Accurate diagnosis of brain tumors using artificial intelligence

Nauwkeurige diagnose van hersentumoren met behulp van kunstmatige intelligentieCancer (2022). DOI: 10.3390/cancers14102363″ width=”800″ height=”374″/>

General diagram of the proposed radiomic and radiophysiomic approach, showing key steps: MRI data collection; calculation of imaging biomarkers; extraction of radiomic features, including segmentation of tumors and edema, data filtering and feature extraction; reduction and selection of the most relevant characteristics; development and validation of ML-based classification models; and testing the performance of the best performing classifiers. Credit: cancers (2022). DOI: 10.3390/cancers14102363

The classification of brain tumors – and thus the choice of optimal treatment options – can become more accurate and precise through the use of artificial intelligence in combination with physiological imaging. This is the result of an extensive study published in cancers and conducted by the Karl Landsteiner University of Health Sciences (KL Krems). Multi-class machine learning methods were used to analyze and classify brain tumors using physiological magnetic resonance imaging data. The results were then compared with classifications made by human experts. Artificial intelligence was found to be superior in areas such as accuracy, precision and misclassification, while professionals outperformed in sensitivity and specificity.

Brain tumors can be easily detected by magnetic resonance imaging (MRI), but their exact classification is difficult. Yet that is exactly what is crucial for choosing the best possible treatment options. Now an international team led by KL Krems has used data from modern MRI methods as the basis for: machine learning (ML) protocols and assessed the use of artificial intelligence to classify brain tumors. They found that classification using artificial intelligence can be superior to that of trained professionals in certain areas.

More MRI, more data

The team led by Prof. Andreas Stadlbauer, a scientist at the Central Institute for Medical Radiology Diagnostics at St. Pölten University Hospital, used both advanced and physiological MRI data for the study. Both methods provide more insight into the structure and metabolism of a brain tumor and have allowed for a better classification for some time. But the price to pay for such a differentiated image is vast amounts of data that must be expertly assessed. “We have now analyzed whether and how a artificial intelligence the use of ML can be enabled to support trained professionals in this mammoth task,” explains Prof. Stadlbauer. “And the results are promising. When it comes to accuracy, precision and avoiding misclassification, an AI can classify brain tumors well using MRI data.”

To achieve their impressive result, the team trained nine well-known Multiclass ML algorithms using MRI data from 167 previous patients who identified one of the five most common brain tumors and had accurate classification using histology. A total of 135 so-called classifiers were generated in a complex protocol. These are mathematical functions that assign the material to be examined to specific categories. “Unlike previous studies, we also took into account data from physiological MRIs,” explains Prof. Stadlbauer. “This included details about the vascular architecture of the tumors and their formation of new blood vessels, as well as the supply of oxygen to the tumor tissue.”


The team called the combination of data from different MRI methods with multiclass ML “radiophysiomics”. It is a term that is likely to catch on quickly, as the potential of this approach became apparent in the second part of the project, the test phase. Herein, the now trained multiclass ML algorithms were fed with corresponding MRI data from 20 current brains tumor patients and the results of the classifications thus obtained were compared with those of two certified radiologists. In doing so, the top two ML algorithms (referred to as “adaptive boosting” and “random forest”) outperformed the human assessment results in terms of accuracy and precision. Also, these ML algorithms resulted in less misclassification than by the professionals (5 vs 6). On the other hand, when it came to the sensitivity and specificity of the assessment, the human assessments proved more accurate than the AI ​​tested.

“This also makes it clear,” says Prof. Stadlbauer, “that the ML approach should not replace classification by qualified personnel, but complement it. Moreover, the time and effort required for this approach is currently still very high. But it presents an opportunity whose potential should be further pursued for everyday clinical use.” All in all, this study once again shows the focus of research at KL Krems on fundamental findings with real clinical added value.

Retrospective MRI analysis reveals pathophysiological process for early detection of recurrent glioblastoma

More information:
Andreas Stadlbauer et al, Radiophysiomics: Classification of brain tumors by machine learning and physiological MRI data, cancers (2022). DOI: 10.3390/cancers14102363

Provided by Karl Landsteiner University

Quote: Accurate Diagnosis of Brain Tumors Using Artificial Intelligence (2022, June 21) Retrieved June 21, 2022 from

This document is copyrighted. Other than fair dealing for personal study or research, nothing may be reproduced without written permission. The content is provided for informational purposes only.

Leave a Comment

Your email address will not be published.