At-rest primary motor cortex GABA was higher in mild-to-moderate Parkinson’s disease (32 patients) vs 28 controls on 7T MRS, but the difference did not survive FDR correction (pFDR=0.06). Inhibitory/excitatory ratio was also higher. Glutamate did not differ. Findings support 7T…
A new k-space-based AI model (LAPANet) accurately estimates cardiac MRI motion from just 2 Cartesian lines/frame or 3 radial spokes/frame, achieving sub-5 ms temporal resolution. This bypasses image reconstruction and avoids undersampling artifacts, enabling real-time motion tra…
A prostate cancer DLA tested on DLR-accelerated bpMRI showed similar per-patient performance. However, with C-T2WI+DLR-DWI, detected lesions fell (99 vs 131), FP/case dropped (0.478 vs 0.722, P=0.003), and IoU improved (0.194 vs 0.153, P=0.014).
FDA clears DRTECH Corporation's EXSYS DEXi digital X-ray detectors (models S40S, S40P, S40T, S50H) under 510(k) for general radiography.
A decision tree model using multi-phase CT radiomic features achieved AUC 0.977 (95% CI 0.934–1.000) in lesions with washout >60% and 0.983 (95% CI 0.962–1.000) in those without, distinguishing adrenal lipid-poor adenomas from pheochromocytomas.
Saturation or ipsilateral systematic plus targeted biopsy match standard combined biopsy for clinically significant prostate cancer detection in PI-RADS 3-5, with fewer biopsies. Targeted biopsy alone insufficient; systematic biopsy still needed for PI-RADS 5.
Radiology Partners Employer Profile - American College of Radiology Career Center
<![CDATA[Mushing Stuff Around a Dinner Plate: The Modern Approach to Health-Care Spending]]>
MPI is an emerging imaging modality providing real-time, quantitative visualization with no background and high sensitivity. This review covers MPI system architectures, rational tracer engineering via a Core-Shell-Function framework, and clinical applications such as cell track…
Can MRI predict tumor behavior in desmoid-type fibromatosis during active surveillance?
Assistive Artificial Intelligence Hold Promise in Improving Radiologist Workflow Efficiency.
Retrieval of a Prematurely Detached Coil by Intentional Thrombus Formation within the Microcatheter.
In 29 patients, an extended reality (XR) hand ruler achieved optimal PICC tip position in all cases with mean error 0.7 cm. Procedure time and radiation were noninferior to controls. Radiologists gave high ease-of-use ratings (mean Likert 1).
Corrigendum to 'Explainability-informed benchmarking of two deep learning models for organ-at-risk segmentation in MR-guided adaptive radiotherapy' [Journal of Medical Imaging and Radiation Sciences 57/3 (2026) 102200].
A deep learning radiomics model (TabResNet) on fat-suppressed T2-weighted MRI achieved AUC 0.8913 for differentiating benign vs malignant parotid tumors, outperforming Random Forest (AUC 0.8641) in 102 patients.
With enhanced prompting, GPT-5 graded cochlear hydrops on MRI with 87% accuracy and vestibular with 80%, but human-AI collaboration improved junior physicians' accuracy from 36% to 59%.
An ensemble deep learning model combining AP and lateral skull X-rays detected infant skull fractures with an AUC of 0.938 and accuracy of 91.6% on external validation, potentially reducing unnecessary CT scans.
Abbreviated breast MRI shows no significant sensitivity/specificity difference vs full MRI for screening extremely dense breasts, reducing scan and reading times. Best as an indication-specific strategy; limitations for lobular cancer, small lesions, and non-mass enhancement.
A two-stage deep learning system (YOLOv11 detection then EfficientNetV2-M classification) identified twelve implant brands on panoramic and periapical radiographs with 96.23% accuracy (95% CI 94.5–97.7%) and an end-to-end precision-aware F1-score of 88.2%.
In women with spontaneous intracerebral hemorrhage (ICH), a decision tree model using midline shift, blend sign, and time to first CT predicted hematoma expansion with an AUC of 0.803 (training) and sensitivity 81.8% (test), outperforming HEP model.
A stacking model integrating DCE-MRI habitat, radiomics, and clinical features differentiated luminal from non-luminal breast cancer with AUC 0.840 (external test set n=120), outperforming habitat-only (0.830) and radiomics-only (0.805) models.
Current AI imaging models for immunotherapy response are risk-stratification tools, not diagnostic classifiers of hyperprogressive disease (HPD). HPD diagnosis requires at least three imaging time points to capture growth acceleration relative to baseline trajectory. Most eviden…
Multimodal AI fusion of imaging, pathology, clinical records, and omics data can improve diagnostic sensitivity, tumor grading, and prediction of immunotherapy response and survival, but adoption is limited by fragmentation, bias, explainability, and regulation.
GPT-4o achieved 92.9% accuracy in labeling chest X-ray reports as diseased vs no disease. Training ConvNeXt-Tiny with these labels yielded AUC 0.832 (95% CI 0.801-0.863) for automated screening on MIMIC-CXR, outperforming EfficientNet-B1 (p=0.014). LLM-derived weak supervision i…
In a patient with knee injury, an adductor canal block led to saphenous (sensory) neuropathy, but denervation edema in the vastus medialis indicated motor branch involvement. Fluid tracking into the femoral triangle likely compressed multiple femoral nerve branches.
Cape Radiology Group Employer Profile - American College of Radiology Career Center
Recent Jobs - kaiser permanente/the permanente medical group, inc.
FLAIR alone gave 57.4% accuracy for WHO grade, ASL 81.5% for final diagnosis (glioblastoma vs astrocytoma), and T1w 66.7% for IDH status. Combining all sequences did not significantly improve performance, supporting streamlined MRI protocols.
An isotropic spherical region enlargement method using SO(3)-equivariant deep learning improved Dice score for the smallest sulcus by 11.16% over baseline, with performance gains negatively correlated with ROI size, confirming its utility for labeling small cortical sulci.
In 595 panoramic radiographs, EM2AI achieved 99.48% accuracy vs ThakaaMed’s 96.81% for tooth detection and numbering (excluding third molars). EM2AI was significantly more sensitive in upper anterior, lower anterior, and lower posterior regions, with no specificity difference. E…
