dcemriS4 (version 0.51)

dcemri.spline: Bayesian P-Splines for Dynamic Contrast-Enhanced MRI Data

Description

Quantitative analysis of DCE-MRI typically involves the convolution of an arterial input function (AIF) with a nonlinear pharmacokinetic model of the contrast agent concentration. This function takes a semi-parametric penalized spline smoothing approach, with which the AIF is convolved with a set of B-splines to produce a design matrix using locally adaptive smoothing parameters based on Bayesian penalized spline models (P-splines).

Usage

## S3 method for class 'array':
dcemri.spline(conc, time, img.mask, time.input=time,
              model="weinmann", aif="tofts.kermode", user=NULL,
              aif.observed=NULL, nriters=500, thin=5,
              burnin=100, ab.hyper=c(1e-5,1e-5),
              ab.tauepsilon=c(1,1/1000), k=4, p=25, rw=2,
              knots=NULL, nlr=FALSE, t0.compute=FALSE,
              samples=FALSE, multicore=FALSE, verbose=FALSE, 
              response=FALSE, fitted=FALSE, 
              ...) 
dcemri.spline.single(conc, time, D, time.input, p, rw, knots, k,
                     A, t0.compute=FALSE, nlr=FALSE, nriters=500,
                     thin=5, burnin=100, ab.hyper=c(1e-5,1e-5),
                     ab.tauepsilon=c(1,1/1000), silent=0,
                     multicore=FALSE, model=NULL,
                     model.func=NULL, model.guess=NULL,
                     samples=FALSE, B=NULL)

Arguments

conc

Matrix or array of concentration time series (last dimension must be time).

time

Time in minutes.

img.mask

Mask matrix or array. Voxels with mask = 0 will be excluded.

time.input

Time in minutes for observed arterial input function (default = time).

aif

is a character string that identifies the parameters of the arterial input function. Acceptable values are: tofts.kermode, fritz.hansen or observed. If observed you must provide the obse

user

...

aif.observed

is the user-defined vector of arterial concentrations observed at time.input (only for aif=observed).

silent

...

multicore

(logical) use the multicore package.

verbose

(logical) allows text-based feedback during execution of the function (default = FALSE).

samples

If TRUE output includes samples drawn from the posterior distribution for all parameters.

model.func

...

model.guess

...

nlr

If TRUE, a response model is fitted to the estimated response function.

model

Only if nlr = TRUE Response model fitted to the estimated response function. Acceptable values include: "AATH" or "weinmann" (default).

ab.hyper

Hyper priors for adaptive smoothness parameter

ab.tauepsilon

Hyper-prior parameters for observation error Gamma prior.

p

Number of knots of B-Spline basis.

t0.compute

If TRUE, the onset time will be estimated from response function.

k

Order of B-Splines.

knots

Vector of knots. Use this if you need unequally spaced knots.

rw

Order of random walk prior. Acceptable values are 1 and 2.

nriters

Total number of iterations.

thin

Thining factor.

burnin

Number of iterations for burn-in.

response

If TRUE, the response functions per voxel are returned.

fitted

If TRUE, then fitted time curved per voxel are returned.

A

...

B

...

D

...

...

...

Value

• The maximum of the response function Fp for the masked region is provided by default. Where appropriate, response functions, fitted functions, and parameter estimates (along with their standard errors) are provided. All multi-dimensional arrays are provided in nifti format.

Details

See Schmid et al. (2009) for more details.

References

Schmid, V., Whitcher, B., Padhani, A.R. and G.-Z. Yang (2009) A semi-parametric technique for the quantitative analysis of dynamic contrast-enhanced MR images based on Bayesian P-splines, IEEE Transactions on Medical Imaging, 28 (6), 789-798.

See Also

dcemri.bayes, dcemri.lm, dcemri.map

Examples

data("buckley")
xi <- seq(5, 300, by=5)
img <- array(t(breast$data)[,xi], c(13,1,1,60))
mask <- array(TRUE, dim(img)[1:3])
time <- buckley$time.min[xi]

## Generate AIF params using the orton.exp function from Buckley's AIF
aif <- buckley$input[xi]

fit.spline <- dcemri.spline(img, time, mask, aif="fritz.hansen",
                            nriters=250, nlr=TRUE)
fit.spline.aif <- dcemri.spline(img, time, mask, aif="observed",
                                aif.observed=aif, nriters=250,
                                nlr=TRUE)

plot(breast$ktrans, fit.spline$ktrans, xlim=c(0,1), ylim=c(0,1),
     xlab=expression(paste("True ", K^{trans})),
     ylab=expression(paste("Estimated ", K^{trans})))
points(breast$ktrans, fit.spline.aif$ktrans, pch=2)
abline(0, 1, lwd=1.5, col="red")
legend("right", c("fritz.hansen", "observed"), pch=1:2)

cbind(breast$ktrans, fit.spline$ktrans[,,1], fit.spline.aif$ktrans[,,1])