Computer-assisted cystoscopy diagnosis of bladder cancer

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Objectives: One of the most reliable methods for diagnosing bladder cancer is cystoscopy. Depending on the findings, this may be followed by a referral to a more experienced urologist or a biopsy and histological analysis of suspicious lesion. In this work, we explore whether computer-assisted triage of cystoscopy findings can identify low-risk lesions and reduce the number of referrals or biopsies, associated complications, and costs, although reducing subjectivity of the procedure and indicating when the risk of a lesion being malignant is minimal.

Materials and methods: Cystoscopy images taken during routine clinical patient evaluation and supported by biopsy were interpreted by an expert clinician. They were further subjected to an automated image analysis developed to best capture cancer characteristics. The images were transformed and divided into segments, using a specialised color segmentation system. After the selection of a set of highly informative features, the segments were separated into 4 classes: healthy, veins, inflammation, and cancerous. The images were then classified as healthy and diseased, using a linear discriminant, the naïve Bayes, and the quadratic linear classifiers. Performance of the classifiers was measured by using receiver operation characteristic curves.

Results: The classification system developed here, with the quadratic classifier, yielded 50% false-positive rate and zero false-negative rate, which means, that no malignant lesions would be missed by this classifier.

Conclusions: Based on criteria used for assessment of cystoscopy images by medical specialists and features that human visual system is less sensitive to, we developed a computer program that carries out automated analysis of cystoscopy images. Our program could be used as a triage to identify patients who do not require referral or further testing.

LanguageEnglish
Article number8
Pages8.e9-8.e15
Number of pages7
JournalUrologic Oncology: Seminars and Original Investigations
Volume36
Issue number1
DOIs
Publication statusPublished - Jan 2018

Fingerprint

Computer-Assisted Diagnosis
Cystoscopy
Urinary Bladder Neoplasms
Referral and Consultation
Triage
Biopsy
Veins
Software
Color
Inflammation
Costs and Cost Analysis
Neoplasms

Keywords

  • Cystoscopy
  • Early detection of cancer
  • Endoscopy
  • Image interpretation
  • Urinary bladder
  • Urinary bladder neoplasms
  • Computer-assisted

Cite this

@article{05f8263b2b7b44f09c3f0621725dc643,
title = "Computer-assisted cystoscopy diagnosis of bladder cancer",
abstract = "Objectives: One of the most reliable methods for diagnosing bladder cancer is cystoscopy. Depending on the findings, this may be followed by a referral to a more experienced urologist or a biopsy and histological analysis of suspicious lesion. In this work, we explore whether computer-assisted triage of cystoscopy findings can identify low-risk lesions and reduce the number of referrals or biopsies, associated complications, and costs, although reducing subjectivity of the procedure and indicating when the risk of a lesion being malignant is minimal.Materials and methods: Cystoscopy images taken during routine clinical patient evaluation and supported by biopsy were interpreted by an expert clinician. They were further subjected to an automated image analysis developed to best capture cancer characteristics. The images were transformed and divided into segments, using a specialised color segmentation system. After the selection of a set of highly informative features, the segments were separated into 4 classes: healthy, veins, inflammation, and cancerous. The images were then classified as healthy and diseased, using a linear discriminant, the na{\"i}ve Bayes, and the quadratic linear classifiers. Performance of the classifiers was measured by using receiver operation characteristic curves.Results: The classification system developed here, with the quadratic classifier, yielded 50{\%} false-positive rate and zero false-negative rate, which means, that no malignant lesions would be missed by this classifier.Conclusions: Based on criteria used for assessment of cystoscopy images by medical specialists and features that human visual system is less sensitive to, we developed a computer program that carries out automated analysis of cystoscopy images. Our program could be used as a triage to identify patients who do not require referral or further testing.",
keywords = "Cystoscopy, Early detection of cancer, Endoscopy, Image interpretation, Urinary bladder, Urinary bladder neoplasms, Computer-assisted",
author = "Gosnell, {Martin E.} and Polikarpov, {Dmitry M.} and Goldys, {Ewa M.} and Zvyagin, {Andrei V.} and Gillatt, {David A.}",
year = "2018",
month = "1",
doi = "10.1016/j.urolonc.2017.08.026",
language = "English",
volume = "36",
pages = "8.e9--8.e15",
journal = "Urologic Oncology: Seminars and Original Investigations",
issn = "1078-1439",
publisher = "Elsevier",
number = "1",

}

Computer-assisted cystoscopy diagnosis of bladder cancer. / Gosnell, Martin E.; Polikarpov, Dmitry M.; Goldys, Ewa M.; Zvyagin, Andrei V.; Gillatt, David A.

In: Urologic Oncology: Seminars and Original Investigations, Vol. 36, No. 1, 8, 01.2018, p. 8.e9-8.e15.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Computer-assisted cystoscopy diagnosis of bladder cancer

AU - Gosnell, Martin E.

AU - Polikarpov, Dmitry M.

AU - Goldys, Ewa M.

AU - Zvyagin, Andrei V.

AU - Gillatt, David A.

PY - 2018/1

Y1 - 2018/1

N2 - Objectives: One of the most reliable methods for diagnosing bladder cancer is cystoscopy. Depending on the findings, this may be followed by a referral to a more experienced urologist or a biopsy and histological analysis of suspicious lesion. In this work, we explore whether computer-assisted triage of cystoscopy findings can identify low-risk lesions and reduce the number of referrals or biopsies, associated complications, and costs, although reducing subjectivity of the procedure and indicating when the risk of a lesion being malignant is minimal.Materials and methods: Cystoscopy images taken during routine clinical patient evaluation and supported by biopsy were interpreted by an expert clinician. They were further subjected to an automated image analysis developed to best capture cancer characteristics. The images were transformed and divided into segments, using a specialised color segmentation system. After the selection of a set of highly informative features, the segments were separated into 4 classes: healthy, veins, inflammation, and cancerous. The images were then classified as healthy and diseased, using a linear discriminant, the naïve Bayes, and the quadratic linear classifiers. Performance of the classifiers was measured by using receiver operation characteristic curves.Results: The classification system developed here, with the quadratic classifier, yielded 50% false-positive rate and zero false-negative rate, which means, that no malignant lesions would be missed by this classifier.Conclusions: Based on criteria used for assessment of cystoscopy images by medical specialists and features that human visual system is less sensitive to, we developed a computer program that carries out automated analysis of cystoscopy images. Our program could be used as a triage to identify patients who do not require referral or further testing.

AB - Objectives: One of the most reliable methods for diagnosing bladder cancer is cystoscopy. Depending on the findings, this may be followed by a referral to a more experienced urologist or a biopsy and histological analysis of suspicious lesion. In this work, we explore whether computer-assisted triage of cystoscopy findings can identify low-risk lesions and reduce the number of referrals or biopsies, associated complications, and costs, although reducing subjectivity of the procedure and indicating when the risk of a lesion being malignant is minimal.Materials and methods: Cystoscopy images taken during routine clinical patient evaluation and supported by biopsy were interpreted by an expert clinician. They were further subjected to an automated image analysis developed to best capture cancer characteristics. The images were transformed and divided into segments, using a specialised color segmentation system. After the selection of a set of highly informative features, the segments were separated into 4 classes: healthy, veins, inflammation, and cancerous. The images were then classified as healthy and diseased, using a linear discriminant, the naïve Bayes, and the quadratic linear classifiers. Performance of the classifiers was measured by using receiver operation characteristic curves.Results: The classification system developed here, with the quadratic classifier, yielded 50% false-positive rate and zero false-negative rate, which means, that no malignant lesions would be missed by this classifier.Conclusions: Based on criteria used for assessment of cystoscopy images by medical specialists and features that human visual system is less sensitive to, we developed a computer program that carries out automated analysis of cystoscopy images. Our program could be used as a triage to identify patients who do not require referral or further testing.

KW - Cystoscopy

KW - Early detection of cancer

KW - Endoscopy

KW - Image interpretation

KW - Urinary bladder

KW - Urinary bladder neoplasms

KW - Computer-assisted

UR - http://www.scopus.com/inward/record.url?scp=85029792669&partnerID=8YFLogxK

U2 - 10.1016/j.urolonc.2017.08.026

DO - 10.1016/j.urolonc.2017.08.026

M3 - Article

VL - 36

SP - 8.e9-8.e15

JO - Urologic Oncology: Seminars and Original Investigations

T2 - Urologic Oncology: Seminars and Original Investigations

JF - Urologic Oncology: Seminars and Original Investigations

SN - 1078-1439

IS - 1

M1 - 8

ER -