+ Site Statistics
+ Search Articles
+ PDF Full Text Service
How our service works
Request PDF Full Text
+ Follow Us
Follow on Facebook
Follow on Twitter
Follow on LinkedIn
+ Subscribe to Site Feeds
Most Shared
PDF Full Text
+ Translate
+ Recently Requested

Variable importance in nonlinear kernels (VINK): classification of digitized histopathology

Variable importance in nonlinear kernels (VINK): classification of digitized histopathology

Medical Image Computing and Computer-Assisted Intervention 16(Pt 2): 238-245

Quantitative histomorphometry is the process of modeling appearance of disease morphology on digitized histopathology images via image-based features (e.g., texture, graphs). Due to the curse of dimensionality, building classifiers with large numbers of features requires feature selection (which may require a large training set) or dimensionality reduction (DR). DR methods map the original high-dimensional features in terms of eigenvectors and eigenvalues, which limits the potential for feature transparency or interpretability. Although methods exist for variable selection and ranking on embeddings obtained via linear DR schemes (e.g., principal components analysis (PCA)), similar methods do not yet exist for nonlinear DR (NLDR) methods. In this work we present a simple yet elegant method for approximating the mapping between the data in the original feature space and the transformed data in the kernel PCA (KPCA) embedding space; this mapping provides the basis for quantification of variable importance in nonlinear kernels (VINK). We show how VINK can be implemented in conjunction with the popular Isomap and Laplacian eigenmap algorithms. VINK is evaluated in the contexts of three different problems in digital pathology: (1) predicting five year PSA failure following radical prostatectomy, (2) predicting Oncotype DX recurrence risk scores for ER+ breast cancers, and (3) distinguishing good and poor outcome p16+ oropharyngeal tumors. We demonstrate that subsets of features identified by VINK provide similar or better classification or regression performance compared to the original high dimensional feature sets.

Please choose payment method:

(PDF emailed within 1 workday: $29.90)

Accession: 056847185

Download citation: RISBibTeXText

PMID: 24579146

Related references

Learned random-walk kernels and empirical-map kernels for protein sequence classification. Journal of Computational Biology 16(3): 457-474, 2009

Variable positions of the sacral auricular surface: classification and importance. Neurosurgical Focus 28(3): E12, 2010

Content-based image retrieval of digitized histopathology in boosted spectrally embedded spaces. Journal of Pathology Informatics 6: 41, 2015

Automatic grading of prostate cancer in digitized histopathology images: Learning from multiple experts. Medical Image Analysis 50: 167-180, 2018

Analysis of Cellular Feature Differences of Astrocytomas with Distinct Mutational Profiles Using Digitized Histopathology Images. Conference Proceedings 2018: 4644-4647, 2018

Computer-Aided Prostate Cancer Diagnosis From Digitized Histopathology: A Review on Texture-Based Systems. IEEE Reviews in Biomedical Engineering 8: 98-113, 2016

Standard errors and confidence intervals for variable importance in random forest regression, classification, and survival. Statistics in Medicine 2018, 2018

A comparison of cervical histopathology variability using whole slide digitized images versus glass slides: experience with a statewide registry. Human Pathology 44(11): 2542-2548, 2014

Classification of digitized faces. Bulletin of the Psychonomic Society 20(3): 145, 1982

Histopathology of alternaria tenuis infected black pointed kernels of wheat. Botanical Bulletin of Academia Sinica 28(2): 123-130, 1987

knnAUC: an open-source R package for detecting nonlinear dependence between one continuous variable and one binary variable. Bmc Bioinformatics 19(1): 448, 2018

Nonlinear kernels of the human ERG. Biological Cybernetics 35(3): 145-160, 1979

Histopathology of Drechslera maydis infected maize kernels from tribal areas of Rajasthan. Indian Phytopathology 39(3): 432-434, 1986

A digitized systematic classification for ecosystems. 1980

The Pulse Extension Method for the Variable Neighborhood Distance Transformation of 2- and 3-Dimensional Digitized Pictures. Systems and Computers in Japan 21(4): 1-14, 1990