TY - GEN
T1 - Probabilistic Deep Learning with Adversarial Training and Volume Interval Estimation - Better Ways to Perform and Evaluate Predictive Models for White Matter Hyperintensities Evolution
AU - Rachmadi, Muhammad Febrian
AU - Valdés-Hernández, Maria del C.
AU - Maulana, Rizal
AU - Wardlaw, Joanna
AU - Makin, Stephen
AU - Skibbe, Henrik
N1 - Acknowledgements. Funds from JSPS (Kakenhi Grant-in-Aid for Research Activity Start-up, Project No. 20K23356) (MFR); Row Fogo Charitable Trust (Grant No. BROD.FID3668413) (MCVH); Wellcome Trust (patient recruitment, scanning, primary study Ref No. WT088134/Z/09/A); Fondation Leducq (Perivascular Spaces Transatlantic Network of Excellence); EU Horizon 2020 (SVDs@Target); and the MRC UK Dementia Research Institute at the University of Edinburgh (Wardlaw programme) are gratefully acknowledged. This research was also supported by the program for Brain Mapping by Integrated Neurotechnologies for Disease Studies (Brain/MINDS) from the Japan Agency for Medical Research and Development AMED (JP21dm0207001).
PY - 2021
Y1 - 2021
N2 - Predicting disease progression always involves a high degree of uncertainty. White matter hyperintensities (WMHs) are the main neuroradiological feature of small vessel disease and a common finding in brain scans of dementia patients and older adults. In predicting their progression previous studies have identified two main challenges: 1) uncertainty in predicting the areas/boundaries of shrinking and growing WMHs and 2) uncertainty in the estimation of future WMHs volume. This study proposes the use of a probabilistic deep learning model called Probabilistic U-Net trained with adversarial loss for capturing and modelling spatial uncertainty in brain MR images. This study also proposes an evaluation procedure named volume interval estimation (VIE) for improving the interpretation of and confidence in the predictive deep learning model. Our experiments show that the Probabilistic U-Net with adversarial training improved the performance of non-probabilistic U-Net in Dice similarity coefficient for predicting the areas of shrinking WMHs, growing WMHs, stable WMHs, and their average by up to 3.35%, 2.94%, 0.47%, and 1.03% respectively. It also improved the volume estimation by 11.84% in the “Correct Prediction in Estimated Volume Interval” metric as per the newly proposed VIE evaluation procedure.
AB - Predicting disease progression always involves a high degree of uncertainty. White matter hyperintensities (WMHs) are the main neuroradiological feature of small vessel disease and a common finding in brain scans of dementia patients and older adults. In predicting their progression previous studies have identified two main challenges: 1) uncertainty in predicting the areas/boundaries of shrinking and growing WMHs and 2) uncertainty in the estimation of future WMHs volume. This study proposes the use of a probabilistic deep learning model called Probabilistic U-Net trained with adversarial loss for capturing and modelling spatial uncertainty in brain MR images. This study also proposes an evaluation procedure named volume interval estimation (VIE) for improving the interpretation of and confidence in the predictive deep learning model. Our experiments show that the Probabilistic U-Net with adversarial training improved the performance of non-probabilistic U-Net in Dice similarity coefficient for predicting the areas of shrinking WMHs, growing WMHs, stable WMHs, and their average by up to 3.35%, 2.94%, 0.47%, and 1.03% respectively. It also improved the volume estimation by 11.84% in the “Correct Prediction in Estimated Volume Interval” metric as per the newly proposed VIE evaluation procedure.
KW - Progression prediction
KW - Volume interval estimation
KW - White matter hyperintensities
UR - http://www.scopus.com/inward/record.url?scp=85116834944&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-87602-9_16
DO - 10.1007/978-3-030-87602-9_16
M3 - Published conference contribution
AN - SCOPUS:85116834944
SN - 9783030876012
VL - 12928
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 168
EP - 180
BT - Predictive Intelligence in Medicine, PRIME 2021
A2 - Rekik, Islem
A2 - Adeli, Ehsan
A2 - Park, Sang Hyun
A2 - Schnabel, Julia
PB - Springer
CY - Cham
T2 - 4th International Workshop on Predictive Intelligence in Medicine, PRIME 2021, held in conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021
Y2 - 1 October 2021 through 1 October 2021
ER -