Table 2. Recursive partitioning analysis (VI = vascularity index, Gr2–3 = WHO Grade 2–3).
| Variables | % Gr 2–3 with variable Vs. without variable | P(Chi-square or Fisher) |
|---|---|---|
| Prior CVA: yes versus no | 73% (8/11) vs 26% (18/68) | p = 0.005 |
| A. Prior CVA | 73% (8/11) | Subsets: 0.055 *< p <0.52 |
| B. No prior CVA | 26%(18/68) | |
| C. Vascularity = 4 vs 1–3 | 54% (7/13) vs 20% (11/54) | 0.032 |
| D. i. Vascularity 4 | 54% (7/13) | Subsets: 0.10 *< p <0.61 |
| ii. Vascularity ≤3 | 20% (11/54) | |
| E. Post Menopausal:Y/N | 10% (3/30) vs 35% (7/20) | p = 0.03 |
| i. VI ≤3 Post Menopausal | 10% (3/30) | |
| ii. VI ≤3 Other | 35% (7/20) | |
| Four Group Model | ||
| 1. Prior CVA | 73% (8/11) | |
| 2. No CVA, VI = 4 | 54% (7/13) | |
| 3. No CVA, VI ≤3, not PostMenopausal | 35% (7/20) | |
| 4. No CVA, VI ≤3, and PostMenopausal | 10% (3/30) |
*
Subset analysis: closest variable for CVA subset was Site: skull base 0/2 vs other 8/9 p = 0.055, closest subsets for no prior CVA and VI-4 were brain invasion by imaging and presence of other meningiomas, both p = 0.103.