Lung Cancer Risk Estimation with Incomplete Data: A Joint Missing Imputation Perspective

@inproceedings{Gao2021LungCR,
  title={Lung Cancer Risk Estimation with Incomplete Data: A Joint Missing Imputation Perspective},
  author={Riqiang Gao and Yucheng Tang and Kaiwen Xu and Ho Hin Lee and Stephen A. Deppen and Kim L. Sandler and Pierre P. Massion and Thomas A. Lasko and Yuankai Huo and Bennett A. Landman},
  booktitle={MICCAI},
  year={2021}
}
  • Riqiang Gao, Yucheng Tang, +7 authors B. Landman
  • Published in MICCAI 2021
  • Computer Science, Engineering
Data from multi-modality provide complementary information in clinical prediction, but missing data in clinical cohorts limits the number of subjects in multi-modal learning context. Multi-modal missing imputation is challenging with existing methods when 1) the missing data span across heterogeneous modalities (e.g., image vs. non-image); or 2) one modality is largely missing. In this paper, we address imputation of missing data by modeling the joint distribution of multi-modal data. Motivated… Expand

Figures and Tables from this paper

References

SHOWING 1-10 OF 29 REFERENCES
GAIN: Missing Data Imputation using Generative Adversarial Nets
TLDR
This work proposes a novel method for imputing missing data by adapting the well-known Generative Adversarial Nets (GAN) framework and calls it GAIN, which significantly outperforms state-of-the-art imputation methods. Expand
MissForest - non-parametric missing value imputation for mixed-type data
TLDR
In this comparative study, missForest outperforms other methods of imputation especially in data settings where complex interactions and non-linear relations are suspected and the out-of-bag imputation error estimates of missForest prove to be adequate in all settings. Expand
MIWAE: Deep Generative Modelling and Imputation of Incomplete Data Sets
TLDR
This work presents a simple technique to train DLVMs when the training set contains missing-at-random data, and develops Monte Carlo techniques for single and multiple imputation using a DLVM trained on an incomplete data set. Expand
Flexible Imputation of Missing Data
TLDR
The problem of missing data concepts of MCAR, MAR and MNAR simple solutions that do not (always) work multiple imputation in a nutshell and some dangers, some do's and some don'ts are covered. Expand
CT-Realistic Lung Nodule Simulation from 3D Conditional Generative Adversarial Networks for Robust Lung Segmentation
TLDR
This work develops a 3D generative adversarial network (GAN) that effectively learns lung nodule property distributions in 3D space and proposes a novel multi-mask reconstruction loss to improve realism and blending with the background. Expand
Multi-Task Deep Model With Margin Ranking Loss for Lung Nodule Analysis
TLDR
A Multi-Task deep model with Margin Ranking loss (referred as MTMR-Net) for automated lung nodule analysis is proposed and the internal relationship between two tasks is explicitly explored in a cause-and-effect manner within this multi-task deep model. Expand
Adversarial Feature Learning
TLDR
Bidirectional Generative Adversarial Networks are proposed as a means of learning the inverse mapping of GANs, and it is demonstrated that the resulting learned feature representation is useful for auxiliary supervised discrimination tasks, competitive with contemporary approaches to unsupervised and self-supervised feature learning. Expand
Learning from Irregularly-Sampled Time Series: A Missing Data Perspective
TLDR
An encoder-decoder framework for learning from generic indexed sequences based on variational autoencoders and generative adversarial networks is introduced and continuous convolutional layers that can efficiently interface with existing neural network architectures are introduced. Expand
Generative Adversarial Nets
We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and aExpand
CT-GAN: Malicious Tampering of 3D Medical Imagery using Deep Learning
TLDR
This paper shows how an attacker can use deep-learning to add or remove evidence of medical conditions from volumetric (3D) medical scans using a 3D conditional GAN framework and demonstrates how the framework (CT-GAN) can be automated. Expand
...
1
2
3
...