Rapid visualization of PD-L1 expression level in glioblastoma immune microenvironment via machine learning cascade-based Raman histopathology

•This study achieved label-free visualization of PD-L1 expression at tissue level for the first time.•MLC-Raman histopathology achieved an accuracy of 0.99 in classifying distinct PD-L1 expression cells in GBM IME.•MLC-Raman histopathology could identify glioma, CD8+ T cells, macrophages and normal cells.•The imaging results exhibited strong concordance with MxIF and traditional pathologists' scoring.•The entire process from signal collection to visualization could be completed within 30 minutes.