Combining radiomic features with a miRNA classifier may improve prediction of malignant pathology for pancreatic intraductal papillary mucinous neoplasms

Florida Pancreas Collaborative

Research output: Contribution to journalArticle

30 Citations (Scopus)

Abstract

Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cancer precursors incidentally discovered by cross-sectional imaging. Consensus guidelines for IPMN management rely on standard radiologic features to predict pathology, but they lack accuracy. Using a retrospective cohort of 38 surgically-resected, pathologically-confirmed IPMNs (20 benign; 18 malignant) with preoperative computed tomography (CT) images and matched plasma-based 'miRNA genomic classifier (MGC)' data, we determined whether quantitative 'radiomic' CT features (+/- the MGC) can more accurately predict IPMN pathology than standard radiologic features 'high-risk' or 'worrisome' for malignancy. Logistic regression, principal component analyses, and cross-validation were used to examine associations. Sensitivity, specificity, positive and negative predictive value (PPV, NPV) were estimated. The MGC, 'high-risk,' and 'worrisome' radiologic features had area under the receiver operating characteristic curve (AUC) values of 0.83, 0.84, and 0.54, respectively. Fourteen radiomic features differentiated malignant from benign IPMNs (p < 0.05) and collectively had an AUC=0.77. Combining radiomic features with the MGC revealed an AUC=0.92 and superior sensitivity (83%), specificity (89%), PPV (88%), and NPV (85%) than other models. Evaluation of uncertainty by 10-fold cross-validation retained an AUC > 0.80 (0.87 (95% CI:0.84-0.89)). This proof-ofconcept study suggests a noninvasive radiogenomic approach may more accurately predict IPMN pathology than 'worrisome' radiologic features considered in consensus guidelines.

Original languageEnglish (US)
Pages (from-to)85785-85797
Number of pages13
JournalOncotarget
Volume7
Issue number52
DOIs
StatePublished - 2016

Fingerprint

MicroRNAs
Pathology
Neoplasms
Tomography
Guidelines
Principal Component Analysis
Pancreatic Neoplasms
ROC Curve
Area Under Curve
Logistic Models
Sensitivity and Specificity

Keywords

  • MiRNA
  • Pancreas
  • Pre-malignant lesions
  • Radiomics
  • Risk stratification

ASJC Scopus subject areas

  • Oncology

Cite this

Combining radiomic features with a miRNA classifier may improve prediction of malignant pathology for pancreatic intraductal papillary mucinous neoplasms. / Florida Pancreas Collaborative.

In: Oncotarget, Vol. 7, No. 52, 2016, p. 85785-85797.

Research output: Contribution to journalArticle

@article{e466e28b69c34200a1c787963bf3cdbc,
title = "Combining radiomic features with a miRNA classifier may improve prediction of malignant pathology for pancreatic intraductal papillary mucinous neoplasms",
abstract = "Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cancer precursors incidentally discovered by cross-sectional imaging. Consensus guidelines for IPMN management rely on standard radiologic features to predict pathology, but they lack accuracy. Using a retrospective cohort of 38 surgically-resected, pathologically-confirmed IPMNs (20 benign; 18 malignant) with preoperative computed tomography (CT) images and matched plasma-based 'miRNA genomic classifier (MGC)' data, we determined whether quantitative 'radiomic' CT features (+/- the MGC) can more accurately predict IPMN pathology than standard radiologic features 'high-risk' or 'worrisome' for malignancy. Logistic regression, principal component analyses, and cross-validation were used to examine associations. Sensitivity, specificity, positive and negative predictive value (PPV, NPV) were estimated. The MGC, 'high-risk,' and 'worrisome' radiologic features had area under the receiver operating characteristic curve (AUC) values of 0.83, 0.84, and 0.54, respectively. Fourteen radiomic features differentiated malignant from benign IPMNs (p < 0.05) and collectively had an AUC=0.77. Combining radiomic features with the MGC revealed an AUC=0.92 and superior sensitivity (83{\%}), specificity (89{\%}), PPV (88{\%}), and NPV (85{\%}) than other models. Evaluation of uncertainty by 10-fold cross-validation retained an AUC > 0.80 (0.87 (95{\%} CI:0.84-0.89)). This proof-ofconcept study suggests a noninvasive radiogenomic approach may more accurately predict IPMN pathology than 'worrisome' radiologic features considered in consensus guidelines.",
keywords = "MiRNA, Pancreas, Pre-malignant lesions, Radiomics, Risk stratification",
author = "{Florida Pancreas Collaborative} and Permuth, {Jennifer B.} and Jung Choi and Yoganand Balarunathan and Jongphil Kim and Chen, {Dung Tsa} and Lu Chen and Sonia Orcutt and Doepker, {Matthew P.} and Kenneth Gage and Geoffrey Zhang and Kujtim Latifi and Sarah Hoffe and Kun Jiang and Domenico Coppola and Centeno, {Barbara A.} and Anthony Magliocco and Qian Li and Jose Trevino and Nipun Merchant and Robert Gillies and Mokenge Malafa",
year = "2016",
doi = "10.18632/oncotarget.11768",
language = "English (US)",
volume = "7",
pages = "85785--85797",
journal = "Oncotarget",
issn = "1949-2553",
publisher = "Impact Journals",
number = "52",

}

TY - JOUR

T1 - Combining radiomic features with a miRNA classifier may improve prediction of malignant pathology for pancreatic intraductal papillary mucinous neoplasms

AU - Florida Pancreas Collaborative

AU - Permuth, Jennifer B.

AU - Choi, Jung

AU - Balarunathan, Yoganand

AU - Kim, Jongphil

AU - Chen, Dung Tsa

AU - Chen, Lu

AU - Orcutt, Sonia

AU - Doepker, Matthew P.

AU - Gage, Kenneth

AU - Zhang, Geoffrey

AU - Latifi, Kujtim

AU - Hoffe, Sarah

AU - Jiang, Kun

AU - Coppola, Domenico

AU - Centeno, Barbara A.

AU - Magliocco, Anthony

AU - Li, Qian

AU - Trevino, Jose

AU - Merchant, Nipun

AU - Gillies, Robert

AU - Malafa, Mokenge

PY - 2016

Y1 - 2016

N2 - Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cancer precursors incidentally discovered by cross-sectional imaging. Consensus guidelines for IPMN management rely on standard radiologic features to predict pathology, but they lack accuracy. Using a retrospective cohort of 38 surgically-resected, pathologically-confirmed IPMNs (20 benign; 18 malignant) with preoperative computed tomography (CT) images and matched plasma-based 'miRNA genomic classifier (MGC)' data, we determined whether quantitative 'radiomic' CT features (+/- the MGC) can more accurately predict IPMN pathology than standard radiologic features 'high-risk' or 'worrisome' for malignancy. Logistic regression, principal component analyses, and cross-validation were used to examine associations. Sensitivity, specificity, positive and negative predictive value (PPV, NPV) were estimated. The MGC, 'high-risk,' and 'worrisome' radiologic features had area under the receiver operating characteristic curve (AUC) values of 0.83, 0.84, and 0.54, respectively. Fourteen radiomic features differentiated malignant from benign IPMNs (p < 0.05) and collectively had an AUC=0.77. Combining radiomic features with the MGC revealed an AUC=0.92 and superior sensitivity (83%), specificity (89%), PPV (88%), and NPV (85%) than other models. Evaluation of uncertainty by 10-fold cross-validation retained an AUC > 0.80 (0.87 (95% CI:0.84-0.89)). This proof-ofconcept study suggests a noninvasive radiogenomic approach may more accurately predict IPMN pathology than 'worrisome' radiologic features considered in consensus guidelines.

AB - Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cancer precursors incidentally discovered by cross-sectional imaging. Consensus guidelines for IPMN management rely on standard radiologic features to predict pathology, but they lack accuracy. Using a retrospective cohort of 38 surgically-resected, pathologically-confirmed IPMNs (20 benign; 18 malignant) with preoperative computed tomography (CT) images and matched plasma-based 'miRNA genomic classifier (MGC)' data, we determined whether quantitative 'radiomic' CT features (+/- the MGC) can more accurately predict IPMN pathology than standard radiologic features 'high-risk' or 'worrisome' for malignancy. Logistic regression, principal component analyses, and cross-validation were used to examine associations. Sensitivity, specificity, positive and negative predictive value (PPV, NPV) were estimated. The MGC, 'high-risk,' and 'worrisome' radiologic features had area under the receiver operating characteristic curve (AUC) values of 0.83, 0.84, and 0.54, respectively. Fourteen radiomic features differentiated malignant from benign IPMNs (p < 0.05) and collectively had an AUC=0.77. Combining radiomic features with the MGC revealed an AUC=0.92 and superior sensitivity (83%), specificity (89%), PPV (88%), and NPV (85%) than other models. Evaluation of uncertainty by 10-fold cross-validation retained an AUC > 0.80 (0.87 (95% CI:0.84-0.89)). This proof-ofconcept study suggests a noninvasive radiogenomic approach may more accurately predict IPMN pathology than 'worrisome' radiologic features considered in consensus guidelines.

KW - MiRNA

KW - Pancreas

KW - Pre-malignant lesions

KW - Radiomics

KW - Risk stratification

UR - http://www.scopus.com/inward/record.url?scp=85007494966&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85007494966&partnerID=8YFLogxK

U2 - 10.18632/oncotarget.11768

DO - 10.18632/oncotarget.11768

M3 - Article

C2 - 27589689

AN - SCOPUS:85007494966

VL - 7

SP - 85785

EP - 85797

JO - Oncotarget

JF - Oncotarget

SN - 1949-2553

IS - 52

ER -