Computer-aided detection thresholds for digital chest radiography interpretation in tuberculosis diagnostic algorithms

Fiona Vanobberghen; Alfred Kipyegon Keter; Bart K M Jacobs; Tracy R Glass; Lutgarde Lynen; Irwin Law; Keelin Murphy; Bram van Ginneken; Irene Ayakaka; Alastair van Heerden; Llang Maama; Klaus Reither

doi:10.1183/23120541.00508-2023

Computer-aided detection thresholds for digital chest radiography interpretation in tuberculosis diagnostic algorithms

Fiona Vanobberghen, Alfred Kipyegon Keter, Bart K M Jacobs, Tracy R Glass, Lutgarde Lynen, Irwin Law, Keelin Murphy, Bram van Ginneken, Irene Ayakaka, Alastair van Heerden, Llang Maama, Klaus Reither

Research output: Contribution to journal › A1: Web of Science-article › peer-review

Abstract

OBJECTIVES: Use of computer-aided detection (CAD) software is recommended to improve tuberculosis screening and triage, but threshold determination is challenging if reference testing has not been performed in all individuals. We aimed to determine such thresholds through secondary analysis of the 2019 Lesotho national tuberculosis prevalence survey.

METHODS: Symptom screening and chest radiographs were performed in participants aged ≥15 years; those symptomatic or with abnormal chest radiographs provided samples for Xpert MTB/RIF and culture testing. Chest radiographs were processed using CAD4TB version 7. We used six methodological approaches to deal with participants who did not have bacteriological test results to estimate pulmonary tuberculosis prevalence and assess diagnostic accuracy.

RESULTS: Among 17 070 participants, 5214 (31%) had their tuberculosis status determined; 142 had tuberculosis. Prevalence estimates varied between methodological approaches (0.83-2.72%). Using multiple imputation to estimate tuberculosis status for those eligible but not tested, and assuming those not eligible for testing were negative, a CAD4TBv7 threshold of 13 had a sensitivity of 89.7% (95% CI 84.6-94.8) and a specificity of 74.2% (73.6-74.9), close to World Health Organization (WHO) target product profile criteria. Assuming all those not tested were negative produced similar results.

CONCLUSIONS: This is the first study to evaluate CAD4TB in a community screening context employing a range of approaches to account for unknown tuberculosis status. The assumption that those not tested are negative - regardless of testing eligibility status - was robust. As threshold determination must be context specific, our analytically straightforward approach should be adopted to leverage prevalence surveys for CAD threshold determination in other settings with a comparable proportion of eligible but not tested participants.

Original language	English
Article number	00508-2023
Journal	ERJ Open Research
Volume	10
Issue number	1
Number of pages	13
ISSN	2312-0541
DOIs	https://doi.org/10.1183/23120541.00508-2023
Publication status	Published - 2024

Access to Document

10.1183/23120541.00508-2023Licence: CC BY-NC

Cite this

Vanobberghen, F., Keter, A. K., Jacobs, B. K. M., Glass, T. R., Lynen, L., Law, I., Murphy, K., van Ginneken, B., Ayakaka, I., van Heerden, A., Maama, L., & Reither, K. (2024). Computer-aided detection thresholds for digital chest radiography interpretation in tuberculosis diagnostic algorithms. ERJ Open Research, 10(1), Article 00508-2023. https://doi.org/10.1183/23120541.00508-2023

@article{b95bfcb534944929aec6d591ba6f32fe,

title = "Computer-aided detection thresholds for digital chest radiography interpretation in tuberculosis diagnostic algorithms",

abstract = "OBJECTIVES: Use of computer-aided detection (CAD) software is recommended to improve tuberculosis screening and triage, but threshold determination is challenging if reference testing has not been performed in all individuals. We aimed to determine such thresholds through secondary analysis of the 2019 Lesotho national tuberculosis prevalence survey.METHODS: Symptom screening and chest radiographs were performed in participants aged ≥15 years; those symptomatic or with abnormal chest radiographs provided samples for Xpert MTB/RIF and culture testing. Chest radiographs were processed using CAD4TB version 7. We used six methodological approaches to deal with participants who did not have bacteriological test results to estimate pulmonary tuberculosis prevalence and assess diagnostic accuracy.RESULTS: Among 17 070 participants, 5214 (31%) had their tuberculosis status determined; 142 had tuberculosis. Prevalence estimates varied between methodological approaches (0.83-2.72%). Using multiple imputation to estimate tuberculosis status for those eligible but not tested, and assuming those not eligible for testing were negative, a CAD4TBv7 threshold of 13 had a sensitivity of 89.7% (95% CI 84.6-94.8) and a specificity of 74.2% (73.6-74.9), close to World Health Organization (WHO) target product profile criteria. Assuming all those not tested were negative produced similar results.CONCLUSIONS: This is the first study to evaluate CAD4TB in a community screening context employing a range of approaches to account for unknown tuberculosis status. The assumption that those not tested are negative - regardless of testing eligibility status - was robust. As threshold determination must be context specific, our analytically straightforward approach should be adopted to leverage prevalence surveys for CAD threshold determination in other settings with a comparable proportion of eligible but not tested participants.",

author = "Fiona Vanobberghen and Keter, {Alfred Kipyegon} and Jacobs, {Bart K M} and Glass, {Tracy R} and Lutgarde Lynen and Irwin Law and Keelin Murphy and {van Ginneken}, Bram and Irene Ayakaka and {van Heerden}, Alastair and Llang Maama and Klaus Reither",

note = "FTX; DOAJ; (CC BY NC)",

year = "2024",

doi = "10.1183/23120541.00508-2023",

language = "English",

volume = "10",

journal = "ERJ Open Research",

issn = "2312-0541",

publisher = "European Respiratory Society",

number = "1",

}

TY - JOUR

T1 - Computer-aided detection thresholds for digital chest radiography interpretation in tuberculosis diagnostic algorithms

AU - Vanobberghen, Fiona

AU - Keter, Alfred Kipyegon

AU - Jacobs, Bart K M

AU - Glass, Tracy R

AU - Lynen, Lutgarde

AU - Law, Irwin

AU - Murphy, Keelin

AU - van Ginneken, Bram

AU - Ayakaka, Irene

AU - van Heerden, Alastair

AU - Maama, Llang

AU - Reither, Klaus

N1 - FTX; DOAJ; (CC BY NC)

PY - 2024

Y1 - 2024

N2 - OBJECTIVES: Use of computer-aided detection (CAD) software is recommended to improve tuberculosis screening and triage, but threshold determination is challenging if reference testing has not been performed in all individuals. We aimed to determine such thresholds through secondary analysis of the 2019 Lesotho national tuberculosis prevalence survey.METHODS: Symptom screening and chest radiographs were performed in participants aged ≥15 years; those symptomatic or with abnormal chest radiographs provided samples for Xpert MTB/RIF and culture testing. Chest radiographs were processed using CAD4TB version 7. We used six methodological approaches to deal with participants who did not have bacteriological test results to estimate pulmonary tuberculosis prevalence and assess diagnostic accuracy.RESULTS: Among 17 070 participants, 5214 (31%) had their tuberculosis status determined; 142 had tuberculosis. Prevalence estimates varied between methodological approaches (0.83-2.72%). Using multiple imputation to estimate tuberculosis status for those eligible but not tested, and assuming those not eligible for testing were negative, a CAD4TBv7 threshold of 13 had a sensitivity of 89.7% (95% CI 84.6-94.8) and a specificity of 74.2% (73.6-74.9), close to World Health Organization (WHO) target product profile criteria. Assuming all those not tested were negative produced similar results.CONCLUSIONS: This is the first study to evaluate CAD4TB in a community screening context employing a range of approaches to account for unknown tuberculosis status. The assumption that those not tested are negative - regardless of testing eligibility status - was robust. As threshold determination must be context specific, our analytically straightforward approach should be adopted to leverage prevalence surveys for CAD threshold determination in other settings with a comparable proportion of eligible but not tested participants.

AB - OBJECTIVES: Use of computer-aided detection (CAD) software is recommended to improve tuberculosis screening and triage, but threshold determination is challenging if reference testing has not been performed in all individuals. We aimed to determine such thresholds through secondary analysis of the 2019 Lesotho national tuberculosis prevalence survey.METHODS: Symptom screening and chest radiographs were performed in participants aged ≥15 years; those symptomatic or with abnormal chest radiographs provided samples for Xpert MTB/RIF and culture testing. Chest radiographs were processed using CAD4TB version 7. We used six methodological approaches to deal with participants who did not have bacteriological test results to estimate pulmonary tuberculosis prevalence and assess diagnostic accuracy.RESULTS: Among 17 070 participants, 5214 (31%) had their tuberculosis status determined; 142 had tuberculosis. Prevalence estimates varied between methodological approaches (0.83-2.72%). Using multiple imputation to estimate tuberculosis status for those eligible but not tested, and assuming those not eligible for testing were negative, a CAD4TBv7 threshold of 13 had a sensitivity of 89.7% (95% CI 84.6-94.8) and a specificity of 74.2% (73.6-74.9), close to World Health Organization (WHO) target product profile criteria. Assuming all those not tested were negative produced similar results.CONCLUSIONS: This is the first study to evaluate CAD4TB in a community screening context employing a range of approaches to account for unknown tuberculosis status. The assumption that those not tested are negative - regardless of testing eligibility status - was robust. As threshold determination must be context specific, our analytically straightforward approach should be adopted to leverage prevalence surveys for CAD threshold determination in other settings with a comparable proportion of eligible but not tested participants.

U2 - 10.1183/23120541.00508-2023

DO - 10.1183/23120541.00508-2023

M3 - A1: Web of Science-article

C2 - 38196890

SN - 2312-0541

VL - 10

JO - ERJ Open Research

JF - ERJ Open Research

IS - 1

M1 - 00508-2023

ER -

Computer-aided detection thresholds for digital chest radiography interpretation in tuberculosis diagnostic algorithms

Abstract

Access to Document

Fingerprint

Cite this