Cancer Imaging Phenomics Toolkit (CaPTk)  1.0.1.20170207
Application Usage

EGFRvIII Surrogate Index

This evaluates the EGFRvIII status in individual primary glioblastoma patients, by quantitative pattern analysis of the spatial heterogeneity of peritumoral perfusion imaging dynamics from pre-operative Dynamic Susceptibility Contrast Magnetic Resonance Imaging (DSC-MRI) scans, through the Peritumoral Heterogeneity Index (PHI / φ-index) [1,2].

REQUIRED IMAGES:

  1. Post-contrast T1-weighted (T1-Gd): To annotate the immediate peritumoral region of interest (ROI)
  2. T2-weighted Fluid Attenuated Inversion Recovery (T2-FLAIR): To annotate the distant peritumoral ROI
  3. Dynamic susceptibility contrast-enhanced MRI (DSC-MRI): To perform the analysis

USAGE:

Glioblastoma Infiltration Index (Recurrence)

This presents the imaging signatures of deeply infiltrating tumor which largely agree with subsequent recurrence in de novo glioblastoma patients, via multi-parametric imaging pattern analysis that enhances the spatial heterogeneity of peritumoral edema. [3,4]

REQUIRED IMAGES:

  1. T1-weighted (T1)
  2. Post-contrast T1-weighted (T1-Gd)
  3. T2-weighted (T2)
  4. T2-weighted Fluid Attenuated Inversion Recovery (T2-FLAIR)
  5. Diffusion Tensor Imaging (DTI) images: AX, B0, FA, RAD, TR
  6. Dynamic susceptibility contrast-enhanced MRI (DSC-MRI)

USAGE:

NOTE: Currently, the user only has the option to train a new classifier based on their data and do testing on that trained classifier. In the future, a classifier trained on a large cohort will be provided.

Survival Prediction Index

This tool extracts and employs distinctive imaging biomarkers predictive of an individual patient’s survival to predict patients’ survival in de novo glioblastoma patients via multi-parametric MR imaging pattern analysis which might assist in personalized treatment. [5-7]

REQUIRED IMAGES:

  1. T1-weighted (T1)
  2. Post-contrast T1-weighted (T1-Gd)
  3. T2-weighted (T2)
  4. T2-weighted Fluid Attenuated Inversion Recovery (T2-FLAIR)
  5. Diffusion Tensor Imaging (DTI) images: AX, FA, RAD, TR
  6. Dynamic susceptibility contrast-enhanced MRI (DSC-MRI)
  7. Segmentations for the following tissues:
    • Ventricles (label 10)
    • Peritumoral Edema (label 100)
    • Non-Enhancing core of tumor (label 175)
    • Enhancing tumor (label 200)

USAGE:

Training process:

Each sub-folder must hold images with filenames that include the corresponding modality, such as t1, t1ce for T1 and T2 images and "labels" tag, in the name for segmentation, as in "AAAC0_t1ce_pp.nii"

Testing process:

NOTE: Currently, the user only has the option to train a new classifier based on their data and do testing on that trained classifier. In the future, a classifier trained on a large cohort will be provided.

Confetti

This is a method for automated extraction of white matter tracts of interest in a consistent and comparable manner over a large group of subjects without drawing the inclusion and exclusion regions of interest (ROI), facilitating an easy correspondence between different subjects, as well as providing a representation that is robust to edema, mass effect, and tract infiltration [8-10].

REQUIRED IMAGES:

  1. Parcellation of the brain into 87 Desikan/Freesurfer gray matter (GM) regions [11]
  2. 87 Track Density Images (TDI): For each region (among 87 GM regions), voxel-wise map of number of fibers connecting to the region
  3. Streamlines (fibers) to be clustered: Either in Trackvis (.trk) or Camino (.Bfloat) format.

USAGE:

WhiteStripe

This algorithm normalizes conventional magnetic resonance images [13] by detecting a latent subdistribution of normal tissue and linearly scaling the histogram of the images.

REQUIRED IMAGES:

  1. Inhomogeneity-corrected (N3 or N4) T1-weighted or T2-weighted images, ideally either skull-stripped or rigidly aligned to MNI space.

USAGE:

NOTE: WhiteStripe uses KernelFit library from Lentner.

Radiomics Analysis of Lung Cancer (SBRT Lung)

“Radiomics Analysis of Lung Cancer” calculates quantitative imaging measures including Intensity statistics, Gray Level Co-occurrence Matrix (GLCM), Gray Level Run-Length Matrix (GLRLM), Local Binary Patterns (LBPs), and shape features from PET/CT scans of lung cancer patients for predicting clinical outcomes, such as treatment response and patient survival using pattern recognition and machine learning techniques [14].

REQUIRED IMAGES:

  1. CT image
  2. PET image (coregistered to the CT image)

USAGE:

NOTE: SBRT uses a pre-trained model for estimation; in the future we will provide a mechanism to do training on own data.

References:

[1] S. Bakas, H. Akbari, J. Pisapia, M. Rozycki, D. M. O'Rourke, C. Davatzikos, "Identification of Imaging Signatures of the Epidermal Growth Factor Receptor Variant III (EGFRvIII) in Glioblastoma", Neuro-Oncology, 17(Suppl.5):V154, 2015, doi: 10.1093/neuonc/nov225.05

[2] S. Bakas, H. Akbari, J. Pisapia, M. Martinez-Lage, M. Rozycki, S. Rathore, N. Dahmane, D. M. O’Rourke, C. Davatzikos. "In vivo detection of EGFRvIII in glioblastoma via perfusion magnetic resonance imaging signature consistent with deep peritumoral infiltration: the φ-index", Clinical Cancer Research, 2017 (Under Review)

[3] Akbari H, Macyszyn L, Da X, Bilello M, Wolf RL, Martinez-Lage M, Biros G, Alonso-Basanta M, O'Rourke DM, Davatzikos C. Imaging Surrogates of Infiltration Obtained Via Multiparametric Imaging Pattern Analysis Predict Subsequent Location of Recurrence of Glioblastoma. Neurosurgery. 2016 Apr 1; 78(4):572-80.

[4] Akbari H, Macyszyn L, Da X, Wolf RL, Bilello M, Verma R, O’Rourke DM, Davatzikos C. Pattern analysis of dynamic susceptibility contrast-enhanced MR imaging demonstrates peritumoral tissue heterogeneity. Radiology. 2014 Jun 19;273(2):502-10.

[5] Macyszyn L, Akbari H, Pisapia JM, Da X, Attiah M, Pigrish V, Bi Y, Pal S, Davuluri RV, Roccograndi L, Dahmane N. Imaging patterns predict patient survival and molecular subtype in glioblastoma via machine learning techniques. Neuro-oncology. 2016 Mar 1;18(3):417-25.

[6] Akbari H, Macyszyn L, Da X, Wolf RL, Bilello M, Verma R, O’Rourke DM, Davatzikos C. Pattern analysis of dynamic susceptibility contrast-enhanced MR imaging demonstrates peritumoral tissue heterogeneity. Radiology. 2014 Jun 19;273(2):502-10.

[7] H. Akbari, L. Macyszyn, J. Pisapia, X. Da, M. Attiah, Y. Bi, S. Pal, R. Davuluri, L. Roccograndi, N. Dahmane, R. Wolf, M. Bilello, D. O’ Rourke, C. Davatzikos, Survival Prediction in Glioblastoma Patients Using Multi-parametric MRI Biomarkers and Machine Learning Methods, ASNR 2015; American Society of Neuroradiology; O-525, pp. 2042-2044 (http://www.asnr.org/sites/default/files/proceedings/2015_Proceedings.pdf)

[8] B. Tunç, M. Ingalhalikar, D. Parker, J. Lecoeur, R. L. Wolf, L. Macyszyn, S. Brem, R. Verma, Individualized Map of White Matter Pathways: Connectivity-based Paradigm for Neurosurgical Planning, Neurosurgery, Vol. 79 (4), pp. 568-77, 2016.

[9] B. Tunç, W. A. Parker, M. Ingalhalikar, R. Verma, Automated tract extraction via atlas based Adaptive Clustering, NeuroImage, Vol. 102 (2), pp. 596-607, 2014.

[10] B. Tunç, A. R. Smith, D. Wasserman, X. Pennec, W. M. Wells, R. Verma, K. M. Pohl, Multinomial Probabilistic Fiber Representation for Connectivity Driven Clustering, Information Processing in Medical Imaging (IPMI), 2013.

[11] Fischl B, Sereno MI, Dale AM: Cortical surface-based analysis. II: Inflation, flattening, and a surface-based coordinate system. Neuroimage 9:195–207, 1999.

[12] Desikan RS, Segonne F, Fischl B, Quinn B, Dickerson B, Blacker D, Buckner R, Dale A, Maguire R, Hyman B, Albert M, Killiany R: An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 31, 2006.

[13] R.T. Shinohara, E.M. Sweeney, J. Goldsmith, et al. Statistical normalization techniques for magnetic resonance imaging. Neuroimage Clinical, 2014

[14] H. Li, M. Galperin-Aizenberg, D. Pryma, C. Simone, Y. Fan, “Predicting treatment response and survival of early-stage non-small cell lung cancer patients treated with stereotactic body radiation therapy using unsupervised two-way clustering of radiomic features”, The 2017 International Workshop on Pulmonary Imaging (Submitted)