Quality control of prostate 1 H MRSI data

A novel automated pipeline to assess MR spectroscopy quality control: Comparing current standards and manual assessment

BACKGROUND AND PURPOSE

The absence of a consensus data quality control (DQC) process inhibits the widespread adoption of MR spectroscopy. Poor DQC can lead to unreliable clinical diagnosis and irreproducible research conclusions. Currently, manual visual assessment or the standard quantitative metrics of signal-to-noise, linewidth, and model fit are used as classifiers, but these measures may not be sufficient. To supplement standard metrics, this paper proposes a novel automated DQC pipeline named Visual Evaluative Control Technology Of Resonance Spectroscopy (VECTORS).

METHODS

Manual DQC ratings were conducted on 7180 spectra obtained from 110 young adults using short-echo chemical shift imaging at 3 Tesla. Four reviewers conducted manual ratings on the presence of artifacts and location of metabolites. The ratings were labor intensive, taking over 180 hours. VECTORS was developed to quantify their DQC criteria, detecting artifacts that present as duplicate peaks, vertical shifts, and glutamine + glutamate and myoinositol peak shapes. Run on the same data using a standard laptop, VECTORS only took 2 hours.

RESULTS

The manual ratings were not monotonic to the standard quantitative metrics. VECTORS correctly flagged spectra that the manual ratings missed. VECTORS accurately flagged an additional 126 poor DQ spectra that consensus cutoffs of the standard quantitative metrics deemed good DQ.

CONCLUSION

Standard quantitative metrics may not account for all DQC artifacts as they are not monotonic to the manual ratings. However, manual ratings are labor intensive, subjective, and irreproducible. VECTORS addresses these issues and should be used in conjunction with standard quantitative metrics.

A multivariate curve resolution analysis of multicenter proton spectroscopic imaging of the prostate for cancer localization and assessment of aggressiveness

In this study, we investigated the potential of the multivariate curve resolution alternating least squares (MCR-ALS) algorithm for analyzing three-dimensional (3D) 1 H-MRSI data of the prostate in prostate cancer (PCa) patients. MCR-ALS generates relative intensities of components representing spectral profiles derived from a large training set of patients, providing an interpretable model. Our objectives were to classify magnetic resonance (MR) spectra, differentiating tumor lesions from benign tissue, and to assess PCa aggressiveness. We included multicenter 3D 1 H-MRSI data from 106 PCa patients across eight centers. The patient cohort was divided into a training set (N = 63) and an independent test set (N = 43). Singular value decomposition determined that MR spectra were optimally represented by five components. The profiles of these components were extracted from the training set by MCR-ALS and assigned to specific tissue types. Using these components, MCR-ALS was applied to the test set for a quantitative analysis to discriminate tumor lesions from benign tissue and to assess tumor aggressiveness. Relative intensity maps of the components were reconstructed and compared with histopathology reports. The quantitative analysis demonstrated a significant separation between tumor and benign voxels (t-test, p < 0.001). This result was achieved including voxels with low-quality MR spectra. A receiver operating characteristic analysis of the relative intensity of the tumor component revealed that low- and high-risk tumor lesions could be distinguished with an area under the curve of 0.88. Maps of this component properly identified the extent of tumor lesions. Our study demonstrated that MCR-ALS analysis of 1 H-MRSI of the prostate can reliably identify tumor lesions and assess their aggressiveness. It handled multicenter data with minimal preprocessing and without using prior knowledge or quality control. These findings indicate that MCR-ALS can serve as an automated tool to assess the presence, extent, and aggressiveness of tumor lesions in the prostate, enhancing diagnostic capabilities and treatment planning of PCa patients.

Developments in proton MR spectroscopic imaging of prostate cancer

In this paper, we review the developments of 1H-MR spectroscopic imaging (MRSI) methods designed to investigate prostate cancer, covering key aspects such as specific hardware, dedicated pulse sequences for data acquisition and data processing and quantification techniques. Emphasis is given to recent advancements in MRSI methodologies, as well as future developments, which can lead to overcome difficulties associated with commonly employed MRSI approaches applied in clinical routine. This includes the replacement of standard PRESS sequences for volume selection, which we identified as inadequate for clinical applications, by sLASER sequences and implementation of 1H MRSI without water signal suppression. These may enable a new evaluation of the complementary role and significance of MRSI in prostate cancer management.

Absolute choline tissue concentration mapping for prostate cancer localization and characterization using 3D 1 H MRSI without water-signal suppression

PURPOSE

Until now, 1 H MRSI of the prostate has been performed with suppression of the large water signal to avoid distortions of metabolite signals. However, this signal can be used for absolute quantification and spectral corrections. We investigated the feasibility of water-unsuppressed MRSI in patients with prostate cancer for water signal-mediated spectral quality improvement and determination of absolute tissue levels of choline.

METHODS

Eight prostate cancer patients scheduled for radical prostatectomy underwent multi-parametric MRI at 3 T, including 3D water-unsuppressed semi-LASER MRSI. A postprocessing algorithm was developed to remove the water signal and its artifacts and use the extracted water signal as intravoxel reference for phase and frequency correction of metabolite signals and for absolute metabolite quantification.

RESULTS

Water-unsuppressed MRSI with dedicated postprocessing produced water signal and artifact-free MR spectra throughout the prostate. In all patients, the absolute choline tissue concentration was significantly higher in tumorous than in benign tissue areas (mean ± SD: 7.2 ± 1.4 vs 3.8 ± 0.7 mM), facilitating tumor localization by choline mapping. Tumor tissue levels of choline correlated better with the commonly used (choline + spermine + creatine)/citrate ratio (r = 0.78 ± 0.1) than that of citrate (r = 0.21 ± 0.06). The highest maximum choline concentrations occurred in high-risk cancer foci.

CONCLUSION

This report presents the first successful water-unsuppressed MRSI of the whole prostate. The water signal enabled amelioration of spectral quality and absolute metabolite quantification. In this way, choline tissue levels were identified as tumor biomarker. Choline mapping may serve as a tool in prostate cancer localization and risk scoring in multi-parametric MRI for diagnosis and biopsy procedures.

Quality control of 3D MRSI data in glioblastoma: Can we do without the experts?

PURPOSE

Proton magnetic resonance spectroscopic imaging (1H MRSI) is a noninvasive technique for assessing tumor metabolism. Manual inspection is still the gold standard for quality control (QC) of spectra, but it is both time-consuming and subjective. The aim of the present study was to assess automatic QC of glioblastoma MRSI data using random forest analysis.

METHODS

Data for 25 patients, acquired prospectively in a preradiotherapy examination, were submitted to postprocessing with syngo.MR Spectro (VB40A; Siemens) or Java-based magnetic resonance user interface (jMRUI) software. A total of 28 features were extracted from each spectrum for the automatic QC. Three spectroscopists also performed manual inspections, labeling each spectrum as good or poor quality. All statistical analyses, with addressing unbalanced data, were conducted with R 3.6.1 (R Foundation for Statistical Computing; https://www.r-project.org).

RESULTS

The random forest method classified the spectra with an area under the curve of 95.5%, sensitivity of 95.8%, and specificity of 81.7%. The most important feature for the classification was Residuum_Lipids_Versus_Fit, obtained with syngo.MR Spectro.

CONCLUSION

The automatic QC method was able to distinguish between good- and poor-quality spectra, and can be used by radiation oncologists who are not spectroscopy experts. This study revealed a novel set of MRSI signal features that are closely correlated with spectral quality.

A Single-Arm, Multicenter Validation Study of Prostate Cancer Localization and Aggressiveness With a Quantitative Multiparametric Magnetic Resonance Imaging Approach

OBJECTIVES

The aims of this study were to assess the discriminative performance of quantitative multiparametric magnetic resonance imaging (mpMRI) between prostate cancer and noncancer tissues and between tumor grade groups (GGs) in a multicenter, single-vendor study, and to investigate to what extent site-specific differences affect variations in mpMRI parameters.

MATERIALS AND METHODS

Fifty patients with biopsy-proven prostate cancer from 5 institutions underwent a standardized preoperative mpMRI protocol. Based on the evaluation of whole-mount histopathology sections, regions of interest were placed on axial T2-weighed MRI scans in cancer and noncancer peripheral zone (PZ) and transition zone (TZ) tissue. Regions of interest were transferred to functional parameter maps, and quantitative parameters were extracted. Across-center variations in noncancer tissues, differences between tissues, and the relation to cancer grade groups were assessed using linear mixed-effects models and receiver operating characteristic analyses.

RESULTS

Variations in quantitative parameters were low across institutes (mean [maximum] proportion of total variance in PZ and TZ, 4% [14%] and 8% [46%], respectively). Cancer and noncancer tissues were best separated using the diffusion-weighted imaging-derived apparent diffusion coefficient, both in PZ and TZ (mean [95% confidence interval] areas under the receiver operating characteristic curve [AUCs]; 0.93 [0.89-0.96] and 0.86 [0.75-0.94]), followed by MR spectroscopic imaging and dynamic contrast-enhanced-derived parameters. Parameters from all imaging methods correlated significantly with tumor grade group in PZ tumors. In discriminating GG1 PZ tumors from higher GGs, the highest AUC was obtained with apparent diffusion coefficient (0.74 [0.57-0.90], P < 0.001). The best separation of GG1-2 from GG3-5 PZ tumors was with a logistic regression model of a combination of functional parameters (mean AUC, 0.89 [0.78-0.98]).

CONCLUSIONS

Standardized data acquisition and postprocessing protocols in prostate mpMRI at 3 T produce equivalent quantitative results across patients from multiple institutions and achieve similar discrimination between cancer and noncancer tissues and cancer grade groups as in previously reported single-center studies.

Nuclear magnetic resonance spectroscopy of human body fluids and in vivo magnetic resonance spectroscopy: Potential role in the diagnosis and management of prostate cancer

Prostate cancer is the most common solid organ cancer in men, and the second most common cause of male cancer-related mortality. It has few effective therapies, and is difficult to diagnose accurately. Prostate-specific antigen (PSA), which is currently the most effective diagnostic tool available, cannot reliably discriminate between different pathologies, and in fact only around 30% of patients found to have elevated levels of PSA are subsequently confirmed to actually have prostate cancer. As such, there is a desperate need for more reliable diagnostic tools that will allow the early detection of prostate cancer so that the appropriate interventions can be applied. Nuclear magnetic resonance (NMR) spectroscopy and magnetic resonance spectroscopy (MRS) are 2 high throughput, noninvasive analytical procedures that have the potential to enable differentiation of prostate cancer from other pathologies using metabolomics, by focusing specifically on certain metabolites which are associated with the development of prostate cancer cells and its progression. The value that this type of approach has for the early detection, diagnosis, prognosis, and personalized treatment of prostate cancer is becoming increasingly apparent. Recent years have seen many promising developments in the fields of NMR spectroscopy and MRS, with improvements having been made to hardware as well as to techniques associated with the acquisition, processing, and analysis of related data. This review focuses firstly on proton NMR spectroscopy of blood serum, urine, and expressed prostatic secretions in vitro, and then on 1- and 2-dimensional proton MRS of the prostate in vivo. Major advances in these fields and methodological principles of data collection, acquisition, processing, and analysis are described along with some discussion of related challenges, before prospects that proton MRS has for future improvements to the clinical management of prostate cancer are considered.

Simple and broadly applicable automatic quality control for 3D 1 H MR spectroscopic imaging data of the prostate

PURPOSE

Quality control (QC) is a prerequisite for clinical MR spectroscopic imaging (MRSI) to avoid that bad spectra hamper data interpretation. The aim of this work was to present a simple automatic QC for prostate ¹ H MRSI that can handle data obtained with different commonly used pulse sequences, echo times, field strengths, and MR platforms.

METHODS

A QC method was developed with a ratio (Qratio) where the numerator and the denominator are functions of several signal heights, logically combined for their positive or negative contribution to spectral quality. This Qratio was tested on 4 data sets obtained at 1.5, 3, and 7T, with and without endorectal coil and different localization sequences and echo times. Spectra of 25,248 voxels in 26 prostates were labeled as acceptable or unacceptable by MRS experts as gold standard. A threshold value was determined for Qratio from a subset of voxels, labeled in consensus by 4 experts, for an optimal accuracy to separate spectra.

RESULTS

Applying this Qratio threshold to the remaining test voxels, an automatic separation of good and bad spectra was possible with an accuracy of 0.88, similar to manual separation between the 2 classes. Qratio values were used to generate maps representing spectral quality on a binary or continuous scale.

CONCLUSION

Automated QC of prostate ¹ H MRSI by Qratio is fast, simple, easily transferable and more practical than supervised feature extraction methods and therefore easy to integrate into different clinical MR systems. Moreover, quality maps can be generated to read the reliability of spectra in each voxel.

Improving labeling efficiency in automatic quality control of MRSI data

PURPOSE

To improve the efficiency of the labeling task in automatic quality control of MR spectroscopy imaging data.

METHODS

28'432 short and long echo time (TE) spectra (1.5 tesla; point resolved spectroscopy (PRESS); repetition time (TR)= 1,500 ms) from 18 different brain tumor patients were labeled by two experts as either accept or reject, depending on their quality. For each spectrum, 47 signal features were extracted. The data was then used to run several simulations and test an active learning approach using uncertainty sampling. The performance of the classifiers was evaluated as a function of the number of patients in the training set, number of spectra in the training set, and a parameter α used to control the level of classification uncertainty required for a new spectrum to be selected for labeling.

RESULTS

The results showed that the proposed strategy allows reductions of up to 72.97% for short TE and 62.09% for long TE in the amount of data that needs to be labeled, without significant impact in classification accuracy. Further reductions are possible with significant but minimal impact in performance.

CONCLUSION

Active learning using uncertainty sampling is an effective way to increase the labeling efficiency for training automatic quality control classifiers. Magn Reson Med 78:2399-2405, 2017. © 2017 International Society for Magnetic Resonance in Medicine.

Hierarchical non-negative matrix factorization applied to three-dimensional 3 T MRSI data for automatic tissue characterization of the prostate

In this study non-negative matrix factorization (NMF) was hierarchically applied to simulated and in vivo three-dimensional 3 T MRSI data of the prostate to extract patterns for tumour and benign tissue and to visualize their spatial distribution. Our studies show that the hierarchical scheme provides more reliable tissue patterns than those obtained by performing only one NMF level. We compared the performance of three different NMF implementations in terms of pattern detection accuracy and efficiency when embedded into the same kind of hierarchical scheme. The simulation and in vivo results show that the three implementations perform similarly, although one of them is more robust and better pinpoints the most aggressive tumour voxel(s) in the dataset. Furthermore, they are able to detect tumour and benign tissue patterns even in spectra with lipid artefacts. Copyright © 2016 John Wiley & Sons, Ltd.

Automatic quality control in clinical (1)H MRSI of brain cancer

MRSI grids frequently show spectra with poor quality, mainly because of the high sensitivity of MRS to field inhomogeneities. These poor quality spectra are prone to quantification and/or interpretation errors that can have a significant impact on the clinical use of spectroscopic data. Therefore, quality control of the spectra should always precede their clinical use. When performed manually, quality assessment of MRSI spectra is not only a tedious and time-consuming task, but is also affected by human subjectivity. Consequently, automatic, fast and reliable methods for spectral quality assessment are of utmost interest. In this article, we present a new random forest-based method for automatic quality assessment of (1)H MRSI brain spectra, which uses a new set of MRS signal features. The random forest classifier was trained on spectra from 40 MRSI grids that were classified as acceptable or non-acceptable by two expert spectroscopists. To account for the effects of intra-rater reliability, each spectrum was rated for quality three times by each rater. The automatic method classified these spectra with an area under the curve (AUC) of 0.976. Furthermore, in the subset of spectra containing only the cases that were classified every time in the same way by the spectroscopists, an AUC of 0.998 was obtained. Feature importance for the classification was also evaluated. Frequency domain skewness and kurtosis, as well as time domain signal-to-noise ratios (SNRs) in the ranges 50-75 ms and 75-100 ms, were the most important features. Given that the method is able to assess a whole MRSI grid faster than a spectroscopist (approximately 3 s versus approximately 3 min), and without loss of accuracy (agreement between classifier trained with just one session and any of the other labelling sessions, 89.88%; agreement between any two labelling sessions, 89.03%), the authors suggest its implementation in the clinical routine. The method presented in this article was implemented in jMRUI's SpectrIm plugin.

Mapping of prostate cancer by 1H MRSI

In many studies, it has been demonstrated that (1)H MRSI of the human prostate has great potential to aid prostate cancer management, e.g. in the detection and localisation of cancer foci in the prostate or in the assessment of its aggressiveness. It is particularly powerful in combination with T2 -weighted MRI. Nevertheless, the technique is currently mainly used in a research setting. This review provides an overview of the state-of-the-art of three-dimensional MRSI, including the specific hardware required, dedicated data acquisition sequences and information on the spectral content with background on the MR-visible metabolites. In clinical practice, it is important that relevant MRSI results become available rapidly, reliably and in an easy digestible way. However, this functionality is currently not fully available for prostate MRSI, which is a major obstacle for routine use by inexperienced clinicians. Routine use requires more automation in the processing of raw data than is currently available. Therefore, we pay specific attention in this review on the status and prospects of the automated handling of prostate MRSI data, including quality control. The clinical potential of three-dimensional MRSI of the prostate is illustrated with literature examples on prostate cancer detection, its localisation in the prostate, its role in the assessment of cancer aggressiveness and in the selection and monitoring of therapy.