potential. This manuscript explores the integration of large language models (LLMs) in Radiology education and training. These advanced AI tools, trained on vast data sets, excel in processing and generating human-like text, and have even demonstrated the ability to pass medical board exams. In radiology, LLMs enhance clinical education by providing interactive training environments that improve diagnostic skills and structured reporting. They also support research by streamlining literature reviews and automating data analysis, thus boosting productivity. However, their integration raises significant challenges, including the risk of over-reliance on AI, ethical concerns related to patient privacy, and potential biases in AI-generated content. This commentary from the Early Career Committee of the Society for Advanced Body Imaging (SABI) offers insights into the current applications and future possibilities of LLMs in Radiology education while being mindful of their limitations and ethical implications to optimize their use in the health care system.

Purpose: To compare the scan timing adequacy for the pancreatic phase between fixed and tailored scan delay in the pancreatic protocol CT with a bolus-tracking technique.

Materials and methods: This retrospective study included patients who underwent pancreatic protocol CT using a fixed scan delay of 20 s from January 2020 to November 2022 (conventional group) and those using a tailored scan delay from January 2023 to July 2024 (tailored group). Tailored scan delay was identified to be the same as the time from contrast injection to reaching to trigger threshold of 100 HU (Time TRIG ). The scan delay ratio (SDR) was calculated by dividing the scan delay by Time TRIG . Two radiologists assessed the scan timing adequacy for the pancreatic phase and classified it into 3 categories: early, appropriate, and late. The SDR and scan timing adequacy were compared between the conventional and tailored groups.

Results: This study involved 128 patients (75 men; median age, 71 y), including 63 and 65 in the conventional and tailored groups, respectively. The median SDR was significantly different between the two groups (1.2 and 1.0 in the conventional and tailored groups; P <0.001). The proportion of appropriate scan timing for the pancreatic phase was higher in the tailored group (55/65; 84%) than in the conventional group (47/63; 75%); however, no statistical significance was observed ( P = 0.36).

Conclusions: The tailored scan delay tended to provide a higher rate of appropriate scan timing for the pancreatic phase compared with the conventional protocol using a fixed scan delay of 20 s.

Purpose: To describe imaging features of metanephric adenomas, assess the reliability of a diagnosis with image-guided percutaneous renal mass biopsy, and evaluate patient survival outcomes.

Materials and methods: In this IRB-approved, HIPAA-compliant retrospective study, our institution's radiology report database was searched for the term "metanephric adenoma" from 2010 to 2020. Patient information, imaging mass characteristics, and percutaneous biopsy technique and complications were recorded. Analyses of per-tumor growth rate, per-procedure diagnostic rates, and per-patient disease-specific and metastasis-free survival were evaluated.

Results: The database search yielded 8 tumors (mean diameter 2.0 cm, range 1.0 to 3.1 cm) in 8 patients (median age 60.5 y, range 40 to 66 y; 6 women) who underwent percutaneous biopsies and had imaging available for review. All tumors (8/8) were solitary, well-defined, and hypoenhancing on post-contrast images. For those with available MR, 100% (5/5) demonstrated restricted diffusion. On unenhanced CT, 62.5% (5/8) were hyperdense. The mean tumor growth rate was 0.7 mm/y (range: -0.1 to 3 mm/y) with a median imaging follow-up of 83.4 months (range: 1.6 to 198.0 mo). Specific diagnosis of metanephric adenoma on the first percutaneous biopsy was found in 75% (6/8) of patients; with repeat biopsy in 2 patients confirming metanephric adenoma. Per-patient survival outcome after a median clinical follow-up of 151.8 months (range: 1.6 to 250.6 mo) showed 100% disease-specific and metastasis-free survival.

Conclusions: Metanephric adenomas are usually solitary, well-defined, and hypoenhancing masses on imaging, hyperattenuating compared with the renal parenchyma on noncontrast CT, and with restricted diffusion on MR. Image-guided percutaneous biopsy results of this tumor are reliable and safe.

Objective: This study aims to explore a grading diagnostic method for the binary classification of meniscal tears based on magnetic resonance imaging radiomics. We hypothesize that a radiomics model can accurately grade meniscal injuries in the knee joint. By extracting T2-weighted imaging features, a radiomics model was developed to distinguish meniscal tears from nontear abnormalities.

Materials and methods: This retrospective study included imaging data from 100 patients at our institution between May 2022 and May 2024. The study subjects were patients with knee pain or functional impairment, excluding those with severe osteoarthritis, infections, meniscal cysts, or other relevant conditions. The patients were randomly allocated to the training group and test group in a 4:1 ratio. Sagittal fat-suppressed T2-weighted imaging sequences were utilized to extract radiomic features. Feature selection was performed using the minimum Redundancy Maximum Relevance (mRMR) method, and the final model was constructed using the Least Absolute Shrinkage and Selection Operator (LASSO) regression. Model performance was evaluated on both the training and test sets using receiver operating characteristic curves, sensitivity, specificity, and accuracy.

Results: The results showed that the model achieved area under the curve values of 0.95 and 0.94 on the training and test sets, respectively, indicating high accuracy in distinguishing meniscal injury from noninjury. In confusion matrix analysis, the sensitivity, specificity, and accuracy of the training set were 88%, 92%, and 87%, respectively, while the test set showed sensitivity, specificity, and accuracy of 89%, 82%, and 85%, respectively.

Conclusions: Our radiomics model demonstrates high accuracy in distinguishing meniscal tears from nontear abnormalities, providing a reliable tool for clinical decision-making. Although the model demonstrated slightly lower specificity in the test set, its overall performance was good with high diagnostic capabilities. Future research could incorporate more clinical data to optimize the model and further improve diagnostic accuracy.

Objective: To evaluate the value of dual-energy computed tomography (DECT) parameters for the quantitative diagnosis of acute pancreatitis (AP) and classification of its severity.

Methods: Patients with AP underwent a plain CT scan and three contrast-enhanced DECT scans. We analyzed the group differences in iodine concentration (IC) and slope of the spectral Hounsfield unit curve (λ HU ) of the 3-phase enhanced scans (arterial, venous, and delayed phases).

Results: The study included 60 AP patients (38 males and 22 females; mean age: 47.43±13.47 y). On the basis of the CT severity index (CTSI), the patients were divided into 2 groups: group A (mild AP, n=26) and group B (moderate/severe AP, n=34). IC and λ HU in the arterial and venous phases were all significantly higher in group A than in group B ( P <0.001) and could effectively differentiate the 2 groups. The areas under the curve were 0.753 (95% CI: 0.624-0.855), 0.799 (95% CI: 0.676-0.892), 0.774 (95% CI: 0.647-0.872), and 0.842 (95% CI: 0.724-0.923) for IC at arterial and venous phases and λ HU at arterial and venous phases, respectively. These parameters decreased with the increase of CTSI, showing significant negative correlations, with r were -0.512 (95% CI: -0.678 to -0.297), -0.492 (95% CI: -0.663 to -0.272), -0.552 (95% CI: -0.707 to -0.346), -0.569 (95% CI: -0.719 to -0.368) for IC at arterial and venous phases and λ HU at arterial and venous phases, respectively ( P <0.001).

Conclusions: DECT imaging can quantitatively analyze AP, and the IC and λ HU can be used to distinguish mild and severe cases, adding functional information to the CT morphology to determine the severity and prognosis of the disease.

Objective: This study aims to clarify the frequency of renal parenchymal defects and deformations in each subtype of perirenal liposarcomas and to compare the differences between well-differentiated and non-well-differentiated types.

Methods: Patients with perirenal liposarcomas seen between July 2004 and June 2024 were included. Two radiologists blinded to the subtypes retrospectively evaluated CT or MR images for renal parenchymal defects and deformations. Frequencies of these findings were compared between well-differentiated versus non-well-differentiated types using the Fisher test.

Results: Forty-two patients (mean age: 66.3±11.5 y; 15 men) with perirenal liposarcomas were included. Renal parenchymal defects and deformations were observed in 0 (0%) and 1 (7.7%) of 13 well-differentiated, 5 (29.4%) and 6 (35.3%) of 17 dedifferentiated, 3 (37.5%) and 0 (0%) of 8 myxoid, and 1 (25.0%) and 1 (25.0%) of 4 pleomorphic types, respectively. Non-well-differentiated liposarcomas had higher frequencies of renal parenchymal defects and deformations compared with well-differentiated liposarcomas [9 of 29 (31.0%) vs. 0 of 13 (0%), P =0.038 and 7 of 29 (24.1%) vs. 1 of 13 (7.7%), P =0.398].

Conclusion: Renal parenchymal defects can be occasionally observed (31.0%) in non-well-differentiated perirenal liposarcomas unlike well-differentiated liposarcomas.

Objectives: (a) To investigate the relationship between tunnel volume (TV) and morphologic parameters of interatrial septum (IAS) in cases with type 3 and type 4 IAS; (b) To investigate the relationship between TV of the IAS and ischemic gliotic foci in brain MRI.

Materials and methods: We retrospectively reviewed the images of 301 cases who underwent CCTA in our center between 2020 and 2022. TV, tunnel length (TL), opening diameter of the right (ODRAE) and left atrium entrance (ODLAE), interatrial groove (IAG) diameter, and free flap length (FFL) were measured. The presence, number, and distribution of ischemic gliotic foci were examined in patients who had undergone brain MRI in the last 5 years before the CCTA. Pearson χ 2 , the Fisher Exact, Mann-Whitney U , linear regression analysis, Kruskal-Wallis test, and the Spearman correlation tests were used for statistical analysis of the data.

Results: A shorter FFL was related to the higher IAS type and increased likelihood of jet flow ( P =0.013). The correlation between wide IAG diameter and FFL was statistically significant ( P =0.003, r =0.22). The correlation between TV and ODRAE and ODLAE was also statistically significant (P <0.001, r =0.364, P <0.001, r =0.332, respectively). In type 3 and type 4 IAS, TV was associated with an increased number of ischemic gliotic foci ( P =0.008) and bilateral distribution ( P =0.006) on brain MRI.

Conclusion: Measurement of TL, ODRAE, ODLAE, and tunnel diameter in symptomatic cases with type 3 and type 4 IAS is crucial in determining the appropriate treatment approach. By adding the TV to the defined parameters, we thought that this innovation would contribute to invasive and noninvasive treatment management.

Objective: To construct and validate the optimal model for preoperative prediction of proliferative HCC based on habitat-derived radiomics features of Gd-EOB-DTPA-Enhanced MRI.

Methods: A total of 187 patients who underwent Gd-EOB-DTPA-enhanced MRI before curative partial hepatectomy were divided into training (n=130, 50 proliferative and 80 nonproliferative HCC) and validation cohort (n=57, 25 proliferative and 32 nonproliferative HCC). Habitat subregion generation was performed using the Gaussian Mixture Model (GMM) clustering method to cluster all pixels to identify similar subregions within the tumor. Radiomic features were extracted from each tumor subregion in the arterial phase (AP) and hepatobiliary phase (HBP). Independent sample t tests, Pearson correlation coefficient, and Least Absolute Shrinkage and Selection Operator (LASSO) algorithm were performed to select the optimal features of subregions. After feature integration and selection, machine-learning classification models using the sci-kit-learn library were constructed. Receiver Operating Characteristic (ROC) curves and the DeLong test were performed to compare the identified performance for predicting proliferative HCC among these models.

Results: The optimal number of clusters was determined to be 3 based on the Silhouette coefficient. 20, 12, and 23 features were retained from the AP, HBP, and the combined AP and HBP habitat (subregions 1, 2, 3) radiomics features. Three models were constructed with these selected features in AP, HBP, and the combined AP and HBP habitat radiomics features. The ROC analysis and DeLong test show that the Naive Bayes model of AP and HBP habitat radiomics (AP-HBP-Hab-Rad) archived the best performance. Finally, the combined model using the Light Gradient Boosting Machine (LightGBM) algorithm, incorporating the AP-HBP-Hab-Rad, age, and AFP (Alpha-Fetoprotein), was identified as the optimal model for predicting proliferative HCC. For the training and validation cohort, the accuracy, sensitivity, specificity, and AUC were 0.923, 0.880, 0.950, 0.966 (95% CI: 0.937-0.994) and 0.825, 0.680, 0.937, 0.877 (95% CI: 0.786-0.969), respectively. In its validation cohort of the combined model, the AUC value was statistically higher than the other models ( P <0.01).

Conclusions: A combined model, including AP-HBP-Hab-Rad, serum AFP, and age using the LightGBM algorithm, can satisfactorily predict proliferative HCC preoperatively.

Objective: In CT imaging of severely obese patients, demanding clinical tasks like liver imaging may be constrained by scanner radiation output limits. This may impose an unavoidable increase in image noise and loss of image quality. In such patients, scan parameters may be restricted, leading to excessive x-ray tube heating and increased scan times that degrade exam and image consistency relative to other patients. In this study, the performance of dual-source (DS) CT with enhanced radiation output capacity was quantified relative to conventional single-source (SS) CT. The focus was on abdominopelvic imaging in severely obese patients (BMI >45 kg/m 2 ).

Methods: Abdominopelvic portal venous phase CT exams performed using DSCT were compared with exams using SSCT. General usage characteristics of the DSCT protocol were analyzed for >3000 exams over a 42-month period. More specifically, a total of 95 matched SS and DS scan pairs for the same patients were assessed in detail. The tube voltage, reconstruction method, and scanner platform were consistent in matched SS and DS scans, and changes in patient weight, diameter, and water equivalent diameter were <5%. Image global noise (GN), radiation dose (CTDI vol ), and key scan parameters were compared between matched SS and DS exams.

Results: The median (IQR) patient BMI was 48.4 kg/m 2 (45.9-52.1 kg/m 2 ). In the matched scan pairs, SS scans had a median (IQR) CTDI vol of 36.5 mGy (35.2-42.9 mGy) and median (IQR) GN of 14.1 HU (12.6-15.9 HU). DS scans had a significantly increased median (IQR) CTDI vol of 62.5 mGy (55.8-69.8 mGy) and reduced median (IQR) GN of 11.4 HU (10.6-12.4 HU; both P <0.001). Relative to SSCT, the DSCT protocol also enabled faster scan times at equal CTDI vol , lower tube current per x-ray tube, and improved GN consistency throughout axial slices.

Conclusion: It is feasible to utilize a DSCT protocol to significantly increase radiation output, bringing image noise characteristics in line with the general patient population in abdominopelvic imaging of severely obese patients. The DSCT protocol offers a more straightforward option to attain consistency in a group of patients where achieving diagnostic CT quality has proved challenging.