Radiomics models - cerr/CERR GitHub Wiki
Using radiomics models in CERR
SITE |
Model |
Features |
Reference |
Ovarian |
Survival for high grade Ovarain cancer |
CluDiss (inter-tumor heterogeneity |
Veeraraghavan et al, bioRxiv 531046 (2019) |
Breast, Brain |
Radiation necrosis in T1-w MRI Brain, Subtype classification in DCE-MRI Breast |
ColLAGe (gradient orientations) |
Prasanna et al, Scientific Reports volume 6, article no. 37241 (2016) |
Head & Neck |
H&N Hypoxia |
P90 of FDG PET SUV, LRHGLE of CT |
Crispin-Ortuzar M et al, Radiother Oncol., 27(1):36-42 (2018) |
Lung, H&N |
Survival for NSCLC and HNSCC |
Statistics Energy, Shape Compactness, Grey Level Nonuniformity, wavelet ‘Grey Level Nonuniformity HLH’ |
Aerts et al, Nature Communications volume 5, article no. 4006 (2014) |
% Example 1: Aerts et al (Nature Communications 2014;5:4006. doi:10.1038/ncomms5006)
global planC;
structNum = 1; % structure number coresponding to the tumor
featureS = getNcomms5006Feature(structNum,planC)
% Example 2: Crispin Ortuzar et al (Radiotherapy and Oncology, Volume 127, Issue 1, 2018, Pages 36-42. https://doi.org/10.1016/j.radonc.2017.11.025)
global planC
ctStructNumV = 15;
petStructNumV = 14;
featureS = getHypoxiaCrispinOrtuzarFeatures(ctStructNumV,petStructNumV,planC);
P90_norm = (featureS(1).P90-11.5446)/5.1479;
LRHGLE_norm = (featureS(1).lrhgle - 7254.5)/1122.1;
TBRmax = 1.9061 + 0.32381*P90_norm+0.13032*LRHGLE_norm;
% Example 3: Prasanna et al (Scientific Reports, volume 6, Article number: 37241 (2016))
scanNum = 1;
structNum = 1;
domDirPatchRadV = [3 3 0];
haralTextPatchRadV = [3 3 0];
numGrayLevels = 64;
dim2d3dFlag = '2d';
hWait = NaN;
coLlAGe3M = getCollageFeature(scanNum, structNum, domDirPatchRadV,...
haralTextPatchRadV, numGrayLevels, dim2d3dFlag, hWait, planC);
% Example 4: Veeraraghavan et al (bioRxiv 531046; doi: https://doi.org/10.1101/531046)
global planC
structNumV = [1 2 3]; % tumor indices
featureS = getInterTumorHeteroFeature(structNumV,planC)