AI May Help Improve Brain Cancer Diagnosis
Traditionally, brain tumors are diagnosed by computer topography scans (CT), magnetic resonance imaging (MRI) or a neurologic exam. However, researchers from the Karl Landsteiner University for Health Sciences (KL Krems) have now discovered a more precise and accurate way of diagnosing brain cancer by using artificial intelligence (AI) combined with physiological imaging.
The researchers used several multiclass machine learning techniques to study data sourced from MRIs and classified different brain tumors. They then compared the results produced through machine learning with classifications that were made by human specialists.
Although the experts were better at specificity and sensitivity, the researchers discovered that the AI was superior in terms of precision, accuracy and spotting misclassification.
This shows that machine learning can be used to improve brain cancer diagnosis. Although MRIs can be used to easily spot a brain tumor, making a classification with MRI data only can be difficult, which in turn makes choosing the optimal treatment option more challenging. The team from KL Krems sought to fix this issue by finding a more accurate way of making brain cancer diagnoses and classifications.
The researchers, led by Professor Andreas Stadlbauer, leveraged both physiological and advanced MRI data and used both datasets to gain a deeper understanding of the metabolism and structures of a brain tumor as well as find better ways to classify tumors. However, this would involve assessing an inordinate amount of data, hence the use of artificial intelligence and machine learning.
Stadlbauer explained that the study was instrumental in proving that machine learning can indeed be used to analyze MRI data and make more accurate brain tumor classifications.
First, the research team had to train nine multiclass machine learning algorithms using MRI data sourced from 167 patients who had one of the five most prevalent types of brain tumors. They then used histology, which is the study of the anatomy, structure, and role of animal and plant tissues, to teach the AIs precise classification.
Stadlbauer noted that his study also used data sourced from physiological MRIs, something that previous studies had not done. This included details on oxygen supply to tumors, how the tumors form new vessels and vascular structures of tumors.
After comparing classifications made by the AI with those made by human professionals, the research team found that the AI’s classifications were more accurate and precise.
Still, the team cautions against using AIs as a substitute for qualified personnel. Rather, machine learning could be used as a tool to supplement brain cancer diagnosis and find the most optimal treatments.
The researchers published their findings in the “Cancers” journal.
As some teams focus on improving brain cancer diagnosis, others at companies such as CNS Pharmaceuticals Inc. (NASDAQ: CNSP) are looking to find new treatments that will improve patient outcomes. The convergence of these different efforts bodes well for the future of cancer care across the globe.
NOTE TO INVESTORS: The latest news and updates relating to CNS Pharmaceuticals Inc. (NASDAQ: CNSP) are available in the company’s newsroom at https://ibn.fm/CNSP
Please see full terms of use and disclaimers on the BioMedWire website applicable to all content provided by BMW, wherever published or re-published: http://BMW.fm/Disclaimer
Traditionally, brain tumors are diagnosed by computer topography scans (CT), magnetic resonance imaging (MRI) or a neurologic exam. However, researchers from the Karl Landsteiner University for Health Sciences (KL Krems) have now discovered a more precise and accurate way of diagnosing brain cancer by using artificial intelligence (AI) combined with physiological imaging.
The researchers used several multiclass machine learning techniques to study data sourced from MRIs and classified different brain tumors. They then compared the results produced through machine learning with classifications that were made by human specialists.
Although the experts were better at specificity and sensitivity, the researchers discovered that the AI was superior in terms of precision, accuracy and spotting misclassification.
This shows that machine learning can be used to improve brain cancer diagnosis. Although MRIs can be used to easily spot a brain tumor, making a classification with MRI data only can be difficult, which in turn makes choosing the optimal treatment option more challenging. The team from KL Krems sought to fix this issue by finding a more accurate way of making brain cancer diagnoses and classifications.
The researchers, led by Professor Andreas Stadlbauer, leveraged both physiological and advanced MRI data and used both datasets to gain a deeper understanding of the metabolism and structures of a brain tumor as well as find better ways to classify tumors. However, this would involve assessing an inordinate amount of data, hence the use of artificial intelligence and machine learning.
Stadlbauer explained that the study was instrumental in proving that machine learning can indeed be used to analyze MRI data and make more accurate brain tumor classifications.
First, the research team had to train nine multiclass machine learning algorithms using MRI data sourced from 167 patients who had one of the five most prevalent types of brain tumors. They then used histology, which is the study of the anatomy, structure, and role of animal and plant tissues, to teach the AIs precise classification.
Stadlbauer noted that his study also used data sourced from physiological MRIs, something that previous studies had not done. This included details on oxygen supply to tumors, how the tumors form new vessels and vascular structures of tumors.
After comparing classifications made by the AI with those made by human professionals, the research team found that the AI’s classifications were more accurate and precise.
Still, the team cautions against using AIs as a substitute for qualified personnel. Rather, machine learning could be used as a tool to supplement brain cancer diagnosis and find the most optimal treatments.
The researchers published their findings in the “Cancers” journal.
As some teams focus on improving brain cancer diagnosis, others at companies such as CNS Pharmaceuticals Inc. (NASDAQ: CNSP) are looking to find new treatments that will improve patient outcomes. The convergence of these different efforts bodes well for the future of cancer care across the globe.
NOTE TO INVESTORS: The latest news and updates relating to CNS Pharmaceuticals Inc. (NASDAQ: CNSP) are available in the company’s newsroom at https://ibn.fm/CNSP
Please see full terms of use and disclaimers on the BioMedWire website applicable to all content provided by BMW, wherever published or re-published: http://BMW.fm/Disclaimer
