Magnetic resonance image-based brain tumour segmentation methods: A systematic review

Abstract

Background Image segmentation is an essential step in the analysis and subsequent characterisation of brain tumours through magnetic resonance imaging. In the literature, segmentation methods are empowered by open-access magnetic resonance imaging datasets, such as the brain tumour segmentation dataset. Moreover, with the increased use of artificial intelligence methods in medical imaging, access to larger data repositories has become vital in method development. Purpose To determine what automated brain tumour segmentation techniques can medical imaging specialists and clinicians use to identify tumour components, compared to manual segmentation. Methods We conducted a systematic review of 572 brain tumour segmentation studies during 2015–2020. We reviewed segmentation techniques using T1-weighted, T2-weighted, gadolinium-enhanced T1-weighted, fluid-attenuated inversion recovery, diffusion-weighted and perfusion-weighted magnetic resonance imaging sequences. Moreover, we assessed physics or mathematics-based methods, deep learning methods, and software-based or semi-automatic methods, as applied to magnetic resonance imaging techniques. Particularly, we synthesised each method as per the utilised magnetic resonance imaging sequences, study population, technical approach (such as deep learning) and performance score measures (such as Dice score). Statistical tests We compared median Dice score in segmenting the whole tumour, tumour core and enhanced tumour. Results We found that T1-weighted, gadolinium-enhanced T1-weighted, T2-weighted and fluid-attenuated inversion recovery magnetic resonance imaging are used the most in various segmentation algorithms. However, there is limited use of perfusion-weighted and diffusion-weighted magnetic resonance imaging. Moreover, we found that the U-Net deep learning technology is cited the most, and has high accuracy (Dice score 0.9) for magnetic resonance imaging-based brain tumour segmentation. Conclusion U-Net is a promising deep learning technology for magnetic resonance imaging-based brain tumour segmentation. The community should be encouraged to contribute open-access datasets so training, testing and validation of deep learning algorithms can be improved, particularly for diffusion- and perfusion-weighted magnetic resonance imaging, where there are limited datasets available.

Artificial Intelligence with Great Potential in Medical Informatics: A Brief Review

“…Computational pathology focuses on multiple data sources, such as pathological and tissue image information, using AI models to perform tasks such as detection, diagnosis, prediction, and prognosis, and showcasing clinically applicable knowledge to patients (Verghese et al, 2023). These studies on image processing, also known as imageomics (He et al, 2021), not only include pathological slice information but also X-ray images , magnetic resonance images (Bhalodiya et al, 2022), and so on.…”

Informatics Opportunities and Challenges in Medical Imaging: A Journey

“…Image segmentation is a critical step in the analysis and subsequent diagnostic/prognostic characterisation of brain tumours, using MR-based functional imaging approaches. The team has conducted a systematic review of 572 brain tumour segmentation studies, as reported during the period 2015-2020 [32]. The review assessed "physics or mathematics-based methods, deep learning methods, and software-based or semiautomatic methods, as applied to magnetic resonance imaging techniques" [32], including T1-weighted, T2-weighted, gadolinium-enhanced T1-weighted, fluidattenuated inversion recovery, diffusion-weighted and perfusion-weighted MR imaging.…”

“…The team has conducted a systematic review of 572 brain tumour segmentation studies, as reported during the period 2015-2020 [32]. The review assessed "physics or mathematics-based methods, deep learning methods, and software-based or semiautomatic methods, as applied to magnetic resonance imaging techniques" [32], including T1-weighted, T2-weighted, gadolinium-enhanced T1-weighted, fluidattenuated inversion recovery, diffusion-weighted and perfusion-weighted MR imaging. The performance of each segmentation method was assessed through its median Dice score (initially proposed as image segmentation performance metric by Zijdenbos et al [33]).…”

Segmentación automática del glioblastoma y su edema periférico utilizando Otsu, Chan-Vese y una red neuronal

“…mediante herramientas computacionales como la minería de datos y la inteligencia artificial (Vogelbaum et al 2022). Estos métodos, en lugar de guiar o reemplazar a los radiólogos, pueden proporcionarles asistencia terapéutica, aumentando potencialmente la eficiencia y mejorando la eficacia (Bhalodiya et al 2022). Entre estas nuevas ciencias, la radiómica tiene como objetivo adquirir, de forma específica para cada paciente, información cuantitativa que el operador no puede obtener a simple vista (Mu et al 2022).…”

“…Existe mucha literatura sobre el uso de la radiómica en la segmentación, diagnóstico, pronóstico, evaluación del tratamiento, detección de pseudoprogresión y predicción de recurrencia/supervivencia de GBM (Aftab et al 2022), existiendo hoy en día una tendencia a combinar imágenes multimodales, análisis de subregiones tumorales y enfoques de aprendizaje automático para lograr medicina personalizada (Zhang et al 2022). La segmentación automática de tumores cerebrales por radiómica resulta relevante para la caracterización y planificación del volumen objetivo en radioterapia de precisión (Bhalodiya et al 2022). Sin embargo, la segmentación de GBM sigue siendo un problema abierto que actualmente es el objetivo de la Universidad de Pensilvania en su desafío anual de segmentación de tumores cerebrales BraTS (Menze et al 2015), el último en 2021 (Baid et al 2021).…”