On 76 test cases, the AI framework achieved Cobb angle MAE 2.4°, Lenke curve type accuracy 0.895, and reduced review time from 8.2 to 1.5 min/case.
PSAD remained below 0.11 ng/mL² across all ages in 1037 healthy German men, contrasting with continuously rising PSA. 95th percentile reference: PV 51.39 mL, PSA 2.79 ng/mL, PSAD 0.071 ng/mL².
Noncontrast CT-based unimodal imaging AI detected anterior circulation large vessel occlusion with sensitivity 0.78 (95% credible interval 0.68-0.86) and specificity 0.88, versus expert readers' sensitivity 0.62 and similar specificity; low-certainty retrospective evidence.
SIR Foundation research consensus panel calls for standardized acute DVT intervention endpoints: biologically informed patient selection, anatomic/physiologic metrics, validated severity tools, longitudinal biomarkers, and economic/QoL measures to advance precision and reduce po…
Vision–language models as an integrative layer for clinical artificial intelligence in radiology: A systems-level perspective
AI can transform radiology education by enabling automated needs assessments, personalized learning pathways, and synthetic board-style question generation, though cost, bias, and privacy concerns must be addressed.
Radiologists can review thoracic manifestations of obstetric and gynecologic disease, from preeclampsia and embolism to metastatic tumors and benign syndromes like Meigs syndrome, to aid timely diagnosis.
Invited Commentary: AI for Education, Education for AI: Rethinking Radiology Training.
Sex as an Overlooked Factor Influencing Blood Volume and Iodine Enhancement.
Chest Radiographic Patterns of Diffuse Lung Disease: A Systematic Approach.
CSF shunt failure: 14% at 1 month, 40% at 1 year, 50% at 2 years. Early complications: tip misplacement, improper connection; late: fracture, erosion, overdrainage. Obstruction, infection, or migration possible anytime.
Understanding lateral pelvic anatomy on MRI is crucial for detecting deep endometriosis and guiding surgical planning. A multidisciplinary approach improves preoperative staging and helps avoid misdiagnosis.
Role of Dual-Energy CT in Cholelithiasis and Associated Complications.
Radiologists should recognize local, regional, and distant patterns of lung cancer recurrence on post-treatment chest CT. Disseminated forms include aerogenous, miliary, pleural, and lymphangitic metastases. Early detection of these patterns can improve patient care.
Invited Commentary: Surveillance Imaging for Recurrence in Lung Cancer: The Expected, the Unexpected, and the Unknown.
Accessory Cavitated Uterine Mass.
CT first-line for early prostate post-procedure complications (hemorrhage, collections); fluoroscopy for urinary leaks/fistulas/strictures; MRI (+PSMA PET) for late complications & recurrence. Knowing expected vs unexpected findings avoids delayed diagnosis.
This RadioGraphics review details how advanced neuroimaging is crucial for planning and guiding high-intensity focused ultrasound, laser interstitial thermal therapy, and deep brain stimulation treatments for movement disorders and epilepsy.
This dermatologic primer helps breast radiologists evaluate common skin abnormalities by integrating clinical and imaging findings to guide biopsy decisions.
Invited Commentary: Sharper than a Scalpel: How Imaging Redefines What It Means to Operate on the Brain.
MRI of Lesions Growing Along the Pituitary Stalk.
Invited Commentary: Early Detection of Pancreatic Cancer: Are We Up for the Challenge?
Radiology Board Examinations: A Fundamental Shift.
A knowledge-distilled diffusion model with meta-guidance improved cone-beam CT image quality, achieving MAE 13.22 HU, SSIM 0.9516, and PSNR 30.35 dB, significantly outperforming baselines (p<0.01).
A logistic regression model using MRI and amyloid PET features classified tau positivity in amyloid-positive cohorts with AUC 0.92, accuracy 85%, sensitivity 83%, specificity 85% on external validation. MCI subjects with predicted tau positivity progressed to AD faster (p<10^-6).
Mixed Reality-Assisted Planning Enables Successful Coil Embolization of Anatomically Complex Peridevice Leak After LAAC.
3D radiomic texture analysis of quantitative muscle MRI reveals distinct fat infiltration patterns: higher entropy and contrast in Charcot-Marie-Tooth type 1A vs myotonic dystrophy type 1, reflecting reticular vs confluent fat replacement. CMT1A also showed steeper disto-proxima…
Editorial. The unspoken curriculum: grief across the arc of a neurosurgical career.
In 105 children with perinatal stroke, brain age prediction from T1 MRI showed accelerated aging (more positive brain age gaps) versus matched controls, especially in arterial ischemic stroke. Motor scores negatively correlated with brain age gaps, with sex-specific differences.
A deep learning algorithm (XProstate) for prostate cancer MRI detection performed similarly to radiologists in external validation: heat map AUC 0.76 (0.70-0.82) vs PI-RADS 0.73 (0.68-0.79), p=0.38. 202 patients, MR-US fusion biopsy gold standard.
AI synthetic CT from MR showed soft-tissue HU differences <50 HU and centroid shifts <2 mm vs reference CT in brain and prostate RT, but bone HU compression reached 347±14 HU in femoral heads.
A radiomics-deep learning nomogram using planar whole-body bone scintigraphy predicted MDP-avid bone metastases in non-small cell lung cancer with an external validation AUC of 0.870 (95% CI 0.840-0.900).
A new diffusion model (EqDiff-CT) synthesized head and neck CT from CBCT with substantial gains in structural fidelity and Hounsfield unit accuracy compared to CycleGAN and DDPM, potentially improving CBCT-guided treatment planning.
LLM summaries of clinical history for radiology were more comprehensive (37.14% top rating vs clinician, P<.001) and factual (68.05%, P<.001). The proprietary LLM was ranked most useful for protocoling (40.87% rank 1) and interpretation (44.61% rank 1).