Introduction: Meningiomas are the most common primary central nervous system tumors, for which surgical resection is the first line treatment option. Preoperative knowledge of the WHO grade and consistency can greatly influence the therapeutic plan, which nowadays can only be determined by histological examination. The complex geometry of tumors can be characterized by their shape using various metrics. Fractal geometry plays an important role in the characterization of irregular, rough shapes.
Methods: We performed a retrospective clinical study of patients who underwent surgery for convexity and falco-tentorial meningiomas. For all patients, the presence of T1, ceT1, T2, FLAIR sequences of the MRI series were required. Tumors were segmented via ITKSNAP software, then fractal analysis (employing the sliding-window method to determine fractal dimension and lacunarity index), t-test, logistic regression and ROC analysis were performed.
Results: Forty-eight patients (34 females and 14 males) met the selection criteria. 72.9% had convexity and 27.1 parasagittal meningiomas, 56.2% were WHO grade1, 43.8% WHO grade 2. Fractal dimension seemed to be suitable for estimating WHO grade (p=0.009), lacunarity index to separate the consistency property (p=0.025). For the prediction of histopathology, our formula had an AUC of 0.841, and we could predict the consistency with an AUC of 0.763.
Conclusions: Based on the results of this study, fractal dimension and lacunarity index seem to be suitable for estimating the WHO grade and consistency of meningiomas.