Classification of Astrocytoma using MLP and LSTM

Astrocytoma is a common type of glioma that can range in severity of cancer. Precisely classifying gliomas into high-grade glioma (HGG) and low-grade glioma (LGG) forms is essential for treatment planning and patient prognosis. By combining a Long Short-Term Memory (LSTM) network for sequential analysis of MRI slices with a Multi-Layer Perceptron (MLP) for spatial feature extraction, this study investigates a novel method for astrocytoma classification utilizing a hybrid deep learning model. FLAIR, T1, T1ce, and T2 scans are among the multimodal MRI data processed by the algorithm, which also integrates patient metadata for improved classification. Using the BraTS2020 dataset, our trials show that this design works well, delivering excellent classification accuracy.