Advanced Imaging and Biomarker Discovery for Early Diagnosis in Skull Base Pathology

Early detection of skull base (SB) tumors and inflammatory lesions remains difficult because of deep location, complex anatomy, and subtle clinical prodromes. This study systematically evaluates contemporary advanced imaging techniques and emerging molecular biomarkers to define an evidence-based framework for earlier SB disease detection.

A Preferred Reporting Items for Systematic Reviews and Meta-Analyses-conformant search of PubMed, Embase, Scopus, and Cochrane Library (January 2015-June 2025) identified original human studies that reported diagnostic performance of (1) advanced magnetic resonance imaging (MRI), (2) positron-emission tomography (PET) or optical imaging, (3) radiomics/radiogenomics, or (4) fluid-based biomarkers in SB pathology. Methodological quality was rated with QUADAS-2. Data were extracted on sensitivity, specificity, and incremental diagnostic yield versus conventional computed tomography/MRI; meta-analysis was precluded by heterogeneity, so results were synthesized descriptively.

Seventy-four studies (4,934 patients) met inclusion criteria. Readout-segmented echo-planar diffusion-weighted imaging (DWI) increased lesion detection at the anterior cranial fossa by 17% over single-shot DWI. Diffusion tensor tractography improved cranial nerve visualization and altered operative corridors in 29% of cases. Susceptibility-, perfusion-, and spectroscopy-based MRI differentiated benign from malignant SB lesions with pooled sensitivity 85 to 92%. PET radiotracers outperformed 18F-FDG in low-contrast SB regions; 11C-methionine and 68Ga-FAPI achieved sensitivities of 88 to 93%, whereas 2-18F-tyrosine PET correctly mapped residual meningioma beyond Gd-MRI in 38% of postradiation cases. Stimulated Raman histology coupled with contrastive AI yielded 96.6% intraoperative accuracy for SB tumor classification. Radiomic signatures on multiparametric MRI distinguished chordoma from chondrosarcoma with area under the curve 0.93 and predicted postoperative progression in primary SB chordoma (C-index 0.80). Multiomics analyses identified recurrent genomic drivers (PBRM1 mutation, chromosome-1q gain) and metabolic markers (asparagine synthetase overexpression) associated with subclinical SB tumor presence. Circulating tumor DNA, H3K27M, and TBXT mutations, and exosomal miRNA panels were detectable in cerebrospinal fluid months before radiographic progression.

Advanced MRI (particularly readout-segmented echo-planar diffusion-weighted imaging and tractography), amino acid and fibroblast-activation PET, AI-augmented optical imaging, and liquid biopsy assays collectively enhance early SB lesion detection well beyond conventional imaging. Integrating radiomics with fluid biomarkers and targeted PET could enable a risk-adapted surveillance algorithm and expedite minimally invasive intervention. Prospective multimodal studies with standardized reference standards are now warranted.