struct.quant: Structural quantitation summaries of tumor uptake. Arguments:...
In ericwol/mia: Medical Imaging Analysis Tools

Structural quantitation summaries of tumor uptake. Arguments: out = output of hetE() zdim = slice thickness (defaults to 1) xdim = voxel transverse dimensions (defaults to 1) Ex: zdim=xx$Thickness[kk]; xdim=xx$InPlaneDim[kk] bndT = cutoff quantile threshold defining the start of the boundary rim (defaults to 75th quantile of VOI radial coordinates) re.sign = if TRUE, flips gradient curve sign for better interpretation (a negative output gradient value would then indicate a decreasing rate) sm.reg = if TRUE, gradients are computed from a smoothed unimodal regression fit ddf = corresponding number of smooth.spline df's (between 5-10 usually works well) Values: u: radial coordinates (sorted in increasing order) yh.i: isotonic fit for u values yh.b: bitonic fit for u values fractional.void = F1 volume = F2 contrast.ratio = F3 max = F4 boundary.gradient = F5 cv = F6 y = model uptake (where y = out$v[order(v[,1],3]) gradients = all gradient values evaluated at locations u (expressed in the units of the input uptake values) (where u is a normalized version of sort(out$v[,1])) normalized.gradients = gradients/F4 (unitless, normalized to (-1,1)) w.gradients = uptake-weighted normalized gradients, bnd.gradients = gradients at boundary normalized.bnd.gradients = normalized gradients at boundary w.bnd.gradients = uptake-weighted (normalized) gradients at boundary