An integrative latent class model of heterogeneous data modalities for diagnosing kidney obstruction

Abstract

Radionuclide imaging plays a critical role in the diagnosis and management of kidney obstruction. However, most practicing radiologists in US hospitals have insufficient time and resources to acquire training and experience needed to interpret radionuclide images, leading to increased diagnostic errors. To tackle this problem, Emory University embarked on a study that aims to develop a computer-assisted diagnostic (CAD) tool for kidney obstruction by mining and analyzing patient data comprised of renogram curves, ordinal expert ratings on the obstruction status, pharmacokinetic variables, and demographic information. The major challenges here are the heterogeneity in data modes and the lack of gold standard for determining kidney obstruction. In this article, we develop a statistically principled CAD tool based on an integrative latent class model that leverages heterogeneous data modalities available for each patient to provide accurate prediction of kidney obstruction. Our integrative model consists of three sub-models (multilevel functional latent factor regression model, probit scalar-on-function regression model, and Gaussian mixture model), each of which is tailored to the specific data mode and depends on the unknown obstruction status (latent class). An efficient MCMC algorithm is developed to train the model and predict kidney obstruction with associated uncertainty. Extensive simulations are conducted to evaluate the performance of the proposed method. An application to an Emory renal study demonstrates the usefulness of our model as a CAD tool for kidney obstruction.

Keywords: Bayesian prediction; Function-on-scalar regression; Heterogeneous data modalities; Integrative modeling; Latent class model; Scalar-on-function regression.

Fig. 1

The top panel depicts the mean baseline (left) and mean post-furosemide (right) renogram curves, stratified by the unanimous interpretation of the three nuclear medicine experts in the Emory renal study (black solid line for non-obstructed; red/black dashed line for equivocal; and green/gray solid line for obstructed). The bottom panel shows the baseline (left) and post-furosemide (right) renogram curves of 12 selected kidneys from the Emory renal study. Herein, 4 are unanimously interpreted as non-obstructed, 4 are unanimously interpreted as equivocal, and 4 are unanimously interpreted as obstructed by the three nuclear medicine experts (same coloring and line type as in the top panel).

Fig. 2

Predictive probabilities of kidney obstruction and intervals representing the 5th and 95th percentiles of the posterior MCMC samples $\{\widehat{P}(C_{\text{new}}=1|\mathcal{D},{\mathcal{D}}_{\text{new}},{\Theta }^{[q]});q=1,\dots ,Q\}$ for 50 testing kidneys. Black and green (gray in print) colors, respectively, denote non-obstructed and obstructed kidneys based on the semi-gold standard (review panel of nuclear medicine experts). Labels #6, #12, and #50 are indices of kidneys whose diagnoses differed between the proposed model and the semi-gold standard.