Predicting clinical outcomes from large scale cancer genomic profiles with deep survival models

Yousefi, Sahar; Amrollahi, Fatemeh; Amgad, Mohamed; Dong, Coco; Lewis, Joshua E.; Song, Congzheng; Gutman, David; Halani, Sameer H.; Vega, José E. Velázquez; Brat, Daniel J.; Cooper, Lee

doi:10.1101/131367

Cited by 24 publications

(25 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Interrelated multi-omics factors including genetic polymorphisms, somatic mutations, epigenetic modifications, and alterations in expression of RNAs and proteins collectively contribute to cancer pathogenesis and progression. Clinical omics data from large cancer cohorts provide an unprecedented opportunity to elucidate the complex molecular basis of cancers [14][15][16] and to develop machine learning based predictive analytics for precision oncology [17][18][19][20][21][22] . However, data inequality among racial groups continues to be conspicuous in recent large-scale genomics-focused biomedical research programs 7,23,24 .…”

Section: Clinical Omics Data Inequalities Among Racial Groupsmentioning

confidence: 99%

Deep transfer learning for reducing health care disparities arising from biomedical data inequality

Gao

Cui

2020

Preprint

View full text Add to dashboard Cite

As artificial intelligence (AI) is increasingly involved in biomedical research and clinical decisions, developing unbiased AI models that work equally well for all racial and ethnic groups is of crucial importance to health disparity prevention and reduction. However, the long-standing data inequality between different racial/ethnic groups is generating new health disparities through datadriven, algorithm-based biomedical research, and clinical decisions. Here we found from a large set of machine learning experiments on cancer omics data that the current prevalent schemes of multiethnic machine learning are prone to generate significant model performance disparities between racial groups. We showed that this performance disparity is caused by data inequality and data distribution discrepancy between racial groups. We also found that transfer learning can improve the machine learning model performance on the data-disadvantaged racial groups, and thus, provides an approach to reduce the health disparities arising from the long-standing data inequality among racial groups.

show abstract

Section: Clinical Omics Data Inequalities Among Racial Groupsmentioning

confidence: 99%

Deep transfer learning for reducing health care disparities arising from biomedical data inequality

Gao

Cui

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Other examples include survival prognosis from multi-omics data in hepatocellular carcinoma, 25 breast cancer, 26 or multiple data sets from the TCGA. 27 In order to deal with the "curse of dimensionality" inherent to the very high number of features (consider that gene expression data can span ~ 20,000 genes), investigators have implemented various feature selection strategies to reduce dimensionality, 26 including using neural networks themselves. 24 Of note, DL methods have also been shown able to integrate multi-omics data spanning RNA and micro RNA sequencing, methylation data, and copy number alterations.…”

Section: Omics Datamentioning

confidence: 99%

Artificial Intelligence and Mechanistic Modeling for Clinical Decision Making in Oncology

Benzekry

2020

Clin Pharma and Therapeutics

View full text Add to dashboard Cite

The amount of “big” data generated in clinical oncology, whether from molecular, imaging, pharmacological, or biological origin, brings novel challenges. To mine efficiently this source of information, mathematical models able to produce predictive algorithms and simulations are required, with applications for diagnosis, prognosis, drug development, or prediction of the response to therapy. Such mathematical and computational constructs can be subdivided into two broad classes: biologically agnostic, statistical models using artificial intelligence techniques, and physiologically based, mechanistic models. In this review, recent advances in the applications of such methods in clinical oncology are outlined. These include machine learning applied to big data (omics, imaging, or electronic health records), pharmacometrics and quantitative systems pharmacology, as well as tumor kinetics and metastasis modeling. Focus is set on studies with high potential of clinical translation, and particular attention is given to cancer immunotherapy. Perspectives are given in terms of combinations of the two approaches: “mechanistic learning.”

show abstract

“…The dataset of 120 is equivalently small for detection and no validation metrics was considered. [24] demonstrated how Deep Learning and Bayesian optimization methods were used in predicting clinical outcomes from large scale cancer genomic profiles for survival analysis, and described a framework for interpreting deep survival models using a risk back propagation technique. The framework was implemented in Python for training, evaluation and interpretation of deep survival models.…”

Section: Related Workmentioning

confidence: 99%

A Logical Approach for Empirical Risk Minimization in Machine Learning for Data Stratification

Taiwo¹,

Awodele²,

Kuyoro³

2017

RJMCS

View full text Add to dashboard Cite

The data-driven methods capable of understanding, mimicking and aiding the information processing tasks of Machine Learning (ML) have been applied in an increasing range over the past years in diverse areas at a very high rate, and had achieved great success in predicting and stratifying given data instances of a problem domain. There has been generalization on the performance of the classifier to be the optimal based on the existing performance benchmarks such as accuracy, speed, time to learn, number of features, comprehensibility, robustness, scalability and interpretability. However, these benchmarks alone do not guarantee the successful adoption of an algorithm for prediction and stratification since there may be an incurring risk in its adoption. Therefore, this paper aims at developing a logical approach for using Empirical Risk Minimization (ERM) technique to determine the machine learning classifier with the minimum risk function for data stratification. The generalization on the performance of optimal algorithm was tested on BayesNet, Multilayered perceptron, Projective Adaptive Resonance Theory (PART) and Logistic Model Trees algorithms based on existing performance benchmarks such as correctly classified instances, time to build, kappa statistics, sensitivity and specificity to determine the algorithms with great performances. The study showed that PART and Logistic Model Trees algorithms perform well than others. Hence, a logical approach to apply Empirical Risk Minimization technique on PART and Logistic Model Trees algorithms is shown to give a detailed procedure of determining their empirical risk function to aid the decision of choosing an algorithm to be the best fit classifier for data stratification. This therefore serves as a benchmark for selecting an optimal algorithm for stratification and prediction alongside other benchmarks.

show abstract

Predicting clinical outcomes from large scale cancer genomic profiles with deep survival models

Cited by 24 publications

References 31 publications

Deep transfer learning for reducing health care disparities arising from biomedical data inequality

Deep transfer learning for reducing health care disparities arising from biomedical data inequality

Artificial Intelligence and Mechanistic Modeling for Clinical Decision Making in Oncology

A Logical Approach for Empirical Risk Minimization in Machine Learning for Data Stratification

Contact Info

Product

Resources

About