Liquid Biopsy versus CT: Comparison of Tumor Burden Quantification in 1065 Patients with Metastases Open Access 

Radiology, published 11/26/24 | Estimated Read Time: 25 minutes

Groundbreaking study of 1065 cancer patients with 56,288 metastatic lesions reveals that tumor fraction from liquid biopsy correlates only weakly with CT-measured tumor burden (R² = 0.17, P < .001), with 43% of patients showing low TF despite having large tumor volumes >151 cm³. While TF >10% demonstrated good specificity (80%) but limited sensitivity (57%) for detecting high tumor burden, the study found that liver metastases were significantly associated with contributory liquid biopsies, suggesting tumor location may be more important than overall volume for DNA shedding patterns.

Office-Based Diagnostic Imaging Interpreted by Nonphysician Practitioners: Characteristics, Recent Trends, and State Variation

Journal of the American College of Radiology, published 11/20/24 | Estimated Read Time: 15 minutes

Landmark analysis of Medicare claims (2013-2022), NPPs' share of office-based imaging interpretations more than doubled to 5.47%, with radiography showing the highest increase from 4.3% to 9.2% of all interpretations. The study reveals that just 5.5% of NPPs perform these interpretations, primarily in primary care and orthopedic practices, with male NPPs having 61% higher odds of interpreting imaging than female NPPs (OR: 1.61, 95% CI: 1.53-1.68) and dramatic state-level variations ranging from 20.88% NPP interpretation share in Alaska to 0.27% in Hawaii for radiography. 

Recent advances in nuclear medicine and their role in inflammatory arthritis: focus on the emerging role of FAPI PET/CT Open Access 

Skeletal Radiology, published 11/25/24 | Estimated Read Time: 30 minutes

FAPI PET/CT emerges as a game-changing imaging tool for inflammatory arthritis, with studies showing it detected 244 affected joints in rheumatoid arthritis patients, identifying 6.1% more affected joints than traditional FDG PET/CT, and demonstrated 100% sensitivity and 72% specificity in diagnosing periprosthetic joint infections. In a pivotal psoriatic arthritis study, FAPI PET/CT identified early disease activity in 80.6% of patients with increased tracer uptake in 318 joints and 369 entheses, with 11 out of 13 patients progressing to clinical PsA.

PiSlope: A New CT Metric for Quantifying Airway Remodeling in Chronic Obstructive Pulmonary Disease 

Radiology, published 11/26/24 | Estimated Read Time: 25 minutes

PiSlope is a novel CT metric for measuring airway remodeling in COPD that outperforms traditional Pi10 measurements by accounting for spatial variations across the entire airway tree. In a comprehensive study of 8,199 smokers and former smokers followed for a median of 3,826 days, lower PiSlope values were independently associated with worse clinical outcomes, including greater mortality (adjusted HR 0.88 per SD increase, 95% CI: 0.83-0.92, P < .001), faster FEV1 decline (-2.46 mL/year per SD decrease, 95% CI: -4.29 to -0.62, P = .009), and higher exacerbation rates (IRR 0.93, 95% CI: 0.89-0.98, P = .009), even after adjusting for Pi10 measurements. While the research shows promise in transforming how we assess airway disease progression, key limitations include lack of spirometry gating during CT acquisition and potential scanner variability across study sites.

MRI differentiation of CNS fungal infections: correlating imaging features with microbiological and histopathological findings

Neuroradiology, published 11/25/24 | Estimated Read Time: 25 minutes

In a study of 50 CNS fungal infection cases, researchers identified key MRI features that could predict fungal pathogen types, with dematiaceous molds showing significant correlations with dual rim sign (47.3%, p=0.04) and intracavitary susceptibility foci (52.6%, p=0.009) on SWI sequences. Among the study population (37 males, 13 females, mean age 39.3 years), Aspergillus was the most common pathogen (48%), with infections showing distinct patterns: yeasts associated with leptomeningitis (p<0.001), while larger fungi like Aspergillus and mucorales demonstrated macro-vessel occlusion patterns.

Cryoablation for treatment of peripheral lung metastases from colorectal cancer: a bicenter retrospective study 

European Radiology, published 11/22/24 | Estimated Read Time: 25 minutes

Bicenter study with N=73 shows cryoablation achieves 95% local progression-free survival at 1 year and 89% at 2-4 years for peripheral colorectal cancer lung metastases, while allowing a median 14-month chemotherapy-free interval (IQR: 5.0-21.5). Technical success was achieved in 100% of procedures with minimal complications (79% grade 1), though outcomes were significantly better in patients without liver metastases or synchronous lung metastases at initial diagnosis (p<0.05).

Interventional Radiology Treatments to Address Chronic Pelvic Pain in Females

Cardiovascular and Interventional Radiology, published 11/22/24 | Estimated Read Time: 30 minutes

Study demonstrates high success rates for established procedures like UAE for fibroids (96% technical success) and ovarian vein embolization for pelvic venous disease (92-95% clinical success with VAS scores reduced from 7 to 1). While highlighting promising early results for emerging treatments like cryoablation for abdominal wall endometriosis and sclerotherapy for peritoneal inclusion cysts (90% long-term success), the authors note that randomized controlled trials are still needed to strengthen the evidence base, particularly for adenomyosis where current data shows up to 50% of patients require hysterectomy within 3 years despite initial treatment. 

Restoring mobility: roles of percutaneous consolidation for pelvic ring bone lesions—a multicenter study

European Radiology, published 11/21/24 | Estimated Read Time: 20 minutes

Multicenter study of 143 procedures shows percutaneous consolidation for pelvic ring bone tumors achieved a 99% success rate for postoperative weight-bearing ambulation, with 82% of patients walking within 24 hours and 57% discharged within one day. The technique demonstrated excellent durability with 96.5% long-term consolidation success and only a 13% complication rate (mostly asymptomatic cement leakage), though the study was limited by its retrospective design and 4.5-month median follow-up.

Combining Radioembolization and Immune Checkpoint Inhibitors for the Treatment of Hepatocellular Carcinoma: The Quest for Synergy

Journal of Vascular and Interventional Radiology, published 11/22/24 | Estimated Read Time: 20 minutes

Comprehensive review examines the emerging combination of Y-90 radioembolization with immune checkpoint inhibitors for HCC treatment, where early trials demonstrate promising median overall survival of up to 27.3 months in advanced disease and objective response rates of 30-40% with a favorable safety profile (only 2-5 patients experiencing grade 3/4 adverse events across most trials). While these results appear encouraging, the authors note significant limitations in current evidence, as most trials have been small (20-40 patients), single-arm studies without personalized dosimetry, highlighting critical knowledge gaps in optimal patient selection, timing, and dosing strategies that need to be addressed through larger randomized trials.

A Decade Long Analysis of Healthcare Disparities and Uterine Artery Embolization: An Exploration of Social Determinants of Health

Journal of Vascular and Interventional Radiology, published 11/22/24 | Estimated Read Time: 25 minutes

Groundbreaking 10-year study analyzing 212,532 patients reveals that Uterine Artery Embolization (UAE) demonstrated greater-than-expected utilization among Non-Hispanic Black patients (45.6% vs expected 32.6%, p<0.0001) and equitable distribution across income quartiles, with Medicaid cases increasing significantly from 18.2% to 28.6% (p<0.001) over the decade. However, the study identified critical gaps in rural healthcare access, with only 1.2% of UAE procedures performed in rural hospitals despite serving 12.9% of the US population, though findings were limited by the study's focus on inpatient procedures only.

Effect of Artificial Intelligence as a Second Reader on the Lung Nodule Detection and Localization Accuracy of Radiologists and Non-radiology Physicians in Chest Radiographs: A Multicenter Reader Study Open Access 

Academic Radiology, published 11/25/24 | Estimated Read Time: 25 minutes

Multicenter study demonstrates that AI assistance significantly improved lung nodule detection accuracy for both radiologists and non-radiology physicians, with the area under the alternative free-response ROC curve increasing from 0.73 to 0.81 (p<.001) and missed nodules decreasing by 46.4% (from median 56 to 30 nodules out of 175), without significantly impacting false positive rates. The study of 300 chest radiographs and 15 readers showed particularly strong improvements for emergency physicians and pulmonologists (AFROC improvement of 0.12), though its sequential reading design without a washout period was noted as a limitation.

Artificial intelligence in respiratory pandemics—ready for disease X? A scoping review Open Access 

European Radiology, published 11/21/24 | Estimated Read Time: 25 minutes

Review spanning 96 studies across four major respiratory pandemics – researchers found consistent delays in establishing coordinated imaging data collection systems, with COVID-19's RICORD database taking over 12 months to become accessible despite analyzing 240 CT scans, while the subsequent STOIC dataset expanded to 10,735 CTs. The authors discovered that previous pandemic responses were consistently reactive rather than proactive, exemplified by MERS where despite being classified as a WHO priority pathogen since 2013, no successful international imaging data collection system exists even after 2,578 confirmed cases across 27 countries. This systematic analysis in European Radiology proposes an actionable framework for future pandemic preparedness, particularly critical for AI implementation, which showed improved reliability once sufficient data became available during COVID-19 but was initially hindered by lack of high-quality training data early in outbreaks.

Radiomics-based automated machine learning for differentiating focal liver lesions on unenhanced computed tomography  Open Access 

Abdominal Radiology, published 11/22/24 | Estimated Read Time: 30 minutes

Researchers achieved near-perfect accuracy (AUC 0.93-1.00) in distinguishing between four types of focal liver lesions using an automated machine learning model on unenhanced CT scans, significantly outperforming radiologists' diagnostic rates of 0.60-0.96. The model, tested on 260 patients across two medical centers, demonstrated exceptional performance by combining radiomics and clinical features, with the fusion model achieving AUCs of 1.00 for hepatic cysts, 0.97 for hemangiomas, 0.92 for abscesses, and 0.93 for malignancies in external validation.

Deep Learning Applied to Diffusion-weighted Imaging for Differentiating Malignant from Benign Breast Tumors without Lesion Segmentation

Radiology: Artificial Intelligence, published 11/20/24 | Estimated Read Time: 10 minutes

Deep learning approach using a small 2D CNN matched radiologists' performance in differentiating breast tumors on DWI alone, achieving an AUC of 0.88 (vs. radiologists' 0.86) across 334 breast lesions, with the key advantage of requiring no manual lesion segmentation. The model demonstrated comparable specificity (81.4% vs. 72.1%) and sensitivity (85.9% vs. 98.8%) to radiologists' assessments of full multiparametric breast MRI, though the single-center study design and lack of external validation warrant further investigation before clinical implementation. 

AI-Assisted Post Contrast Brain MRI: Eighty Percent Reduction in Contrast Dose Open Access 

Academic Radiology, published 11/25/24 | Estimated Read Time: 25 minutes

Deep learning algorithms achieved an 80% reduction in gadolinium contrast dose for brain MRI while maintaining diagnostic quality, with DL-reconstructed images showing higher SNR and vessel conspicuity scores (P < 0.05) and achieving a structural similarity index of 86 ± 12.1% compared to full-dose images across 101 patients. While the technology demonstrated statistical non-inferiority for lesion visualization across all three independent neuroradiologist readers (P < 0.001), researchers noted limitations including occasional reduced vascular enhancement and pseudo-enhancement artifacts, suggesting larger studies across more diverse pathologies are needed before clinical implementation.

Upcoming Conferences and Approaching Abstract Submission Deadlines

If you have any comments, questions, or feedback, please respond to this email!

Thank you,

Emily Barnard