Longitudinal analysis
| Change marker prediction | ||||||||||
| Dementia conversion (n=15) | Decline in Global | Decline in PS | ||||||||
| Cohort | β (SE) | P value | HR (95% CI) | AUC | β (SE) | P value | Adj. R | β (SE) | P value | Adj. R |
| Serum NfL (pg/mL) (log 10) | 0.178 (0.301) | 0.554 | 1.195 (0.663 to 2.155) | 0.705 | −0.051 (0.104) | 0.625 | 0.070 | 0.035 (0.108) | 0.751 | −0.009 |
| MDPH (mm2/s) | −0.813 (0.297) | 0.006 | 0.444 (0.248 to 0.794) | 0.788 | 0.260 (0.102) | 0.012 | 0.134 | 0.049 (0.110) | 0.655 | −0.008 |
| PSMD (mm2/s) | 0.594 (0.293) | 0.043 | 1.812 (1.020 to 3.219) | 0.761 | −0.362 (0.100) | 0.001 | 0.199 | −0.230 (0.108) | 0.036 | 0.040 |
Change in DTI but not in NfL predicts cognitive decline and dementia conversion.
Values show standardised regression coefficients: β and SEs for predictor variables in regression models of dementia conversion and decline in cognition.Significant at p value <0.05.
Adj R2, adjusted explained variance; AUC, area under the curve; DTI, diffusion tensor imaging; Global, global cognition; MDPH, mean diffusivity normalised peak height; NfL, neurofilament light chain; PS, processing speed; PSMD, peak width of skeletonised mean diffusivity.