Comparing radiomic classifiers and classifier ensembles for detection of peripheral zone prostate tumors on T2-weighted MRI

A multi-site study

Satish E. Viswanath, Prathyush V. Chirra, Michael C. Yim, Neil M Rofsky, Andrei S. Purysko, Mark A. Rosen, B. Nicolas Bloch, Anant Madabhushi

Research output: Contribution to journalArticle

Abstract

Background: For most computer-aided diagnosis (CAD) problems involving prostate cancer detection via medical imaging data, the choice of classifier has been largely ad hoc, or been motivated by classifier comparison studies that have involved large synthetic datasets. More significantly, it is currently unknown how classifier choices and trends generalize across multiple institutions, due to heterogeneous acquisition and intensity characteristics (especially when considering MR imaging data). In this work, we empirically evaluate and compare a number of different classifiers and classifier ensembles in a multi-site setting, for voxel-wise detection of prostate cancer (PCa) using radiomic texture features derived from high-resolution in vivo T2-weighted (T2w) MRI. Methods: Twelve different supervised classifier schemes: Quadratic Discriminant Analysis (QDA), Support Vector Machines (SVMs), naïve Bayes, Decision Trees (DTs), and their ensemble variants (bagging, boosting), were compared in terms of classification accuracy as well as execution time. Our study utilized 85 prostate cancer T2w MRI datasets acquired from across 3 different institutions (1 for discovery, 2 for independent validation), from patients who later underwent radical prostatectomy. Surrogate ground truth for disease extent on MRI was established by expert annotation of pre-operative MRI through spatial correlation with corresponding ex vivo whole-mount histology sections. Classifier accuracy in detecting PCa extent on MRI on a per-voxel basis was evaluated via area under the ROC curve. Results: The boosted DT classifier yielded the highest cross-validated AUC (= 0.744) for detecting PCa in the discovery cohort. However, in independent validation, the boosted QDA classifier was identified as the most accurate and robust for voxel-wise detection of PCa extent (AUCs of 0.735, 0.683, 0.768 across the 3 sites). The next most accurate and robust classifier was the single QDA classifier, which also enjoyed the advantage of significantly lower computation times compared to any of the other methods. Conclusions: Our results therefore suggest that simpler classifiers (such as QDA and its ensemble variants) may be more robust, accurate, and efficient for prostate cancer CAD problems, especially in the context of multi-site validation.

Original languageEnglish (US)
Article number0308
JournalBMC Medical Imaging
Volume19
Issue number1
DOIs
StatePublished - Feb 28 2019

Fingerprint

Prostate
Prostatic Neoplasms
Discriminant Analysis
Neoplasms
Area Under Curve
Decision Trees
Diagnostic Imaging
Prostatectomy
ROC Curve
Histology

Keywords

  • Classifiers
  • Comparison
  • MRI
  • Prostate cancer
  • Radiomics

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

Comparing radiomic classifiers and classifier ensembles for detection of peripheral zone prostate tumors on T2-weighted MRI : A multi-site study. / Viswanath, Satish E.; Chirra, Prathyush V.; Yim, Michael C.; Rofsky, Neil M; Purysko, Andrei S.; Rosen, Mark A.; Bloch, B. Nicolas; Madabhushi, Anant.

In: BMC Medical Imaging, Vol. 19, No. 1, 0308, 28.02.2019.

Research output: Contribution to journalArticle

Viswanath, Satish E. ; Chirra, Prathyush V. ; Yim, Michael C. ; Rofsky, Neil M ; Purysko, Andrei S. ; Rosen, Mark A. ; Bloch, B. Nicolas ; Madabhushi, Anant. / Comparing radiomic classifiers and classifier ensembles for detection of peripheral zone prostate tumors on T2-weighted MRI : A multi-site study. In: BMC Medical Imaging. 2019 ; Vol. 19, No. 1.
@article{67fe853a8a4c48a7a435e046bd253141,
title = "Comparing radiomic classifiers and classifier ensembles for detection of peripheral zone prostate tumors on T2-weighted MRI: A multi-site study",
abstract = "Background: For most computer-aided diagnosis (CAD) problems involving prostate cancer detection via medical imaging data, the choice of classifier has been largely ad hoc, or been motivated by classifier comparison studies that have involved large synthetic datasets. More significantly, it is currently unknown how classifier choices and trends generalize across multiple institutions, due to heterogeneous acquisition and intensity characteristics (especially when considering MR imaging data). In this work, we empirically evaluate and compare a number of different classifiers and classifier ensembles in a multi-site setting, for voxel-wise detection of prostate cancer (PCa) using radiomic texture features derived from high-resolution in vivo T2-weighted (T2w) MRI. Methods: Twelve different supervised classifier schemes: Quadratic Discriminant Analysis (QDA), Support Vector Machines (SVMs), na{\"i}ve Bayes, Decision Trees (DTs), and their ensemble variants (bagging, boosting), were compared in terms of classification accuracy as well as execution time. Our study utilized 85 prostate cancer T2w MRI datasets acquired from across 3 different institutions (1 for discovery, 2 for independent validation), from patients who later underwent radical prostatectomy. Surrogate ground truth for disease extent on MRI was established by expert annotation of pre-operative MRI through spatial correlation with corresponding ex vivo whole-mount histology sections. Classifier accuracy in detecting PCa extent on MRI on a per-voxel basis was evaluated via area under the ROC curve. Results: The boosted DT classifier yielded the highest cross-validated AUC (= 0.744) for detecting PCa in the discovery cohort. However, in independent validation, the boosted QDA classifier was identified as the most accurate and robust for voxel-wise detection of PCa extent (AUCs of 0.735, 0.683, 0.768 across the 3 sites). The next most accurate and robust classifier was the single QDA classifier, which also enjoyed the advantage of significantly lower computation times compared to any of the other methods. Conclusions: Our results therefore suggest that simpler classifiers (such as QDA and its ensemble variants) may be more robust, accurate, and efficient for prostate cancer CAD problems, especially in the context of multi-site validation.",
keywords = "Classifiers, Comparison, MRI, Prostate cancer, Radiomics",
author = "Viswanath, {Satish E.} and Chirra, {Prathyush V.} and Yim, {Michael C.} and Rofsky, {Neil M} and Purysko, {Andrei S.} and Rosen, {Mark A.} and Bloch, {B. Nicolas} and Anant Madabhushi",
year = "2019",
month = "2",
day = "28",
doi = "10.1186/s12880-019-0308-6",
language = "English (US)",
volume = "19",
journal = "BMC Medical Imaging",
issn = "1471-2342",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - Comparing radiomic classifiers and classifier ensembles for detection of peripheral zone prostate tumors on T2-weighted MRI

T2 - A multi-site study

AU - Viswanath, Satish E.

AU - Chirra, Prathyush V.

AU - Yim, Michael C.

AU - Rofsky, Neil M

AU - Purysko, Andrei S.

AU - Rosen, Mark A.

AU - Bloch, B. Nicolas

AU - Madabhushi, Anant

PY - 2019/2/28

Y1 - 2019/2/28

N2 - Background: For most computer-aided diagnosis (CAD) problems involving prostate cancer detection via medical imaging data, the choice of classifier has been largely ad hoc, or been motivated by classifier comparison studies that have involved large synthetic datasets. More significantly, it is currently unknown how classifier choices and trends generalize across multiple institutions, due to heterogeneous acquisition and intensity characteristics (especially when considering MR imaging data). In this work, we empirically evaluate and compare a number of different classifiers and classifier ensembles in a multi-site setting, for voxel-wise detection of prostate cancer (PCa) using radiomic texture features derived from high-resolution in vivo T2-weighted (T2w) MRI. Methods: Twelve different supervised classifier schemes: Quadratic Discriminant Analysis (QDA), Support Vector Machines (SVMs), naïve Bayes, Decision Trees (DTs), and their ensemble variants (bagging, boosting), were compared in terms of classification accuracy as well as execution time. Our study utilized 85 prostate cancer T2w MRI datasets acquired from across 3 different institutions (1 for discovery, 2 for independent validation), from patients who later underwent radical prostatectomy. Surrogate ground truth for disease extent on MRI was established by expert annotation of pre-operative MRI through spatial correlation with corresponding ex vivo whole-mount histology sections. Classifier accuracy in detecting PCa extent on MRI on a per-voxel basis was evaluated via area under the ROC curve. Results: The boosted DT classifier yielded the highest cross-validated AUC (= 0.744) for detecting PCa in the discovery cohort. However, in independent validation, the boosted QDA classifier was identified as the most accurate and robust for voxel-wise detection of PCa extent (AUCs of 0.735, 0.683, 0.768 across the 3 sites). The next most accurate and robust classifier was the single QDA classifier, which also enjoyed the advantage of significantly lower computation times compared to any of the other methods. Conclusions: Our results therefore suggest that simpler classifiers (such as QDA and its ensemble variants) may be more robust, accurate, and efficient for prostate cancer CAD problems, especially in the context of multi-site validation.

AB - Background: For most computer-aided diagnosis (CAD) problems involving prostate cancer detection via medical imaging data, the choice of classifier has been largely ad hoc, or been motivated by classifier comparison studies that have involved large synthetic datasets. More significantly, it is currently unknown how classifier choices and trends generalize across multiple institutions, due to heterogeneous acquisition and intensity characteristics (especially when considering MR imaging data). In this work, we empirically evaluate and compare a number of different classifiers and classifier ensembles in a multi-site setting, for voxel-wise detection of prostate cancer (PCa) using radiomic texture features derived from high-resolution in vivo T2-weighted (T2w) MRI. Methods: Twelve different supervised classifier schemes: Quadratic Discriminant Analysis (QDA), Support Vector Machines (SVMs), naïve Bayes, Decision Trees (DTs), and their ensemble variants (bagging, boosting), were compared in terms of classification accuracy as well as execution time. Our study utilized 85 prostate cancer T2w MRI datasets acquired from across 3 different institutions (1 for discovery, 2 for independent validation), from patients who later underwent radical prostatectomy. Surrogate ground truth for disease extent on MRI was established by expert annotation of pre-operative MRI through spatial correlation with corresponding ex vivo whole-mount histology sections. Classifier accuracy in detecting PCa extent on MRI on a per-voxel basis was evaluated via area under the ROC curve. Results: The boosted DT classifier yielded the highest cross-validated AUC (= 0.744) for detecting PCa in the discovery cohort. However, in independent validation, the boosted QDA classifier was identified as the most accurate and robust for voxel-wise detection of PCa extent (AUCs of 0.735, 0.683, 0.768 across the 3 sites). The next most accurate and robust classifier was the single QDA classifier, which also enjoyed the advantage of significantly lower computation times compared to any of the other methods. Conclusions: Our results therefore suggest that simpler classifiers (such as QDA and its ensemble variants) may be more robust, accurate, and efficient for prostate cancer CAD problems, especially in the context of multi-site validation.

KW - Classifiers

KW - Comparison

KW - MRI

KW - Prostate cancer

KW - Radiomics

UR - http://www.scopus.com/inward/record.url?scp=85062299611&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062299611&partnerID=8YFLogxK

U2 - 10.1186/s12880-019-0308-6

DO - 10.1186/s12880-019-0308-6

M3 - Article

VL - 19

JO - BMC Medical Imaging

JF - BMC Medical Imaging

SN - 1471-2342

IS - 1

M1 - 0308

ER -