TY - JOUR
T1 - A deep-learning-based unsupervised model on esophageal manometry using variational autoencoder
AU - Kou, Wenjun
AU - Carlson, Dustin A.
AU - Baumann, Alexandra J.
AU - Donnan, Erica
AU - Luo, Yuan
AU - Pandolfino, John E.
AU - Etemadi, Mozziyar
N1 - Publisher Copyright:
© 2021 Elsevier B.V.
PY - 2021/2
Y1 - 2021/2
N2 - High-resolution manometry (HRM) is the primary method for diagnosing esophageal motility disorders and its interpretation and classification are based on variables (features) from data of each swallow. Modeling and learning the semantics directly from raw swallow data could not only help automate the feature extraction, but also alleviate the bias from pre-defined features. With more than 32-thousand raw swallow data, a generative model using the approach of variational auto-encoder (VAE) was developed, which, to our knowledge, is the first deep-learning-based unsupervised model on raw esophageal manometry data. The VAE model was reformulated to include different types of loss motivated by domain knowledge and tuned with different hyper-parameters. Training of the VAE model was found sensitive on the learning rate and hence the evidence lower bound objective (ELBO) was further scaled by the data dimension. Case studies showed that the dimensionality of latent space have a big impact on the learned semantics. In particular, cases with 4-dimensional latent variables were found to encode various physiologically meaningful contraction patterns, including strength, propagation pattern as well as sphincter relaxation. Cases with so-called hybrid L2 loss seemed to better capture the coherence of contraction/relaxation transition. Discriminating capability was further evaluated using simple linear discriminative analysis (LDA) on predicting swallow type and swallow pressurization, which yields clustering patterns consistent with clinical impression. The current work on modeling and understanding swallow-level data will guide the development of study-level models for automatic diagnosis as the next stage.
AB - High-resolution manometry (HRM) is the primary method for diagnosing esophageal motility disorders and its interpretation and classification are based on variables (features) from data of each swallow. Modeling and learning the semantics directly from raw swallow data could not only help automate the feature extraction, but also alleviate the bias from pre-defined features. With more than 32-thousand raw swallow data, a generative model using the approach of variational auto-encoder (VAE) was developed, which, to our knowledge, is the first deep-learning-based unsupervised model on raw esophageal manometry data. The VAE model was reformulated to include different types of loss motivated by domain knowledge and tuned with different hyper-parameters. Training of the VAE model was found sensitive on the learning rate and hence the evidence lower bound objective (ELBO) was further scaled by the data dimension. Case studies showed that the dimensionality of latent space have a big impact on the learned semantics. In particular, cases with 4-dimensional latent variables were found to encode various physiologically meaningful contraction patterns, including strength, propagation pattern as well as sphincter relaxation. Cases with so-called hybrid L2 loss seemed to better capture the coherence of contraction/relaxation transition. Discriminating capability was further evaluated using simple linear discriminative analysis (LDA) on predicting swallow type and swallow pressurization, which yields clustering patterns consistent with clinical impression. The current work on modeling and understanding swallow-level data will guide the development of study-level models for automatic diagnosis as the next stage.
KW - Artificial intelligence
KW - Esophageal diagnosis
KW - Generative modeling
KW - High-resolution manometry
UR - http://www.scopus.com/inward/record.url?scp=85099212889&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85099212889&partnerID=8YFLogxK
U2 - 10.1016/j.artmed.2020.102006
DO - 10.1016/j.artmed.2020.102006
M3 - Article
C2 - 33581826
AN - SCOPUS:85099212889
SN - 0933-3657
VL - 112
JO - Artificial Intelligence In Medicine
JF - Artificial Intelligence In Medicine
M1 - 102006
ER -