Multi-modal Disease Classification in Incomplete Datasets Using Geometric Matrix Completion

Abstract

In large population-based studies and in clinical routine, tasks like disease diagnosis and progression prediction are inherently based on a rich set of multi-modal data, including imaging and other sensor data, clinical scores, phenotypes, labels and demographics. However, missing features, rater bias and inaccurate measurements are typical ailments of real-life medical datasets. Recently, it has been shown that deep learning with graph convolution neural networks (GCN) can outperform traditional machine learning in disease classification, but missing features remain an open problem. In this work, we follow up on the idea of modeling multi-modal disease classification as a matrix completion problem, with simultaneous classification and non-linear imputation of features. Compared to methods before, we arrange subjects in a graphstructure and solve classification through geometric matrix completion, which simulates a heat diffusion process that is learned and solved with a recurrent neural network. We demonstrate the potential of this method on the ADNI-based TADPOLE dataset and on the task of predicting the transition from MCI to Alzheimer's disease. With an AUC of 0.950 and classification accuracy of 87%, our approach outperforms standard linear and non-linear classifiers, as well as several state-of-the-art results in related literature, including a recently proposed GCN-based approach.

Using Base-ml to Learn Classification of Common Vestibular Disorders on DizzyReg Registry Data

“…Importantly, no single method is guaranteed to perform best on all datasets ( 30 ), which is why it is recommendable to test multiple algorithms and let their performances be compared and critically reviewed by a domain expert, instead of deciding on a single algorithm a priori. Therefore, as described in the introduction, we compare several linear, non-linear and neural-network based ML algorithms, along with a novel graph deep learning method that we recently proposed ( 6 , 12 , 13 ). Details on all classifier models and their parametrization are given in section Overview of Selected Classification Algorithms.…”

“…For better transparency, several methods can and should be investigated at the same time, subject to a comparable data pre-processing and cross-validation strategy. To this end, we compare several linear, non-linear and neural-network based ML algorithms, along with a novel graph deep learning method that we recently proposed ( 6 , 12 , 13 ). Following insights from multiple classification experiments for diagnostic decision support in our research over the last few years ( 4 , 6 , 13 , 14 ), we also provide a multi-faceted analysis of algorithm outcomes, including an examination of class imbalance, multiple classification metrics, patient feature distributions, and feature importances as rated by the classifiers.…”

Modern machine-learning can support diagnostic differentiation of central and peripheral acute vestibular disorders

“…We further proposed GMC [36] (denoted in the following as SingleGMC) to alleviate the common problem of missing values in medical data sets [37]. Recently, we have combined these ideas into multigraph matrix completion (MultiGMC) [38]. Here, we apply both the original SingleGMC approach [36] and MultiGMC to our data set.…”

“…In MultiGMC, instead of taking the sum, we use them as two separate graphs. We learn separate patient representations within these two graphs (a single spectral convolutional layer per graph) and aggregate them via selfattention, before computing the classification posterior [38]. The calculation of accuracy, F1-score, and ROC-AUC is performed as for LR/RF/ANN.…”

“…The models LR, RF, and ANN were based on implementations in the scikit-learn machine-learning library [39], while GMC [36] and MultiGMC [38] are custom implementations, based on PyTorch [40].…”