Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Facilitating clinically relevant skin tumor diagnostics with spectroscopy-driven machine learning

Andersson, Emil LU orcid ; Hult, Jenny LU orcid ; Troein, Carl LU orcid ; Stridh, Magne LU ; Sjögren, Benjamin LU ; Pekar-Lukacs, Agnes LU ; Hernandez-Palacios, Julio ; Edén, Patrik LU ; Persson, Bertil LU and Olariu Annell, Victor LU , et al. (2024) In iScience 27(5).
Abstract
In the dawning era of artificial intelligence (AI), health care stands to undergo a significant transformation with the increasing digitalization of patient data. Digital imaging, in particular, will serve as an important platform for AI to aid decision making and diagnostics. A growing number of studies demonstrate the potential of automatic pre-surgical skin tumor delineation, which could have tremendous impact on clinical practice. However, current methods rely on having ground truth images in which tumor borders are already identified, which is not clinically possible. We report a novel approach where hyperspectral images provide spectra from small regions representing healthy tissue and tumor, which are used to generate prediction... (More)
In the dawning era of artificial intelligence (AI), health care stands to undergo a significant transformation with the increasing digitalization of patient data. Digital imaging, in particular, will serve as an important platform for AI to aid decision making and diagnostics. A growing number of studies demonstrate the potential of automatic pre-surgical skin tumor delineation, which could have tremendous impact on clinical practice. However, current methods rely on having ground truth images in which tumor borders are already identified, which is not clinically possible. We report a novel approach where hyperspectral images provide spectra from small regions representing healthy tissue and tumor, which are used to generate prediction maps using artificial neural networks (ANNs), after which a segmentation algorithm automatically identifies the tumor borders. This circumvents the need for ground truth images, since an ANN model is trained with data from each individual patient, representing a more clinically relevant approach. (Less)
Please use this url to cite or link to this publication:
@article{073753ab-9338-4cb0-bae2-4cdf4148d66f,
  abstract     = {{In the dawning era of artificial intelligence (AI), health care stands to undergo a significant transformation with the increasing digitalization of patient data. Digital imaging, in particular, will serve as an important platform for AI to aid decision making and diagnostics. A growing number of studies demonstrate the potential of automatic pre-surgical skin tumor delineation, which could have tremendous impact on clinical practice. However, current methods rely on having ground truth images in which tumor borders are already identified, which is not clinically possible. We report a novel approach where hyperspectral images provide spectra from small regions representing healthy tissue and tumor, which are used to generate prediction maps using artificial neural networks (ANNs), after which a segmentation algorithm automatically identifies the tumor borders. This circumvents the need for ground truth images, since an ANN model is trained with data from each individual patient, representing a more clinically relevant approach.}},
  author       = {{Andersson, Emil and Hult, Jenny and Troein, Carl and Stridh, Magne and Sjögren, Benjamin and Pekar-Lukacs, Agnes and Hernandez-Palacios, Julio and Edén, Patrik and Persson, Bertil and Olariu Annell, Victor and Malmsjö, Malin and Merdasa, Aboma}},
  issn         = {{2589-0042}},
  keywords     = {{Hyperspectral imaging; Machine learning; Skin tumours}},
  language     = {{eng}},
  month        = {{05}},
  number       = {{5}},
  publisher    = {{Elsevier}},
  series       = {{iScience}},
  title        = {{Facilitating clinically relevant skin tumor diagnostics with spectroscopy-driven machine learning}},
  url          = {{http://dx.doi.org/10.1016/j.isci.2024.109653}},
  doi          = {{10.1016/j.isci.2024.109653}},
  volume       = {{27}},
  year         = {{2024}},
}