Optimal classification of long echo time in vivo magnetic resonance spectra in the detection of recurrent brain tumors

A comparison of non-negative matrix underapproximation methods for the decomposition of magnetic resonance spectroscopy data from human brain tumors

Magnetic resonance spectroscopy (MRS) is an MR technique that provides information about the biochemistry of tissues in a noninvasive way. MRS has been widely used for the study of brain tumors, both preoperatively and during follow-up. In this study, we investigated the performance of a range of variants of unsupervised matrix factorization methods of the non-negative matrix underapproximation (NMU) family, namely, sparse NMU, global NMU, and recursive NMU, and compared them with convex non-negative matrix factorization (C-NMF), which has previously shown a good performance on brain tumor diagnostic support problems using MRS data. The purpose of the investigation was 2-fold: first, to ascertain the differences among the sources extracted by these methods; and second, to compare the influence of each method in the diagnostic accuracy of the classification of brain tumors, using them as feature extractors. We discovered that, first, NMU variants found meaningful sources in terms of biological interpretability, but representing parts of the spectrum, in contrast to C-NMF; and second, that NMU methods achieved better classification accuracy than C-NMF for the classification tasks when one class was not meningioma.

Analyzing magnetic resonance imaging data from glioma patients using deep learning

The quantitative analysis of images acquired in the diagnosis and treatment of patients with brain tumors has seen a significant rise in the clinical use of computational tools. The underlying technology to the vast majority of these tools are machine learning methods and, in particular, deep learning algorithms. This review offers clinical background information of key diagnostic biomarkers in the diagnosis of glioma, the most common primary brain tumor. It offers an overview of publicly available resources and datasets for developing new computational tools and image biomarkers, with emphasis on those related to the Multimodal Brain Tumor Segmentation (BraTS) Challenge. We further offer an overview of the state-of-the-art methods in glioma image segmentation, again with an emphasis on publicly available tools and deep learning algorithms that emerged in the context of the BraTS challenge.

Binary Markov Random Fields and interpretable mass spectra discrimination

For mass spectra acquired from cancer patients by MALDI or SELDI techniques, automated discrimination between cancer types or stages has often been implemented by machine learning algorithms. Nevertheless, these techniques typically lack interpretability in terms of biomarkers. In this paper, we propose a new mass spectra discrimination algorithm by parameterized Markov Random Fields to automatically generate interpretable classifiers with small groups of scored biomarkers. A dataset of 238 MALDI colorectal mass spectra and two datasets of 216 and 253 SELDI ovarian mass spectra respectively were used to test our approach. The results show that our approach reaches accuracies of 81% to 100% to discriminate between patients from different colorectal and ovarian cancer stages, and performs as well or better than previous studies on similar datasets. Moreover, our approach enables efficient planar-displays to visualize mass spectra discrimination and has good asymptotic performance for large datasets. Thus, our classifiers should facilitate the choice and planning of further experiments for biological interpretation of cancer discriminating signatures. In our experiments, the number of mass spectra for each colorectal cancer stage is roughly half of that for each ovarian cancer stage, so that we reach lower discrimination accuracy for colorectal cancer than for ovarian cancer.

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.

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, Tukey's 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.

A hybrid method of application of independent component analysis to in vivo 1H MR spectra of childhood brain tumours

Independent component analysis (ICA) can automatically extract individual metabolite, macromolecular and lipid (MMLip) components from a series of in vivo MR spectra. The traditional feature extraction (FE)-based ICA approach is limited, in that a large sample size is required and a combination of metabolite and MMLip components can appear in the same independent component. The alternative ICA approach, based on blind source separation (BSS), is weak when dealing with overlapping peaks. Combining the advantages of both BSS and FE methods may lead to better results. Thus, we propose an ICA approach involving a hybrid of the BSS and FE techniques for the automated decomposition of a series of MR spectra. Experiments were performed on synthesised and patient in vivo childhood brain tumour MR spectra datasets. The hybrid ICA method showed an improvement in the decomposition ability compared with BSS-ICA or FE-ICA, with an increased correlation between the independent components and simulated metabolite and MMLip signals. Furthermore, we were able to automatically extract metabolites from the patient MR spectra dataset that were not in commonly used basis sets (e.g. guanidinoacetate).

Identifying malignant transformations in recurrent low grade gliomas using high resolution magic angle spinning spectroscopy

OBJECTIVE

The objective of this study was to determine whether metabolic parameters derived from ex vivo analysis of tissue samples are predictive of biologic characteristics of recurrent low grade gliomas (LGGs). This was achieved by exploring the use of multivariate pattern recognition methods to generate statistical models of the metabolic characteristics of recurrent LGGs that correlate with aggressive biology and poor clinical outcome.

METHODS

Statistical models were constructed to distinguish between patients with recurrent gliomas that had undergone malignant transformation to a higher grade and those that remained grade 2. The pattern recognition methods explored in this paper include three filter-based feature selection methods (chi-square, gain ratio, and two-way conditional probability), a genetic search wrapper-based feature subset selection algorithm, and five classification algorithms (linear discriminant analysis, logistic regression, functional trees, support vector machines, and decision stump logit boost). The accuracy of each pattern recognition framework was evaluated using leave-one-out cross-validation and bootstrapping.

MATERIALS

The population studied included fifty-three patients with recurrent grade 2 gliomas. Among these patients, seven had tumors that transformed to grade 4, twenty-four had tumors that transformed to grade 3, and twenty-two had tumors that remained grade 2. Image-guided tissue samples were obtained from these patients using surgical navigation software. Part of each tissue sample was examined by a pathologist for histological features and for consistency with the tumor grade diagnosis. The other part of the tissue sample was analyzed with ex vivo nuclear magnetic resonance (NMR) spectroscopy.

RESULTS

Distinguishing between recurrent low grade gliomas that transformed to a higher grade and those that remained grade 2 was achieved with 96% accuracy, using areas of the ex vivo NMR spectrum corresponding to myoinositol, 2-hydroxyglutarate, hypo-taurine, choline, glycerophosphocholine, phosphocholine, glutathione, and lipid. Logistic regression and decision stump boosting models were able to distinguish between recurrent gliomas that transformed to a higher grade and those that did not with 100% training accuracy (95% confidence interval [93-100%]), 96% leave-one-out cross-validation accuracy (95% confidence interval [87-100%]), and 96% bootstrapping accuracy (95% confidence interval [95-97%]). Linear discriminant analysis, functional trees, and support vector machines were able to achieve leave-one-out cross-validation accuracy above 90% and bootstrapping accuracy above 85%. The three feature ranking methods were comparable in performance.

CONCLUSIONS

This study demonstrates the feasibility of using quantitative pattern recognition methods for the analysis of metabolic data from brain tissue obtained during the surgical resection of gliomas. All pattern recognition techniques provided good diagnostic accuracies, though logistic regression and decision stump boosting slightly outperform the other classifiers. These methods identified biomarkers that can be used to detect malignant transformations in individual low grade gliomas, and can lead to a timely change in treatment for each patient.

Pattern recognition methods in (1)H MRS monitoring in vivo of normal appearing cerebellar tissue after treatment of posterior fossa tumors

The objective of this study was to investigate the metabolic responses of normal appearing cerebellar tissue after posterior fossa tumor treatment, and to identify characteristics of the particular treatment method. Moreover, this work examined the metabolic alterations of normal appearing tissue induced by a particular tumor state including resection, stagnation, progression, and recurrence. The studied group consisted of 29 patients treated for posterior fossa tumors. All of them were irradiated with a total dose of 54 Gy at 1.8 Gy/fraction (median values). In addition, 13 underwent chemotherapy, 25 underwent total tumor resection, 18 were tumor-free in control examinations, 5 had a stable disease, and tumor progression or recurrence was observed in 2 and 4 cases, respectively. The 69 spectra, acquired using a MRI/MRS 2T system, were analyzed using Partial Least Squares Discriminant Analysis (PLS-DA) with orthogonal signal correction (OSC) spectral filtering. A significantly elevated spectral region (0.97-1.55 ppm) was observed in patients after total resection in comparison to non-operated subjects. Patients treated with chemotherapy showed an elevated band between 1.15-1.75 and 2.7-3.0 ppm and had decreases in the remaining parts of the spectra. Increases in lactate and decreases in the remaining metabolites were characteristic for the tumor progression/recurrence group. Pattern recognition methods coupled with MRS revealed significant treatment-dependent alterations in normal appearing cerebellar tissue, as well as metabolic changes induced by tumor progression/recurrence.

A comparison of random forest and its Gini importance with standard chemometric methods for the feature selection and classification of spectral data

Background

Regularized regression methods such as principal component or partial least squares regression perform well in learning tasks on high dimensional spectral data, but cannot explicitly eliminate irrelevant features. The random forest classifier with its associated Gini feature importance, on the other hand, allows for an explicit feature elimination, but may not be optimally adapted to spectral data due to the topology of its constituent classification trees which are based on orthogonal splits in feature space.

Results

We propose to combine the best of both approaches, and evaluated the joint use of a feature selection based on a recursive feature elimination using the Gini importance of random forests' together with regularized classification methods on spectral data sets from medical diagnostics, chemotaxonomy, biomedical analytics, food science, and synthetically modified spectral data. Here, a feature selection using the Gini feature importance with a regularized classification by discriminant partial least squares regression performed as well as or better than a filtering according to different univariate statistical tests, or using regression coefficients in a backward feature elimination. It outperformed the direct application of the random forest classifier, or the direct application of the regularized classifiers on the full set of features.

Conclusion

The Gini importance of the random forest provided superior means for measuring feature relevance on spectral data, but – on an optimal subset of features – the regularized classifiers might be preferable over the random forest classifier, in spite of their limitation to model linear dependencies only. A feature selection based on Gini importance, however, may precede a regularized linear classification to identify this optimal subset of features, and to earn a double benefit of both dimensionality reduction and the elimination of noise from the classification task.

MRI for identification of progression in brain tumors: from morphology to function

For monitoring of brain tumors, it is crucial to identify progression or treatment failure early during follow-up to change treatment schemes and, thereby, optimize patient outcome. In the past years, several areas within the field of magnetic resonance (MR) have seen considerable advances: modern contrast media, advanced morphologic approaches and several functional techniques, for example, in the visualization of tumor perfusion or tumor cell metabolism. This review presents these recent advances by introducing the different techniques and outlining their benefit for identification of progression in brain tumors, with a focus on gliomas, metastases and meningiomas. After radiotherapy, MR spectroscopy helps to more accurately discriminate between radiation necrosis and glioma progression. In low-grade gliomas, perfusion MR techniques enable a more sensitive detection of anaplastic transformation than conventional MRI. Modern contrast media, as well as diffusion tensor imaging, allow for an improved tumor delineation and assessment of tumor extension. We will also highlight the biological background of these techniques, their applicability and current limitations. In conclusion, modern MRI techniques have been developed that are on the doorstep to be integrated in clinical routine.

Multiproject-multicenter evaluation of automatic brain tumor classification by magnetic resonance spectroscopy

JUSTIFICATION

Automatic brain tumor classification by MRS has been under development for more than a decade. Nonetheless, to our knowledge, there are no published evaluations of predictive models with unseen cases that are subsequently acquired in different centers. The multicenter eTUMOUR project (2004-2009), which builds upon previous expertise from the INTERPRET project (2000-2002) has allowed such an evaluation to take place.

MATERIALS AND METHODS

A total of 253 pairwise classifiers for glioblastoma, meningioma, metastasis, and low-grade glial diagnosis were inferred based on 211 SV short TE INTERPRET MR spectra obtained at 1.5 T (PRESS or STEAM, 20-32 ms) and automatically pre-processed. Afterwards, the classifiers were tested with 97 spectra, which were subsequently compiled during eTUMOUR.

RESULTS

In our results based on subsequently acquired spectra, accuracies of around 90% were achieved for most of the pairwise discrimination problems. The exception was for the glioblastoma versus metastasis discrimination, which was below 78%. A more clear definition of metastases may be obtained by other approaches, such as MRSI + MRI.

CONCLUSIONS

The prediction of the tumor type of in-vivo MRS is possible using classifiers developed from previously acquired data, in different hospitals with different instrumentation under the same acquisition protocols. This methodology may find application for assisting in the diagnosis of new brain tumor cases and for the quality control of multicenter MRS databases.

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

(1)H MRS is becoming an accurate, non-invasive technique for initial examination of brain masses. We investigated if the combination of single-voxel (1)H 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.

Spectrum separation resolves partial-volume effect of MRSI as demonstrated on brain tumor scans

Magnetic resonance spectroscopic imaging (MRSI) is currently used clinically in conjunction with anatomical MRI to assess the presence and extent of brain tumors and to evaluate treatment response. Unfortunately, the clinical utility of MRSI is limited by significant variability of in vivo spectra. Spectral profiles show increased variability because of partial coverage of large voxel volumes, infiltration of normal brain tissue by tumors, innate tumor heterogeneity, and measurement noise. We address these problems directly by quantifying the abundance (i.e. volume fraction) within a voxel for each tissue type instead of the conventional estimation of metabolite concentrations from spectral resonance peaks. This 'spectrum separation' method uses the non-negative matrix factorization algorithm, which simultaneously decomposes the observed spectra of multiple voxels into abundance distributions and constituent spectra. The accuracy of the estimated abundances is validated on phantom data. The presented results on 20 clinical cases of brain tumor show reduced cross-subject variability. This is reflected in improved discrimination between high-grade and low-grade gliomas, which demonstrates the physiological relevance of the extracted spectra. These results show that the proposed spectral analysis method can improve the effectiveness of MRSI as a diagnostic tool.

Mimicking the human expert: pattern recognition for an automated assessment of data quality in MR spectroscopic images

Besides the diagnostic evaluation of a spectrum, the assessment of its quality and a check for plausibility of its information remains a highly interactive and thus time-consuming process in MR spectroscopic imaging (MRSI) data analysis. In the automation of this quality control, a score is proposed that is obtained by training a machine learning classifier on a representative set of spectra that have previously been classified by experts into evaluable data and nonevaluable data. In the first quantitative evaluation of different quality measures on a test set of 45,312 long echo time spectra in the diagnosis of brain tumor, the proposed pattern recognition (using the random forest classifier) separated high- and low-quality spectra comparable to the human operator (area-under-the-curve of the receiver-operator-characteristic, AUC>0.993), and performed better than decision rules based on the signal-to-noise-ratio (AUC<0.934) or the estimated Cramér-Rao-bound on the errors of a spectral fitting (AUC<0.952). This probabilistic assessment of the data quality provides comprehensible confidence images and allows filtering the input of any subsequent data processing, i.e., quantitation or pattern recognition, in an automated fashion. It thus can increase robustness and reliability of the final diagnostic evaluation and allows for the automation of a tedious part of MRSI data analysis.

Quantification and classification of high-resolution magic angle spinning data for brain tumor diagnosis

The goal of this work is to propose a complete protocol (preprocessing, processing and classification) for classifying brain tumors with proton high-resolution magic-angle spinning ((1)H HR-MAS) data. The different steps of the procedure are detailed and discussed. Feature extraction techniques such as peak integration, including also the automated quantitation method AQSES, were combined with linear (LDA) and non-linear (least-squares support vector machine or LS-SVM) classifiers. Classification accuracy was assessed using a stratified random sampling scheme. The results suggest that LS-SVM performs better than LDA while AQSES performs better than the standard peak integration feature extraction method.

Automated estimation of tumor probability in prostate magnetic resonance spectroscopic imaging: pattern recognition vs quantification

Despite its diagnostic value and technological availability, (1)H NMR spectroscopic imaging (MRSI) has not found its way into clinical routine yet. Prerequisite for the clinical application is an automated and reliable method for the diagnostic evaluation of MRS images. In the present paper, different approaches to the estimation of tumor probability from MRSI in the prostate are assessed. Two approaches to feature extraction are compared: quantification (VARPRO, AMARES, QUEST) and subspace methods on spectral patterns (principal components, independent components, nonnegative matrix factorization, partial least squares). Linear as well as nonlinear classifiers (support vector machines, Gaussian processes, random forests) are applied and discussed. Quantification-based approaches are much more sensitive to the choice and parameterization of the quantification algorithm than to the choice of the classifier. Furthermore, linear methods based on magnitude spectra easily achieve equal performance and also allow for biochemical interpretation in combination with subspace methods. Nonlinear methods operating directly on magnitude spectra achieve the best results but are less transparent than the linear methods.