top of page

First 3D model of black skin cancer developed for digital pathology

Scientists from the German Cancer Research Center (DKFZ), together with dermatopathologists, have succeeded in creating the first digital 3D model of black skin cancer, melanoma. This allows dermatologists to "scroll" through the tumor. In the future, this could improve assessment and enable more precise prognoses.


Skin cancer diagnoses today are made by dermatopathologists on two-dimensional tissue sections under the microscope. With the advent of digital pathology and artificial intelligence, tissue sections are increasingly being digitized, opening up entirely new possibilities for integrating their information.


Gewebe eines Pigmentmals ; © DKFZ/Brinker


"Tumors are three-dimensional, so it makes sense to digitally recreate the 3D structure to improve cancer diagnostics," says Titus Brinker, dermatologist and junior research group leader at DKFZ. In collaboration with experienced dermatopathologists, he and his team have succeeded in creating the first digital 3D model of black skin cancer, melanoma, and evaluating its implications for digital pathology together with experts.


A malignant melanoma of the skin was sliced into three-micrometer-thin sections using a microtome, which were subsequently digitized by a slide scanner. Brinker's team adapted open-source software to reconstruct a 3D model from these sections. Nine pathologists from four different countries who had at least ten years of experience in the histological diagnosis of melanoma then tested the model.


The vast majority of experts rated the digital 3D model positively. Scrolling through the tumor enables rapid diagnosis of many tissue layers. The better representation of the anatomy was considered an advantage, as was the possibility of evaluating different tissue layers simultaneously. The experts saw the high tissue consumption and still low resolution due to a lack of computer power as limitations.


Group leader Brinker is optimistic: "We assume that increasing automation in pathology will support the relevance of such digital 3D models in the future."



Alexander Kurz, Dieter Krahl, Heinz Kutzner, Raymond Barnhill, Antonio Perasole, Maria Teresa Fernandez Figuers, Gerardo Ferrara, Stephan A. Braun, Hans Starz, Mar Llamas-Velasco, Jochen Sven Utikal, Stefan Fröhling Christof von Kalle, Jakob Nikolas Kather, Lucas Schneider, Titus J. Brinker:

European Journal of Cancer, 2023, 113294, DOI: https://doi.org/10.1016/j.ejca.2023.113294







コメント


bottom of page