Magnetic resonance imaging (MRI) is the established gold standard for the staging of primary bone tumors; however, conventional protocols typically use a site-specific model supplemented by ionizing radiation for systemic screening. Recent advancements in whole-body MRI (WB-MRI) have enabled a "marrow-first" diagnostic approach, allowing for the comprehensive assessment of the primary intraosseous lesion and the entire skeleton in a single, radiation-free study. Technical developments, including Dixon fat suppression and diffusion-weighted imaging (DWI), have significantly enhanced the detection of occult skip metastases, which occur in ∼7.5% of high-grade osteosarcomas and 4.5% of Ewing sarcomas. Beyond malignant staging, WB-MRI offers distinct advantages for the longitudinal surveillance of syndromic conditions such as Ollier disease and hereditary multiple exostoses, where it facilitates a shift toward patient-initiated, symptom-led monitoring. Despite these advances, challenges remain, specifically the "chest gap" in pulmonary nodule detection and the potential for false positives from hematopoietic marrow mimics. This review discusses the technical principles, clinical applications, and diagnostic pitfalls of WB-MRI in bone tumors, emphasizing its role in optimizing staging accuracy while minimizing cumulative radiation exposure.
OBJECTIVE:Intratumoral heterogeneity may limit the representativeness of biopsy-based Ki-67 assessment in breast cancer. We therefore developed and validated a habitat-guided 2.5D deep learning (DL) model based on multiparametric MRI for noninvasive preoperative prediction of high versus low Ki-67 expression. METHODS:This retrospective study enrolled 333 patients with invasive breast carcinoma from 2 distinct MRI vendor cohorts (Siemens, training set, n=233; United Imaging, independent test set, n=100). All patients underwent preoperative multiparametric MRI, including DCE-MRI and DWI. Hemodynamic parametric maps (wash-in, wash-out) and ADC maps were generated and subsequently clustered using a K-means algorithm (k=3) to create a functional habitat mask that quantitatively encodes intratumoral heterogeneity. A 7-channel 2.5D input tensor was then constructed by concatenating the central habitat-guided slice with its 6 adjacent anatomic slices. A ResNet18 backbone was trained to classify high (≥20%) versus low Ki-67 expression. The model's performance was rigorously evaluated against conventional 2D DL, clinical, and combined (DL+clinical) models using AUC, the DeLong test, and decision curve analysis (DCA). RESULTS:In the challenging independent cross-vendor test set, our habitat-guided DL25D model demonstrated superior performance, achieving an AUC of 0.821 (95% CI: 0.736-0.906) and a sensitivity of 0.804. It significantly outperformed both the conventional DL2D model (AUC: 0.654, P=0.002) and the clinical model (AUC: 0.686, P=0.019). The incorporation of clinical variables failed to yield further improvement (combined model AUC: 0.837, P=0.483 vs. DL25D; NRI=0.021, P>0.05). DCA confirmed the superior net clinical benefit of our approach across a wide spectrum of threshold probabilities. Importantly, Grad-CAM visualizations revealed that the habitat-guided model strategically focused its attention on intratumoral core regions, whereas the conventional 2D model was distracted by tumor margins and background tissue. CONCLUSIONS:The habitat-guided 2.5D deep learning model showed potential as a noninvasive imaging adjunct for preoperative Ki-67 status prediction in breast cancer. Multicenter prospective validation is required before clinical use.
This case series describes an imaging finding seen in patients following lymphangiography-lace-like hyperdensities outlining the pulmonic lymphatic pathways-termed the spiderweb sign. In a retrospective review of patients who underwent lymphangiograms at a large tertiary care hospital from January 2022 to December 2025, 11 patients had CT thorax images within 1 week of the procedure, and 4 of these patients demonstrated the spiderweb sign. Specifically, the spiderweb sign manifests as hyperdense linear opacities outlining the interlobular septa, peribronchovascular interstitium, and visceral pleura. We present the 4 representative cases illustrating the imaging findings. For the cohort, procedural parameters were collected to assess associations with the presence or absence of the imaging finding, but no definitive association was found. While the exact mechanism remains unclear, these findings parallel patterns seen in prior studies and appear clinically benign. An additional literature review of pulmonary findings following lymphangiography was conducted to contextualize the spiderweb sign. Awareness of these postprocedural imaging findings as expected findings rather than developing pathology is essential to prevent misinterpretation. Larger multicenter studies are needed to clarify procedural factors and the underlying pathophysiology.
OBJECTIVE:To compare the diagnostic performance of high-resolution computed tomography (HRCT) and non-echo-planar diffusion-weighted magnetic resonance imaging (non-EPI DWI-MRI) for middle ear cholesteatoma and to assess the added value of their combined use in preoperative evaluation. METHODS:Imaging data sets from 85 subjects (65 with surgically or histopathologically confirmed cholesteatoma, 20 controls) were retrospectively analyzed. HRCT, non-EPI DWI-MRI, and combined HRCT+DWI-MRI interpretations were compared against reference standards. Diagnostic metrics included sensitivity, specificity, accuracy, localization error, and interobserver agreement (κ). RESULTS:HRCT demonstrated high sensitivity (95.4%) for detecting bony erosions and ossicular abnormalities, whereas non-EPI DWI-MRI showed greater specificity (85.0%) in distinguishing cholesteatoma from other soft tissues. The combined approach yielded optimal diagnostic performance (sensitivity 93.8%, specificity 92.0%, accuracy 93.2%) and the lowest localization error (mean 1.8 mm). Interobserver agreement was excellent (κ=0.88). CONCLUSION:High-resolution computed tomography and non-echo-planar diffusion-weighted MRI provide complementary diagnostic information in the evaluation of middle ear cholesteatoma. Their combined interpretation may improve diagnostic confidence and radiologic assessment of lesion extent; however, the direct impact on surgical outcomes requires further prospective validation.
OBJECTIVE:Chronic pancreatitis (CP) is a risk factor for pancreatic ductal adenocarcinoma (PDAC), with a cumulative risk of ~4% over 20 years. Most prior studies have evaluated the risk of PDAC among patients with CP. The frequency and imaging characteristics of chronic calcific pancreatitis (CCP) in patients presenting with PDAC are unknown. METHODS:A retrospective review of the clinical history and staging computed tomography (CT) scans of patients diagnosed with resectable or borderline resectable/locally advanced PDAC between 2011 and 2022 was conducted. CCP and ductal dilatation were defined by the presence of pancreatic calcifications and a maximal pancreatic ductal diameter >5 mm on CT, respectively. Two independent radiologists reviewed all imaging studies, and discrepancies were adjudicated by a third radiologist. RESULTS:Among 1008 patients with PDAC, CCP was found in 34 (3.4%), with a mean age of 69.9±8 years, 50% men, 91% White, and 59% with a history of smoking, with 26.95±18.5 pack-years. Only 7 (21%) patients had a documented history of chronic abdominal pain before PDAC diagnosis. The presence of ductal dilation, pseudocysts, or pancreatic atrophy was observed in nearly 40%, 10%, and 40% of patients, respectively, and 7 (21%) showed none of these features. Ductal dilation was associated with higher levels of CA 19-9 (349.8 U/mL [IQR; 238.3, 455.0] vs. 90.0 U/mL [IQR; 58.3, 162.2], P=0.003). Intratumoral and extratumoral-only calcifications were found in 14 (41%) and 20 (59%) patients, respectively. Extratumoral-only calcifications were associated with pancreatic atrophy (P=0.02). Radiologist inter-rater agreement was highest for tumor location, calcification location, maximal size, and number, and side branch dilatation, and was lowest for the percentage of pancreas with abnormal findings and PD contour. CONCLUSION:CCP is found in 3% of patients presenting with PDAC, and most have primary painless disease. Further studies are required to determine optimal PDAC screening strategies and surveillance in patients found to have primary painless CCP.
As applications of MRI in the abdomen and pelvis have increased, access to MRI scanners has become more crucial. 0.55 T MRI scanners can potentially improve access by reducing the costs of MR systems and simplifying their siting requirements. These savings come with important tradeoffs, most notably decreased SNR efficiency. Decreases in SNR can be ameliorated to some degree by shorter T1 (faster repetition time), longer T2* (reduced susceptibility artifacts), and higher gadolinium relaxivity, which are found at lower fields. Reduced frequency separation between fat and water presents a particular challenge in imaging the abdomen and pelvis. In this article, we review how MRI at 0.55 T affects abdominal and pelvic imaging and discuss clinical examples of the liver, kidney, bile ducts, uterus, and ovary. We illustrate how metal artifacts are reduced at 0.55 T and discuss how deep learning image reconstructions, combined with free-breathing techniques, can overcome some of the limitations imposed by reduced SNR. We discuss future challenges and opportunities in this field, leveraging technological advances from other fields to bring MRI to more people who need it.
Prospectively ECG-triggered sequential coronary CT angiography (CCTA) is prone to step artifacts between cardiac segments imaged during different heartbeats, which can obscure plaque and impair automated tools such as CT-fractional flow reserve. We evaluated a vendor-specific step artifact reduction (SAR) algorithm, alone and combined with motion-compensated reconstruction (MCR), in sequential CCTA. We retrospectively analyzed 79 patients scanned on a 128-slice dual-layer spectral CT system. Reconstructions compared were clinical standard, SAR, and SAR+MCR. Lumen continuity across the step interface in the left anterior descending (LAD) and right coronary arteries (RCA) was quantified by the Jaccard index (JI), and whole-heart continuity by histogram entropy of subtraction images; Cohen d is reported for effect sizes. Across 165 coronary step segments, SAR increased mean JI from 0.38 to 0.58 (+0.20; 95% CI: 0.16-0.23; P <0.01; d =0.84), similar in both vessels (LAD +0.20, d =0.82; RCA +0.19, d =0.87). Entropy decreased from 3.81 to 3.51 (-0.30; P <0.01; d =-3.12), indicating fewer step-related structures. In 25 motion-affected RCA segments, SAR alone increased the JI from 0.35 to 0.49, and adding MCR raised it to 0.70 (+0.22 vs. SAR alone; P <0.01; d =0.89). SAR improved vessel and whole-heart continuity on dual-layer spectral CT, with additional benefit from MCR in motion-affected segments.
BACKGROUND:To develop a combined radiomics nomogram model and a clinical model based on dual-layer detector spectral computed tomography (DLCT) for predicting pathologic subtypes and Ki-67 index status in patients with non-small cell lung cancer (NSCLC). METHODS:A total of 405 patients with NSCLC who underwent DLCT scans and Ki-67 testing from 2 medical centers were enrolled in the study. Patients from center 1 were divided into training (n=233) and internal validation (n=83) sets, while center 2 (n=89) formed the external validation dataset. The most valuable radiomics features from dual-phase enhanced CT images and their corresponding iodine images were extracted to construct radiomics models. Clinical factors were screened by univariate and multivariate logistic regression analysis for the construction of clinical models. Subsequently, 2 nomograms that incorporated the radiomics score and clinical factors were established and tested for the 2 tasks, respectively. RESULTS:For type classification, the clinical model, radiomics model, and nomogram had AUCs of 0.863, 0.853, and 0.931, respectively, in the internal validation datasets and 0.862, 0.832, and 0.910, respectively, in the external validation datasets. For Ki-67 classification, the clinical model, radiomics model, and nomogram were available in the internal validation datasets at 0.718, 0.724, and 0.808, respectively, and the AUCs in the external validation dataset were 0.734, 0.721, and 0.791, respectively. Decision curve analysis demonstrated that the nomogram had greater net benefits than did the clinical model and the radiomics model in both terms of type classification and Ki-67 classification. CONCLUSION:The nomograms integrating clinical models and radiomics models all showed good performance not only for type classification but also for Ki-67 classification in NSCLC patients, which could facilitate clinical decision-making.
OBJECTIVE:To study the utility of intratumoral and peritumoral radiomic models in distinguishing between benign and malignant thyroid micronodules. METHODS:CT-enhanced images of 607 patients with thyroid micronodules were obtained, and an additional 100 patients with micronodules were collected as an independent test group on the basis of the sequence of their examination times. The peritumoral region of interest (ROI) was delineated by expanding outward by 1 and 2 mm. Radiomics features were extracted from the 3-dimensional volume of interest (VOI), which included intratumoral, peritumoral 1 mm, peritumoral 2 mm, intratumoral+peritumoral 1 mm and 2 mm in the arterial phase, venous phase, and combined arteriovenous phase. All of the data were randomly divided into training and validation groups at an 8:2 ratio. The minimum redundancy maximum relevance (mRMR) and the least absolute shrinkage and selection operator (LASSO) method was utilized for regression dimension reduction. Radscore, clinical, and combined models were constructed. A nomogram was used to assess the performance of the fusion model, and an external test group was utilized to test the model's performance. RESULTS:The intratumoral+peritumoral 1 mm radiomic model based on arterial-venous phase fusion had the best prediction performance (AUC training group 0.838; 95% CI: 0.802-0.873; validation group 0.827; 95% CI: 0.755-0.899; external test group 0.770; 95% CI: 0675-0.864). The combined model constructed with the radscore and clinical model was the optimal model (AUC training group 0.859; 95% CI: 0.826-0.892; validation group 0.876; 95% CI: 0817-0.935; external test group 0.810; 95% CI: 0.723-0.897). CONCLUSION:Both the intratumoral and peritumoral radiomic models had good diagnostic efficacy in differentiating benign and malignant thyroid micronodules.
Objective: Needle artifact redistribution technique (Needle-ART) is a recently developed image artifact-reduction method for CT-guided needle procedures. The principle of this technique is to create an angle between the needle and imaging plane by tilting the CT gantry, then reformatting images along the needle plane, thus redistributing the artifact out of the viewing plane. The purpose of this study is to evaluate the feasibility, artifact reduction, and clinical applicability of Needle-ART for the placement of biopsy introducer needles into the liver, kidneys, and lungs of a deceased animal model. Methods: A paired before-and-after study was performed in 2 deceased swine to evaluate the effectiveness of Needle-ART in reducing metal artifacts during CT-guided interventions. A 17G co-axial needle was inserted into the liver, kidney, and lung at 6 sites per organ and scanned at each site with both 0-degree gantry tilt (standard axial as the control group) and 6-degree tilted gantry on Needle-ART. A total of 48 scans with 24 control and 24 Needle-ART were collected. Images were reformatted to align with the needle trajectory. Collected images were evaluated for overall artifact severity by 3 blinded radiologists. Results: Needle-ART significantly reduced metal artifacts across all quantitative metrics, which was consistent with blinded reader assessments. Mean artifact length was reduced by 35.5% ( P =0.0019) in liver, 65.9% ( P =0.003) in kidney, and 59.4% ( P =0.0021) in lung. Median artifact width decreased by 30% ( P =0.028) in liver, 24.3% ( P =0.028) in kidney, and 8.3% ( P =0.017) in lung. Median artifact intensity was reduced by 38.8% ( P =0.011) in liver, 35.4% ( P =0.028) in kidney, and 23.6% ( P =0.021) in lung. The radiologists scored conventional CT images (with non-tilted gantry) significantly worse artifacts compared with Needle-ART (with tilted gantry) ( P =0.000041). Conclusions: The off-axis reconstruction method is a novel technique that effectively reduces artifacts by acquiring scans with the CT gantry tilted 6 degrees relative to the plane containing the needle and target.
OBJECTIVE:This study aimed to compare the diagnostic performance of consolidation-to-tumor ratio (CTR) and tumor disappearance rate (TDR) for differentiating minimally invasive adenocarcinoma (MIA) from invasive adenocarcinoma (IAC) in patients with part-solid nodules (PSNs), and to identify their optimal cutoff values as well as clinical application value. METHODS:A retrospective analysis was performed on 206 patients with pathologically confirmed pulmonary adenocarcinoma from PSNs who underwent chest CT and surgical resection between January 2023 and December 2025. They were divided into the MIA group and IAC group based on postoperative pathology. Two experienced radiologists independently measured and calculated CTR and TDR on 1-mm-thick CT images. ICC evaluated interobserver consistency; independent-samples t test or Mann-Whitney U test compared group differences; ROC curve analyzed diagnostic efficacy; multivariate Logistic regression identified independent diagnostic factors. RESULTS:CTR and TDR had excellent interobserver consistency (ICC=0.912 and 0.905, both P <0.001). IAC group had significantly higher CTR [(0.60±0.13) vs. (0.32±0.11), P <0.001] and lower TDR [(0.34±0.16) vs. (0.65±0.14), P <0.001] than MIA group. ROC analysis showed AUC of CTR was 0.893 (95% CI: 0.857-0.935, cutoff=0.48, sensitivity=85.7%, specificity=83.5%) and TDR was 0.902 (95% CI: 0.862-0.938, cutoff=0.52, sensitivity=85.8%, specificity=87.6%), with no significant AUC difference ( P =0.375). Both CTR (OR=18.785, P <0.001) and TDR (OR=0.126, P <0.001) were independent diagnostic factors. CONCLUSION:TDR and CTR have comparable diagnostic efficacy in differentiating minimally invasive adenocarcinoma (MIA) from invasive adenocarcinoma (IAC) in part-solid pulmonary nodules. Both TDR and CTR are independent diagnostic factors for distinguishing MIA from IAC.
Magnetic resonance imaging (MRI) at low magnetic field strengths (under 1 T) has seen renewed interest, driven by technological advances that enhance accessibility, affordability, and versatility. Improvements in design, gradient performance, and artificial intelligence (AI)-assisted reconstruction enable diagnostic-quality neuroimaging at well under 1 T, disrupting the once-peerless dominance of high-field imaging. Modern low-field systems offer advantages in patient comfort, siting flexibility, and/or portability. Reduced infrastructure and safety constraints enable scanning in diverse settings, including outpatient, emergency, and intensive care environments. Lower susceptibility improves imaging near air-bone interfaces, metallic implants, and the skull base, while markedly lowering specific absorption rate (SAR) and acoustic noise. Clinical neurological applications extend across the brain, spine, and skull base and include pediatric or fetal imaging, where improved tolerance and reduced motion artifacts are particularly useful. Lower signal-to-noise ratio (SNR) and limited spatial resolution remain challenges. AI-enhanced reconstruction, optimized pulse sequences, and higher relaxivity of contrast agents help mitigate these drawbacks. Portable low-field MRI also enables point-of-care neuroimaging and intraoperative use, offering real-time insights to expedite diagnosis. However, it is critical to acknowledge that, while low-field applications have found important roles in terms of quality, speed, and performance, higher-field-strength scanners will remain the diagnostic standard for the foreseeable future.
OBJECTIVE:To evaluate the diagnostic performance of left ventricular (LV) volumetric parameters derived from routine abdominal MRI for detecting left ventricular hypertrophy (MRI-LVH) in patients with autosomal dominant polycystic kidney disease (ADPKD) who were previously diagnosed with LVH by echocardiography (Echo-LVH). METHODS:This retrospective study reviewed 156 ADPKD patients (27 with LVH). The LV wall (LVW) and cavity were manually segmented on T2-weighted abdominal MRI images to calculate 3 height-adjusted parameters: LV wall volume (ht-LVWV), LV cavity volume (ht-LVCV), and total LV volume (ht-LVV). Diagnostic performance was assessed using receiver operating characteristic curve analysis, with Echo-LVH as the reference standard. Subgroup analyses were performed across sex, chronic kidney disease stage, Mayo Imaging Classification, and blood pressure status. Sensitivity analysis was adjusted for body surface area, age, sex, and systolic blood pressure. RESULTS:Ht-LVWV demonstrated high diagnostic accuracy for Echo-LVH, with an area under the curve (AUC) of 0.82 (95% CI: 0.74-0.90), sensitivity=92%, and specificity=64% at a threshold of 60 mL/m. Ht-LVV also showed good performance (AUC=0.70), while ht-LVCV had inadequate diagnostic performance (AUC=0.61). Ht-LVWV maintained robust performance across all clinical subgroups. After adjusting for confounders, the AUC for ht-LVWV improved to 0.92 (95% CI: 0.85-0.98). CONCLUSION:Ht-LVWV measured on routine abdominal MRI is a promising biomarker for detecting LVH in ADPKD patients, with performance independent of key clinical variables. These findings support the potential for a diagnosis of MRI-LVH from standard ADPKD abdominal imaging protocols that are obtained routinely in the evaluation of ADPKD for renal and cardiovascular risk assessment in ADPKD.
OBJECTIVE:To develop and validate a prediction model combining Gd-EOB-DTPA-enhanced magnetic resonance imaging (MRI) diffusion-weighted imaging (DWI) parameters and clinicopathologic features for preoperative prediction of microvascular invasion (MVI) in hepatocellular carcinoma (HCC). METHODS:This study applied a retrospective method to collect preoperative MRI imaging data of patients with HCC from January 2021 to June 2025. All 279 patients (mean age of 58, M:F=203:76, 195 cases in training set and 84 cases in validation set) underwent MRI by Gd-EOB-DTPA and received DWI imaging scan before surgery. The MRI imaging features were observed, and the apparent diffusion coefficient (ADC) of the tumor solid region and tumor-to-liver parenchyma relative intensity ratio (RIR) were measured. LASSO logistic regression algorithm and multivariate analysis method were conducted to analyze the correlation between preoperative MRI imaging features and MVI. The prediction model was established, and the efficiency evaluation of the model was performed. RESULTS:Among 279 patients, 66 cases (23.66%) were MVI-positive. The study subjects were assigned to a training set (195 cases) and a validation set (84 cases) at a 7:3 ratio. Twelve variables were selected by LASSO regression. Multivariate analysis identified RIR transitional phase (OR=0.21), tumor size (OR=4.52), and ADC (OR=0.63) as independent predictors of MVI (P=0.029, <0.001, 0.032). The model achieved an AUC of 0.85 in the training set and 0.83 in the validation set. The negative predictive value (NPV) was 0.94 and 0.90 in the training and validation sets, respectively, while the positive predictive value (PPV) was limited to 0.49 and 0.40. Calibration curve and decision curve analyses demonstrated good consistency and clinical utility. CONCLUSION:On the basis of clinicopathologic features, MRI imaging features, and DWI parameters, this study preliminarily constructs a prediction model for positive MVI risk in HCC patients. The model exhibits good discrimination and a high NPV for effectively ruling out MVI, but its limited PPV warrants cautious interpretation of positive predictions due to a high false-positive rate.
OBJECTIVE:Osteoporosis is a major global health burden and often remains undiagnosed until fragility fractures occur. Dual-energy x-ray absorptiometry (DXA) is underutilized in routine practice. Computed tomography (CT) scans obtained for other indications offer an opportunity for opportunistic bone quality assessment without additional radiation exposure. This systematic review and meta-analysis evaluated: (i) the overall diagnostic performance [area under the receiver operating characteristic curve (AUC)] of artificial intelligence (AI) algorithms for osteoporosis/nonosteoporosis screening using CT; (ii) AUC for fracture/nonfracture screening; (iii) the most frequently studied body part; and (iv) the predominant machine learning (ML) and deep learning (DL) algorithms. METHODS:Following PRISMA guidelines, 2 reviewers searched PubMed, Web of Science, Scopus, and Google Scholar for English-language studies (2011 to 2025). Eligible studies used DXA as the reference standard and reported diagnostic performance metrics. Studies lacking patient-level AI assessment, using alternative reference standards, and incorrect analysis methods were excluded. Risk of bias was assessed using QUADAS-2. A random-effects meta-analysis pooled sensitivity, specificity, and AUC. Subgroup analysis was conducted by algorithm type and study size. Heterogeneity was quantified using the I2 and Q-statistic. RESULTS:Of 1258 screened, 31 studies were included [26/31 (83.87%) osteoporosis screening; 5/31 (16.13%) fracture screening]. For osteoporosis screening, the abdomen/lumbar spine (L-spine) was the most frequently evaluated body part [14/26 (53.85%) studies]. Within this subgroup, osteoporosis/nonosteoporosis (normal/osteopenia) [5/14 (35.71%) studies] was the most common outcome. For fracture screening, the abdomen/L-spine was the most assessed body part [4/5 (80%) studies]. Within this subgroup, fracture/nonfracture [4/4 (100%)] was the most common outcome. The pooled AUC for abdomen/L-spine osteoporosis/nonosteoporosis screening was 0.931 (95% CI: 0.926-0.936), and for abdomen/L-spine fracture/nonfracture screening was 0.863 (95% CI: 0.73-0.936). ML was used in 21/26 (80.77%) studies and DL in 4/26 (15.38%) studies for osteoporosis screening. Support vector machine (SVM) was the most common ML (13/21, 61.90%), while custom convolutional neural networks (CNNs) (3/4, 75%) predominated in DL. ML was used in 3/5 (60%) studies and DL in 1/5 (20%) studies for fracture screening. SVM was the most common ML (n=3/5, 60%), while custom CNN (1/1, 100%) predominated DL. Heterogeneity was lower for abdomen/L-spine osteoporosis/nonosteoporosis screening (I2=47.24%) than for fracture/nonfracture screening (I2=64.8%). CONCLUSIONS:AI algorithms applied to abdomen/L-spine demonstrate excellent performance for opportunistic osteoporosis/nonosteoporosis screening and fracture/nonfracture screening, supporting integration into routine CT interpretation to improve early detection and prevention.
OBJECTIVE:In patients with acute ischemic stroke, CT perfusion (CTP)-derived infarct core is valuable for prognostication, triage and transfer decision-making, and for informing studies of emerging therapeutic targets. In this study, we compare the accuracy and variability of the infarct core predicted by 2 FDA-cleared CTP programs, using diffusion-weighted MRI (DWI) as the reference standard. METHODS:We analyzed 61 stroke patients who underwent admission CTP and DWI within 90 minutes. Infarct core was estimated using relative cerebral blood flow thresholds and compared with DWI-derived ground truth. After coregistration of CTP and DWI, Dice similarity coefficients were calculated to quantify the topographic concordance of the infarct core. RESULTS:The CTP-determined infarct core volumes were significantly correlated with DWI but demonstrated substantial variability and bias. Our results showed limited topographic overlap, with median dice scores of 0.367 (CTP-A) and 0.289 (CTP-B). However, in patients with larger infarcts (volumes ≥50 mL), CTP provides more reliable estimates of spatial agreement, reaching median Dice scores of 0.61 (CTP-A) and 0.47 (CTP-B). CONCLUSIONS:Although these findings show a generally low accuracy of CTP estimation, they suggest that CTP may offer clinically meaningful insights for decision-making in the large-core setting and inform the design of future trials.
Hepatobiliary emergencies include both traumatic and nontraumatic pathologies that can affect the gallbladder, bile ducts, and liver. Due to the complexity and urgency of these conditions, advanced imaging techniques are required for rapid, accurate diagnosis. Traumatic liver emergencies were historically treated surgically; however, with the implementation of contrast-enhanced computed tomography (CT), management options have evolved. Contrast-enhanced CT imaging can provide additional insight into the extent and severity of injuries, enabling clinicians to decide between operative and nonoperative management with greater confidence. Furthermore, nontraumatic biliary emergencies present diagnostic challenges due to the diversity in their clinical presentations. When biliary pathology is suspected, the use of ultrasound, CT, and magnetic resonance imaging (MRI) aids in differentiating between conditions such as acute cholecystitis, acalculous cholecystitis, gangrenous cholecystitis, and emphysematous cholecystitis, thereby directing management strategies. Moreover, liver pathologies such as liver abscesses, hemorrhagic liver lesions, and vascular disorders demand integrated imaging assessments. The combination of ultrasound, CT, and MRI facilitates the differentiation between pyogenic and parasitic abscesses, evaluation of spontaneous liver ruptures, and diagnosis of portal vein thrombosis and Budd-Chiari syndrome, guiding clinical interventions. In addition, biloma and cholangitis, while less frequent, require precise diagnosis. The use of multimodality imaging enables early detection and intervention while mitigating potential complications. In conclusion, this review highlights the key role of advanced imaging in hepatobiliary emergencies and provides a comprehensive overview to help clinicians navigate complex scenarios effectively.
OBJECTIVE:To investigate the potential of magnetic resonance imaging (MRI) radiomics-based machine learning (ML) models in predicting local lymph node metastasis (LNM) at initial diagnosis in patients with laryngeal squamous cell carcinoma (LSCC). METHODS:This retrospective single-center study included 192 patients with pathologically confirmed LSCC who underwent preoperative contrast-enhanced T1-weighted neck MRI followed by curative surgical resection and lymphadenectomy. Tumor volumes were manually segmented, and 107 standardized radiomic features, along with clinical variables, were extracted. After collinearity reduction and feature selection, the most predictive features were used to train a random forest classifier. Model robustness was assessed through repeated cross-validation, and diagnostic performance was evaluated on an independent test set using area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. RESULTS:LNM was present in 78 patients (40.6%). T stage and supraglottic involvement showed significant associations with LNM (P<0.001 and P=0.020, respectively). Feature selection consistently identified 6 radiomic features, with "Sphericity" emerging as the most predictive across all runs. The final radiomics model achieved a mean AUC of 0.90 (95% CI: 0.85-0.94), with sensitivities of 0.96, specificities of 0.82, and an accuracy of 0.87. CONCLUSIONS:MRI-based radiomics demonstrates potential for the preoperative prediction of regional LNM in laryngeal cancer by capturing quantitative imaging features related to tumor heterogeneity, shape, and morphology. However, these findings should be interpreted with caution, given the retrospective single-center design and lack of external validation, and require confirmation in larger multicenter studies.
OBJECTIVE:Outline the expected course of normal sternal healing at different timepoints after surgery. MATERIALS AND METHODS:A total of 270 chest CT examinations in 228 unique patients performed between January 2018 and 2022 were retrospectively identified in patients who had undergone a first-time median sternotomy for elective surgery within the preceding 18 months. These examinations were divided into 4 intervals based on time since surgery-0 to 2 months, 2 to 6 months, 6 to 12 months, and >12 months. Basic patient data, surgery type, and imaging findings and measurements for the sternum and mediastinum were characterized and summarized. RESULTS:Osseous callus was initially rare (manubrium 5.1%, sternal body 7.8% to 8.7%) but increased over time, reaching 87.2% in the manubrium and 78.7% to 85.1% in the sternal body beyond 12 months, with consistently lower prevalence in the manubrium at intermediate timepoints. The sternal gap measurements were widest at 2 to 6 months and smallest at 0 to 2 months and >12 months. Pneumomediastinum was exclusively seen at 0 to 2 months (36.2% of examinations). Mediastinal fat stranding (95.7%) and lymphadenopathy (23.3%) were most common at 0 to 2 months and persisted in a minority of patients at later timepoints. CONCLUSIONS:Post-sternotomy healing is prolonged and variable, with the manubrium demonstrating slower callus formation and persistently wider sternal gaps than the sternal body. Early postoperative findings such as mediastinal fat stranding, lymphadenopathy, and osseous resorptive changes are common and may reflect normal healing rather than pathology, with callus typically evident after 6 months.
OBJECTIVE:Lung cancer is the leading cause of cancer-related deaths worldwide, and lung nodules serve as early indicators. This study aimed to evaluate computed tomography (CT) texture features to differentiate minimally invasive adenocarcinoma (MIA) from invasive adenocarcinoma (IAC) in subsolid nodules (SSNs). METHODS:This retrospective study included patients with lung adenocarcinoma presenting as subsolid nodules. Based on preoperative pathologic findings, patients were categorized into MIA and IAC groups. Morphologic characteristics-including location, size, density, shape, and lobulation-were extracted from preoperative thin-section CT scans. CT texture features were quantified using MaZda software through histogram parameters (mean, standard deviation, skewness, and kurtosis) and gray-level co-occurrence matrix metrics (contrast, entropy, inverse difference moment, and autocorrelation). RESULTS:Among 298 patients, 134 had MIA and 164 had IAC. Significant differences were found in CT morphologic features: IAC demonstrated larger size (11.65±3.17 mm vs. 9.32±2.41 mm, P<0.001), a higher prevalence of mixed ground-glass nodules (58.56% vs. 21.35%, P<0.001), more irregular shapes (64.86% vs. 32.58%, P<0.001), and more frequent lobulation (54.95% vs. 20.22%, P<0.001). Key texture features distinguishing IAC from MIA included higher mean attenuation (-448.69±46.52 HU vs. -562.47±145.28 HU, P<0.001) and entropy (6.95±0.22 vs. 6.66±0.38, P<0.001). The predictive nomogram integrating these features achieved an area under the curve of 0.831 in the training set and 0.893 in the validation set. CONCLUSION:Quantitative CT texture analysis-particularly entropy and mean attenuation-serves as a valuable tool for differentiating MIA from IAC in SSNs. This noninvasive approach enhances preoperative risk stratification and supports personalized surgical planning.
