RESUMEN
Meningiomas rank among the most common intracranial tumors, and surgery stands as the primary treatment modality for meningiomas. The precise subtyping and diagnosis of meningiomas, both before and during surgery, play a pivotal role in enabling neurosurgeons choose the optimal surgical program. In this study, we utilized multiphoton microscopy (MPM) based on 2-photon excited fluorescence and second-harmonic generation to identify 5 common meningioma subtypes. The morphological features of these subtypes were depicted using the MPM multichannel mode. Additionally, we developed 2 distinct programs to quantify collagen content and blood vessel density. Furthermore, the lambda mode of the MPM characterized architectural and spectral features, from which 3 quantitative indicators were extracted. Moreover, we employed machine learning to differentiate meningioma subtypes automatically, achieving high classification accuracy. These findings demonstrate the potential of MPM as a noninvasive diagnostic tool for meningioma subtyping and diagnosis, offering improved accuracy and resolution compared with traditional methods.