Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

A fully automated AI-based method for tumour detection and quantification on [18F]PSMA-1007 PET–CT images in prostate cancer

Trägårdh, Elin LU orcid ; Ulén, Johannes ; Enqvist, Olof LU ; Larsson, Måns ; Valind, Kristian LU orcid ; Minarik, David LU and Edenbrandt, Lars (2025) In EJNMMI Physics 12. p.1-18
Abstract

Background: In this study, we further developed an artificial intelligence (AI)-based method for the detection and quantification of tumours in the prostate, lymph nodes and bone in prostate-specific membrane antigen (PSMA)-targeting positron emission tomography with computed tomography (PET–CT) images. Methods: A total of 1064 [18F]PSMA-1007 PET–CT scans were used (approximately twice as many compared to our previous AI model), of which 120 were used as test set. Suspected lesions were manually annotated and used as ground truth. A convolutional neural network was developed and trained. The sensitivity and positive predictive value (PPV) were calculated using two sets of manual segmentations as reference. Results were also... (More)

Background: In this study, we further developed an artificial intelligence (AI)-based method for the detection and quantification of tumours in the prostate, lymph nodes and bone in prostate-specific membrane antigen (PSMA)-targeting positron emission tomography with computed tomography (PET–CT) images. Methods: A total of 1064 [18F]PSMA-1007 PET–CT scans were used (approximately twice as many compared to our previous AI model), of which 120 were used as test set. Suspected lesions were manually annotated and used as ground truth. A convolutional neural network was developed and trained. The sensitivity and positive predictive value (PPV) were calculated using two sets of manual segmentations as reference. Results were also compared to our previously developed AI method. The correlation between manually and AI-based calculations of total lesion volume (TLV) and total lesion uptake (TLU) were calculated. Results: The sensitivities of the AI method were 85% for prostate tumour/recurrence, 91% for lymph node metastases and 61% for bone metastases (82%, 86% and 70% for manual readings and 66%, 88% and 71% for the old AI method). The PPVs of the AI method were 85%, 83% and 58%, respectively (63%, 86% and 39% for manual readings, and 69%, 70% and 39% for the old AI method). The correlations between manual and AI-based calculations of TLV and TLU ranged from r = 0.62 to r = 0.96. Conclusion: The performance of the newly developed and fully automated AI-based method for detecting and quantifying prostate tumour and suspected lymph node and bone metastases increased significantly, especially the PPV. The AI method is freely available to other researchers (www.recomia.org).

(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
keywords
Artificial intelligence, CNN, PET–CT, Prostate cancer, PSMA
in
EJNMMI Physics
volume
12
article number
78
pages
1 - 18
publisher
Springer Science and Business Media B.V.
external identifiers
  • pmid:40833689
  • scopus:105013762784
ISSN
2197-7364
DOI
10.1186/s40658-025-00786-9
language
English
LU publication?
yes
additional info
Publisher Copyright: © The Author(s) 2025.
id
43744fd8-8f9f-4fc4-9a1f-774d8d2e5f60
date added to LUP
2025-09-15 14:39:49
date last changed
2025-09-29 16:46:53
@article{43744fd8-8f9f-4fc4-9a1f-774d8d2e5f60,
  abstract     = {{<p>Background: In this study, we further developed an artificial intelligence (AI)-based method for the detection and quantification of tumours in the prostate, lymph nodes and bone in prostate-specific membrane antigen (PSMA)-targeting positron emission tomography with computed tomography (PET–CT) images. Methods: A total of 1064 [<sup>18</sup>F]PSMA-1007 PET–CT scans were used (approximately twice as many compared to our previous AI model), of which 120 were used as test set. Suspected lesions were manually annotated and used as ground truth. A convolutional neural network was developed and trained. The sensitivity and positive predictive value (PPV) were calculated using two sets of manual segmentations as reference. Results were also compared to our previously developed AI method. The correlation between manually and AI-based calculations of total lesion volume (TLV) and total lesion uptake (TLU) were calculated. Results: The sensitivities of the AI method were 85% for prostate tumour/recurrence, 91% for lymph node metastases and 61% for bone metastases (82%, 86% and 70% for manual readings and 66%, 88% and 71% for the old AI method). The PPVs of the AI method were 85%, 83% and 58%, respectively (63%, 86% and 39% for manual readings, and 69%, 70% and 39% for the old AI method). The correlations between manual and AI-based calculations of TLV and TLU ranged from r = 0.62 to r = 0.96. Conclusion: The performance of the newly developed and fully automated AI-based method for detecting and quantifying prostate tumour and suspected lymph node and bone metastases increased significantly, especially the PPV. The AI method is freely available to other researchers (www.recomia.org).</p>}},
  author       = {{Trägårdh, Elin and Ulén, Johannes and Enqvist, Olof and Larsson, Måns and Valind, Kristian and Minarik, David and Edenbrandt, Lars}},
  issn         = {{2197-7364}},
  keywords     = {{Artificial intelligence; CNN; PET–CT; Prostate cancer; PSMA}},
  language     = {{eng}},
  pages        = {{1--18}},
  publisher    = {{Springer Science and Business Media B.V.}},
  series       = {{EJNMMI Physics}},
  title        = {{A fully automated AI-based method for tumour detection and quantification on [<sup>18</sup>F]PSMA-1007 PET–CT images in prostate cancer}},
  url          = {{http://dx.doi.org/10.1186/s40658-025-00786-9}},
  doi          = {{10.1186/s40658-025-00786-9}},
  volume       = {{12}},
  year         = {{2025}},
}