Non model based analysis of DCE MRI

Temporal changes in MR signal intensity after contrast injection produce a dynamic curve S(t) at each voxel. The relative signal enhancement (RSE) at time t after contrast uptake is computed by normalizing with respect to signal baseline (S(0)):

   RSE = (S(t) - S(0))/ S(0)

S(0) is computed by averaging signal intensities prior to the start of contrast uptake.

Semi-quantitative DCE-MR imaging metrics

CERR provides tools to compute semi-quantitative DCE-MR imaging metrics from the relative signal enhancement (RSE) curve. The following parameters are supported:

Parameter	Description
Peak enhancement (PE)	Peak relative signal enhancement (RSE)
Time-to-peak (TTP)	Time at which PE occurs, measured from the start of uptake.
Time to half-peak (TTHP)	Time at which half PE occurs, measured from start of uptake.
Signal at half-peak (SHP)	Signal at TTHP.
Wash-in slope (WIS)	PE divided by TTP .
Wash-out slope (WOS)	Difference between PE and RSE at the last time point (Tend), divided by difference between Tend and TTP. Set to zero if the PE ooccurs at Tend.
Initial gradient (IG)	Gradient estimated through linear regression of all RSEs between 20% and 80% PE.
Wash-out gradient (WOG)	Gradient calculated by linear regression of RSE between PE and 1 min after PE
Signal enhancement ratio (SER)	RSE at 0.5 min divided by RSE at 2.5 min from start of uptake.
Integrated area under the curve (IAUCt):	Area under the RSE curve until a stipulated time point "t" from the start of uptake. CERR returns IAUC at t=30s, 60s, 90s, 120s, 150s, 180s, TTP & TTHP. "NaN" values are returned where 't' exceeds the available time points.

Maps of the above parameters are generated for analysis/visualization & summary statistics are returned.

Usage

Requirements

The CERR archive must contain only the relevant DCE-MR images to be analyzed, along with segmentations. Scans acquired at the different time points should be merged into a single CERR file.

Code

  
global planC;
strNum = 1;          % Structure of interest.
shift  = [];         % No. time points to start of uptake. Leave empty for interactive selection.
createMapsFlag = 1;  % Set flag to generate parameter maps for visualization.
                     % Set to 0 to turn off. 

%Example-1 (using default settings)

% Compute parameters
[planC,statT] = getSemiQuantMetricsForDCEMR(planC,strNum,shift,...
    createMapsFlag);

%Example-2 (with optional inputs)
  
tempSmoothFlag = 1;       %Flag for turning on (1) or off (0) temporal smoothing of the time–signal kinetic curve
                          % Note: Default = 0 if not specified.
resampFlag = 1;           %Flag for turning on (1) or off (0) interpolation of the signal at a uniform sample rate of 1/0.01s
                          % Note: Default = 0 if not specified.
fSize = 5;
fSigma = 2;
filtV =  [fSize,fSigma];  %Parameters (filter size & sigma) for spatial smoothing of images prior to analysis
                          %Note: Default: no smoothing if filtV is not specified.
  
%Compute parameters        
[planC, statT] = getSemiQuantMetricsForDCEMR(planC,strNum,shift,...
    createMapsFlag,tempSmoothFlag,resampFlag,filtV);

Running the analyis

1. Run getSemiQuantMetricsForDCEMR as in the examples above.
2. A pop-window is displayed, showing a plot of the average signal intensity within the selected ROI vs. time points in the acquisition. Enter the no. of time points to skip until the start of the uptake curve when prompted.

3. Maps of the computed parameters are stored in planC as pseudo-scans for display using the CERR Viewer. Table statT contains summary statistics for each parameter over the entire ROI.