Brain MRI Study Assesses Impact of AI in Differentiating Radionecrosis from Neoplastic Progression of Metastasis
Convolutional neural network-enabled segmentation of brain MRI offered a 25.7 percent higher specificity than a radiomic model for differentiating radionecrosis and metastatic progression in patients treated with stereotactic radiosurgery for brain metastases.
While overlapping features on magnetic resonance imaging (MRI) may make it challenging to differentiate between radionecrosis and metastatic progression in patients treated with stereotactic radiosurgery (SRS), artificial intelligence (AI) may have an impact in resolving this diagnostic challenge.
In a recent study presented at the Radiological Society of North America (RSNA) conference, researchers reviewed preoperative MRI data and histological findings for 96 patients (mean age of 56.4) with brain metastases who had stereotactic radiosurgery and subsequent neurosurgery for suspected recurrence or symptomatic disease recurrence. There was a total of 104 brain lesions in the cohort with 29 cases revealing sole radionecrosis, according to the study.
The use of adjunctive AI may enhance MRI capability for differentiating between radionecrosis and metastatic progression in patients treated with stereotactic radiosurgery, according to recent research presented at the RSNA conference. (Images courtesy of RSNA.)
The researchers found that the combined assessment of 3D T1 post-contrast MRI (T1c) and 3D T2 fluid-attenuated inversion recovery (FLAIR) imaging assessment via a radiomics model yielded an area under the receiver operating characteristic curve (AUROC) of 81 percent, sensitivity of 86.7 percent, accuracy of 79 percent and specificity of 60 percent.
In contrast, the study authors noted that combined assessment of T1ce and FLAIR imaging via convolutional neural network-enabled brain segmentation yielded similar sensitivity (87.8 percent) but an 8 percent higher AUROC (89 percent), 7.9 percent higher accuracy (86.9 percent) and 25.7 percent higher specificity (85.7 percent) for differentiating between radionecrosis and metastatic progression.
“Integrating artificial intelligence models with conventional MRI offers a promising non-invasive approach, potentially revolutionizing diagnostic methodologies in this challenging clinical context,” wrote lead study author Gaia Ressa, M.D., who is affiliated with the Department of Biomedical Sciences at Humanitas University in Milan, Italy, and colleagues.
(Editor’s note: For additional coverage from the RSNA conference, click here.)
While noting that post-surgical histological determination is currently the gold standard for differentiating between radionecrosis and metastatic progression, the researchers maintained that their study findings may have a significant impact upon the monitoring of patients who have undergone stereotactic radiosurgery for brain metastases.
“(Artificial intelligence models) have the potential to enhance patient management strategies and reduce unnecessary neurosurgical interventions,” emphasized Ressa and colleagues.
Reference
1. Ressa G, Levi R, Savini G, et al. Artificial intelligence accurately differentiates radionecrosis from true progression in brain metastasis treated with stereotactic radiosurgery: analysis of 104 histologically assessed lesions. Poster presented at the Radiological Society of North America (RSNA) 2024 110th Scientific Assembly and Annual Meeting Dec. 1-5, 2024. Available at: https://www.rsna.org/annual-meeting .
Can Abbreviated Breast MRI Have an Impact in Assessing Post-Neoadjuvant Chemotherapy Response?
April 24th 2025New research presented at the Society for Breast Imaging (SBI) conference suggests that abbreviated MRI is comparable to full MRI in assessing pathologic complete response to neoadjuvant chemotherapy for breast cancer.
Clarius Mobile Health Unveils Anterior Knee Feature for Handheld Ultrasound
April 23rd 2025The T-Mode Anterior Knee feature reportedly offers a combination of automated segmentation and real-time conversion of grayscale ultrasound images into color-coded visuals that bolster understanding for novice ultrasound users.
What is the Best Use of AI in CT Lung Cancer Screening?
April 18th 2025In comparison to radiologist assessment, the use of AI to pre-screen patients with low-dose CT lung cancer screening provided a 12 percent reduction in mean interpretation time with a slight increase in specificity and a slight decrease in the recall rate, according to new research.
Meta-Analysis Shows Merits of AI with CTA Detection of Coronary Artery Stenosis and Calcified Plaque
April 16th 2025Artificial intelligence demonstrated higher AUC, sensitivity, and specificity than radiologists for detecting coronary artery stenosis > 50 percent on computed tomography angiography (CTA), according to a new 17-study meta-analysis.
