A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms

Research output: Contribution to journal › Journal article › Research › peer-review

Standard

A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms. / Carlsen, Esben Andreas; Lindholm, Kristian; Hindsholm, Amalie; Gæde, Mathias; Ladefoged, Claes Nøhr; Loft, Mathias; Johnbeck, Camilla Bardram; Langer, Seppo Wang; Oturai, Peter; Knigge, Ulrich; Kjaer, Andreas; Andersen, Flemming Littrup.

In: EJNMMI Research, Vol. 12, No. 1, 30, 2022, p. 1-10.

Research output: Contribution to journal › Journal article › Research › peer-review

Harvard

Carlsen, EA, Lindholm, K, Hindsholm, A, Gæde, M, Ladefoged, CN, Loft, M, Johnbeck, CB, Langer, SW, Oturai, P, Knigge, U, Kjaer, A & Andersen, FL 2022, 'A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms', EJNMMI Research, vol. 12, no. 1, 30, pp. 1-10. https://doi.org/10.1186/s13550-022-00901-2

APA

Carlsen, E. A., Lindholm, K., Hindsholm, A., Gæde, M., Ladefoged, C. N., Loft, M., Johnbeck, C. B., Langer, S. W., Oturai, P., Knigge, U., Kjaer, A., & Andersen, F. L. (2022). A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms. EJNMMI Research, 12(1), 1-10. [30]. https://doi.org/10.1186/s13550-022-00901-2

Vancouver

Carlsen EA, Lindholm K, Hindsholm A, Gæde M, Ladefoged CN, Loft M et al. A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms. EJNMMI Research. 2022;12(1):1-10. 30. https://doi.org/10.1186/s13550-022-00901-2

Author

Carlsen, Esben Andreas ; Lindholm, Kristian ; Hindsholm, Amalie ; Gæde, Mathias ; Ladefoged, Claes Nøhr ; Loft, Mathias ; Johnbeck, Camilla Bardram ; Langer, Seppo Wang ; Oturai, Peter ; Knigge, Ulrich ; Kjaer, Andreas ; Andersen, Flemming Littrup. / A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms. In: EJNMMI Research. 2022 ; Vol. 12, No. 1. pp. 1-10.

Bibtex

@article{903609cb5eea4d008087b03c285b266a,
title = "A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms",
abstract = "Background: Segmentation of neuroendocrine neoplasms (NENs) in [64Cu]Cu-DOTATATE positron emission tomography makes it possible to extract quantitative measures useable for prognostication of patients. However, manual tumor segmentation is cumbersome and time-consuming. Therefore, we aimed to implement and test an artificial intelligence (AI) network for tumor segmentation. Patients with gastroenteropancreatic or lung NEN with [64Cu]Cu-DOTATATE PET/CT performed were included in our training (n = 117) and test cohort (n = 41). Further, 10 patients with no signs of NEN were included as negative controls. Ground truth segmentations were obtained by a standardized semiautomatic method for tumor segmentation by a physician. The nnU-Net framework was used to set up a deep learning U-net architecture. Dice score, sensitivity and precision were used for selection of the final model. AI segmentations were implemented in a clinical imaging viewer where a physician evaluated performance and performed manual adjustments. Results: Cross-validation training was used to generate models and an ensemble model. The ensemble model performed best overall with a lesion-wise dice of 0.850 and pixel-wise dice, precision and sensitivity of 0.801, 0.786 and 0.872, respectively. Performance of the ensemble model was acceptable with some degree of manual adjustment in 35/41 (85%) patients. Final tumor segmentation could be obtained from the AI model with manual adjustments in 5 min versus 17 min for ground truth method, p < 0.01. Conclusion: We implemented and validated an AI model that achieved a high similarity with ground truth segmentation and resulted in faster tumor segmentation. With AI, total tumor segmentation may become feasible in the clinical routine.",
keywords = "Artificial intelligence, Neuroendocrine neoplasms, Prognostication, Tumor segmentation, [Cu]Cu-DOTATATE PET",
author = "Carlsen, {Esben Andreas} and Kristian Lindholm and Amalie Hindsholm and Mathias G{\ae}de and Ladefoged, {Claes N{\o}hr} and Mathias Loft and Johnbeck, {Camilla Bardram} and Langer, {Seppo Wang} and Peter Oturai and Ulrich Knigge and Andreas Kjaer and Andersen, {Flemming Littrup}",
note = "Publisher Copyright: {\textcopyright} 2022, The Author(s).",
year = "2022",
doi = "10.1186/s13550-022-00901-2",
language = "English",
volume = "12",
pages = "1--10",
journal = "EJNMMI Research",
issn = "2191-219X",
publisher = "SpringerOpen",
number = "1",

}

RIS

TY - JOUR

T1 - A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms

AU - Carlsen, Esben Andreas

AU - Lindholm, Kristian

AU - Hindsholm, Amalie

AU - Gæde, Mathias

AU - Ladefoged, Claes Nøhr

AU - Loft, Mathias

AU - Johnbeck, Camilla Bardram

AU - Langer, Seppo Wang

AU - Oturai, Peter

AU - Knigge, Ulrich

AU - Kjaer, Andreas

AU - Andersen, Flemming Littrup

N1 - Publisher Copyright: © 2022, The Author(s).

PY - 2022

Y1 - 2022

N2 - Background: Segmentation of neuroendocrine neoplasms (NENs) in [64Cu]Cu-DOTATATE positron emission tomography makes it possible to extract quantitative measures useable for prognostication of patients. However, manual tumor segmentation is cumbersome and time-consuming. Therefore, we aimed to implement and test an artificial intelligence (AI) network for tumor segmentation. Patients with gastroenteropancreatic or lung NEN with [64Cu]Cu-DOTATATE PET/CT performed were included in our training (n = 117) and test cohort (n = 41). Further, 10 patients with no signs of NEN were included as negative controls. Ground truth segmentations were obtained by a standardized semiautomatic method for tumor segmentation by a physician. The nnU-Net framework was used to set up a deep learning U-net architecture. Dice score, sensitivity and precision were used for selection of the final model. AI segmentations were implemented in a clinical imaging viewer where a physician evaluated performance and performed manual adjustments. Results: Cross-validation training was used to generate models and an ensemble model. The ensemble model performed best overall with a lesion-wise dice of 0.850 and pixel-wise dice, precision and sensitivity of 0.801, 0.786 and 0.872, respectively. Performance of the ensemble model was acceptable with some degree of manual adjustment in 35/41 (85%) patients. Final tumor segmentation could be obtained from the AI model with manual adjustments in 5 min versus 17 min for ground truth method, p < 0.01. Conclusion: We implemented and validated an AI model that achieved a high similarity with ground truth segmentation and resulted in faster tumor segmentation. With AI, total tumor segmentation may become feasible in the clinical routine.

AB - Background: Segmentation of neuroendocrine neoplasms (NENs) in [64Cu]Cu-DOTATATE positron emission tomography makes it possible to extract quantitative measures useable for prognostication of patients. However, manual tumor segmentation is cumbersome and time-consuming. Therefore, we aimed to implement and test an artificial intelligence (AI) network for tumor segmentation. Patients with gastroenteropancreatic or lung NEN with [64Cu]Cu-DOTATATE PET/CT performed were included in our training (n = 117) and test cohort (n = 41). Further, 10 patients with no signs of NEN were included as negative controls. Ground truth segmentations were obtained by a standardized semiautomatic method for tumor segmentation by a physician. The nnU-Net framework was used to set up a deep learning U-net architecture. Dice score, sensitivity and precision were used for selection of the final model. AI segmentations were implemented in a clinical imaging viewer where a physician evaluated performance and performed manual adjustments. Results: Cross-validation training was used to generate models and an ensemble model. The ensemble model performed best overall with a lesion-wise dice of 0.850 and pixel-wise dice, precision and sensitivity of 0.801, 0.786 and 0.872, respectively. Performance of the ensemble model was acceptable with some degree of manual adjustment in 35/41 (85%) patients. Final tumor segmentation could be obtained from the AI model with manual adjustments in 5 min versus 17 min for ground truth method, p < 0.01. Conclusion: We implemented and validated an AI model that achieved a high similarity with ground truth segmentation and resulted in faster tumor segmentation. With AI, total tumor segmentation may become feasible in the clinical routine.

KW - Artificial intelligence

KW - Neuroendocrine neoplasms

KW - Prognostication

KW - Tumor segmentation

KW - [Cu]Cu-DOTATATE PET

UR - http://www.scopus.com/inward/record.url?scp=85130936426&partnerID=8YFLogxK

U2 - 10.1186/s13550-022-00901-2

DO - 10.1186/s13550-022-00901-2

M3 - Journal article

C2 - 35633448

AN - SCOPUS:85130936426

VL - 12

SP - 1

EP - 10

JO - EJNMMI Research

JF - EJNMMI Research

SN - 2191-219X

IS - 1

M1 - 30

ER -

ID: 316874621