Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Histology-guided 3D virtual staining of microCT-imaged lung tissue via deep learning

Almagro-Pérez, Cristina ; Peruzzi, Niccolò LU orcid ; Galambos, Csaba ; Song, Andrew H ; Brunnström, Hans LU orcid ; Gawlik, Kinga I LU ; Stampanoni, Marco ; Tran-Lundmark, Karin LU and Lovric, Goran (2026) In Journal of the Royal Society, Interface 23(239).
Abstract

Histologically stained tissue sections are considered the gold standard for studying microscopic anatomy and diagnosing disease in clinical practice. However, the processes of sectioning and staining are laborious, and the overall method relies on two-dimensional analysis. In contrast, X-ray-based virtual histology offers the advantage of virtual sectioning while retaining the full three-dimensional (3D) volumetric representation of the tissue. Nevertheless, its greyscale nature limits its specificity compared to conventional histological stains and creates an additional barrier for pathologists, whose training is primarily based on colour-stained histology. In this work, we present a histology-guided enhancement platform that can... (More)

Histologically stained tissue sections are considered the gold standard for studying microscopic anatomy and diagnosing disease in clinical practice. However, the processes of sectioning and staining are laborious, and the overall method relies on two-dimensional analysis. In contrast, X-ray-based virtual histology offers the advantage of virtual sectioning while retaining the full three-dimensional (3D) volumetric representation of the tissue. Nevertheless, its greyscale nature limits its specificity compared to conventional histological stains and creates an additional barrier for pathologists, whose training is primarily based on colour-stained histology. In this work, we present a histology-guided enhancement platform that can integrate the 3D information provided by synchrotron radiation phase-contrast microCT (PCµCT) with the rich visual features characteristic of histological stains. We introduce a multistage PCµCT-histology co-registration method combined with a virtual staining deep neural network and demonstrate successful virtual histological staining of PCµCT human and mouse lung tissue that closely resembles standard histology. We evaluate our strategy on multiple histological stains and apply it to identify 3D collagen-based remodelling of pulmonary arteries in patients with pulmonary hypertension. Overall, we expect our work to facilitate the integration of PCµCT as a clinical tool for 3D analysis of biological tissues and support non-destructive 3D pathology for disease biomarker exploration.

(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
Animals, Humans, X-Ray Microtomography/methods, Lung/diagnostic imaging, Imaging, Three-Dimensional/methods, Deep Learning, Mice, Staining and Labeling
in
Journal of the Royal Society, Interface
volume
23
issue
239
article number
20251186
publisher
The Royal Society of Canada
external identifiers
  • pmid:42303257
ISSN
1742-5662
DOI
10.1098/rsif.2025.1186
language
English
LU publication?
yes
additional info
© 2026 The Authors.
id
d2284c4a-15f6-41b6-b7f9-e434d84f789d
date added to LUP
2026-06-17 09:10:05
date last changed
2026-06-17 09:16:29
@article{d2284c4a-15f6-41b6-b7f9-e434d84f789d,
  abstract     = {{<p>Histologically stained tissue sections are considered the gold standard for studying microscopic anatomy and diagnosing disease in clinical practice. However, the processes of sectioning and staining are laborious, and the overall method relies on two-dimensional analysis. In contrast, X-ray-based virtual histology offers the advantage of virtual sectioning while retaining the full three-dimensional (3D) volumetric representation of the tissue. Nevertheless, its greyscale nature limits its specificity compared to conventional histological stains and creates an additional barrier for pathologists, whose training is primarily based on colour-stained histology. In this work, we present a histology-guided enhancement platform that can integrate the 3D information provided by synchrotron radiation phase-contrast microCT (PCµCT) with the rich visual features characteristic of histological stains. We introduce a multistage PCµCT-histology co-registration method combined with a virtual staining deep neural network and demonstrate successful virtual histological staining of PCµCT human and mouse lung tissue that closely resembles standard histology. We evaluate our strategy on multiple histological stains and apply it to identify 3D collagen-based remodelling of pulmonary arteries in patients with pulmonary hypertension. Overall, we expect our work to facilitate the integration of PCµCT as a clinical tool for 3D analysis of biological tissues and support non-destructive 3D pathology for disease biomarker exploration.</p>}},
  author       = {{Almagro-Pérez, Cristina and Peruzzi, Niccolò and Galambos, Csaba and Song, Andrew H and Brunnström, Hans and Gawlik, Kinga I and Stampanoni, Marco and Tran-Lundmark, Karin and Lovric, Goran}},
  issn         = {{1742-5662}},
  keywords     = {{Animals; Humans; X-Ray Microtomography/methods; Lung/diagnostic imaging; Imaging, Three-Dimensional/methods; Deep Learning; Mice; Staining and Labeling}},
  language     = {{eng}},
  month        = {{06}},
  number       = {{239}},
  publisher    = {{The Royal Society of Canada}},
  series       = {{Journal of the Royal Society, Interface}},
  title        = {{Histology-guided 3D virtual staining of microCT-imaged lung tissue via deep learning}},
  url          = {{http://dx.doi.org/10.1098/rsif.2025.1186}},
  doi          = {{10.1098/rsif.2025.1186}},
  volume       = {{23}},
  year         = {{2026}},
}