ABSTRACT and FLAIR, and within each compartment:

ABSTRACT BACKGROUND ANDPURPOSE: Activation of the AKT pathway has a significant role in cellularproliferation, migration, and apoptosis, and is associated with poor prognosisin GBM. Biopsy and genetic sequencing are used to diagnose this mutation, whichis not a routine procedure in the majority of hospitals and clinics. In thiswork, we explored the feasibility of directional gradients (Gabor) and localintensity statistics (Haralick, Law) texture features obtained from differenttumor-specific sub-compartments (enhancing, non-enhancing, infiltrating edges,and necrotic regions) on T1, T2 and FLAIR sequence in non-invasively predictingAKT pathway mutational status from routinely acquired MRI.

 MATERIALS AND METHODS: A retrospective study of brain tumor MRimaging performed on images from TCIA. 75 patient studies were analyzed. Brainlesions on MR imaging were manually annotated by an expert neuroradiologist.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Aset of three dimensional radiomic features was extracted for every lesion oneach MR imaging sequence: gadolinium T1WI, T2WI, and FLAIR, and within eachcompartment: enhancing tumor, edema, necrosis. Rank-sum tests were used to findfeatures that correlated with the AKT pathway status of the patient studies.RESULTS: 5radiomic texture features with the most mutual information were Gabor, Sobel,and Law texture features, predominantly from the enhancing tumor region in FLAIRsequenced images. CONCLUSIONS: Ourpreliminary results suggest that radiomic features, particularly in thenecrotic region of the tumor may provide complementary diagnostic informationon routine MR imaging sequences that may improve the distinction between thetumors with activated and wild type AKT pathways in patients with GBM.                    IntroductionGlioblastoma (GBM) is the mostcommon and aggressive form of cancer that begins in the brain. The disease isdifficult to detect without MR imaging due to the non-specific nature of thesymptoms, which begin mildly and can progress rapidly as the diseaseprogresses.1  Without treatment, survival is typically 3months.

  With treatment, survivalincreases to between 12 and 15 months, with only 3 to 5 % surviving past 5years past diagnosis.  The AKT/PI3K signaling pathway isinvolved in cellular proliferation, and in many cancers, this pathway isoveractive. PTEN acts as an antagonist to the AKT pathway by dephosphorylatingPIP3 to PIP2, which inhibits AKT’s ability to bind to the cell membrane,decreasing cellular proliferation.2 PTEN is the most commonlydisrupted tumor suppressor, with mutations found in 20 – 40% of GBM tumors. PTENis activated by phosphatidylinositol 3-kinase (PI3K), a product of PIK3CA3.

If either PTEN or PIK3CAhave a mutation resulting in a loss of function the AKT pathway is activated,which can cause therapy resistance and tumorigenesis1. Mutational status iscurrently determined through DNA sequencing of surgically resected specimens orbiopsy samples. Unfortunately, this requires invasive intervention, as well asbeing prone to sampling errors (as gene profiling is assayed only on a small portionof tissue). GBM is routinely diagnosed usingMagnetic Resonance Imaging (MRI).

While MRI provides structural and functionalnon-invasive characterization of the tumor, it is not currently effective incapturing the underlying tumor pathology, and molecular heterogeneity. Todefinitively diagnose GBM, a stereotactic biopsy or craniotomy with pathologicconfirmation is required. Biopsy provides valuable genetic information, but itis invasive and does not capture the heterogeneity of the tumor. Radiomics is the use ofdata-characterization algorithms to extract large sums of quantitative texturaland shape information from medical images. While radiomics is applicable tomany medical conditions, it is the most developed and commonly applied foroncology.4 Radiogenomics is the termused to refer to finding the correlation between cancer imaging features andgenetic expression information5. This approach has had someprevious success in determining the associated genetics with MRI phenotypes inGBM.6 Other studies have found thatvolumetric MRI features are significantly predictive of mutation status in TP53,RB1, NF1, EGFR, and PDGFRA in GBM cases.

We are not aware of previous studiesassociating the AKT pathway with MRI features in GBM. The purpose of this study is todetermine if radiogenomic analysis of the routinely generated MR images cangive insight to the AKT pathway status in cases of GBM. We wish to assesswhether these radiomic features present themselves differently within thedifferent regions of GBM (enhancing tumor, edema, necrosis) and across 3routine multiparametric MR images (Gd-T1WI, T2WI, FLAIR). In this retrospectivestudy, we identified radiomic features in the necrotic region in FLAIR and T2images that had significant correlation with AKT pathway status among a cohortof 58 GBM patients. The eventual goal of this study is to develop a tool thatis more readily available world-wide than biopsy and genetic sequencing todiagnose genetic variants in glioblastoma.

Genetic sequencing is also expensiveand not readily available to all GBM patients around the world.  In this study, we find radiomic features thatare the most correlated with AKT pathway mutational status, and assess thepredictive value of those features using a random forest analysis. MATERIALS AND METHODSStudy PopulationThe cohort for this study consistsof 75 retrospectively analyzed pre-treatment multi-parametric MRI scans fromThe Cancer Imaging Archive (TCIA). The set included 30 individuals with anactivated AKT pathway, and 45 with a wildtype AKT pathway (Table 1). Inclusioncriteria in the training cohort were: (1) the availability of all three of theroutine MRI sequences (T1, T2, FLAIR) for patients, pre-treatment, in thetraining cohort, (2) MRI scans with diagnostic image quality, and (3) theavailability of AKT pathway status, including mutational data for PTEN andPIK3CA. Table 1,Distribution of age, gender and race amongst cases with wild type and activatedAKT pathways. p-values were generated using Chi-squared tests.

Stratified by AKT Pathway Status Wild Type Activated p n 45 30   age (mean(sd)) 58.07(13.21) 62.43(12.

62) 0.158 gender = male (%) 29(64.4) 16(53.3) 0.47 race (%)   0.

659     Asian 1(2.2) 0(0.0)     Black or African American 2(4.4) 2(6.7)     White 42(93.3) 28(93.3)   Image RegistrationFor every patientscan, 3D Slicer 4.5 was used to co-register T2w and FLAIR with reference to Gd Gadolinium(GD)-T1w using 3D affine registration with 12 degrees of freedom encodingrotation.

Every MRI slice was resampled to a uniform pixel spacing of 0.5 x 0.5mm2 and interpolated to a 1 mm slice thickness to account forresolution variability. Annotation of each 2-dimensional slice with visibletumor was performed by an expert reader in 3 sections: edema, enhancing tumor,and tumor necrosis. Tumor necrosis was represented by hypo-intense regionsgenerally located centrally in the tumoral region in Gd-T1w sequences.Hyper-intense regions in FLAIR sequences correlate with greater interstitialleakage and lower cell density, signifying edema.

3D Radiomic Feature ExtractionFor eachpatient, 13 first order statistics, 13 second order statistics, 39 Haralickfeatures (captures tumor heterogeneity), 96 Gabor features (captures structuraldetail at different scales and orientations) , 152 Law’s energy features(captures presence of spots, edges, waves and ripples in an image), 13 CoLlAGe7 features, and 38 shapefeatures were extracted in the context of three dimensions for each compartment(tumor necrosis, enhancing tumor, edema) on a per-voxel basis for each MRIsequence (FLAIR, Gd-T1w, T2w). All feature calculations were performed using in-housesoftware implemented in MATLAB R 2016b platform (Mathworks, Natick, MA).AKT Pathway ActivationClassification               Themutation status of PIK3CA and PTEN were determined for each patient using datafrom TumorPortal.8 Mutation status wasdetermined through DNA sequencing. Patients with a tumor that had a PTENinactivating mutation, or a PIK3CA activating mutation were classified as havingan activated AKT pathway. Feature Selection and StatisticalAnalysisThe cohort wasdivided into a random set of 37 cases to train the model and 38 to test. Thestandard deviation and median were computed for each feature, and thesestatistics were normalized and compared between patients with wild type AKTpathway status and activated AKT pathway status.

Feature statistics were rankedby computing the mutual information between each feature and AKT pathwaystatus, employing three-fold cross validation. The five features thatconsistently had the most mutual information were used in a QuadraticDiscriminant Analysis (QDA) classification. The classifier was then applied to the independent test set and assessedfor sensitivity and specificity in predicting AKT pathway status. RESULTSThe mutualinformation analysis reported that the five features that shared the mostinformation with AKT status activation were one Gabor feature from the necroticregion in the T1 sequence, in addition to two intensity features and twoHaralick features from the enhancing tumor region in the FLAIR sequence (Table2).

In the training set, the QDA classifier had an accuracyrate of 0.7329. The overall prediction accuracy for the testing set was 0.7368(Sensitivity = 0.

90, Specificity = 0.844). Table 2:Table of the five  features with thehighest mutual information with AKT pathway status, and the sequence andsegment that those features were found. Feature Name Sequence Region Feature Type median-XY Orient=0.

5236, XZ Orient=0, Bandwidth=1, Wavelength=1.4142 T1 Necrosis Gabor median-raw_intensity FLAIR  Enhancing Tumor Intensity std-raw_intensity FLAIR Enhancing Tumor Intensity median-mean ws=5 FLAIR Enhancing Tumor Haralick median-median ws=5 FLAIR Enhancing Tumor Haralick   DISCUSSIONWe found thattexture features in the enhancing tumor region of a GBM tumor could becorrelated with whether the AKT pathway is activated or wildtype. Contrast enhancementin MRI imaging is indicative of the localized deterioration of the blood-brainbarrier, which is generally associated with high-grade gliomas.

9 Previous studies have foundthat the AKT pathway influences the vascular endothelial growth factor (VEGF)which affects the permeability of the blood brain barrier, so it is not surprisingthat the AKT pathway may be associated with the enhancing tumor region in someway. 2  This analysis was limited in sample size. Withmore patients, feature selection could be more thorough and classificationcould be more accurate, so this is something to explore in future work. The AKTpathway activation status was characterized using the mutation of PTEN and/orPIK3CA as markers.

This characterization method is not perfect and could allowfor some misclassification of activated and wildtype individuals that may skewresults.     CONCLUSIONSIn this study,we investigated the distinction in AKT pathway status between 3D MRI featuresin GBM across the regions (edema, enhancing tumor, necrosis) of the tumor. Ourresults suggest that radiomic analysis on routinely acquired MR imaging mayenable discrimination between wildtype and activated AKT pathways in the enhancingtumor region of GBM, specifically intensity and Haralick features in the FLAIRsequenced images.  REFERENCES 1.

            Li, X. et al.PI3K/Akt/mTOR signaling pathway and targeted therapy for glioblastoma.

Oncotarget7, 33440–33450 (2016).2.            Kilic,E. et al. The phosphatidylinositol-3 kinase/Akt pathway mediates VEGF’sneuroprotective activity and induces blood brain barrier permeability afterfocal cerebral ischemia. FASEB J.

Off. Publ. Fed. Am.

Soc. Exp. Biol. 20,1185–1187 (2006).3.            Samuels,Y. & Waldman, T.

Oncogenic Mutations of PIK3CA in Human Cancers. Curr.Top. Microbiol. Immunol. 347, 21–41 (2010).4.            Gillies,R.

J., Kinahan, P. E. & Hricak, H.

Radiomics: Images Are More than Pictures,They Are Data. Radiology 278, 563–577 (2015).5.            Mazurowski,M. A. Radiogenomics: what it is and why it is important.

J. Am. Coll.

Radiol. JACR 12, 862–866 (2015).6.            Gevaert,O. et al. Glioblastoma multiforme: exploratory radiogenomic analysis byusing quantitative image features.

Radiology 273, 168–174 (2014).7.            Prasanna,P.

, Tiwari, P. & Madabhushi, A. Co-occurrence of Local Anisotropic GradientOrientations (CoLlAGe): A new radiomics descriptor. Sci. Rep. 6,37241 (2016).

8.            Lawrence,M. S. et al. Discovery and saturation analysis of cancer genes across 21tumor types. Nature 505, 495–501 (2014).

9.            Mabray,M. C., Barajas, R.

F. & Cha, S. Modern Brain Tumor Imaging. Brain TumorRes.

Treat. 3, 8–23 (2015).

x

Hi!
I'm Mary!

Would you like to get a custom essay? How about receiving a customized one?

Check it out