- COVID Brain Changes: A UK-based research team revealed that COVID can lead to substantial changes in brain tissues. The researchers examined brain MRIs from 785 individuals who were scanned before and after COVID’s peak (aged 51–81), including 401 people who contracted COVID between their two scans. The COVID group’s MRIs showed longitudinal reductions in gray matter in the orbitofrontal cortex and parahippocampal gyrus, tissue damage in regions associated with the primary olfactory cortex, and reductions in brain size. These declines were even evident when excluding the 15 patients who were hospitalized.
- Nuance’s Ambient Enhancement: Nuance Communications announced an upcoming expansion to its PowerScribe One reporting platform’s ambient AI capabilities, intended to streamline and enhance radiologists’ structured reporting output. The PowerScribe One platform’s Ambient Mode (currently used to create structured reports from free-form dictations) will add new Auto Impression functionality that turns report findings into auto-generated impression sections and follow-up recommendations.
- TB & LC Screening Alliance: Qure.ai, AstraZeneca, and global health non-profit PATH launched an alliance to expand lung cancer screening in low and middle-income countries. The partners will integrate lung cancer screening into tuberculosis screening pathways, leveraging Qure.ai’s algorithms for CXR-based lung disease detection, AstraZeneca’s lung cancer expertise, and PATH’s health program advocacy and execution. The partnership builds upon Qure.ai and AstraZeneca’s ongoing lung cancer screening alliance, and continues Qure.ai’s efforts to improve disease detection in developing countries.
- Liver Lesion Ultrasound AI: A new Radiology Artificial Intelligence study detailed an ultrasound AI model that was able to accurately detect and characterize focal liver lesions. The France-based team developed the model using B-mode abdominal ultrasound exams from 1,026 patients (2,551 images) and tested it with exams from 48 patients (155 images). The model matched or exceeded the performance of three clinicians (two experts) for detecting (90% specificity, 97% sensitivity), localizing (80%), and characterizing lesions as benign or malignant (81% specificity, 82% sensitivity).
- Enlitic Evidence: Fresh off revealing its pivot from traditional pixel-based diagnostic AI to a focus on standardizing radiology data, Enlitic provided the first evidence of how its Curie Standardize hanging protocol solution can help real world workflows. Over a three-year proof of concept study, UT Health San Antonio found that Curie Standardize would reduce its average CT (head, chest, abdomen) study organization time from 25 to 9 seconds. That adds up when multiplied over 480k annual studies, and as this recent radiologist Twitter discussion shows, it solves a problem multiple rads are looking for.
- FibroScan GO Launch: Echosens announced the launch of its FibroScan GO elastography-based liver screening system. The FibroScan GO is smaller and lower cost than Echosens’ other FibroScan devices, giving the company a more accessible solution for office-based practice settings and potentially allowing more PCPs to assess liver health (stiffness, fibrosis, steatosis) at the point of care.
- Neurosurgery MRI VR: An Italian research team developed an MRI-based mixed-reality solution that they believe might improve brain tumor removals and avoid damage to critical brain structures. The prototype solution automatically analyzes multimodal presurgical MRI scans (structural MRI, DTI, TOF) and transfers the results to a Microsoft HoloLens-integrated system, allowing neurosurgeons to utilize 3D holograms for planning and surgical navigation.
- DeepTek Funding: India-based AI startup DeepTek completed a $10M funding round that it will use to fund its global expansion, nearly 3.5 years after landing an early investment from NTT Data. DeepTek might not be a household name in the west, but its AI-integrated cloud PACS and AI-based population health screening products target two of the more intriguing radiology software categories.
- Ultrasound for SRM Surveillance: A study published in Urology suggests that ultrasound could be useful for active surveillance of small renal masses, potentially as an alternative to CT/MR cross-sectional imaging. Researchers analyzed data from 3,046 patients from 14 institutions who received ultrasound and cross-sectional imaging, and separated them into surgical (1,464 patients) and small renal mass cohorts (SRM, 1,582 patients, 1,921 imaging pairs). Among the SRM patients, 75% of the US and CT/MR measurements were within 0.5 cm and only 7.8% were different by more than 1 cm.
- Exo Works: Handheld ultrasound startup Exo announced the launch of its new POCUS workflow solution, Exo Works, intended to streamline patient data documentation and support collaboration, credentialing, and program tracking. Exo Works enables physicians to document, review, bill, and manage QA from their web-connected PCs or mobile devices and integrates directly with “virtually any” POCUS systems and most common EMRs and PACS systems.
- Cost of Incidentalomas: A new Clinical Imaging study revealed that the vast majority of a Dutch sarcoma center’s incidentalomas were later found to be “clinically irrelevant,” suggesting that the cost of detecting incidentalomas might rival the benefits. Incidentalomas were found in 28 of 221 patients (10.4%) who were referred to the specialized sarcoma center in 2018-19, but 23 of those were later labeled benign or low risk (82% of incidentalomas). These “irrelevant” incidentalomas led to multiple downstream medical procedures, resulting in an average cost of €1,857 per case (not to mention additional psychological distress).
