Article
Intraoperative microscopic autofluorescence characterisation in brain tumours based on combined Stimulated Raman Histology (SRH) and Two-Photon fluorescence
Mikroskopische Charakterisierung von Autofluoreszenz in Hirntumoren mithilfe von intraoperativer stimulierter Raman Histologie und Zwei-Photonen Fluoreszenz
Search Medline for
Authors
-
Gina Fürtjes
- Universitätsklinikum Köln, Allgemeine Neurochirurgie, Köln, Deutschland; Helmholtz Zentrum München, Helmholtz Pioneer Campus, München, Deutschland
-
David Reinecke
- Universitätsklinikum Köln, Allgemeine Neurochirurgie, Köln, Deutschland
-
Niklas von Spreckelsen
- Universitätsklinikum Köln, Allgemeine Neurochirurgie, Köln, Deutschland
-
Christian Freudiger
- Invenio Imaging Inc., Santa Clara, CA, Vereinigte Staaten
-
Adrian Ion-Margineanu
- Invenio Imaging Inc., Santa Clara, CA, Vereinigte Staaten
-
Florian Khalid
- Invenio Imaging Inc., Santa Clara, CA, Vereinigte Staaten
-
Christian Mawrin
- Universitätsklinikum Magdeburg, Institut für Neuropathologie, Magdeburg, Deutschland
-
Andriy Chmyrov
- Helmholtz Zentrum München, Helmholtz Pioneer Campus, München, Deutschland
-
Oliver Bruns
- Helmholtz Zentrum München, Helmholtz Pioneer Campus, München, Deutschland
-
Roland Goldbrunner
- Universitätsklinikum Köln, Allgemeine Neurochirurgie, Köln, Deutschland
-
Volker Neuschmelting
- Universitätsklinikum Köln, Allgemeine Neurochirurgie, Köln, Deutschland
| Published: | May 25, 2022 |
|---|
Outline
Text
Objective: Biological tissues show a varying extent of autofluorescence lowering the signal-to-noise-ratio and complicating the detection of fluorophores in fluorescent guided surgery. Yet, data on the characterization of autofluorescence in brain tissue and tumors is scarce. This study aims to assess autofluorescence of the brain and its tumors by using stimulated Raman histology (SRH) combined with 2-photon fluorescence, a new label-free microscopy technique that enables virtual hematoxylin-and-eosin-stained-like images with corresponding fluorescence images out of unprocessed tissues.
Methods: Small unlabeled brain and tumor samples (1-3 mm3) were squashed onto a slide, SRH and fluorescence images were acquired on random areas of 2x2 mm under the same conditions by a dual wavelength laser (790 nm and 1020 nm) for excitation and detecting emission in 570-613 nm and 603-678 nm, respectively. The mean fluorescence intensity per image was evaluated by using ROIs based on heatmaps created by a residual convolutional neural network (CNN) that reliably differentiates between tumor, non-tumor brain tissue and low quality SRH image. Semiquantitative analysis focused on the fluorescence pattern of the tissues in relation to the clinical and pathological data.
Results: We analyzed SRH and corresponding autofluorescence images of 397 samples deriving from 82 consecutive patients that underwent brain tumor surgery in 07-08/2021 at our center. In healthy brain tissue, we found an increased mean autofluorescence signal in the white (12.13 +/-3.42, n=30) compared to the gray matter (4.62 +/-2.51, n=10, p<0.05) and in the cerebrum (11.83 +/-3.29, n=33) versus the cerebellum (2.82 +/-0.93, n=7, p<0.05), respectively. The signal of carcinoma metastases, meningiomas, gliomas and pituitary adenomas were significantly lower (each p<0.05) compared to the autofluorescence in the cerebrum and dura, whereas they showed a significantly higher signal (except of meningiomas) (each p<0.05) compared to the cerebellum. Melanomas were found to have a significantly higher fluorescent signal (p<0.05) compared to both, cerebrum and cerebellum.
Conclusion: Our data indicate that autofluorescence in the brain strongly depends on localization and differs significantly among various brain tumors and needs to be taken into account for fluorescent labeled guided surgery. Further studies are needed whether the resulting contrast may pave the way to enable label-free autofluorescence guided surgery.
Figure 1 [Fig. 1]
