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…
Primitive basal vein of Rosenthal (BVR) drainage (type 3) was more frequent in PMNSAH than controls (47.4% vs 24.4%), but BVR variants did not predict hemorrhage severity or neurological outcome. Age was independently associated with worse outcomes.
ASFNR clinical state-of-practice: For posterior circulation large vessel occlusion stroke, use noncontrast CT to exclude hemorrhage, CT angiography to identify occlusion, and MRI with DWI as reference standard. Imaging scores like PC-ASPECTS help select patients for thrombectomy.
An AI algorithm for CTA detected anterior circulation large vessel occlusions with 95.8% sensitivity and 88.6% specificity. In a multicenter study of 379 patients, mean time-to-notification was just 3.30 minutes, supporting its use as a rapid triage tool.
7T MRI more than doubles venous cerebral microbleed (vCMB) detection in CAA vs 3T (1173 vs 385), with higher vCMB ratio (0.70 vs 0.33). But weak cross-scanner correlation (r=0.32) precludes interchangeability; 7T may better detect small lesions.
Conservative management predominated (94%) in 573 renal trauma patients. Embolization achieved 100% technical success, nephrectomy rate 3%. 30-day all-cause mortality 2.3%, renal-specific 0.2%.
For 100 kVp ultra-low-dose chest CT, a machine learning phantom-less QCT model reduced bone mineral density error vs standard 120 kVp model (2.39 vs 16.68 mg/cm³, p<0.0001) and agreed with DXA in 88.8%.
Why AI struggles with rare catastrophic disease: lessons from the AORTA-AI study of acute aortic syndrome.
A narrative review finds deep learning models, including CNNs and GANs, can segment and classify lumbar spine radiographs, with lightweight architectures potentially deployable in resource-limited environments. However, real-world validation and clinical integration challenges p…
A multimodal signature combining CT radiomics, deep learning, and pathomics achieved a C-index of 0.821 for progression-free survival prediction in chondrosarcoma, outperforming individual signatures (validation cohort).
A radiomics-based support vector machine (SVM) model using parotid ultrasound distinguished Sjögren's disease from healthy controls with AUC 0.99, accuracy 0.94, sensitivity 0.86, specificity 0.96, outperforming radiologists' accuracy of 0.62–0.72.
A hierarchical deep learning model on X-rays detected acute osteoporotic vertebral fractures with 91% sensitivity (95% CI 84.8-95.0%) and 98.8% NPV.
In HCC patients on atezolizumab/bevacizumab, AI-assisted CT volumetry showed high-volume ascites (≥613.5 mL) independently predicted worse overall survival (HR 4.01, p<0.001), while low-volume ascites did not.
In a survey of 215 Society of Breast Imaging radiologists, 47% had already implemented AI for mammography detection. However, AI users reported significantly fewer reductions in recall rates (34.7% vs 59.3% anticipated by non-users, p=0.003), biopsy rates (9.1% vs 36.4%, p<0.001…
Super-resolution reconstruction of 5-mm chest CT lifted artificial intelligence lung nodule sensitivity from 31.4% to 61.0% and positive predictive value from 42.9% to 80.0% in 96 colorectal cancer patients (20% hallucinated nodules).
A lightweight volumetric AI foundation model, SAT3D, trained on 17,075 3D volume-mask pairs, generalized across 11 public tumour segmentation datasets, including out-of-distribution scans; a 3D-Slicer plugin supports interactive use.
UAB startup LifeNuclear is developing TheraGuide, a digital platform to replace complex paper-based radiation safety instructions with interactive checklists, multilingual support, and personalized timelines after radiopharmaceutical therapy.
ITM receives Complete Response Letter for ITM-11
ACR applauds new legislation on Medicare physician payments
Many women interested in AI-based risk assessment for breast cancer
Teledyne will acquire Varex Imaging for $1.1B, adding X-ray tubes and photon-counting detectors to its medical imaging portfolio. The deal, expected to close early 2027, may accelerate next-gen imaging technology for radiologists.
From Dyspnea to Suspected Breast Cancer: A Post-traumatic Hematoma as a Diagnostic Pitfall
<![CDATA[Breast MRI Study Reveals 29 Percent Reduction in Scan Time with Deep Learning Reconstruction and Multi-Shot DWI]]>
RadNet reported a 21.2% aggregate jump in advanced imaging volumes (MRI, CT, PET/CT) in Q2 compared to last year, leading the company to raise its 2026 revenue forecast to over $2.42 billion. Same-center advanced imaging volumes grew 9.6%, outpacing routine exams.
<![CDATA[The Weight of Numbers in Radiology]]>
Pediatric radiologist Fariba Goodarzian, MD, appointed radiologist-in-chief at Children’s Hospital Los Angeles, where she has worked >25 years; will oversee MRI, CT, nuclear medicine, and interventional radiology.
In a phantom study, a 30-40% data extraction percentage for data-driven gating combined with deep learning PET reconstruction (AiCE-i) yielded optimal balance between image noise and respiratory motion effects.
Gemini 3.0 achieved mean Dice 0.935 for zero-shot STS segmentation (radiologist inter-rater Dice 0.949). Tumor size from masks had ICC 0.976 vs 0.922 for direct LLM estimate (p=0.027), with bias -0.5 mm vs -5.0 mm. Segmentation-derived measurements are more reliable.
Imaging evaluation of liver metastases is evolving from purely morphologic assessment to biologically informed techniques. Advanced imaging, radiomics, and artificial intelligence can assess tumor biological heterogeneity, guide therapy, and predict outcomes.
Pairing hip X-ray or MRI with radiology text descriptions boosted multimodal LLMs' detection of femoral head osteonecrosis: AUC 0.91 vs 0.55 for image-only, and staging reliability ICC 0.97 vs 0.51. Multi-image input offered no benefit.
