Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Assessing the robustness of AI lesion risk scores at different exposure settings using an anthropomorphic breast phantom

Alström, L. ; Bjerkén, Anna LU orcid ; Dahlblom, V. LU orcid ; Timberg, P. LU ; Dustler, M. LU orcid and Tingberg, A. LU orcid (2026) In Radiation Protection Dosimetry 202(3-4). p.220-228
Abstract

To assess the robustness of risk scores provided by an artificial intelligence (AI) system for digital mammography (DM), when varying the exposure conditions. An anthropomorphic breast phantom containing a lesion, was imaged with DM at different tube voltages (kV), tube loadings (mAs), and anode/filter combinations (W/Rh, Mo/Mo, and Mo/Rh). The organ doses were extracted from the DICOM header and used as a substitute for average glandular dose. The images were analyzed with an AI system, which provided a lesion risk score which translates to suspicion for malignancy. Correlations between the lesion risk score and the exposure conditions were investigated. In most imaging conditions, weak to moderately strong positive associations... (More)

To assess the robustness of risk scores provided by an artificial intelligence (AI) system for digital mammography (DM), when varying the exposure conditions. An anthropomorphic breast phantom containing a lesion, was imaged with DM at different tube voltages (kV), tube loadings (mAs), and anode/filter combinations (W/Rh, Mo/Mo, and Mo/Rh). The organ doses were extracted from the DICOM header and used as a substitute for average glandular dose. The images were analyzed with an AI system, which provided a lesion risk score which translates to suspicion for malignancy. Correlations between the lesion risk score and the exposure conditions were investigated. In most imaging conditions, weak to moderately strong positive associations between lesion risk scores and kV and mAs, respectively, were reported (varying by anode/filter combinations). When organ dose increased the AI risk scores plateaued, and further increase did not increase the lesion risk score. For typical clinical settings (W/Rh, 27 kV and 71 mAs) the range of lesion risk scores was 33–56 (mean: 42, SD: 9). Greatest reported variability (range: 36–63, mean: 51, SD: 12) was found at 27 kV and 36 mAs (using W/Rh). Images of suboptimal quality may result in inaccurate AI system performance. The unexpectedly large intra-group variability of AI risk scores should be further investigated.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
in
Radiation Protection Dosimetry
volume
202
issue
3-4
pages
9 pages
publisher
Oxford University Press
external identifiers
  • scopus:105032779586
  • pmid:41821441
ISSN
0144-8420
DOI
10.1093/rpd/ncaf166
language
English
LU publication?
yes
additional info
Publisher Copyright: © The Author(s) 2026. Published by Oxford University Press.
id
f5c6a8db-7d8b-447a-bc4c-cfb0a788c56d
date added to LUP
2026-04-30 13:08:09
date last changed
2026-05-01 03:17:25
@article{f5c6a8db-7d8b-447a-bc4c-cfb0a788c56d,
  abstract     = {{<p>To assess the robustness of risk scores provided by an artificial intelligence (AI) system for digital mammography (DM), when varying the exposure conditions. An anthropomorphic breast phantom containing a lesion, was imaged with DM at different tube voltages (kV), tube loadings (mAs), and anode/filter combinations (W/Rh, Mo/Mo, and Mo/Rh). The organ doses were extracted from the DICOM header and used as a substitute for average glandular dose. The images were analyzed with an AI system, which provided a lesion risk score which translates to suspicion for malignancy. Correlations between the lesion risk score and the exposure conditions were investigated. In most imaging conditions, weak to moderately strong positive associations between lesion risk scores and kV and mAs, respectively, were reported (varying by anode/filter combinations). When organ dose increased the AI risk scores plateaued, and further increase did not increase the lesion risk score. For typical clinical settings (W/Rh, 27 kV and 71 mAs) the range of lesion risk scores was 33–56 (mean: 42, SD: 9). Greatest reported variability (range: 36–63, mean: 51, SD: 12) was found at 27 kV and 36 mAs (using W/Rh). Images of suboptimal quality may result in inaccurate AI system performance. The unexpectedly large intra-group variability of AI risk scores should be further investigated.</p>}},
  author       = {{Alström, L. and Bjerkén, Anna and Dahlblom, V. and Timberg, P. and Dustler, M. and Tingberg, A.}},
  issn         = {{0144-8420}},
  language     = {{eng}},
  month        = {{03}},
  number       = {{3-4}},
  pages        = {{220--228}},
  publisher    = {{Oxford University Press}},
  series       = {{Radiation Protection Dosimetry}},
  title        = {{Assessing the robustness of AI lesion risk scores at different exposure settings using an anthropomorphic breast phantom}},
  url          = {{http://dx.doi.org/10.1093/rpd/ncaf166}},
  doi          = {{10.1093/rpd/ncaf166}},
  volume       = {{202}},
  year         = {{2026}},
}