Salvador Tortajada

Accurate classification of childhood brain tumours by in vivo 1H MRS – A multi-centre study

AIMS To evaluate the accuracy of single-voxel Magnetic Resonance Spectroscopy ((1)H MRS) as a non-invasive diagnostic aid for paediatric brain tumours in a multi-national study. Our hypotheses are (1) that automated classification based on (1)H MRS provides an accurate non-invasive diagnosis in multi-centre datasets and (2) using a protocol which increases the metabolite information improves the diagnostic accuracy. METHODS Seventy-eight patients under 16 years old with histologically proven brain tumours from 10 international centres were investigated. Discrimination of 29 medulloblastomas, 11 ependymomas and 38 pilocytic astrocytomas (PILOAs) was evaluated. Single-voxel MRS was undertaken prior to diagnosis (1.5 T Point-Resolved Spectroscopy (PRESS), Proton Brain Exam (PROBE) or Stimulated Echo Acquisition Mode (STEAM), echo time (TE) 20-32 ms and 135-136 ms). MRS data were processed using two strategies, determination of metabolite concentrations using TARQUIN software and automatic feature extraction with Peak Integration (PI). Linear Discriminant Analysis (LDA) was applied to this data to produce diagnostic classifiers. An evaluation of the diagnostic accuracy was performed based on resampling to measure the Balanced Accuracy Rate (BAR). RESULTS The accuracy of the diagnostic classifiers for discriminating the three tumour types was found to be high (BAR 0.98) when a combination of TE was used. The combination of both TEs significantly improved the classification performance (p<0.01, Tukeys test) compared with the use of one TE alone. Other tumour types were classified accurately as glial or primitive neuroectodermal (BAR 1.00). CONCLUSION (1)H MRS has excellent accuracy for the non-invasive diagnosis of common childhood brain tumours particularly if the metabolite information is maximised and should become part of routine clinical assessment for these children.

The effect of combining two echo times in automatic brain tumor classification by MRS

1H MRS is becoming an accurate, non‐invasive technique for initial examination of brain masses. We investigated if the combination of single‐voxel 1H MRS at 1.5 T at two different (TEs), short TE (PRESS or STEAM, 20–32 ms) and long TE (PRESS, 135–136 ms), improves the classification of brain tumors over using only one echo TE. A clinically validated dataset of 50 low‐grade meningiomas, 105 aggressive tumors (glioblastoma and metastasis), and 30 low‐grade glial tumors (astrocytomas grade II, oligodendrogliomas and oligoastrocytomas) was used to fit predictive models based on the combination of features from short‐TEs and long‐TE spectra. A new approach that combines the two consecutively was used to produce a single data vector from which relevant features of the two TE spectra could be extracted by means of three algorithms: stepwise, reliefF, and principal components analysis. Least squares support vector machines and linear discriminant analysis were applied to fit the pairwise and multiclass classifiers, respectively. Significant differences in performance were found when short‐TE, long‐TE or both spectra combined were used as input. In our dataset, to discriminate meningiomas, the combination of the two TE acquisitions produced optimal performance. To discriminate aggressive tumors from low‐grade glial tumours, the use of short‐TE acquisition alone was preferable. The classifier development strategy used here lends itself to automated learning and test performance processes, which may be of use for future web‐based multicentric classifier development studies. Copyright

A generic and extensible automatic classification framework applied to brain tumour diagnosis in HealthAgents

This work was partially funded by the European Commission: HealthAgents (contract no. FP6-2005-IST 027214). Jan Luts is a PhD student supported by an IWT grant. Carlos Saez, Salvador Tortajada and Javier Vicente are PhD students partially supported by the Programa Torres Quevedo from the Ministerio de Educacion y Ciencia, co-founded by the European Social Fund (PTQ-08-01-06817, PTQ-08-01-06802, PTQ05-02-03386). We thank INTERPRET partners for their support and for providing the data used for training some of the classifiers included in the HealthAgents network; in particular we thank C. Majos (IDI-Bellvitge), John Griffiths (SGUL), Arend Heerschap (RU), Witold Gajewicz (MUL), Jorge Calvar (FLENI), Margarida Julia-Sape (UAB) and Carles Arus (UAB). The language revision of this document was funded by the Universidad Politecnica de Valencia. This work has been partially supported by the Health Institute Carlos III through the RETICS Combiomed, RD07/0067/2001.

Incremental Gaussian Discriminant Analysis based on Graybill and Deal weighted combination of estimators for brain tumour diagnosis

In the last decade, machine learning (ML) techniques have been used for developing classifiers for automatic brain tumour diagnosis. However, the development of these ML models rely on a unique training set and learning stops once this set has been processed. Training these classifiers requires a representative amount of data, but the gathering, preprocess, and validation of samples is expensive and time-consuming. Therefore, for a classical, non-incremental approach to ML, it is necessary to wait long enough to collect all the required data. In contrast, an incremental learning approach may allow us to build an initial classifier with a smaller number of samples and update it incrementally when new data are collected. In this study, an incremental learning algorithm for Gaussian Discriminant Analysis (iGDA) based on the Graybill and Deal weighted combination of estimators is introduced. Each time a new set of data becomes available, a new estimation is carried out and a combination with a previous estimation is performed. iGDA does not require access to the previously used data and is able to include new classes that were not in the original analysis, thus allowing the customization of the models to the distribution of data at a particular clinical center. An evaluation using five benchmark databases has been used to evaluate the behaviour of the iGDA algorithm in terms of stability-plasticity, class inclusion and order effect. Finally, the iGDA algorithm has been applied to automatic brain tumour classification with magnetic resonance spectroscopy, and compared with two state-of-the-art incremental algorithms. The empirical results obtained show the ability of the algorithm to learn in an incremental fashion, improving the performance of the models when new information is available, and converging in the course of time. Furthermore, the algorithm shows a negligible instance and concept order effect, avoiding the bias that such effects could introduce.

Prediction of postpartum depression using multilayer perceptrons and pruning

OBJECTIVE The main goal of this paper is to obtain a classification model based on feed-forward multilayer perceptrons in order to improve postpartum depression prediction during the 32 weeks after childbirth with a high sensitivity and specificity and to develop a tool to be integrated in a decision support system for clinicians. MATERIALS AND METHODS Multilayer perceptrons were trained on data from 1397 women who had just given birth, from seven Spanish general hospitals, including clinical, environmental and genetic variables. A prospective cohort study was made just after delivery, at 8 weeks and at 32 weeks after delivery. The models were evaluated with the geometric mean of accuracies using a hold-out strategy. RESULTS Multilayer perceptrons showed good performance (high sensitivity and specificity) as predictive models for postpartum depression. CONCLUSIONS The use of these models in a decision support system can be clinically evaluated in future work. The analysis of the models by pruning leads to a qualitative interpretation of the influence of each variable in the interest of clinical protocols.

Incremental Logistic Regression for Customizing Automatic Diagnostic Models

In the last decades, and following the new trends in medicine, statistical learning techniques have been used for developing automatic diagnostic models for aiding the clinical experts throughout the use of Clinical Decision Support Systems. The development of these models requires a large, representative amount of data, which is commonly obtained from one hospital or a group of hospitals after an expensive and time-consuming gathering, preprocess, and validation of cases. After the model development, it has to overcome an external validation that is often carried out in a different hospital or health center. The experience is that the models show underperformed expectations. Furthermore, patient data needs ethical approval and patient consent to send and store data. For these reasons, we introduce an incremental learning algorithm base on the Bayesian inference approach that may allow us to build an initial model with a smaller number of cases and update it incrementally when new data are collected or even perform a new calibration of a model from a different center by using a reduced number of cases. The performance of our algorithm is demonstrated by employing different benchmark datasets and a real brain tumor dataset; and we compare its performance to a previous incremental algorithm and a non-incremental Bayesian model, showing that the algorithm is independent of the data model, iterative, and has a good convergence.

Extracting MRS discriminant functional features of brain tumors

The current challenge in automatic brain tumor classification based on MRS is the improvement of the robustness of the classification models that explicitly account for the probable breach of the independent and identically distributed conditions in the MRS data points. To contribute to this purpose, a new algorithm for the extraction of discriminant MRS features of brain tumors based on a functional approach is presented. Functional data analysis based on region segmentation (RSFDA) is based on the functional data analysis formalism using nonuniformly distributed B splines according to spectral regions that are highly correlated. An exhaustive characterization of the method is presented in this work using controlled and real scenarios. The performance of RSFDA was compared with other widely used feature extraction methods. In all simulated conditions, RSFDA was proven to be stable with respect to the number of variables selected and with respect to the classification performance against noise and baseline artifacts. Furthermore, with real multicenter datasets classification, RSFDA and peak integration (PI) obtained better performance than the other feature extraction methods used for comparison. Other advantages of the method proposed are its usefulness in selecting the optimal number of features for classification and its simplified functional representation of the spectra, which contributes to highlight the discriminative regions of the MR spectrum for each classification task. Copyright

Application of Artificial Neural Network for Reducing Random Coincidences in PET

Positron Emission Tomography (PET) is based on the detection in coincidence of the two photons created in a positron annihilation. In conventional PET, this coincidence identification is usually carried out through a coincidence electronic unit. An accidental coincidence occurs when two photons arising from different annihilations are classified as a coincidence. Accidental coincidences are one of the main sources of image degradation in PET. Some novel systems allow coincidences to be selected post-acquisition in software, or in real time through a digital coincidence engine in an FPGA. These approaches provide the user with extra flexibility in the sorting process and allow the application of alternative coincidence sorting procedures. In this work a novel sorting procedure based on Artificial Neural Network (ANN) techniques has been developed. It has been compared to a conventional coincidence sorting algorithm based on a time coincidence window. The data have been obtained from Monte-Carlo simulations. A small animal PET scanner has been implemented to this end. The efficiency (the ratio of correct identifications) can be selected for both methods. In one case by changing the actual value of the coincidence window used, and in the other by changing a threshold at the output of the neural network. At matched efficiencies, the ANN-based method always produces a sorted output with a smaller random fraction. In addition, two differential trends are found: the conventional method presents a maximum achievable efficiency, while the ANN-based method is able to increase the efficiency up to unity, the ideal value, at the cost of increasing the random fraction. Images reconstructed using ANN sorted data (no compensation for randoms) present better contrast, and those image features which are more affected by randoms are enhanced. For the image quality phantom used in the paper, the ANN method decreases the spillover ratio by a factor of 18%.

Genomics and metabolomics research for brain tumour diagnosis based on machine learning

The incorporation of new biomedical technologies in the diagnosis and prognosis of cancer is changing medicine to an evidence-based diagnosis. We summarize some studies related to brain tumour research in Europe, based on the metabolic information provided by in vivo Magnetic Resonance Spectroscopy (MRS) and transcriptomic profiling observed by DNA microarrays. The first result presents the improvement in brain tumour diagnosis by combining Long TE and Short TE single voxel MR Spectra. Afterwards, a mixture model for binned and truncated data to characterize and classify MRS is reviewed. The classification of Glioblastomas Multiforme and Meningothelial Meningiomas using single-labeling cDNA-based microarrays was studied as proof of principle in the incorporation of genomic information to clinical diagnosis. Finally, we present a Decision Support System for in-vivo classification of brain tumours were the best inferred classifiers are deployed for their clinical use.

Ranking of Brain Tumour Classifiers Using a Bayesian Approach

This study presents a ranking for classifers using a Bayesian perspective. This ranking framework is able to evaluate the performance of the models to be compared when they are inferred from different sets of data. It also takes into account the performance obtained on samples not used during the training of the classifiers. Besides, this ranking assigns a prior to each model based on a measure of similarity of the training data to a test case. An evaluation consisting of ranking brain tumour classifiers is presented. These multilayer perceptron classifiers are trained with 1H magnetic resonance spectroscopy (MRS) signals following a multiproject multicenter evaluation approach. We demonstrate that such a framework can be effectively applied to the real problem of selecting classifiers for brain tumour classification.