System of joint brain tumor and cortex reconstruction

An apparatus for performing fully automatic brain tumor and tumor-aware cortex segmentation and reconstruction, comprising:

• (a) a computer processor;
• and
• (b) a non-transitory computer-readable memory storing instructions executable by the computer processor;
• (c) wherein said instructions, when executed by the computer processor, perform joint reconstruction and visualization of tumors, white matter, and gray matter surfaces automatically with steps comprising:
• (i) receiving multi-modal MRI data, comprising T1, T1c, T2, and T2-Flair data for a patient;
• (ii) performing 3D tumor segmentation and classification if data is sufficient, and training a model on-the-fly for tumor segmentation and classification if existing data is insufficient;
• (iii) utilizing random forest training for classifying tumor labels by using trained decision trees when training a model on-the-fly;
• (iv) performing a tumor-aware cortex segmentation utilizing tumor labels and adapting to presence of a tumor;
• and
• (v) reconstructing and visualizing both tumor and cortical surfaces utilizing both cortical and tumor labels to output images that provide diagnostic and surgical guidance in discerning between gray matter, white matter, tumor edema, and tumor active core regions.

A method of performing fully automatic brain tumor and tumor-aware cortex segmentation and reconstruction, by steps comprising:

• (a) receiving multi-modal MRI data, which may include T1, T1c, T2, and T2-Flair data, in an electronic device for imaging MRI data to perform joint reconstruction and visualization of tumors, white matter, and gray matter surfaces automatically;
• (b) performing 3D tumor segmentation and classification if data is sufficient, and training a model on-the-fly for tumor segmentation and classification if existing data is insufficient;
• (c) performing a tumor-aware cortex segmentation utilizing tumor labels and adapting to presence of a tumor;
• and
• (d) reconstructing and visualizing both tumor and cortical surfaces utilizing both cortical and tumor labels to output images that provide diagnostic and surgical guidance in discerning between gray matter, white matter, tumor edema, and tumor active core regions.