Automatic differentiation of Grade I and II meningiomas on magnetic resonance image using an asymmetric convolutional neural network

Abstract

The Grade of meningioma has significant implications for selecting treatment regimens ranging from observation to surgical resection with adjuvant radiation. For most patients, meningiomas are diagnosed radiologically, and Grade is not determined unless a surgical procedure is performed. The goal of this study is to train a novel auto-classification network to determine Grade I and II meningiomas using T1-contrast enhancing (T1-CE) and T2-Fluid attenuated inversion recovery (FLAIR) magnetic resonance (MR) images. Ninety-six consecutive treatment naïve patients with pre-operative T1-CE and T2-FLAIR MR images and subsequent pathologically diagnosed intracranial meningiomas were evaluated. Delineation of meningiomas was completed on both MR images. A novel asymmetric 3D convolutional neural network (CNN) architecture was constructed with two encoding paths based on T1-CE and T2-FLAIR. Each path used the same 3 × 3 × 3 kernel with different filters to weigh the spatial features of each sequence separately. Final model performance was assessed by tenfold cross-validation. Of the 96 patients, 55 (57%) were pathologically classified as Grade I and 41 (43%) as Grade II meningiomas. Optimization of our model led to a filter weighting of 18:2 between the T1-CE and T2-FLAIR MR image paths. 86 (90%) patients were classified correctly, and 10 (10%) were misclassified based on their pre-operative MRs with a model sensitivity of 0.85 and specificity of 0.93. Among the misclassified, 4 were Grade I, and 6 were Grade II. The model is robust to tumor locations and sizes. A novel asymmetric CNN with two differently weighted encoding paths was developed for successful automated meningioma grade classification. Our model outperforms CNN using a single path for single or multimodal MR-based classification.

A repository of grade 1 and 2 meningioma MRIs in a public dataset for radiomics reproducibility tests

“…The dataset can be found on http://www.cancerimagingarchive.net/ under the title "Segmentation and Classification of Grade I and II Meningiomas from MRI: an Open Annotated Dataset (Meningioma-SEG-CLASS)". 13 Alternatively, the dataset can also be accessed by digital object identifier (DOI) https://doi.org/…”

“…Magnetic Resonance Imaging (MRI) is the modality of choice in assessing meningiomas, and deep learning models have been developed for auto-differentiation of grade. [7][8][9][10] However, there are limited datasets available for researchers to develop and validate radiomic models. Furthermore, in the limited and small datasets available, thorough pathologic evaluation is unclear and there is a low representation of grade 2 meningiomas.…”

“…Likewise, radiomic features have the potential to predict meningioma grade. Magnetic Resonance Imaging (MRI) is the modality of choice in assessing meningiomas, and deep learning models have been developed for auto‐differentiation of grade 7–10 . However, there are limited datasets available for researchers to develop and validate radiomic models.…”

“…Our group has previously published on auto‐differentiation of grade 1 and 2 meningiomas utilizing an asymmetric 3D convolutional neural network (CNN) architecture using two encoding paths based on the T1post and T2‐FLAIR MRIs. The model was trained and tested on 96 patients (random rotation was used for data augmentation and assessed by tenfold cross‐validation, with each fold consisting of randomly selected 10 testing, 76 training, and 10 validation subjects) and showed improved grading accuracy using both T1post and T2‐FLAIR MRIs with accuracy rates of 90% 10 . Previous studies using a single sequence MRI showed reduced accuracy, 7–9 highlighting the importance of sharing datasets with the community to allow researchers to train models with new datasets and multiple MRI sequences.…”

AI-assisted clinical decision making (CDM) for dose prescription in radiosurgery of brain metastases using three-path three-dimensional CNN

“…AI has pushed the limits of what is possible in the domain of medical image processing, particularly in image registration, detection, segmentation, regression, and classification [9] , [10] , [11] , [12] , [13] . Meanwhile, AI has been reported to improve the quality and efficiency of a large variety of tasks in radiation oncology, such as image enhancement, treatment planning, organ segmentation, quality assurance, and treatment response prediction, as shown in many publications including ours [14] , [15] , [16] , [17] , [18] , [19] .…”