Latest Research
All publications from the Cancer3.AI database, newest first.
Magnetic resonance imaging thoracic organ-at-risk atlas for radiation oncology.
Bainbridge H, et al
Researchers developed a consensus MRI-based atlas to guide radiation oncologists in identifying and outlining critical organs at risk during radiotherapy planning for lung cancer. The study acquired MRI scans from nine patients with early-stage lung cancer using a 1.5 Tesla diagnostic MRI system, employing both T1-weighted and T2-weighted imaging sequences optimized for visualizing different thoracic structures. An international panel of six radiation oncologists and two MRI radiologists collaborated to create standardized contouring recommendations alongside high-resolution annotated MR images. The resulting atlas addresses a key gap in clinical practice, as thoracic oncologists have historically had limited experience interpreting thoracic anatomy on MRI compared to the more familiar CT scans. This resource is expected to improve the accuracy of contouring challenging structures such as the brachial plexus and to support the transition toward fully MRI-guided radiotherapy workflows, ultimately enhancing treatment precision for patients with locally advanced non-small cell lung cancer.
Physics and imaging in radiation oncology
Source →The Use of Topical Lidocaine Versus Lidocaine Injection for Myringotomy and Ventilation Tube Insertion.
Bacan F, et al
A new study published in Medicina examined whether topical lidocaine spray could serve as a less painful alternative to lidocaine injections for adults undergoing myringotomy and ventilation tube insertion, a common minor ear procedure. Fifty adult patients participated, with 29 receiving lidocaine injections and 21 receiving topical lidocaine spray, and all patients rated their post-procedure pain using three validated scales. Results showed that pain levels were significantly lower in the topical anesthesia group, with average visual analogue scale scores of 9.76 mm compared to 23.14 mm in the injection group, and numeric rating scale scores of 1.19 versus 2.41, both differences being statistically significant. These findings suggest that lidocaine spray offers effective pain control for this procedure while avoiding the discomfort and psychological distress associated with needle injections. For clinicians, this represents a practical, less invasive anesthetic option that may improve patient experience and reduce anxiety around minor ear surgeries performed in outpatient settings.
Medicina (Kaunas, Lithuania)
Source →Deep learning-based computer-aided diagnosis for parotid gland tumors on MRI.
Irifune Y, et al
Researchers investigated whether a deep learning-based computer-aided diagnosis (CAD) system could help radiologists more accurately distinguish between benign and malignant tumors of the parotid gland — the largest salivary gland — using MRI scans. The study trained an EfficientNet convolutional neural network on MRI data from 170 histologically confirmed cases and then tested how four readers (experienced radiologists and residents) performed with and without CAD assistance on 134 cases. The AI model itself achieved an accuracy of 85% and an area under the curve (AUC) of 0.93, and when readers used the CAD system, their combined accuracy rose from 76% to 86% and AUC from 0.82 to 0.94. Notably, both experienced radiologists and less-experienced residents benefited equally from the tool, with residents showing particularly strong gains. The system was especially helpful for identifying intermediate- and high-grade malignancies and more advanced-stage tumors, though it provided less benefit for low-grade or early-stage cancers. These findings suggest that AI-assisted MRI reading could meaningfully improve clinical decision-making for parotid gland tumor patients, potentially guiding more appropriate surgical and treatment planning.
Auris, nasus, larynx
Source →Long COVID Syndrome and Associated New Daily Persistent Headache (NDPH) Presenting With Crossed Clonus Response.
Wetmore E, et al
This case report investigates a patient with Long COVID syndrome who developed New Daily Persistent Headache (NDPH) alongside an uncommon neurological sign known as a crossed clonus response, highlighting the diagnostic complexity that can arise following SARS-CoV-2 infection. The study demonstrates that evaluating post-COVID patients with multiple chronic and progressive neurological symptoms requires a structured clinical approach combining a detailed temporal symptom history, a thorough neurological examination, and systematic screening for red flag features. The authors show that this framework can effectively distinguish potentially serious neurological pathology from more benign post-infectious phenomena, reducing the risk of missed diagnoses. The case is clinically significant because crossed clonus is an unusual upper motor neuron sign that, when encountered in a Long COVID context, may signal spinal cord or brainstem involvement requiring urgent further investigation. These findings provide practical, actionable guidance for neurologists and general practitioners managing the growing population of patients with complex post-COVID neurological presentations.
Clinical case reports
Source →Risk factors and exploratory clustering of complications after reconstruction following Mohs surgery: A national NIH All of Us study.
Parekh HN, et al
A large retrospective cohort study using data from the NIH All of Us Research Program examined complication rates and risk factors among 6,062 patients who underwent reconstructive surgery following Mohs micrographic surgery (MMS) for non-melanoma skin cancer. Overall complication rates were 3.2% at 30 days and 11.9% at one year, with older age, male sex, and comorbidities such as hypertension and diabetes identified as significant risk factors linked to complications including cellulitis, hematoma, wound dehiscence, and nerve injury. Primary wound closure was associated with a lower likelihood of one-year complications, while elevated body mass index (BMI) predicted specific adverse outcomes in initial analyses. Researchers applied K-means unsupervised machine learning clustering of patients by age and BMI, revealing four distinct subgroups, with younger patients carrying higher BMI experiencing the greatest one-year complication burden at 17.6%. These findings demonstrate that complication risk following post-MMS reconstruction is multifactorial and amenable to meaningful stratification, providing surgeons with an evidence-based framework for identifying high-risk patients and tailoring individualized postoperative care.
Journal of plastic, reconstructive & aesthetic surgery : JPRAS
Source →