Screening for prostate cancer using multivariate mixed-effects models

Christopher H. Morrell, Larry J. Brant, Shan Sheng, E. Metter

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Using several variables known to be related to prostate cancer, a multivariate classification method is developed to predict the onset of clinical prostate cancer. A multivariate mixed-effects model is used to describe longitudinal changes in prostate-specific antigen (PSA), a free testosterone index (FTI), and body mass index (BMI) before any clinical evidence of prostate cancer. The patterns of change in these three variables are allowed to vary depending on whether the subject develops prostate cancer or not and the severity of the prostate cancer at diagnosis. An application of Bayes' theorem provides posterior probabilities that we use to predict whether an individual will develop prostate cancer and, if so, whether it is a high-risk or a low-risk cancer. The classification rule is applied sequentially one multivariate observation at a time until the subject is classified as a cancer case or until the last observation has been used. We perform the analyses using each of the three variables individually, combined together in pairs, and all three variables together in one analysis. We compare the classification results among the various analyses and a simulation study demonstrates how the sensitivity of prediction changes with respect to the number and type of variables used in the prediction process.

Original languageEnglish (US)
Pages (from-to)1151-1175
Number of pages25
JournalJournal of Applied Statistics
Volume39
Issue number6
DOIs
StatePublished - Jun 1 2012
Externally publishedYes

Fingerprint

Mixed Effects Model
Prostate Cancer
Screening
Cancer
Bayes' Formula
Predict
Classification Rules
Prediction
Posterior Probability
Several Variables
Prostate cancer
Simulation Study
Vary
Demonstrate

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Screening for prostate cancer using multivariate mixed-effects models. / Morrell, Christopher H.; Brant, Larry J.; Sheng, Shan; Metter, E.

In: Journal of Applied Statistics, Vol. 39, No. 6, 01.06.2012, p. 1151-1175.

Research output: Contribution to journalArticle

Morrell, Christopher H. ; Brant, Larry J. ; Sheng, Shan ; Metter, E. / Screening for prostate cancer using multivariate mixed-effects models. In: Journal of Applied Statistics. 2012 ; Vol. 39, No. 6. pp. 1151-1175.
@article{c45746d39e2e484d8aa5772429435080,
title = "Screening for prostate cancer using multivariate mixed-effects models",
abstract = "Using several variables known to be related to prostate cancer, a multivariate classification method is developed to predict the onset of clinical prostate cancer. A multivariate mixed-effects model is used to describe longitudinal changes in prostate-specific antigen (PSA), a free testosterone index (FTI), and body mass index (BMI) before any clinical evidence of prostate cancer. The patterns of change in these three variables are allowed to vary depending on whether the subject develops prostate cancer or not and the severity of the prostate cancer at diagnosis. An application of Bayes' theorem provides posterior probabilities that we use to predict whether an individual will develop prostate cancer and, if so, whether it is a high-risk or a low-risk cancer. The classification rule is applied sequentially one multivariate observation at a time until the subject is classified as a cancer case or until the last observation has been used. We perform the analyses using each of the three variables individually, combined together in pairs, and all three variables together in one analysis. We compare the classification results among the various analyses and a simulation study demonstrates how the sensitivity of prediction changes with respect to the number and type of variables used in the prediction process.",
author = "Morrell, {Christopher H.} and Brant, {Larry J.} and Shan Sheng and E. Metter",
year = "2012",
month = "6",
day = "1",
doi = "10.1080/02664763.2011.644523",
language = "English (US)",
volume = "39",
pages = "1151--1175",
journal = "Journal of Applied Statistics",
issn = "0266-4763",
publisher = "Routledge",
number = "6",

}

TY - JOUR

T1 - Screening for prostate cancer using multivariate mixed-effects models

AU - Morrell, Christopher H.

AU - Brant, Larry J.

AU - Sheng, Shan

AU - Metter, E.

PY - 2012/6/1

Y1 - 2012/6/1

N2 - Using several variables known to be related to prostate cancer, a multivariate classification method is developed to predict the onset of clinical prostate cancer. A multivariate mixed-effects model is used to describe longitudinal changes in prostate-specific antigen (PSA), a free testosterone index (FTI), and body mass index (BMI) before any clinical evidence of prostate cancer. The patterns of change in these three variables are allowed to vary depending on whether the subject develops prostate cancer or not and the severity of the prostate cancer at diagnosis. An application of Bayes' theorem provides posterior probabilities that we use to predict whether an individual will develop prostate cancer and, if so, whether it is a high-risk or a low-risk cancer. The classification rule is applied sequentially one multivariate observation at a time until the subject is classified as a cancer case or until the last observation has been used. We perform the analyses using each of the three variables individually, combined together in pairs, and all three variables together in one analysis. We compare the classification results among the various analyses and a simulation study demonstrates how the sensitivity of prediction changes with respect to the number and type of variables used in the prediction process.

AB - Using several variables known to be related to prostate cancer, a multivariate classification method is developed to predict the onset of clinical prostate cancer. A multivariate mixed-effects model is used to describe longitudinal changes in prostate-specific antigen (PSA), a free testosterone index (FTI), and body mass index (BMI) before any clinical evidence of prostate cancer. The patterns of change in these three variables are allowed to vary depending on whether the subject develops prostate cancer or not and the severity of the prostate cancer at diagnosis. An application of Bayes' theorem provides posterior probabilities that we use to predict whether an individual will develop prostate cancer and, if so, whether it is a high-risk or a low-risk cancer. The classification rule is applied sequentially one multivariate observation at a time until the subject is classified as a cancer case or until the last observation has been used. We perform the analyses using each of the three variables individually, combined together in pairs, and all three variables together in one analysis. We compare the classification results among the various analyses and a simulation study demonstrates how the sensitivity of prediction changes with respect to the number and type of variables used in the prediction process.

UR - http://www.scopus.com/inward/record.url?scp=84859772296&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84859772296&partnerID=8YFLogxK

U2 - 10.1080/02664763.2011.644523

DO - 10.1080/02664763.2011.644523

M3 - Article

VL - 39

SP - 1151

EP - 1175

JO - Journal of Applied Statistics

JF - Journal of Applied Statistics

SN - 0266-4763

IS - 6

ER -