1H MR spectroscopy of brain tumours and masses

Metabolic profiles of brain metastases

Metastasis to the brain is a feared complication of systemic cancer, associated with significant morbidity and poor prognosis. A better understanding of the tumor metabolism might help us meet the challenges in controlling brain metastases. The study aims to characterize the metabolic profile of brain metastases of different origin using high resolution magic angle spinning (HR-MAS) magnetic resonance spectroscopy (MRS) to correlate the metabolic profiles to clinical and pathological information. Biopsy samples of human brain metastases (n = 49) were investigated. A significant correlation between lipid signals and necrosis in brain metastases was observed (p < 0.01), irrespective of their primary origin. The principal component analysis (PCA) showed that brain metastases from malignant melanomas cluster together, while lung carcinomas were metabolically heterogeneous and overlap with other subtypes. Metastatic melanomas have higher amounts of glycerophosphocholine than other brain metastases. A significant correlation between microscopically visible lipid droplets estimated by Nile Red staining and MR visible lipid signals was observed in metastatic lung carcinomas (p = 0.01), indicating that the proton MR visible lipid signals arise from cytoplasmic lipid droplets. MRS-based metabolomic profiling is a useful tool for exploring the metabolic profiles of metastatic brain tumors.

The effect of paramagnetic contrast in choline peak in patients with glioblastoma multiforme might not be significant

BACKGROUND AND PURPOSE

(1)H-MR spectroscopy is a useful tool in brain tumor evaluation. A critical point in obtaining representative spectra is the correct voxel positioning, which can be more accurate after Gd administration. Some experimental data suggested that Gd could cause Cho signal loss. Our aim was to evaluate the effect of Gd in the Cho peak area and width in patients with GBM.

MATERIALS AND METHODS

We performed multivoxel (1)H-MR spectroscopy before and after Gd administration in 18 patients with GBM. Quantification of Cho peak area and width in each voxel was completed, and the Cho mean and maximum values before and after Gd injection were calculated in the tumor and contralateral hemisphere. Choline peak area and width values obtained before and after contrast were compared, considering as separate entities enhancing and nonenhancing tumoral voxels and the contralateral hemisphere.

RESULTS

No statistically significant differences were found for the Cho peak area mean values in the tumoral voxels or contralaterally (P > .05). A tendency for an increase in the Cho peak width mean value was found in the tumoral enhancing voxels (P = .055). A statistically significant decrease was found for the mean value of the maximum Cho peak area in enhancing tumoral voxels (P = .020). No significant differences were found in the nonenhancing tumoral voxels or contralaterally (P > .05).

CONCLUSIONS

The injection of Gd before performing (1)H-MR spectroscopy might not significantly affect the Cho peak area in patients with GBM. The paramagnetic contrast seems to cause a different effect, depending on Gd enhancement.

Functional imaging in adult and paediatric brain tumours

Imaging is a key component in the management of brain tumours, with MRI being the preferred modality for most clinical scenarios. However, although conventional MRI provides mainly structural information, such as tumour size and location, it leaves many important clinical questions, such as tumour type, aggressiveness and prognosis, unanswered. An increasing number of studies have shown that additional information can be obtained using functional imaging methods (which probe tissue properties), and that these techniques can give key information of clinical importance. These techniques include diffusion imaging, which can assess tissue structure, and perfusion imaging and magnetic resonance spectroscopy, which measures tissue metabolite profiles. Tumour metabolism can also be investigated using PET, with 18F-deoxyglucose being the most readily available tracer. This Review discusses these methods and the studies that have investigated their clinical use. A strong emphasis is placed on the measurement of quantitative parameters, which is a move away from the qualitative nature of conventional radiological reporting and presents major challenges, particularly for multicentre studies.

Magnetic Resonance Spectroscopy for Differentiating Pseudo-Progression from True Progression in GBM on Concurrent Chemoradiotherapy

Neoadjuvant chemo-radiation therapy including temozolomide is commonly used for the treatment of gliomas. However, increased lesion size and contrast enhancement are frequently observed following this therapy and this appearance is termed as 'pseudo-progression'. Since conventional imaging is unable to differentiate pseudo-progression from tumour recurrence, we evaluated the utility of MR spectroscopy (MRS) to differentiate these two pathological entities. Longitudinal MRI and MRS studies prior to and within four months post chemo-radiation therapy including diffusion-weighted imaging and single voxel spectroscopy (short and intermediate echo) were performed in 62 glioblastoma (GBM) patients undergoing chemo-radiation therapy. Clinical follow-up demonstrated four cases of pseudo-progression. In this study, results from these four cases and a known case of tumour recurrence are reported. Metabolite ratios and presence or absence of lipids at 1.3 ppm were used to differentiate between pseudo-progression and tumour recurrence. All four cases of pseudo-progression demonstrated elevated lipid signals on MRS. Additionally, an absence of choline or a low choline/NAA ratio was also observed. In comparison, the patient with tumour recurrence showed lower lipid signals and a high choline/NAA ratio. The presence of elevated lipid signals along with low choline/NAA ratios can aid in differentiation of pseudo-progression from tumour recurrence.

Convex non-negative matrix factorization for brain tumor delimitation from MRSI data

Background

Pattern Recognition techniques can provide invaluable insights in the field of neuro-oncology. This is because the clinical analysis of brain tumors requires the use of non-invasive methods that generate complex data in electronic format. Magnetic Resonance (MR), in the modalities of spectroscopy (MRS) and spectroscopic imaging (MRSI), has been widely applied to this purpose. The heterogeneity of the tissue in the brain volumes analyzed by MR remains a challenge in terms of pathological area delimitation.

Methodology/Principal Findings

A pre-clinical study was carried out using seven brain tumor-bearing mice. Imaging and spectroscopy information was acquired from the brain tissue. A methodology is proposed to extract tissue type-specific sources from these signals by applying Convex Non-negative Matrix Factorization (Convex-NMF). Its suitability for the delimitation of pathological brain area from MRSI is experimentally confirmed by comparing the images obtained with its application to selected target regions, and to the gold standard of registered histopathology data. The former showed good accuracy for the solid tumor region (proliferation index (PI)>30%). The latter yielded (i) high sensitivity and specificity in most cases, (ii) acquisition conditions for safe thresholds in tumor and non-tumor regions (PI>30% for solid tumoral region; ≤5% for non-tumor), and (iii) fairly good results when borderline pixels were considered.

Conclusions/Significance

The unsupervised nature of Convex-NMF, which does not use prior information regarding the tumor area for its delimitation, places this approach one step ahead of classical label-requiring supervised methods for discrimination between tissue types, minimizing the negative effect of using mislabeled voxels. Convex-NMF also relaxes the non-negativity constraints on the observed data, which allows for a natural representation of the MRSI signal. This should help radiologists to accurately tackle one of the main sources of uncertainty in the clinical management of brain tumors, which is the difficulty of appropriately delimiting the pathological area.

Specific biomarkers of receptors, pathways of inhibition and targeted therapies: clinical applications

A deeper understanding of the role of specific genes, proteins, pathways and networks in health and disease, coupled with the development of technologies to assay these molecules and pathways in patients, promises to revolutionise the practice of clinical medicine. In particular, the discovery and development of novel drugs targeted to disease-specific alterations could benefit significantly from non-invasive imaging techniques assessing the dynamics of specific disease-related parameters. Here we review the application of imaging biomarkers in the management of patients with brain tumours, especially malignant glioma. This first part of the review focuses on imaging biomarkers of general biochemical and physiological processes related to tumour growth such as energy, protein, DNA and membrane metabolism, vascular function, hypoxia and cell death. These imaging biomarkers are an integral part of current clinical practice in the management of primary brain tumours. The second article of the review discusses the use of imaging biomarkers of specific disease-related molecular genetic alterations such as apoptosis, angiogenesis, cell membrane receptors and signalling pathways. Current applications of these biomarkers are mostly confined to experimental small animal research to develop and validate these novel imaging strategies with future extrapolation in the clinical setting as the primary objective.

Imaging hypoxia in gliomas

Hypoxia plays a central role in tumour development, angiogenesis, growth and resistance to treatment. Owing to constant developments in medical imaging technology, significant advances have been made towards in vitro and in vivo imaging of hypoxia in a variety of tumours, including gliomas of the central nervous system. The aim of this article is to review the literature on imaging approaches currently available for measuring hypoxia in human gliomas and provide an insight into recent advances and future directions in this field. After a brief overview of hypoxia and its importance in gliomas, several methods of measuring hypoxia will be presented. These range from invasive monitoring by Eppendorf polarographic O(2) microelectrodes, positron electron tomography (PET) tracers based on 2-nitroimidazole compounds [(18)F-labelled fluoro-misonidazole ((18)F-MISO) or 1-(2-[((18))F]fluoro-1-[hydroxymethyl]ethoxy)methyl-2-nitroimidazole (FRP-170)], (64)Cu-ATSM Cu-diacetyl-bis(N4-methylthiosemicarbazone) (Cu-ATSM) or (99m)Tc- and (68)Ga-labelled metronidazole (MN) agents to advanced MRI methods, such as blood oxygenation level dependent (BOLD) MRI, oxygen-enhanced MRI, diffusion-weighted MRI (DWI-MRI), dynamic contrast-enhanced MRI (DCE-MRI) and (1)H-magnetic resonance spectroscopy.

Quantifying the CDK inhibitor VMY-1-103's activity and tissue levels in an in vivo tumor model by LC-MS/MS and by MRI

The development of new small molecule-based therapeutic drugs requires accurate quantification of drug bioavailability, biological activity and treatment efficacy. Rapidly measuring these endpoints is often hampered by the lack of efficient assay platforms with high sensitivity and specificity. Using an in vivo model system, we report a simple and sensitive liquid chromatography-tandem mass spectrometry assay to quantify the bioavailability of a recently developed novel cyclin-dependent kinase inhibitor VMY-1-103, a purvalanol B-based analog whose biological activity is enhanced via dansylation. We developed a rapid organic phase extraction technique and validated wide and functional VMY-1-103 distribution in various mouse tissues, consistent with its enhanced potency previously observed in a variety of human cancer cell lines. More importantly, in vivo MRI and single voxel proton MR-Spectroscopy further established that VMY-1-103 inhibited disease progression and affected key metabolites in a mouse model of hedgehog-driven medulloblastoma.

Classification of single-voxel 1H spectra of childhood cerebellar tumors using LCModel and whole tissue representations

In this study, mean tumor spectra are used as the basis functions in LCModel to create a direct classification tool for short echo time (1)H magnetic resonance spectroscopy of pediatric brain tumors. LCModel is a widely used analysis tool designed to fit a linear combination of individual metabolite spectra to in vivo spectra. Here, we have used LCModel to fit mean spectra and corresponding variability components of childhood cerebellar tumors, as calculated using principal component analysis, and assessed for classification accuracy. Classification was performed according to the highest estimated tumor proportion. This method was tested in a leave-one-out analysis discriminating between pediatric brain tumor spectra of medulloblastoma vs. pilocytic astrocytoma and medulloblastoma vs. pilocytic astrocytoma vs. ependymoma. Additionally, the effect of accepting different Cramér-Rao Lower Bound cut-off criteria on classification accuracy and estimated tissue proportions was investigated. The best classification results differentiating medulloblastoma vs. pilocytic astrocytoma and medulloblastoma vs. pilocytic astrocytoma vs. ependymoma were 100 and 87.7%, respectively. These results are comparable to a specialized pattern recognition analysis of this data set and give easy to interpret results in the form of estimated tissue proportions. The method requires minimal user input and is easily transferable across sites and to other magnetic resonance spectroscopy classification problems.

Spectroscopic evaluation of glioma grading at 3T: the combined role of short and long TE

Purpose. To evaluate the diagnostic value of 3T ¹H-MRS in grading cerebral gliomas using short and long echo times. Methods. 1H-MRS was performed on 71 patients with untreated cerebral gliomas. Metabolite ratios of NAA/Cr, Cho/Cr, Cho/NAA, and mI/Cr were calculated for short and long TE and compared between low and high grade gliomas. Lipids were qualitatively evaluated. ROC analysis was performed to obtain the cut-off values for the metabolic ratios presenting statistical difference between the two glioma grades. Results. Intratumoral Cho/Cr at both TEs and long TE Cho/NAA were significantly different between low and high grade gliomas. Peritumoral NAA/Cr of both TEs, as well as long TE Cho/Cr and Cho/NAA ratios, significantly differentiated the two tumor grades. Diagnostic sensitivity of peritumoral short TE NAA/Cr proved to be superior over the other metabolic ratios, whereas intratumoral short TE Cho/Cr reached the highest levels of specificity and accuracy. Overall, short TE 1H-MRS reached higher total sensitivity in predicting glioma grade, over long TE. Conclusion. An advantage was found in using short TE over long TE 1H-MRS in the discrimination of low versus high grade gliomas. Moreover, the results suggested that the peritumoral area of gliomas may be more valuable in predicting glioma grade than using only the intratumoral area.

Compressive sensing could accelerate 1H MR metabolic imaging in the clinic

PURPOSE

To retrospectively evaluate the fidelity of magnetic resonance (MR) spectroscopic imaging data preservation at a range of accelerations by using compressed sensing.

MATERIALS AND METHODS

The protocols were approved by the institutional review board of the university, and written informed consent to acquire and analyze MR spectroscopic imaging data was obtained from the subjects prior to the acquisitions. This study was HIPAA compliant. Retrospective application of compressed sensing was performed on 10 clinical MR spectroscopic imaging data sets, yielding 600 voxels from six normal brain data sets, 163 voxels from two brain tumor data sets, and 36 voxels from two prostate cancer data sets for analysis. The reconstructions were performed at acceleration factors of two, three, four, five, and 10 and were evaluated by using the root mean square error (RMSE) metric, metabolite maps (choline, creatine, N-acetylaspartate [NAA], and/or citrate), and statistical analysis involving a voxelwise paired t test and one-way analysis of variance for metabolite maps and ratios for comparison of the accelerated reconstruction with the original case.

RESULTS

The reconstructions showed high fidelity for accelerations up to 10 as determined by the low RMSE (< 0.05). Similar means of the metabolite intensities and hot-spot localization on metabolite maps were observed up to a factor of five, with lack of statistically significant differences compared with the original data. The metabolite ratios of choline to NAA and choline plus creatine to citrate did not show significant differences from the original data for up to an acceleration factor of five in all cases and up to that of 10 for some cases.

CONCLUSION

A reduction of acquisition time by up to 80%, with negligible loss of information as evaluated with clinically relevant metrics, has been successfully demonstrated for hydrogen 1 MR spectroscopic imaging.

In vivo characterization of several rodent glioma models by 1H MRS

The assessment of metabolites by (1)H MRS can provide information regarding glioma growth, and may be able to distinguish between different glioma models. Rat C6, 9 L/LacZ, F98 and RG2, and mouse GL261, cells were intracerebrally implanted into the respective rodents, and human U87 MG cells were implanted into athymic rats. Ethyl-nitrosourea induction was also used. Glioma metabolites [e.g. total choline (tCho), total creatine (tCr), N-acetylaspartate (NAA), lactate (Lac), glutamine (Gln), glutamate (Glu), aspartate (Asp), guanosine (Gua), mobile lipids and macromolecules (MMs)] were assessed from (1)H MRS using point-resolved spectroscopy (PRESS) [TE = 24 ms; TR = 2500 ms; variable pulse power and optimized relaxation delay (VAPOR) water suppression; 27-μL and 8-μL voxels in rats and mice, respectively] at 7 T. Alterations in metabolites (Totally Automatic Robust Quantitation in NMR, TARQUIN) in tumors were characterized by increases in lipids (Lip1.3: 8.8-54.5 mM for C6 and GL261) and decreases in NAA (1.3-2.0 mM for RG2, GL261 and C6) and tCr (0.8-4.0 mM for F98, RG2, GL261 and C6) in some models. F98, RG2, GL261 and C6 models all showed significantly decreased (p < 0.05) tCr, and RG2, GL261 and C6 models all exhibited significantly decreased (p < 0.05) NAA. The RG2 model showed significantly decreased (p < 0.05) Gln and Glu, the C6 model significantly decreased (p < 0.05) Asp, and the F98 and U87 models significantly decreased (p < 0.05) Gua, compared with controls. The GL261 model showed the greatest alterations in metabolites. (1)H MRS was able to differentiate the metabolic profiles in many of the seven rodent glioma models assessed. These models are considered to resemble certain characteristics of human glioblastomas, and this study may be helpful in selecting appropriate models.

Magnetic resonance spectroscopy of cancer metabolism and response to therapy

Magnetic resonance spectroscopy allows noninvasive in vivo measurements of biochemical information from living systems, ranging from cultured cells through experimental animals to humans. Studies of biopsies or extracts offer deeper insights by detecting more metabolites and resolving metabolites that cannot be distinguished in vivo. The pharmacokinetics of certain drugs, especially fluorinated drugs, can be directly measured in vivo. This review briefly describes these methods and their applications to cancer metabolism, including glycolysis, hypoxia, bioenergetics, tumor pH, and tumor responses to radiotherapy and chemotherapy.

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).

Non-negative matrix factorisation methods for the spectral decomposition of MRS data from human brain tumours

BACKGROUND

In-vivo single voxel proton magnetic resonance spectroscopy (SV 1H-MRS), coupled with supervised pattern recognition (PR) methods, has been widely used in clinical studies of discrimination of brain tumour types and follow-up of patients bearing abnormal brain masses. SV 1H-MRS provides useful biochemical information about the metabolic state of tumours and can be performed at short (< 45 ms) or long (> 45 ms) echo time (TE), each with particular advantages. Short-TE spectra are more adequate for detecting lipids, while the long-TE provides a much flatter signal baseline in between peaks but also negative signals for metabolites such as lactate. Both, lipids and lactate, are respectively indicative of specific metabolic processes taking place. Ideally, the information provided by both TE should be of use for clinical purposes. In this study, we characterise the performance of a range of Non-negative Matrix Factorisation (NMF) methods in two respects: first, to derive sources correlated with the mean spectra of known tissue types (tumours and normal tissue); second, taking the best performing NMF method for source separation, we compare its accuracy for class assignment when using the mixing matrix directly as a basis for classification, as against using the method for dimensionality reduction (DR). For this, we used SV 1H-MRS data with positive and negative peaks, from a widely tested SV 1H-MRS human brain tumour database.

RESULTS

The results reported in this paper reveal the advantage of using a recently described variant of NMF, namely Convex-NMF, as an unsupervised method of source extraction from SV1H-MRS. Most of the sources extracted in our experiments closely correspond to the mean spectra of some of the analysed tumour types. This similarity allows accurate diagnostic predictions to be made both in fully unsupervised mode and using Convex-NMF as a DR step previous to standard supervised classification. The obtained results are comparable to, or more accurate than those obtained with supervised techniques.

CONCLUSIONS

The unsupervised properties of Convex-NMF place this approach one step ahead of classical label-requiring supervised methods for the discrimination of brain tumour types, as it accounts for their increasingly recognised molecular subtype heterogeneity. The application of Convex-NMF in computer assisted decision support systems is expected to facilitate further improvements in the uptake of MRS-derived information by clinicians.

Spatial characteristics of newly diagnosed grade 3 glioma assessed by magnetic resonance metabolic and diffusion tensor imaging

The spatial heterogeneity in magnetic resonance (MR) metabolic and diffusion parameters and their relationship were studied for patients with treatment-naive grade 3 gliomas. MR data were evaluated from 51 patients with newly diagnosed grade 3 gliomas. Anatomic, diffusion, and metabolic imaging data were considered. Variations in metabolite levels, apparent diffusion coefficient (ADC), and fractional anisotropy (FA) were evaluated in regions of gadolinium enhancement and T2 hyperintensity as well as regions with abnormal metabolic signatures. Contrast enhancement was present in only 21 of the 51 patients. When present, the enhancing component of the lesion had higher choline-to-N-acetylaspartate index (CNI), higher choline, lower N-acetylaspartate, similar creatine, similar ADC and FA, and higher lactate/lipid than the nonenhancing lesion. Regions with CNI ≥ 4 had higher choline, lower N-acetylaspartate, higher lactate/lipid, higher ADC, and lower FA than normal-appearing white matter and regions with intermediate CNI values. For lesions that exhibited gadolinium enhancement, the metabolite levels and diffusion parameters in the region of enhancement were consistent with it corresponding to the most abnormal portion of the tumor. For nonenhancing lesions, areas with CNI ≥ 4 were the most abnormal in metabolic and diffusion parameters. This suggests that the region with the highest CNI might provide a good target for biopsies for nonenhancing lesions to obtain a representative histologic diagnosis of its degree of malignancy. Metabolic and diffusion parameter levels may be of interest not only for directing tissue sampling but also for defining the targets for focal therapy and assessing response to therapy.

In vivo 1H MRS for musculoskeletal lesion characterization: which factors affect diagnostic accuracy?

In vivo (1)H MRS is a noninvasive imaging technique for the identification of malignancy. Musculoskeletal lesions vary in their composition, causing field inhomogeneity and magnetic susceptibility effects which may be technical and diagnostic challenges for MRS. This study investigated the factors that affect diagnostic accuracy in the use of MRS for the characterization of musculoskeletal neoplasms. During a 7-year period, 210 consecutive patients with musculoskeletal lesions larger than 1.5 cm in diameter were examined. MRS of a single-voxel point-resolved spectroscopy sequence with TE = 135 ms was undertaken using a 1.5-T scanner. Lesions with a choline signal-to-noise ratio larger than 3.0 were considered to be malignant tumors. The diagnostic accuracy was calculated for all lesions and for subgroups on the basis of lesion type (bone and soft tissue), lesion composition (mixed and solid nonsclerotic), lesion size (≤4, >4-10 and >10 cm), MR scanner (MR scanner 1 and 2) and selected voxel size (≤3, >3-8 and >8 cm(3)). Multivariate logistic regressions were performed to estimate the associations between each factor and diagnostic accuracy. The diagnostic accuracy was 73.3% for all lesions. The accuracy was 54.4% for mixed lesions and 80.4% for solid nonsclerotic lesions (p < 0.001). The diagnostic accuracy was lower for larger lesions [86.8% for lesions of ≤4 cm, 71.6% for lesions of >4-10 cm (p = 0.04) and 63.6% for lesions of >10 cm (p = 0.007)]. There was no difference in diagnostic accuracy for bone versus soft-tissue lesions or as a function of MR scanner or voxel size. By the use of multivariate logistic regression, a solid nonsclerotic lesion was 3.15 times (95% confidence interval, 1.59-6.27) more likely than a mixed lesion to have a diagnosis (p = 0.001). MRS can be used to characterize musculoskeletal lesions, particularly solid nonsclerotic lesions.

Short echo time single voxel 1H magnetic resonance spectroscopy in the diagnosis and characterisation of pineal tumours in children

BACKGROUND

Magnetic resonance spectroscopy (MRS) has been successful in characterising a range of brain tumours and is a useful aid to non-invasive diagnosis. The pineal region poses considerable surgical challenges and a major surgical resection is not required in the management of all tumours. Improved non-invasive assessment of pineal region tumours would be of considerable benefit.

METHODS

Single voxel MRS (TE 30 ms, TR 1500, 1.5 T) was performed on 15 pineal tumours: 5 germinomas, 1 non-germinomatous secreting germ cell tumour (GCT), 2 teratomas, 5 pineoblastomas, 1 pineal parenchymal tumour (PPT) of intermediate differentiation and 1 pineocytoma. Two germinomas outside the pineal gland were also studied. Metabolite, lipid and macromolecule concentrations were determined with LCModel™.

RESULTS

Germ cell tumours had significantly higher lipid and macromolecule concentrations than other tumours (t-test; P < 0.05). The teratomas had significantly lower total choline and creatine levels than germinomas (z test; P < 0.05). Taurine was convincingly detected in germinomas as well as PPTs.

CONCLUSIONS

Magnetic resonance spectroscopy is useful for characterising pineal region tumours, aiding the non-invasive diagnosis and giving additional biological insight.

Influence of common preanalytical variations on the metabolic profile of serum samples in biobanks

A blood pre-centrifugation delay of 24 h at room temperature influenced the proton NMR spectroscopic profiles of human serum. A blood pre-centrifugation delay of 24 h at 4°C did not influence the spectroscopic profile as compared with 4 h delays at either room temperature or 4°C. Five or ten serum freeze-thaw cycles also influenced the proton NMR spectroscopic profiles. Certain common in vitro preanalytical variations occurring in biobanks may impact the metabolic profile of human serum.

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

New biomedical technologies enable the diagnosis of brain tumours by using non-invasive methods. HealthAgents is a European Union-funded research project that aims to build an agent-based distributed decision support system (dDSS) for the diagnosis of brain tumours. This is achieved using the latest biomedical knowledge, information and communication technologies and pattern recognition (PR) techniques. As part of the PR development of HealthAgents, an independent and automatic classification framework (CF) has been developed. This framework has been integrated with the HealthAgents dDSS using the HealthAgents agent platform. The system offers (1) the functionality to search for distributed classifiers to solve specific questions; (2) automatic classification of new cases; (3) instant deployment of new validated classifiers; and (4) the ability to rank a set of classifiers according to their performance and suitability for the case in hand. The CF enables both the deployment of new classifiers using the provided Extensible Markup Language1 classifier specification, and the inclusion of new PR techniques that make the system extensible. These features may enable the rapid integration of PR laboratory results into industrial or research applications, such as the HealthAgents dDSS. Two classification nodes have been deployed and they currently offer classification services by means of dedicated servers connected to the HealthAgents agent platform: one node being located at the Katholieke Universiteit Leuven, Belgium and the other at the Universidad Politécnica de Valencia, Spain. These classification nodes share the current set of brain tumour classifiers that have been trained from in vivo magnetic resonance spectroscopy data. The combination of the CF with a distributed agent system constitutes the basis of the brain tumour dDSS developed in HealthAgents.

The development of a graphical user interface, functional elements and classifiers for the non-invasive characterization of childhood brain tumours using magnetic resonance spectroscopy

Magnetic resonance spectroscopy (MRS) is a non-invasive method, which can provide diagnostic information on children with brain tumours. The technique has not been widely used in clinical practice, partly because of the difficulty of developing robust classifiers from small patient numbers and the challenge of providing decision support systems (DSSs) acceptable to clinicians. This paper describes a participatory design approach in the development of an interactive clinical user interface, as part of a distributed DSS for the diagnosis and prognosis of brain tumours. In particular, we consider the clinical need and context of developing interactive elements for an interface that facilitates the classification of childhood brain tumours, for diagnostic purposes, as part of the HealthAgents European Union project. Previous MRS-based DSS tools have required little input from the clinician user and a raw spectrum is essentially processed to provide a diagnosis sometimes with an estimate of error. In childhood brain tumour diagnosis where there are small numbers of cases and a large number of potential diagnoses, this approach becomes intractable. The involvement of clinicians directly in the designing of the DSS for brain tumour diagnosis from MRS led to an alternative approach with the creation of a flexible DSS that, allows the clinician to input prior information to create the most relevant differential diagnosis for the DSS. This approach mirrors that which is currently taken by clinicians and removes many sources of potential error. The validity of this strategy was confirmed for a small cohort of children with cerebellar tumours by combining two diagnostic types, pilocytic astrocytomas (11 cases) and ependymomas (four cases) into a class of glial tumours which then had similar numbers to the other diagnostic type, medulloblastomas (18 cases). Principal component analysis followed by linear discriminant analysis on magnetic resonance spectral data gave a classification accuracy of 91% for a three-class classifier and 94% for a two-class classifier using a leave-one-out analysis. This DSS provides a flexible method for the clinician to use MRS for brain tumour diagnosis in children.

Five-Year Longitudinal MRI Follow-up and (1)H Single Voxel MRS in 14 patients with Gliomatosis Treated with Temodal, Radiotherapy and Antiangiogenic Therapy

Gliomatosis cerebri (GC) is a challenging tumor, considered to have a poor prognosis and poor response to treatments. The purpose of this study is to better understand glial tumor metabolism and post chemotherapy, radiotherapy and antiangiogenic variations in a longitudinal study to determine cerebral variation in MRS area, amplitude, and ratios of metabolites and spectral profiles during a five year longitudinal follow-up in 14 patients with gliomatosis without initial hyperperfusion and treated with chemotherapy (Temozolomide (Temodal(®))), radiotherapy and subsequent antiangiogenic therapy. The study also aimed to detect changes in infiltration, proliferation, lipids or glycolytic metabolism, as these changes could be monitored longitudinally in humans with glial brain tumors (low and high grade) after therapy, using conventional magnetic resonance imaging (MRI), spectroscopy (MRS) and MR perfusion. Most patients had first initial clinical and MRS improvement and stable MRI. After 12 to 24 chemotherapy treatment cycles MRS usually showed an increase in the Cho/Cr ratio (proliferation) and sometimes contrast enhancements. Later, the patients showed clinical deterioration and radiotherapy was started. There was an improvement with radiotherapy that lasted nine to 18 months. This was followed by a worsening that led to try antiangiogenic therapy. Later in the evolution for three patients with hyperperfusion this symptom disappeared, but proliferation, infiltration and glycolytic metabolism remained at a high level. Spectroscopic and metabolic changes often occur well before clinical deterioration and sometimes before improvement. Therefore, MRS could be more sensitive and could detect changes earlier than MRI and is sometimes predictive. Despite the difficulty, the variability and unknown factors, these repeated measurements give us a better insight into the nature of the different processes, tumor progression and could lead to better understanding of therapeutic response.

Serial magnetic resonance spectroscopy reveals a direct metabolic effect of cediranib in glioblastoma

Proton magnetic resonance spectroscopy is increasingly used in clinical studies of brain tumor to provide information about tissue metabolic profiles. In this study, we evaluated changes in the levels of metabolites predominant in recurrent glioblastoma multiforme (rGBM) to characterize the response of rGBM to antiangiogenic therapy. We examined 31 rGBM patients treated with daily doses of cediranib, acquiring serial chemical shift imaging data at specific time points during the treatment regimen. We defined spectra from three regions of interest (ROI)--enhancing tumor (ET), peritumoral tissue, and normal tissue on the contralateral side (cNT)--in post-contrast T1-weighted images, and normalized the concentrations of N-acetylaspartate (NAA) and choline (Cho) in each ROI to the concentration of creatine in cNT (norCre). We analyzed the ratios of these normalized metabolites (i.e., NAA/Cho, NAA/norCre, and Cho/norCre) by averaging all patients and categorizing two different survival groups. Relative to pretreatment values, NAA/Cho in ET was unchanged through day 28. However, after day 28, NAA/Cho significantly increased in relation to a significant increase in NAA/norCre and a decrease in Cho/norCre; interestingly, the observed trend was reversed after day 56, consistent with the clinical course of GBM recurrence. Notably, receiver operating characteristic analysis indicated that NAA/Cho in tumor shows a high prediction to 6-month overall survival. These metabolic changes in these rGBM patients strongly suggest a direct metabolic effect of cediranib and might also reflect an antitumor response to antiangiogenic treatment during the first 2 months of treatment. Further study is needed to confirm these findings.

Strengths and weaknesses of 1.5T and 3T MRS data in brain glioma classification

Although magnetic resonance spectroscopy (MRS) methods of 1.5Tesla (T) and 3T have been widely applied during the last decade for noninvasive diagnostic purposes, only a few studies have been reported on the value of the information extracted in brain cancer discrimination. The purpose of this study is threefold. First, to show that the diagnostic value of the information extracted from two different MRS scanners of 1.5T and 3T is significantly influenced in terms of brain gliomas discrimination. Second, to statistically evaluate the discriminative potential of publicly known metabolic ratio markers, obtained from these two types of scanners in classifying low-, intermediate-, and high-grade gliomas. Finally, to examine the diagnostic value of new metabolic ratios in the discrimination of complex glioma cases where the diagnosis is both challenging and critical. Our analysis has shown that although the information extracted from 3T MRS scanner is expected to provide better brain gliomas discrimination; some factors like the features selected, the pulse-sequence parameters, and the spectroscopic data acquisition methods can influence the discrimination efficiency. Finally, it is shown that apart from the bibliographical known, new metabolic ratio features such as N-acetyl aspartate/ S, Choline/ S, Creatine/ S , and myo-Inositol/ S play significant role in gliomas grade discrimination.

Neurochemical alterations in adolescent chronic marijuana smokers: a proton MRS study

Converging evidence from neuroimaging and neuropsychological studies indicates that heavy marijuana use is associated with cingulate dysfunction. However, there has been limited human data documenting in vivo biochemical brain changes after chronic marijuana exposure. Previous proton magnetic resonance spectroscopy studies have demonstrated reduced basal ganglia glutamate and dorsolateral prefrontal cortex N-acetyl aspartate levels in adult chronic marijuana users. Similar studies have not been reported in adolescent populations. The present study used proton magnetic resonance spectroscopy to determine whether reductions in glutamate, N-acetyl aspartate and/or other proton metabolite concentrations would be found in the anterior cingulate cortex (ACC) of adolescent marijuana users compared with non-using controls. Adolescent marijuana users (N=17; average age 17.8 years) and similarly aged healthy control subjects (N=17; average age 16.2 years) were scanned using a Siemens 3T Trio MRI system. Proton magnetic resonance spectroscopy data were acquired from a 22.5 mL voxel positioned bilaterally within the ACC. Spectra were fitted using commercial software and all metabolite integrals were normalized to the scaled unsuppressed water integral. Analysis of variance and analysis of covariance were performed to compare between-group metabolite levels. The marijuana-using cohort showed statistically significant reductions in anterior cingulate glutamate (-15%, p<0.01), N-acetyl aspartate (-13%, p=0.02), total creatine (-10%, p<0.01) and myo-inositol (-10%, p=0.03). Within-voxel tissue-type segmentation did not reveal any significant differences in gray/white matter or cerebrospinal fluid content between the two groups. The reduced glutamate and N-acetyl aspartate levels in the adolescent marijuana-using cohort are consistent with precedent human (1)H MRS data, and likely reflect an alteration of anterior cingulate glutamatergic neurotransmission and neuronal integrity within these individuals. The reduced total creatine and myo-inositol levels observed in these subjects might infer altered ACC energetic status and glial metabolism, respectively. These results expand on previous functional MRI data reporting altered cingulate function in individuals with marijuana-abuse.

Imaging in primary central nervous system lymphoma

Primary central nervous system (CNS) lymphoma (PCNSL) accounts for approximately 3% of all primary CNS tumors. Congenital or acquired immunodeficiency is the only established risk factor for PCNSL. Rates decreased slightly in the mid-1990s, concordantly with the decreasing rates of AIDS. However, the incidence has been increasing in the elderly immunocompetent population, and this trend seems to be independent of improvements in diagnostic techniques, and of overall trends in the incidence of brain tumors and systemic lymphomas. This study presents our experience with the imaging features of PCNSL. Computed tomography (CT) and magnetic resonance imaging (MRI) findings were reviewed in a series of 38 cases of pathologically proven PCNSL. The incidence rate of PCNSL was higher in men than in women (58% versus 42%). Mean age at presentation was 63 years; 120 lesions were demonstrated in the 38 patients, with a 53% frequency of tumor multiplicity. Both CT and MR mainly showed solitary or multiple well-defined round or oval-shaped mass lesions, typically hyperdense on unenhanced CT scans, iso to hypointense on T2 MR weighted images. These lesions also showed an increased signal intensity on diffusion-weighted images. Virtually all lesions enhanced after intravenous administration of contrastmedium. On (1)H-magnetic resonance spectroscopy ((1)H-MRS) most lesions presented increased Cho/Cr, Cho/NAA and lactate/Cr ratios when compared to normal gray matter. No changes in the imaging presentation have occurred over the past two decades, apart from lesions now being smaller at diagnosis. Our imaging findings are in agreement with the existing literature data and with the reported increasing trend of multifocal tumors. Our epidemiologic results add value to the existing evidence of increasing incidence rates among the immunocompetent elderly population.

Visualization and recovery of the (bio)chemical interesting variables in data analysis with support vector machine classification

Support vector machines (SVMs) have become a popular technique in the chemometrics and bioinformatics field, and other fields, for the classification of complex data sets. Especially because SVMs are able to model nonlinear relationships, the usage of this technique has increased substantially. This modeling is obtained by mapping the data in a higher-dimensional feature space. The disadvantage of such a transformation is, however, that information about the contribution of the original variables in the classification is lost. In this paper we introduce an innovative method which can retrieve the information about the variables of complex data sets. We apply the proposed method to several benchmark data sets and a metabolomics data set to illustrate that we can determine the contribution of the original variables in SVM classifications. The corresponding visualization of the contribution of the variables can assist in a better understanding of the underlying chemical or biological process.

The INTERPRET Decision-Support System version 3.0 for evaluation of Magnetic Resonance Spectroscopy data from human brain tumours and other abnormal brain masses

BACKGROUND

Proton Magnetic Resonance (MR) Spectroscopy (MRS) is a widely available technique for those clinical centres equipped with MR scanners. Unlike the rest of MR-based techniques, MRS yields not images but spectra of metabolites in the tissues. In pathological situations, the MRS profile changes and this has been particularly described for brain tumours. However, radiologists are frequently not familiar to the interpretation of MRS data and for this reason, the usefulness of decision-support systems (DSS) in MRS data analysis has been explored.

RESULTS

This work presents the INTERPRET DSS version 3.0, analysing the improvements made from its first release in 2002. Version 3.0 is aimed to be a program that 1st, can be easily used with any new case from any MR scanner manufacturer and 2nd, improves the initial analysis capabilities of the first version. The main improvements are an embedded database, user accounts, more diagnostic discrimination capabilities and the possibility to analyse data acquired under additional data acquisition conditions. Other improvements include a customisable graphical user interface (GUI). Most diagnostic problems included have been addressed through a pattern-recognition based approach, in which classifiers based on linear discriminant analysis (LDA) were trained and tested.

CONCLUSIONS

The INTERPRET DSS 3.0 allows radiologists, medical physicists, biochemists or, generally speaking, any person with a minimum knowledge of what an MR spectrum is, to enter their own SV raw data, acquired at 1.5 T, and to analyse them. The system is expected to help in the categorisation of MR Spectra from abnormal brain masses.

Combined use of neuroradiology and 1H-MR spectroscopy may provide an intervention limiting diagnosis of glioblastoma multiforme

PURPOSE

To evaluate the accuracy of (1)H-MR spectroscopy ((1)H-MRS) as an intervention limiting diagnostic tool for glioblastoma multiforme. GBM is the most common and aggressive primary brain tumor, with mean survival under a year. Oncological practice currently requires histopathological diagnosis before radiotherapy.

MATERIALS AND METHODS

Eighty-nine patients had clinical computed tomography (CT) and MR imaging and 1.5T SV SE (1)H-MRS with PRESS localization for neuroradiological diagnosis and tumor classification with spectroscopic and automated pattern recognition analysis (TE 30 ms, TR 2000 ms, spectral width 2500 Hz and 2048 data points, 128-256 signal averages were acquired, depending on voxel size (8 cm(3) to 4 cm(3)). Eighteen patients from a cohort of 89 underwent stereotactic biopsy.

RESULTS

The 18 stereotactic biopsies revealed 14 GBM, 2 grade II astrocytomas, 1 lymphoma, and 1 anaplastic astrocytoma. All 14 biopsied GBMs were diagnosed as GBM by a protocol combining an individual radiologist and an automated spectral pattern recognition program.

CONCLUSION

In patients undergoing stereotactic biopsy combined neuroradiological and spectroscopic evaluation diagnoses GBM with accuracy that could replace the need for biopsy. We do not advocate the replacement of biopsy in all patients; instead our data suggest a specific intervention limiting role for the use of (1)H-MRS in brain tumor diagnosis.

Measurements of diagnostic examination performance using quantitative apparent diffusion coefficient and proton MR spectroscopic imaging in the preoperative evaluation of tumor grade in cerebral gliomas

PURPOSE

Tumor grading is very important both in treatment decision and evaluation of prognosis. While tissue samples are obtained as part of most therapeutic approaches, factors that may result in inaccurate grading due to sampling error (namely, heterogeneity in tissue sampling, as well as tumor-grade heterogeneity within the same tumor specimen), have led to a desire to use imaging better to ascertain tumor grade. The purpose in our study was to evaluate the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), area under the curve (AUC), and accuracy of diffusion-weighted MR imaging (DWI), proton MR spectroscopic imaging (MRSI) or both in grading primary cerebral gliomas.

MATERIALS AND METHODS

We performed conventional MR imaging (MR), DWI, and MRSI in 74 patients with newly diagnosed brain gliomas: 59 patients had histologically verified high-grade gliomas: 37 glioblastomas multiform (GBM) and 22 anaplastic astrocytomas (AA), and 15 patients had low-grade gliomas. Apparent diffusion coefficient (ADC) values of tumor and peritumoral edema, and ADC ratios (ADC in tumor or peritumoral edema to ADC of contralateral white matter, as well as ADC in tumor to ADC in peritumoral edema) were determined from three regions of interest. The average of the mean, maximum, and minimum for ADC variables was calculated for each patient. The metabolite ratios of Cho/Cr and Cho/NAA at intermediate TE were assessed from spectral maps in the solid portion of tumor, peritumoral edema and contralateral normal-appearing white matter. Tumor grade determined with the two methods was then compared with that from histopathologic grading. Logistic regression and receiver operating characteristic (ROC) curve analysis were performed to determine optimum thresholds for tumor grading. Measures of diagnostic examination performance, such as sensitivity, specificity, PPV, NPV, AUC, and accuracy for identifying high-grade gliomas were also calculated.

RESULTS

Statistical analysis demonstrated a threshold minimum ADC tumor value of 1.07 to provide sensitivity, specificity, PPV, and NPV of 79.7%, 60.0%, 88.7%, and 42.9% respectively, in determining high-grade gliomas. Threshold values of 1.35 and 1.78 for peritumoral Cho/Cr and Cho/NAA metabolite ratios resulted in sensitivity, specificity, PPV, and NPV of 83.3%, 85.1%, 41.7%, 97.6%, and 100%, 57.4%, 23.1% and 100% respectively for determining high-grade gliomas. Significant differences were noted in the ADC tumor values and ratios, peritumoral Cho/Cr and Cho/NAA metabolite ratios, and tumoral Cho/NAA ratio between low- and high-grade gliomas. The combination of mean ADC tumor value, maximum ADC tumor ratio, peritumoral Cho/Cr and Cho/NAA metabolite ratios resulted in sensitivity, specificity, PPV, and NPV of 91.5%, 100%, 100% and 60% respectively.

CONCLUSION

Combining DWI and MRSI increases the accuracy of preoperative imaging in the determination of glioma grade. MRSI had superior diagnostic performance in predicting glioma grade compared with DWI alone. The predictive values are helpful in the clinical decision-making process to evaluate the histologic grade of tumors, and provide a means of guiding treatment.

Molecular imaging of cancer: MR spectroscopy and beyond

Proton magnetic resonance spectroscopic imaging is a non-invasive diagnostic tool for the investigation of cancer metabolism. As an adjunct to morphologic and dynamic magnetic resonance imaging, it is routinely used for the staging, assessment of treatment response, and therapy monitoring in brain, breast, and prostate cancer. Recently, its application was extended to other cancerous diseases, such as malignant soft-tissue tumours, gastrointestinal and gynecological cancers, as well as nodal metastasis. In this review, we discuss the current and evolving clinical applications of proton magnetic resonance spectroscopic imaging. In addition, we will briefly discuss other evolving techniques, such as phosphorus magnetic resonance spectroscopic imaging, sodium imaging and diffusion-weighted imaging in cancer assessment.

Ex-vivo HRMAS of adult brain tumours: metabolite quantification and assignment of tumour biomarkers

BACKGROUND

High-resolution magic angle spinning (HRMAS) NMR spectroscopy allows detailed metabolic analysis of whole biopsy samples for investigating tumour biology and tumour classification. Accurate biochemical assignment of small molecule metabolites that are "NMR visible" will improve our interpretation of HRMAS data and the translation of NMR tumour biomarkers to in-vivo studies.

RESULTS

1D and 2D 1H HRMAS NMR was used to determine that 29 small molecule metabolites, along with 8 macromolecule signals, account for the majority of the HRMAS spectrum of the main types of brain tumour (astrocytoma grade II, grade III gliomas, glioblastomas, metastases, meningiomas and also lymphomas). Differences in concentration of 20 of these metabolites were statistically significant between these brain tumour types. During the course of an extended 2D data acquisition the HRMAS technique itself affects sample analysis: glycine, glutathione and glycerophosphocholine all showed small concentration changes; analysis of the sample after HRMAS indicated structural damage that may affect subsequent histopathological analysis.

CONCLUSIONS

A number of small molecule metabolites have been identified as potential biomarkers of tumour type that may enable development of more selective in-vivo 1H NMR acquisition methods for diagnosis and prognosis of brain tumours.

Short echo time MR spectroscopy of brain tumors: grading of cerebral gliomas by correlation analysis of normalized spectral amplitudes

PURPOSE

To process single voxel spectra of low- and high-grade gliomas. To propose correlation analysis of the scatter plots of normalized spectral amplitudes as a pattern recognition tool for the classification (grading) of brain tumors. To propose a spectrum processing approach that improves the differentiation of proton spectra with dominating macromolecule and lipid peaks.

MATERIALS AND METHODS

LCModel was used to process spectra. Mean metabolite concentrations and mean normalized spectra were obtained for normal white matter and for gliomas. The mean spectra of macromolecules and lipids (ML) in the range 1.4-0.9 ppm, and mean difference spectra (DS) without ML and lactate were computed. Correlation analysis of the scatter plot of the patient and mean normalized spectral amplitudes and dispersion of the scatter plot points were used for classification and grading of tumors.

RESULTS

It was found advantageous to perform the classifications using DS spectra. The shape of ML spectrum and concentration of tCr seem to be a good markers for glioma grade.

CONCLUSION

Combining a qualitative comparison of the patient and mean DS spectra of the tumors using correlation analysis of normalized spectra amplitudes with a quantitative comparison of metabolite concentrations is a powerful tool in studying brain lesions.

1H-MRSI pattern perturbation in a mouse glioma: the effects of acute hyperglycemia and moderate hypothermia

MR spectroscopic Imaging (MRSI), with PRESS localization, is used here to monitor the effects of acute hyperglycemia in the spectral pattern of 11 mice bearing GL261 gliomas at normothermia (36.5-37.5 degrees C) and at hypothermia (28.5-29.5 degrees C). These in vivo studies were complemented by ex vivo high resolution magic angle spinning (HR-MAS) analysis of GL261 tumor samples from 6 animals sacrificed by focused microwave irradiation, and blood glucose measurements in 12 control mice. Apparent glucose levels, monitored by in vivo MRSI in brain tumors during acute hyperglycemia, rose to an average of 1.6-fold during hypothermia (p < 0.05), while no significant changes were detected at normothermia, or in control experiments performed at euglycemia, or in normal/peritumoral brain regions. Ex vivo analysis of glioma-bearing mouse brains at hypothermia revealed higher glucose increases in distinct regions during the acute hyperglycemic challenge (up to 6.6-fold at the tumor center), in agreement with maximal in vivo blood glucose changes (5-fold). Phantom studies on taurine plus glucose containing solutions explained the differences between in vivo and ex vivo measurements. Our results also indicate brain tumor heterogeneity in the four animal tumors investigated in response to a defined metabolic challenge.

Histopathological image analysis: a review

Over the past decade, dramatic increases in computational power and improvement in image analysis algorithms have allowed the development of powerful computer-assisted analytical approaches to radiological data. With the recent advent of whole slide digital scanners, tissue histopathology slides can now be digitized and stored in digital image form. Consequently, digitized tissue histopathology has now become amenable to the application of computerized image analysis and machine learning techniques. Analogous to the role of computer-assisted diagnosis (CAD) algorithms in medical imaging to complement the opinion of a radiologist, CAD algorithms have begun to be developed for disease detection, diagnosis, and prognosis prediction to complement the opinion of the pathologist. In this paper, we review the recent state of the art CAD technology for digitized histopathology. This paper also briefly describes the development and application of novel image analysis technology for a few specific histopathology related problems being pursued in the United States and Europe.

Histopathological image analysis: a review

Over the past decade, dramatic increases in computational power and improvement in image analysis algorithms have allowed the development of powerful computer-assisted analytical approaches to radiological data. With the recent advent of whole slide digital scanners, tissue histopathology slides can now be digitized and stored in digital image form. Consequently, digitized tissue histopathology has now become amenable to the application of computerized image analysis and machine learning techniques. Analogous to the role of computer-assisted diagnosis (CAD) algorithms in medical imaging to complement the opinion of a radiologist, CAD algorithms have begun to be developed for disease detection, diagnosis, and prognosis prediction to complement the opinion of the pathologist. In this paper, we review the recent state of the art CAD technology for digitized histopathology. This paper also briefly describes the development and application of novel image analysis technology for a few specific histopathology related problems being pursued in the United States and Europe.

A comparative study of feature extraction and blind source separation of independent component analysis (ICA) on childhood brain tumour 1H magnetic resonance spectra

Independent component analysis (ICA) has the potential of determining automatically the metabolite signals which make up MR spectra. However, the reliability with which this is accomplished and the optimal approach for investigating in vivo MRS have not been determined. Furthermore, the properties of ICA in brain tumour MRS with respect to dataset size and data quality have not been systematically explored. The two common techniques for applying ICA, blind source separation (BSS) and feature extraction (FE) were examined in this study using simulated data and the findings confirmed on patient data. Short echo time (TE 30 ms), low and high field (1.5 and 3 T) in vivo brain tumour MR spectra of childhood astrocytoma, ependymoma and medulloblastoma were generated by using a quantum mechanical simulator with ten metabolite and lipid components. Patient data (TE 30 ms, 1.5 T) were acquired from children with brain tumours. ICA of simulated data shows that individual metabolite components can be extracted from a set of MRS data. The BSS method generates independent components with a closer correlation to the original metabolite and lipid components than the FE method when the number of spectra in the dataset is small. The experiments also show that stable results are achieved with 300 MRS at an SNR equal to 10. The FE method is relatively insensitive to different ranges of full width at half maximum (FWHM) (from 0 to 3 Hz), whereas the BSS method degrades on increasing the range of FWHM. The peak frequency variations do not affect the results within the range of +/-0.08 ppm for the FE method, and +/-0.05 ppm for the BSS method. When the methods were applied to the patient dataset, results consistent with the synthesized experiments were obtained.

Untreated glioblastoma multiforme: increased myo-inositol and glutamine levels in the contralateral cerebral hemisphere at proton MR spectroscopy

PURPOSE

To use localized in vivo proton magnetic resonance (MR) spectroscopy of the contralateral hemisphere in patients with glioblastoma multiforme (GBM) to detect alterations in cerebral metabolites as potential markers of infiltrating GBM cells.

MATERIALS AND METHODS

The study was approved by the ethics committee, and written informed consent was obtained. Twenty-two patients with newly diagnosed and untreated GBM underwent in vivo single-voxel short echo time proton MR spectroscopy with a 3-T MR imaging system. Absolute metabolite concentrations in the hemisphere contralateral to the tumor were compared with data from five patients with low-grade gliomas (LGGs) and from a group of 14 age-matched control subjects by using analysis of variance and subsequent t tests or corresponding nonparametric tests.

RESULTS

In the contralateral hemisphere, MR spectroscopy revealed increased concentrations of myo-inositol and glutamine. Mean myo-inositol levels were significantly increased in patients with GBM (3.6 mmol/L +/- 0.8 [standard deviation]) relative to levels in control subjects (3.1 mmol/L +/- 0.6; P = .03) and tended to be higher relative to levels in patients with LGG (2.7 mmol/L +/- 0.8; P = .09). Mean glutamine concentrations in patients with GBM (3.4 mmol/L +/- 0.9) differed significantly from those in control subjects (2.7 mmol/L +/- 0.7; P = .01); mean concentrations in patients with GBM differed from those in patients with LGG (2.4 mmol/L +/- 0.5; P = .01). There were no significant differences between data in patients with LGG and in control subjects.

CONCLUSION

Increased concentrations of myo-inositol and glutamine in the contralateral normal-appearing white matter of GBM patients are consistent with mild astrocytosis and suggest the detectability of early neoplastic infiltration by using proton MR spectroscopy in vivo.

Nosologic imaging of the brain: segmentation and classification using MRI and MRSI

A new technique is presented to create nosologic images of the brain based on magnetic resonance imaging (MRI) and magnetic resonance spectroscopic imaging (MRSI). A nosologic image summarizes the presence of different tissues and lesions in a single image by color coding each voxel or pixel according to the histopathological class it is assigned to. The proposed technique applies advanced methods from image processing as well as pattern recognition to segment and classify brain tumors. First, a registered brain atlas and a subject-specific abnormal tissue prior, obtained from MRSI data, are used for the segmentation. Next, the detected abnormal tissue is classified based on supervised pattern recognition methods. Class probabilities are also calculated for the segmented abnormal region. Compared to previous approaches, the new framework is more flexible and able to better exploit spatial information leading to improved nosologic images. The combined scheme offers a new way to produce high-resolution nosologic images, representing tumor heterogeneity and class probabilities, which may help clinicians in decision making.

Revealing the metabolic profile of brain tumors for diagnosis purposes

The metabolic behavior of complex brain tumors, like Gliomas and Meningiomas, with respect to their type and grade was investigated in this paper. Towards this direction the smallest set of the most representative metabolic markers for each brain tumor type was identified, using ratios of peak areas of well established metabolites, from (1)H-MRSI (Proton Magnetic Resonance Spectroscopy Imaging) data of 24 patients and 4 healthy volunteers. A feature selection method that embeds Fisher's filter criterion into a wrapper selection scheme was applied; Support Vector Machine (SVM) and Least Squares-SVM (LS-SVM) classifiers were used to evaluate the ratio markers classification significance. The area under the Receiver Operating Characteristic curve (AUROC) was adopted to evaluate the classification significance. It is found that the NAA/CHO, CHO/S, MI/S ratios can be used to discriminate Gliomas and Meningiomas from Healthy tissue with AUROC greater than 0.98. Ratios CHO/S, CRE/S, MI/S, LAC/CRE, ALA/CRE, ALA/S and LIPS/CRE can identify type and grade differences in Gliomas giving AUROC greater than 0.98 apart from the scheme of Gliomas grade II vs grade III where 0.84 was recorded due to high heterogeneity. Finally NAA/CRE, NAA/S, CHO/S, MI/S and ALA/S manage to discriminate Gliomas from Meningiomas providing AUROC exceeding 0.90.

Role of advanced MR imaging modalities in diagnosing cerebral gliomas

The objective of this study was to evaluate the potential role of newly developed, advanced magnetic resonance (MR) imaging techniques (spectroscopy, diffusion and perfusion imaging) in diagnosing brain gliomas, with special reference to histological typing and grading, treatment planning and posttreatment follow-up. Conventional MR imaging enables the detection and localisation of neoplastic lesions, as well as providing, in typical cases, some indication about their nature. However, it has limited sensitivity and specificity in evaluating histological type and grade, delineating margins and differentiating oedema, tumour and treatment side-effects. These limitations can be overcome by supplementing the morphological data obtained with conventional MR imaging with the metabolic, structural and perfusional information provided by new MR techniques that are increasingly becoming an integral part of routine MR studies. Incorporation of such new MR techniques can lead to more comprehensive and precise diagnoses that can better assist surgeons in determining prognosis and planning treatment strategies. In addition, the recent development of new, more effective, treatments for cerebral glioma strongly relies on morphofunctional MR imaging with its ability to provide a biological interpretation of these characteristically heterogeneous tumours.

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.

Identification and characterisation of childhood cerebellar tumours by in vivo proton MRS

(1)H MRS has great potential for the clinical investigation of childhood brain tumours, but the low incidence in, and difficulties of performing trials on, children have hampered progress in this area. Most studies have used a long-TE, thus limiting the metabolite information obtained, and multivariate analysis has been largely unexplored. Thirty-five children with untreated cerebellar tumours (18 medulloblastomas, 12 pilocytic astrocytomas and five ependymomas) were investigated using a single-voxel short-TE PRESS sequence on a 1.5 T scanner. Spectra were analysed using LCModel to yield metabolite profiles, and key metabolite assignments were verified by comparison with high-resolution magic-angle-spinning NMR of representative tumour biopsy samples. In addition to univariate metabolite comparisons, the use of multivariate classifiers was investigated. Principal component analysis was used for dimension reduction, and linear discriminant analysis was used for variable selection and classification. A bootstrap cross-validation method suitable for estimating the true performance of classifiers in small datasets was used. The discriminant function coefficients were stable and showed that medulloblastomas were characterised by high taurine, phosphocholine and glutamate and low glutamine, astrocytomas were distinguished by low creatine and high N-acetylaspartate, and ependymomas were differentiated by high myo-inositol and glycerophosphocholine. The same metabolite features were seen in NMR spectra of ex vivo samples. Successful classification was achieved for glial-cell (astrocytoma + ependymoma) versus non-glial-cell (medulloblastoma) tumours, with a bootstrap 0.632 + error, e(B.632+), of 5.3%. For astrocytoma vs medulloblastoma and astrocytoma vs medulloblastoma vs ependymoma classification, the e(B.632+) was 6.9% and 7.1%, respectively. The study showed that (1)H MRS detects key differences in the metabolite profiles for the main types of childhood cerebellar tumours and that discriminant analysis of metabolite profiles is a promising tool for classification. The findings warrant confirmation by larger multi-centre studies.