Isabella Castiglioni

TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data

The Cancer Genome Atlas (TCGA) research network has made public a large collection of clinical and molecular phenotypes of more than 10 000 tumor patients across 33 different tumor types. Using this cohort, TCGA has published over 20 marker papers detailing the genomic and epigenomic alterations associated with these tumor types. Although many important discoveries have been made by TCGAs research network, opportunities still exist to implement novel methods, thereby elucidating new biological pathways and diagnostic markers. However, mining the TCGA data presents several bioinformatics challenges, such as data retrieval and integration with clinical data and other molecular data types (e.g. RNA and DNA methylation). We developed an R/Bioconductor package called TCGAbiolinks to address these challenges and offer bioinformatics solutions by using a guided workflow to allow users to query, download and perform integrative analyses of TCGA data. We combined methods from computer science and statistics into the pipeline and incorporated methodologies developed in previous TCGA marker studies and in our own group. Using four different TCGA tumor types (Kidney, Brain, Breast and Colon) as examples, we provide case studies to illustrate examples of reproducibility, integrative analysis and utilization of different Bioconductor packages to advance and accelerate novel discoveries.

A Standardized [18F]-FDG-PET Template for Spatial Normalization in Statistical Parametric Mapping of Dementia

Abstract[18F]-fluorodeoxyglucose (FDG) Positron Emission Tomography (PET) is a widely used diagnostic tool that can detect and quantify pathophysiology, as assessed through changes in cerebral glucose metabolism. [18F]-FDG PET scans can be analyzed using voxel-based statistical methods such as Statistical Parametric Mapping (SPM) that provide statistical maps of brain abnormalities in single patients. In order to perform SPM, a “spatial normalization” of an individual’s PET scan is required to match a reference PET template. The PET template currently used for SPM normalization is based on [15O]-H2O images and does not resemble either the specific metabolic features of [18F]-FDG brain scans or the specific morphological characteristics of individual brains affected by neurodegeneration. Thus, our aim was to create a new [18F]-FDG PET aging and dementia-specific template for spatial normalization, based on images derived from both age-matched controls and patients. We hypothesized that this template would increase spatial normalization accuracy and thereby preserve crucial information for research and diagnostic purposes. We investigated the statistical sensitivity and registration accuracy of normalization procedures based on the standard and new template—at the single-subject and group level—independently for subjects with Mild Cognitive Impairment (MCI), probable Alzheimer’s Disease (AD), Frontotemporal lobar degeneration (FTLD) and dementia with Lewy bodies (DLB). We found a significant statistical effect of the population-specific FDG template-based normalisation in key anatomical regions for each dementia subtype, suggesting that spatial normalization with the new template provides more accurate estimates of metabolic abnormalities for single-subject and group analysis, and therefore, a more effective diagnostic measure.

Machine learning on brain MRI data for differential diagnosis of Parkinson's disease and Progressive Supranuclear Palsy

BACKGROUNDnSupervised machine learning has been proposed as a revolutionary approach for identifying sensitive medical image biomarkers (or combination of them) allowing for automatic diagnosis of individual subjects. The aim of this work was to assess the feasibility of a supervised machine learning algorithm for the assisted diagnosis of patients with clinically diagnosed Parkinsons disease (PD) and Progressive Supranuclear Palsy (PSP).nnnMETHODnMorphological T1-weighted Magnetic Resonance Images (MRIs) of PD patients (28), PSP patients (28) and healthy control subjects (28) were used by a supervised machine learning algorithm based on the combination of Principal Components Analysis as feature extraction technique and on Support Vector Machines as classification algorithm. The algorithm was able to obtain voxel-based morphological biomarkers of PD and PSP.nnnRESULTSnThe algorithm allowed individual diagnosis of PD versus controls, PSP versus controls and PSP versus PD with an Accuracy, Specificity and Sensitivity>90%. Voxels influencing classification between PD and PSP patients involved midbrain, pons, corpus callosum and thalamus, four critical regions known to be strongly involved in the pathophysiological mechanisms of PSP.nnnCOMPARISON WITH EXISTING METHODSnClassification accuracy of individual PSP patients was consistent with previous manual morphological metrics and with other supervised machine learning application to MRI data, whereas accuracy in the detection of individual PD patients was significantly higher with our classification method.nnnCONCLUSIONSnThe algorithm provides excellent discrimination of PD patients from PSP patients at an individual level, thus encouraging the application of computer-based diagnosis in clinical practice.

Predictive value of pre-therapy (18)F-FDG PET/CT for the outcome of (18)F-FDG PET-guided radiotherapy in patients with head and neck cancer.

PurposeThe aim of this study was to evaluate the predictive role of pre-therapy fluorodeoxyglucose (FDG) uptake parameters of primary tumour in head and neck cancer (HNC) patients undergoing intensity-modulated radiotherapy (IMRT) with simultaneous integrated boost (SIB) on FDG-positive volume—positron emission tomography (PET) gross tumour volume (PET-GTV).MethodsThis retrospective study included 19 patients (15 men and 4 women, mean age 59.2xa0years, range 23–81xa0years) diagnosed with HNC between 2005 and 2011. Of 19 patients, 15 (79xa0%) had stage III–IV. All patients underwent FDG PET/CT before treatment. Metabolic indexes of primary tumour, including metabolic tumour volume (MTV), maximum and mean standardized uptake value (SUVmax, SUVmean) and total lesion glycolysis (TLG) were considered. Partial volume effect correction (PVC) was performed for SUVmean and TLG estimation. Correlations between PET/CT parameters and 2-year disease-free survival (DFS), local recurrence-free survival (LRFS) and distant metastasis-free survival (DMFS) were assessed. Median patient follow-up was 19.2xa0months (range 4–24xa0months).ResultsMTV, TLG and PVC-TLG predicting patients’ outcome with respect to all the considered local and distant disease control endpoints (LRFS, DMFS and DFS) were 32.4xa0cc, 469.8xa0g and 547.3xa0g, respectively. SUVmean and PVC-SUVmean cut-off values predictive of LRFS and DFS were 10.8 and 13.3, respectively. PVC was able to compensate errors up to 25xa0% in the primary HNC tumour uptake. Moreover, PVC enhanced the statistical significance of the results.ConclusionFDG PET/CT uptake parameters are predictors of patients’ outcome and can potentially identify patients with higher risk of treatment failure that could benefit from more aggressive approaches. Application of PVC is recommended for accurate measurement of PET parameters.

Magnetic resonance imaging biomarkers for the early diagnosis of Alzheimer's disease: a machine learning approach.

Determination of sensitive and specific markers of very early AD progression is intended to aid researchers and clinicians to develop new treatments and monitor their effectiveness, as well as to lessen the time and cost of clinical trials. Magnetic Resonance (MR)-related biomarkers have been recently identified by the use of machine learning methods for the in vivo differential diagnosis of AD. However, the vast majority of neuroimaging papers investigating this topic are focused on the difference between AD and patients with mild cognitive impairment (MCI), not considering the impact of MCI patients who will (MCIc) or not convert (MCInc) to AD. Morphological T1-weighted MRIs of 137 AD, 76 MCIc, 134 MCInc, and 162 healthy controls (CN) selected from the Alzheimers disease neuroimaging initiative (ADNI) cohort, were used by an optimized machine learning algorithm. Voxels influencing the classification between these AD-related pre-clinical phases involved hippocampus, entorhinal cortex, basal ganglia, gyrus rectus, precuneus, and cerebellum, all critical regions known to be strongly involved in the pathophysiological mechanisms of AD. Classification accuracy was 76% AD vs. CN, 72% MCIc vs. CN, 66% MCIc vs. MCInc (nested 20-fold cross validation). Our data encourage the application of computer-based diagnosis in clinical practice of AD opening new prospective in the early management of AD patients.

Integration of mRNA Expression Profile, Copy Number Alterations, and microRNA Expression Levels in Breast Cancer to Improve Grade Definition

Defining the aggressiveness and growth rate of a malignant cell population is a key step in the clinical approach to treating tumor disease. The correct grading of breast cancer (BC) is a fundamental part in determining the appropriate treatment. Biological variables can make it difficult to elucidate the mechanisms underlying BC development. To identify potential markers that can be used for BC classification, we analyzed mRNAs expression profiles, gene copy numbers, microRNAs expression and their association with tumor grade in BC microarray-derived datasets. From mRNA expression results, we found that grade 2 BC is most likely a mixture of grade 1 and grade 3 that have been misclassified, being described by the gene signature of either grade 1 or grade 3. We assessed the potential of the new approach of integrating mRNA expression profile, copy number alterations, and microRNA expression levels to select a limited number of genomic BC biomarkers. The combination of mRNA profile analysis and copy number data with microRNA expression levels led to the identification of two gene signatures of 42 and 4 altered genes (FOXM1, KPNA4, H2AFV and DDX19A) respectively, the latter obtained through a meta-analytical procedure. The 42-based gene signature identifies 4 classes of up- or down-regulated microRNAs (17 microRNAs) and of their 17 target mRNA, and the 4-based genes signature identified 4 microRNAs (Hsa-miR-320d, Hsa-miR-139-5p, Hsa-miR-567 and Hsa-let-7c). These results are discussed from a biological point of view with respect to pathological features of BC. Our identified mRNAs and microRNAs were validated as prognostic factors of BC disease progression, and could potentially facilitate the implementation of assays for laboratory validation, due to their reduced number.

Response to chemotherapy in gastric adenocarcinoma with diffusion‐weighted MRI and 18F‐FDG‐PET/CT: Correlation of apparent diffusion coefficient and partial volume corrected standardized uptake value with histological tumor regression grade

To assess whether changes in diffusion‐weighted MRI (DW‐MRI) and 18F‐fluoro‐2‐deoxyglucose positron emission tomography/computed tomography (18F‐FDG PET/CT), correlate with treatment response to neoadjuvant therapy (NT), as expressed by tumor regression grade (TRG), from locally advanced gastric adenocarcinoma (GA).

Use of Machine Learning to Identify Children with Autism and Their Motor Abnormalities.

In the present work, we have undertaken a proof-of-concept study to determine whether a simple upper-limb movement could be useful to accurately classify low-functioning children with autism spectrum disorder (ASD) aged 2–4. To answer this question, we developed a supervised machine-learning method to correctly discriminate 15 preschool children with ASD from 15 typically developing children by means of kinematic analysis of a simple reach-to-drop task. Our method reached a maximum classification accuracy of 96.7xa0% with seven features related to the goal-oriented part of the movement. These preliminary findings offer insight into a possible motor signature of ASD that may be potentially useful in identifying a well-defined subset of patients, reducing the clinical heterogeneity within the broad behavioral phenotype.

Integrative Analysis with Monte Carlo Cross-Validation Reveals miRNAs Regulating Pathways Cross-Talk in Aggressive Breast Cancer.

In this work an integrated approach was used to identify functional miRNAs regulating gene pathway cross-talk in breast cancer (BC). We first integrated gene expression profiles and biological pathway information to explore the underlying associations between genes differently expressed among normal and BC samples and pathways enriched from these genes. For each pair of pathways, a score was derived from the distribution of gene expression levels by quantifying their pathway cross-talk. Random forest classification allowed the identification of pairs of pathways with high cross-talk. We assessed miRNAs regulating the identified gene pathways by a mutual information analysis. A Fisher test was applied to demonstrate their significance in the regulated pathways. Our results suggest interesting networks of pathways that could be key regulatory of target genes in BC, including stem cell pluripotency, coagulation, and hypoxia pathways and miRNAs that control these networks could be potential biomarkers for diagnostic, prognostic, and therapeutic development in BC. This work shows that standard methods of predicting normal and tumor classes such as differentially expressed miRNAs or transcription factors could lose intrinsic features; instead our approach revealed the responsible molecules of the disease.

Combined analysis of chromosomal instabilities and gene expression for colon cancer progression inference

BackgroundCopy number alterations (CNAs) represent an important component of genetic variations. Such alterations are related with certain type of cancer including those of the pancreas, colon, and breast, among others. CNAs have been used as biomarkers for cancer prognosis in multiple studies, but few works report on the relation of CNAs with the disease progression. Moreover, most studies do not consider the following two important issues. (I) The identification of CNAs in genes which are responsible for expression regulation is fundamental in order to define genetic events leading to malignant transformation and progression. (II) Most real domains are best described by structured data where instances of multiple types are related to each other in complex ways.ResultsOur main interest is to check whether the colorectal cancer (CRC) progression inference benefits when considering both (I) the expression levels of genes with CNAs, and (II) relationships (i.e. dissimilarities) between patients due to expression level differences of the altered genes. We first evaluate the accuracy performance of a state-of-the-art inference method (support vector machine) when subjects are represented only through sets of available attribute values (i.e. gene expression level). Then we check whether the inference accuracy improves, when explicitly exploiting the information mentioned above. Our results suggest that the CRC progression inference improves when the combined data (i.e. CNA and expression level) and the considered dissimilarity measures are applied.ConclusionsThrough our approach, classification is intuitively appealing and can be conveniently obtained in the resulting dissimilarity spaces. Different public datasets from Gene Expression Omnibus (GEO) were used to validate the results.