AI found to be on par with radiologists in diagnosing prostate cancer
An artificial intelligence system developed by UCLA researchers demonstrated comparable results with experienced doctors in reading magnetic resonance imaging scans.
An artificial intelligence system developed by UCLA researchers demonstrated comparable results with experienced doctors in reading magnetic resonance imaging scans.
The AI system—called FocalNet—is a convolutional neural network that is trained to detect prostate cancer lesions and predict their aggressiveness using the Gleason score, a grading system used to determine how aggressive the cancer is and what the best course of treatment will be.
FocalNet, which leverages an algorithm that comprises more than a million trainable variables, characterizes lesion aggressiveness and “fully utilizes distinctive knowledge” from multi-parametric MRI, according to researchers. The system was trained by having it analyze MRI scans from 417 men with prostate cancer.
Also See: Cancer prediction tool combines machine learning, radiomics
A study, published in IEEE Transactions on Medical Imaging, was presented last week at the IEEE International Symposium on Biomedical Imaging in Italy, where it was selected as the runner-up for best paper.
“Multi-parametric MRI (mp-MRI) is considered the best non-invasive imaging modality for diagnosing prostate cancer (PCa),” states the paper. “However, mp-MRI for PCa diagnosis is currently limited by the qualitative or semi-quantitative interpretation criteria, leading to inter-reader variability and a suboptimal ability to assess lesion aggressiveness. Convolutional neural networks are a powerful method to automatically learn the discriminative features for various tasks, including cancer detection.”
Results from the paper indicate that FocalNet was 80.5 percent accurate in reading MRIs, while radiologists with at least 10 years of experience were 83.9 percent accurate.
“With the comparison to the prospective performance of radiologists using the current diagnostic guideline, FocalNet demonstrated comparable detection sensitivity for index lesions and clinically significant lesions, only 3.4 percent and 1.5 percent lower than highly experienced radiologists without statistical significance,” states the paper.
According to the paper’s authors, their research indicates that an AI system like FocalNet could save time and potentially provide diagnostic assistance to less-experienced radiologists to improve their ability to diagnose prostate cancer.
The AI system—called FocalNet—is a convolutional neural network that is trained to detect prostate cancer lesions and predict their aggressiveness using the Gleason score, a grading system used to determine how aggressive the cancer is and what the best course of treatment will be.
FocalNet, which leverages an algorithm that comprises more than a million trainable variables, characterizes lesion aggressiveness and “fully utilizes distinctive knowledge” from multi-parametric MRI, according to researchers. The system was trained by having it analyze MRI scans from 417 men with prostate cancer.
Also See: Cancer prediction tool combines machine learning, radiomicsA study, published in IEEE Transactions on Medical Imaging, was presented last week at the IEEE International Symposium on Biomedical Imaging in Italy, where it was selected as the runner-up for best paper.
“Multi-parametric MRI (mp-MRI) is considered the best non-invasive imaging modality for diagnosing prostate cancer (PCa),” states the paper. “However, mp-MRI for PCa diagnosis is currently limited by the qualitative or semi-quantitative interpretation criteria, leading to inter-reader variability and a suboptimal ability to assess lesion aggressiveness. Convolutional neural networks are a powerful method to automatically learn the discriminative features for various tasks, including cancer detection.”
Results from the paper indicate that FocalNet was 80.5 percent accurate in reading MRIs, while radiologists with at least 10 years of experience were 83.9 percent accurate.
“With the comparison to the prospective performance of radiologists using the current diagnostic guideline, FocalNet demonstrated comparable detection sensitivity for index lesions and clinically significant lesions, only 3.4 percent and 1.5 percent lower than highly experienced radiologists without statistical significance,” states the paper.
According to the paper’s authors, their research indicates that an AI system like FocalNet could save time and potentially provide diagnostic assistance to less-experienced radiologists to improve their ability to diagnose prostate cancer.
More for you
Loading data for hdm_tax_topic #better-outcomes...