Journal of Computer Assisted Tomography最新文献

Objective: Low keV virtual monoenergetic (VME) images are effective in enhancing vessel opacification but require dual-energy CT (DECT), limiting widespread clinical use. Recent advancements in deep learning (DL) enable the generation of VME images from single-energy CT (SECT). However, the performance of the methods has not been evaluated in any clinical use case. The purpose of this work was to assess both objective and subjective image quality of deep learning-based VME images derived from heterogeneous SECT data for pulmonary angiography.

Methods: In this retrospective study, 52 sets of SECT pulmonary angiography images were processed using a deep learning method to estimate material basis images. 40 keV VME images were generated from heterogeneous SECT data using a pretrained physics-constrained Deep-En-Chroma DL model. Two thoracic radiologists, blinded to the image reconstruction method, evaluated pulmonary vessel opacification and overall image quality on DL-VME and SECT images using 5-point Likert scales. Objective image quality was assessed by measuring enhanced vessel contrast and contrast-to-noise ratio (CNR). Statistical analysis was performed using paired t tests and Mann-Whitney U tests.

Results: Compared with SECT, DL-VME images demonstrated significantly higher subjective image quality score and vessel opacification score (P≤0.008). DL-VME yielded a higher average contrast for emboli (1085 vs. 331 HU, P<0.001) and improved CNR (17.8 vs. 11.1, P<0.001). Results of subgroup analysis indicate no significant variation in VME performance across patient sex, scanner model, radiation dose, and tube potential. The vessel opacification scores of both VME and SECT demonstrate dependence on patient weight, with VME providing better vessel opacity for both lighter and heavier patients.

Conclusions: A measure of 40 keV DL-VME derived from SECT effectively enhances both vessel opacification and image quality in CT pulmonary angiography. The image quality advantage of DL-VME over SECT remains robust across variations in data acquisition and patient variables.

Objective: To identify risk factors predicting visceral pleural invasion (VPI) in part-solid and solid nodules through meta-analysis, and to develop a predictive model in an independent cohort.

Methods: The PubMed, Embase, and Web of Science databases were systematically searched to identify studies on pleural-related semantic, nodule semantic, and quantitative computed tomography (CT) features to predict VPI. The pooled odds ratios (ORs) for semantic features and standardized mean differences (SMDs) for quantitative features were calculated to develop a predictive model. A total of 203 patients (147 VPI-negative and 56 VPI-positive) were enrolled in the validation cohort between January and December 2024. The diagnostic performance of the model was assessed using the area under the receiver operating characteristic curve (AUC).

Results: Thirteen studies with 3999 patients were included in this meta-analysis. Several key risk factors were identified to construct the predictive model, including pleural indentation (OR: 3.428, 95% CI: 2.559-4.593), nodule type (OR: 4.867, 95% CI: 3.915-6.051), spiculation (OR: 2.581, 95% CI: 1.640-4.062), lobulation (OR: 1.855, 95% CI: 1.148-2.997), vessel convergence sign (OR: 3.606, 95% CI: 1.698-7.656), and the maximum solid diameter (SMD: 0.894, 95% CI: 0.600-1.188). This model yielded an AUC of 0.892 (95% CI: 0.840-0.931) in the validation cohort.

Conclusions: This meta-analysis involved the construction of an effective model for predicting VPI by integrating pleural indentation, nodule type, spiculation, lobulation, vessel convergence sign, and maximum solid diameter, which could inform preoperative clinical decision-making for subpleural part-solid and solid nodules.

Lung Cancer (LC) screening with low-dose CT (LDCT) is recommended in smokers and former smokers irrespective of cardiovascular disease (CVD) history. Differentiating coronary stents from moderate-to-severe coronary artery calcifications (CACs) on LDCT examinations can be challenging and have clinical implications. We hypothesized that coronary stents are inadequately reported in screening LDCT examinations and conducted a retrospective analysis nested within an LC screening study. Moreover, we devised a postprocessing procedure to differentiate coronary stents and severe CACs in LDCT examinations and assessed its potential. Among 592 participants with detailed CVD history at baseline, the LDCT report mentioned coronary stents in 13 and severe CACs in 60 of 73 (12.3%). Two radiologists independently and blindly reviewed the LDCT examinations, providing a binary judgment (yes/no) on stent presence using images optimized with a window setting optimized to view detail in bone (window width 1800 to 2000 HU and window level 300 to 500 HU) reformatted parallel or perpendicular to the coronary arteries. The review identified stents in 26 subjects and was inconclusive in 2. On images viewed with such "bone" window setting, stents appeared as "tram-track" (large stents) or "pencil-lead" (narrow stents) configurations in parallel views, and as a "rim-of-wheel" shape in perpendicular views. In contrast, CACs lacked these configurations due to their irregular distribution along vessel walls. No difference in attenuation was observed between stents and severe CACs. Concordance with the history of PTCA with stent placement was 42.8% (12/28) for the original LDCT reports and 92.8% (26/28) for the review (P=0.0014). Our study suggests that coronary stents are often under-reported in screening LDCT, but their characteristic shapes on images viewed with a "bone" window setting with multiplanar reformatting can aid in accurate identification.

Objective: This study investigated the associations of radiomics based on baseline MRI with recurrence and disease-free survival (DFS) after neoadjuvant chemotherapy (NAC) in young women with breast cancer.

Materials and methods: In total, 181 women aged 40 years or younger with breast cancer who underwent MRI before NAC were allocated into the training (n=126) and test cohorts (n=55). Three radiomics signatures were built using the intratumoral, peritumoral, and combined regions of MR images. Univariate and multivariate logistic regression were performed to select independent risk factors to construct a clinical model. A nomogram model was developed by integrating the clinical model and the combined radiomics signature. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC). Multivariate Cox regression and Kaplan-Meier analyses were used to determine the associations of various models with DFS.

Results: Among the radiomics signatures, the combined signature best predicted recurrence, with AUCs of 0.899 and 0.849 in the training and test cohorts, respectively. The nomogram model displayed the best performance in predicting recurrence in the training (AUC=0.925) and test cohorts (AUC=0.880). The nomogram model most accurately predicted DFS in the training (C-index=0.872) and test cohorts (C-index=0.846).

Conclusions: The nomogram model based on pretreatment breast MRI could effectively predict breast cancer recurrence in young women undergoing NAC and serve as a potential biomarker for the risk stratification of DFS.

Objective: Minimally invasive parathyroidectomy (MIP) requires high-fidelity localization of parathyroid adenomas through preoperative imaging, commonly 4-dimensional computed tomography (4D-CT). Texture analysis extracts high-order mathematical features from an image and may be applied to 4D-CT for quantitative differentiation of lymph nodes and thyroid nodules from parathyroid adenomas.

Methods: This is a retrospective cohort study of 51 patients diagnosed with PHPT and known parathyroid adenoma and/or thyroid nodule who have undergone preoperative 4D-CT imaging before parathyroidectomy. Three anatomic structures (parathyroid adenoma, lymph node, and thyroid nodule) were manually segmented on 25-second arterial phase axial sections of the 4D-CT scans. Radiomic data were extracted for shape, first-order, and second-order classes (107 total features) for each of the structures in each patient. A series of t tests were conducted to assess for radiomic features with statistically significant differences in lymph nodes or thyroid nodules when compared with parathyroid adenomas. A multivariable logistic regression model for discrimination of parathyroid adenomas was trained on a subset of the data set and assessed on a hold-out test subset.

Results: When comparing parathyroid adenomas and lymph nodes, 14/18 first-order features and 44/75 second-order features were statistically significantly different (P<0.05), of which 13/18 first-order features and 16/75 second-order features were potent discriminators (P<0.0001). No features were significantly different between parathyroid adenomas and thyroid nodules. A multivariable logistic regression model for discrimination of parathyroid adenomas from lymph nodes achieved strong predictive performance (AUC: 0.95, 95% CI: 0.86-1).

Conclusions: Parathyroid adenomas and lymph nodes have statistically distinct radiomic textural signatures on arterial phase 4D-CT, with the most significant differences found in first-order textural features. These findings may facilitate the development of future machine learning models for automated differentiation of parathyroid adenomas, further enhancing uptake of MIP and improving clinical outcomes.

Objective: We aimed to improve preoperative accuracy of tumor staging by evaluating the ability of qualitative and quantitative CT imaging features to predict perirenal fat invasion (PFI) in renal cell carcinoma (RCC).

Methods: This retrospective case-control study included 86 patients with pathologically proven PFI and 169 controls matched for tumor size without PFI who were treated by nephrectomy between January 2016 and December 2020. Two radiologists independently evaluated the qualitative imaging features of tumor complexity, shape, margin, tumor-fascia contact, perirenal vascularity, fascial thickening, septation, stranding, and nodules. We also compared tumor contact length and protruding distance between the groups. Multivariate logistic regression analyses identified significant predictors of PFI, and diagnostic performance metrics of all predictors were assessed to create a combined model that included all significant predictors.

Results: Lobulated shapes and irregular margins were more prevalent in the group with than without PFI (P<0.05). Perirenal increased vascularity, fascial thickening, septation, stranding, and nodularity were also significantly more prevalent in the PFI group (P<0.05 for all). Tumor contact length and protruding distance were significantly greater in the PFI group (P=0.002). Multivariate analysis identified the following independent predictors of PFI: lobulated tumors [odds ratio (OR): 2.03; P=0.042], irregular margin (OR: 3.40; P=0.007), perirenal fascial thickening (OR: 4.20; P<0.001), and contact length >154.2 mm (OR: 3.82; P=0.019). The diagnostic performance of these combined predictors was moderate, with 61.6% sensitivity, 79.3% specificity, and 73.3% accuracy.

Conclusions: Qualitative CT features (lobulated tumors, irregular margins, perirenal thickened fascia) and an objective quantitative parameter (threshold 154.2 mm tumor contact length) were significant independent predictors of perirenal fat invasion in RCC. These findings emphasize the complementary value of combining subjective and objective imaging features to enhance preoperative staging accuracy.

Objective: To evaluate whether our new protocol that uses a saline test injection and a leak detection sensor (LDS) reduces the frequency and amount of contrast media (CM) extravasation during the intravenous CM administration for CT.

Methods: This retrospective study included 20,342 patients who underwent CECT at our hospital from March 2021 to November 2021 (old protocol, direct patient observation, and CM injection pressure monitoring, n=10,529) and from March 2024 to November 2024 (new protocol, old protocol plus saline test injection, and the LDS attachment, n=9813). We compared the frequency and the volume of extravasation between the 2 protocols using the Fisher exact test and the Mann-Whitney U test. We also evaluated the accuracy of the LDS.

Results: Extravasation occurred in 51 patients (age 72.1±12.2 y, 33 men) under the old protocol and in 26 patients (age 73.6±9.0 y, 17 men) with the new protocol. The overall frequency of extravasation and the number of patients with an extravasation volume of at least 20 mL were significantly lower with the new protocol than the old protocol (0.48% vs. 0.26%; 0.16% vs. 0.03% all, P<0.01). The extravasation volume was significantly reduced with the new protocol (14 vs. 6 mL, P<0.01). The sensitivity of the LDS to detect extravasation of 3, 5, 10, and 15 mL was 50%, 88%, 93%, and 100%, respectively; specificity was 99% for all.

Conclusions: Our new protocol reduced the frequency and dose of CM extravasation.

Acquired diverticular disease of the small bowel is often seen in the duodenum. It is an uncommon but under-recognized entity in the jejunum and the ileum. Meckel's diverticulum, a true diverticulum arising in the distal ileum, although the most common congenital abnormality of the gastrointestinal tract, is rare and occurs in about 2% of the population. Most of the time, diverticular disease of the small bowel is asymptomatic. Common complications of small bowel diverticular disease include diverticulitis and hemorrhage. Diverticulitis of the small bowel is an uncommon cause of acute abdomen and may be misdiagnosed if not included in differential considerations based on the imaging features. Certain specific complications can occur related to the location of the diverticulum or due to other factors associated with the diverticulum. For example, obstructive jaundice (Lemmel syndrome) can occur in the setting of a duodenal diverticulum, and intestinal obstruction can occur in the setting of a Meckel's diverticulum. Familiarity with the different imaging manifestations of small bowel diverticular disease complications can help with appropriate diagnoses, thereby improving patient management.

Preoperative computed tomography (CT)-guided localization is widely used to facilitate the surgical resection of pulmonary nodules (PNs) through video-assisted thoracoscopic surgery (VATS). This meta-analysis aims to compare the clinical performance and safety profile of 2 commonly used localization techniques: the anchored needle (AN) and the hook-wire (HW) method. A systematic search was conducted using the PubMed, Wanfang, and Cochrane Library databases to identify relevant comparative studies. Key outcome measures were extracted and analyzed using Stata v12.0 and RevMan v5.3. Seven retrospective studies conducted in China met the inclusion criteria, encompassing a total of 1557 patients. Among these, 889 patients with 961 PNs underwent CT-guided AN localization, whereas 668 patients with 697 PNs received HW localization. Compared with HW, AN localization demonstrated a significantly higher rate of successful localization (P <0.001), lower rates of overall complications (P=0.01), pneumothorax (P=0.003), and pulmonary hemorrhage (P=0.004). Patients in the AN group also reported significantly lower pain scores (P <0.001). Two groups exhibited similar localization (P=0.48) and VATS (P=0.93) time. Notable heterogeneity was observed in localization time, complication rate, VATS time, and pain score (I²=91%, 73%, 94%, and 94%, respectively). No evidence of publication bias was detected across the analyzed outcomes. CT-guided AN localization seems to offer higher successful localization rate and a lower complication rate compared with HW localization for patients undergoing surgical management of PNs.

Objective: Changes in bone marrow fat content measured through relative fat fraction (rFF) obtained from dual-echo gradient-recalled echo (GRE) in- and opposed-phase (IOP) MRI have been proposed to evaluate treatment response for multiple myeloma. However, rFF suffers from several significant limitations that lead to inaccurate fat fraction measurements. In contrast, proton density fat fraction (PDFF) is the most objective and validated MRI metric of tissue fat content, and it is measured through confounder-corrected, multiecho, chemical-shift-encoded (CSE) MRI. The purpose of this study was to evaluate the linearity and bias of bone marrow rFF compared with PDFF.

Methods: This single-center, retrospective study included 100 patients who underwent clinical MRI for liver fat/iron quantification at 1.5T and 3.0T (50 exams/patients for each field strength), which included dual-echo GRE IOP and commercial multiecho CSE MRI (IDEAL-IQ). One region of interest (ROI) was placed in each of the T12, L1, and L2 vertebral bodies. Per-ROI rFF was calculated using (SIP and SOP = signal intensities on IP and OP images, respectively). rFF was correlated with PDFF using linear regression and coefficient of determination (R2). Bland-Altman analysis evaluated rFF bias across the observed range for R2* and PDFF; mean bias and 95% limits of agreement (LOA) were reported.

Results: Bone marrow rFF demonstrated no linearity against PDFF at 1.5T or at 3.0T (R2 = 0.032 and 0.057, respectively). Moreover, bone marrow rFF demonstrated significant bias with respect to PDFF at 1.5T and 3.0T, with significant bias that increases directly with bone marrow fat fraction.

Conclusions: Bone marrow rFF is nonlinear and variably biased compared with PDFF and should not be used in research or clinical settings.