Random field image representations speed up binary discrimination of brain scans and estimate a phenotype glioblastoma cancer cell model

TL;DR

This paper introduces a faster method for analyzing brain scans using random field image representations, which also helps model cancer cells.

Contribution

The novel contribution is using spatially autoregressive models to speed up image analysis and estimate cancer cell phenotypes efficiently.

Findings

AR parameters estimated via OLS regression enable rapid dementia classification of MRI scans.

The RFR method produces a robust cancer cell model despite high noise levels.

The approach reduces data and computational requirements for deep learning in medical image analysis.

Abstract

MRI brain scans alone are not a definitive measure of dementia. Deep-learning algorithms (DLA) and professional human opinion are necessary for diagnosis. Yet, sample sizes are prohibitively large to train a typical DLA, which itself takes considerable computation time to produce diagnostically useful information from contrasting image features. We introduce analytic simplifications to this process to speed it up and reduce data requirements by modeling individual images as solutions of spatially autoregressive (AR) partial difference equations. Image features are the unique individual image AR parameters. Spatially lagged image pixels are explanatory variables for estimating a random-field representation (RFR) of the proposed AR difference equation. RFR model parameters are also those of the image autocorrelation function (ACF). An image pixel matrix-to-vector transformation allows AR parameters to be estimated by ordinary least squares (OLS) regression in millisecond time. Regression degrees of freedom (DOF) -- the number of image pixels -- are unusually large, leading to remarkably precise estimates of AR model parameters. These estimates support a solution of the binary dementia-normal classification of MRI axial brain scans (ADNI and OASIS archives). They also support the AR-RFR process applied to an original microscopic image of a glioblastoma cancer cell. In the face of formidable noise, a sharply defined and robust cancer cell model is estimated, which is an essential tool for cancer - type discrimination exercises and is parametrically plastic enough to serve a wide range of cells.