Automatic and label-free detection of meningioma in dura mater using the combination of multiphoton microscopy and image analysis

Abstract

Aim: Currently, the removal of meningiomas from the dura mater depends solely on cautious visual identification of lesions by a neurosurgeon. Inspired by the requirements for resection, we propose multiphoton microscopy (MPM) based on twophoton-excited fluorescence and second-harmonic generation as a histopathological diagnostic paradigm to assist neurosurgeons in achieving precise and complete resection.Approach: Seven fresh normal human dura mater samples and 10 meningiomainfiltrated dura mater samples, collected from 10 patients with meningioma, were acquired for this study. First, multi-channel mode and lambda mode detection were utilized in the MPM to characterize the architectural and spectral features of normal and meningioma-infiltrated dura mater, respectively. Three imaging algorithms were then employed to quantify the architectural differences between the normal and meningioma-infiltrated dura mater through calculations of the collagen content, orientation, and alignment. Finally, MPM was combined with another customdeveloped imaging algorithm to locate the meningioma within the dura mater and further delineate the tumor boundary.Results: MPM not only detected meningioma cells in the dura mater but also revealed the morphological and spectral differences between normal and meningioma-infiltrated dura mater, providing quantitative information. Furthermore, combined with a self-developed image-processing algorithm, the precise borders of meningiomas in the dura mater could be accurately delineated.Conclusions: MPM can automatically detect meningiomas in the dura mater labelfree. With the development of advanced multiphoton endoscopy, MPM combined with image analysis can provide decision-making support for histopathological diagnosis, as well as offer neurosurgeons more precise intraoperative resection guidance for meningiomas.

Resolution-enhanced multi-core fiber imaging learned on a digital twin for cancer diagnosis

“…The procedure is time-consuming, leading to treatment delay, and has no visual feedback during the surgery, which brings additional risk and complications. 7 – 9 Label-free imaging techniques like autofluorescence 4 – 6 , 10 – 12 and Raman spectroscopy 12 – 16 enable locating target tissue in situ for in vivo tumor diagnosis, 17 – 19 where high spatial resolution plays a critical role. Multi-core fibers (MCFs) are often used in endoscopy since they are flexible and ultra-thin ( ) and provide an efficient way to illuminate and detect in real-time, 20 – 24 which allows minimal invasive access directly to deep tissue for intraoperative imaging.…”