La idea general es calcular la AD signature a partir de la segmentacion de FS. Primero ha de recuperarse el archivo de aparc desde XNAT haciendo algo como,

$ xnat_pullfs.pl -s aparc -d -p PROJECT -o freesurfer_aparc_data.csv

y de aqui usaremos los valores de cortical thickness y volumen de,

• enthorinal cortex
• inferior temporal cortex
• middle temporal cortex
• fusiform cortex

para calcular el ADCS.

Normalmente tambien necesitariamos el volumen de hipocampo y ventricular para caracterizar el AD. Eesto lo podemos sacar del archivo aseg de FS,

$ xnat_pullfs.pl -s aseg -d -p PROJECT -o freesurfer_aseg_data.csv

Pero han de normalizarse por el ICV, preferiblemente con el metodo de los residuos.

library("dplyr")
ifile_aparc = "freesurfer_aparc_data.csv"
ifile_aseg = "freesurfer_aseg_data.csv"
ofile = "resume_data.csv")
x <- read.csv(ifile_aparc)
z <- read.csv(ifile_aseg)
x$c_thick <- x$lh.entorhinal.ThickAvg*x$lh.entorhinal.SurfArea +
x$rh.entorhinal.ThickAvg*x$rh.entorhinal.SurfArea +
	x$lh.inferiortemporal.ThickAvg*x$lh.inferiortemporal.SurfArea +
	x$rh.inferiortemporal.ThickAvg*x$rh.inferiortemporal.SurfArea +
	x$lh.middletemporal.ThickAvg*x$lh.middletemporal.SurfArea +
	x$rh.middletemporal.ThickAvg*x$rh.middletemporal.SurfArea +
	x$lh.fusiform.ThickAvg*x$lh.fusiform.SurfArea +
	x$rh.fusiform.ThickAvg*x$rh.fusiform.SurfArea
 
x$c_sa <- x$lh.entorhinal.SurfArea +
	x$rh.entorhinal.SurfArea +
	x$lh.inferiortemporal.SurfArea +
	x$rh.inferiortemporal.SurfArea +
	x$lh.middletemporal.SurfArea +
	x$rh.middletemporal.SurfArea +
	x$lh.fusiform.SurfArea +
	x$rh.fusiform.SurfArea
 
x$entorhinal_Thickness <- (x$lh.entorhinal.ThickAvg*x$lh.entorhinal.SurfArea +
	x$rh.entorhinal.ThickAvg*x$rh.entorhinal.SurfArea)/(x$lh.entorhinal.SurfArea +  x$rh.entorhinal.SurfArea)
 
x$ADCS <- x$c_thick/x$c_sa
 
a <- lm(x$Cortex_Volume ~ x$eTIV)
b = a$coefficients[[2]]
x$adj_Cortex_Volume <- x$Cortex_Volume - b*(x$eTIV - mean(x$eTIV, na.rm=T))
 
z$hv <- z$Left.Hippocampus + z$Right.Hippocampus
a <- lm(z$hv ~ z$eTIV)
b = a$coefficients[[2]]
z$adjHV <- z$hv - b*(z$eTIV - mean(z$eTIV, na.rm=T))
 
z$vent <- z$X3rd.Ventricle + z$X4th.Ventricle +
	z$Left.Inf.Lat.Vent + z$Left.Lateral.Ventricle +
	z$Right.Inf.Lat.Vent + z$Right.Lateral.Ventricle
a <- lm(z$vent ~ z$eTIV)
b = a$coefficients[[2]]
z$adjVentricles <- z$vent -b*(z$eTIV - mean(z$eTIV, na.rm=T))
 
merge(x, z, by=c("Subject_ID")) -> y
 
x2w <- y[c('Subject_ID','Date.x','entorhinal_Thickness','ADCS','adjHV','adjVentricles','Cortex_Thickness','adj_Cortex_Volume', 'eTIV.x')]
 
x2wr <- rename(x2w, Date = Date.x, adj_Hippocampal_Volume = adjHV, adj_Ventricles_Volume = adjVentricles, eTIV = eTIV.x)
write.csv(x2wr, ofile, row.names=F, quote=F)

All MRI T1w images underwent a Freesurfer 7.2 reconstruction (https://surfer.nmr.mgh.harvard.edu/fswiki/FreeSurferMethodsCitation).

Hippocampal volume was calculated as the sum of right and left hippocampus volume and adjusted by the estimated intracranial volume using the residual method [1]. Ventricle volume was calculated as the sum of the reconstructed ventricles and adjusted by the estimated intracranial volume using the residual method. Total cortex volume was also adjusted by estimated intracranial volume and the same method.

The Alzheimer’s disease cortical signature (ADCS) was expressed by a meta-ROI composed of the of the following areas: entorhinal, inferior temporal, middle temporal, and fusiform. The average thickness of the ADCS ROI was calculated as the mean thickness across these regions weighted by their surface area [2–4]. Entorhinal thickness was also calculated as the mean values between left and right weighted by the surface area. Cortical thickness was just picked up from FS metrics as is.

The possibility of atrophy was studied then as the change of these measurements, per unit of time, across visits 0 and 5.

1. Mathalon DH, Sullivan E v, Rawles JM, Pfefferbaum A. Correction for head size in brain-imaging measurements. Psychiatry Res. 1993;50:121–39.

2. Milà-Alomà M, Brinkmalm A, Ashton NJ, Kvartsberg H, Shekari M, Operto G, et al. CSF Synaptic Biomarkers in the Preclinical Stage of Alzheimer Disease and Their Association With MRI and PET. Neurology. 2021;97:e2065.

3. Jack CR, Wiste HJ, Weigand SD, Therneau TM, Knopman DS, Lowe V, et al. Age-specific and sex-specific prevalence of cerebral β-amyloidosis, tauopathy, and neurodegeneration in cognitively unimpaired individuals aged 50–95 years: a cross-sectional study. The Lancet Neurology. 2017;16:435–44.

4. Jack CR, Wiste HJ, Weigand SD, Therneau TM, Lowe VJ, Knopman DS, et al. Defining imaging biomarker cut points for brain aging and Alzheimer’s disease. Alzheimers Dement. 2017;13:205–16.