Multi-parametric (ADC/PWI/T2-w) image fusion approach for accurate semi-automatic segmentation of tumorous regions in glioblastoma multiforme

Augmented surgical decision-making for glioblastoma: integrating AI tools into education and practice

Surgical decision-making for glioblastoma poses significant challenges due to its complexity and variability. This study investigates the potential of artificial intelligence (AI) tools in improving "decision-making processes" for glioblastoma surgery. A systematic review of literature identified 10 relevant studies, primarily focused on predicting resectability and surgery-related neurological outcomes. AI tools, especially rooted in radiomics and connectomics, exhibited promise in predicting resection extent through precise tumor segmentation and tumor-network relationships. However, they demonstrated limited effectiveness in predicting postoperative neurological due to dynamic and less quantifiable nature of patient-related factors. Recognizing these challenges, including limited datasets and the interpretability requirement in medical applications, underscores the need for standardization, algorithm optimization, and addressing variability in model performance and then further validation in clinical settings. While AI holds potential, it currently does not possess the capacity to emulate the nuanced decision-making process utilized by experienced neurosurgeons in the comprehensive approach to glioblastoma surgery.

Automatic segmentation of brain glioma based on XY-Net

Glioma is a malignant primary brain tumor, which can easily lead to death if it is not detected in time. Magnetic resonance imaging is the most commonly used technique to diagnose gliomas, and precise outlining of tumor areas from magnetic resonance images (MRIs) is an important aid to physicians in understanding the patient's condition and formulating treatment plans. However, relying on radiologists to manually depict tumors is a tedious and laborious task, so it is clinically important to investigate an automated method for outlining glioma regions in MRIs. To liberate radiologists from the heavy task of outlining tumors, we propose a fully convolutional network, XY-Net, based on the most popular U-Net symmetric encoder-decoder structure to perform automatic segmentation of gliomas. We construct two symmetric sub-encoders for XY-Net and build interconnected X-shaped feature map transmission paths between the sub-encoders, while maintaining the feature map concatenation between each sub-encoder and the decoder. Moreover, a loss function composed of the balanced cross-entropy loss function and the dice loss function is used in the training task of XY-Net to solve the class unevenness problem of the medical image segmentation task. The experimental results show that the proposed XY-Net has a 2.16% improvement in dice coefficient (DC) compared to the network model with a single encoder structure, and compare with some state-of-the-art image segmentation methods, XY-Net achieves the best performance. The DC, HD, recall, and precision of our method on the test set are 74.49%, 10.89 mm, 78.06%, and 76.30%, respectively. The combination of sub-encoders and cross-transmission paths enables the model to perform better; based on this combination, the XY-Net achieves an end-to-end automatic segmentation of gliomas on 2D slices of MRIs, which can play a certain auxiliary role for doctors in grasping the state of illness.

Development and Validation of a Deep Learning Model for Brain Tumor Diagnosis and Classification Using Magnetic Resonance Imaging

IMPORTANCE

Deep learning may be able to use patient magnetic resonance imaging (MRI) data to aid in brain tumor classification and diagnosis.

OBJECTIVE

To develop and clinically validate a deep learning system for automated identification and classification of 18 types of brain tumors from patient MRI data.

DESIGN, SETTING, AND PARTICIPANTS

This diagnostic study was conducted using MRI data collected between 2000 and 2019 from 37 871 patients. A deep learning system for segmentation and classification of 18 types of intracranial tumors based on T1- and T2-weighted images and T2 contrast MRI sequences was developed and tested. The diagnostic accuracy of the system was tested using 1 internal and 3 external independent data sets. The clinical value of the system was assessed by comparing the tumor diagnostic accuracy of neuroradiologists with vs without assistance of the proposed system using a separate internal test data set. Data were analyzed from March 2019 through February 2020.

MAIN OUTCOMES AND MEASURES

Changes in neuroradiologist clinical diagnostic accuracy in brain MRI scans with vs without the deep learning system were evaluated.

RESULTS

A deep learning system was trained among 37 871 patients (mean [SD] age, 41.6 [11.4] years; 18 519 women [48.9%]). It achieved a mean area under the receiver operating characteristic curve of 0.92 (95% CI, 0.84-0.99) on 1339 patients from 4 centers' data sets in diagnosis and classification of 18 types of tumors. Higher outcomes were found compared with neuroradiologists for accuracy and sensitivity and similar outcomes for specificity (for 300 patients in the Tiantan Hospital test data set: accuracy, 73.3% [95% CI, 67.7%-77.7%] vs 60.9% [95% CI, 46.8%-75.1%]; sensitivity, 88.9% [95% CI, 85.3%-92.4%] vs 53.4% [95% CI, 41.8%-64.9%]; and specificity, 96.3% [95% CI, 94.2%-98.4%] vs 97.9%; [95% CI, 97.3%-98.5%]). With the assistance of the deep learning system, the mean accuracy of neuroradiologists among 1166 patients increased by 12.0 percentage points, from 63.5% (95% CI, 60.7%-66.2%) without assistance to 75.5% (95% CI, 73.0%-77.9%) with assistance.

CONCLUSIONS AND RELEVANCE

These findings suggest that deep learning system-based automated diagnosis may be associated with improved classification and diagnosis of intracranial tumors from MRI data among neuroradiologists.

Value of Magnetic Resonance Images and Magnetic Resonance Spectroscopy in Diagnosis of Brain Tumors under Fuzzy C-Means Algorithm

This study was aimed to explore the diagnostic value of magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS) in brain tumors under the fuzzy C-means (FCM) algorithm. The two-dimensional FCM hybrid algorithm was improved to be three-dimensional. The MRI images and MRS spectra of 127 patients with brain tumors (low-grade glioma group) and 54 healthy people (healthy group) were analyzed. The results suggested that the membership matrix of the improved algorithm had lower ambiguity, higher segmentation accuracy, closer relationship of intrapixels, and stronger irrelevance of interclass pixels. Through the analysis of gray matter volume, it was found that, compared with the healthy group, the gray matter and white matter volumes in the brain of high-grade glioma were higher, and those of low-grade glioma group were lower. The improved FCM algorithm could obtain a higher accuracy of 88.64% in segmenting images. It had a higher sensitivity to gray matter changes in brain tumors, reaching 92.72%; its specificity was not much different from that of traditional FCM, which were 83.61% and 88.06%, respectively. In the diagnostic value, the area under the curve of mean kurtosis was the largest, which was 0.962 (P < 0.001). The best critical value was 0.4096, which had a greater reference significance for clinical treatment and prognosis. The ratio of choline/N-acetyl-aspartate and the ratio of choline/creatine also showed significant differences in high- and low-grade gliomas (P < 0.05), but the specificity and sensitivity were slightly lower. It also had guiding significance for the grading of gliomas. Overall, the improved FCM algorithm had obvious advantages in the segmentation process of MRI images, which provided help for the clinical diagnosis of brain tumors.

Magnetic resonance image-based brain tumour segmentation methods: A systematic review

Background

Image segmentation is an essential step in the analysis and subsequent characterisation of brain tumours through magnetic resonance imaging. In the literature, segmentation methods are empowered by open-access magnetic resonance imaging datasets, such as the brain tumour segmentation dataset. Moreover, with the increased use of artificial intelligence methods in medical imaging, access to larger data repositories has become vital in method development.

Purpose

To determine what automated brain tumour segmentation techniques can medical imaging specialists and clinicians use to identify tumour components, compared to manual segmentation.

Methods

We conducted a systematic review of 572 brain tumour segmentation studies during 2015-2020. We reviewed segmentation techniques using T1-weighted, T2-weighted, gadolinium-enhanced T1-weighted, fluid-attenuated inversion recovery, diffusion-weighted and perfusion-weighted magnetic resonance imaging sequences. Moreover, we assessed physics or mathematics-based methods, deep learning methods, and software-based or semi-automatic methods, as applied to magnetic resonance imaging techniques. Particularly, we synthesised each method as per the utilised magnetic resonance imaging sequences, study population, technical approach (such as deep learning) and performance score measures (such as Dice score).

Statistical tests

We compared median Dice score in segmenting the whole tumour, tumour core and enhanced tumour.

Results

We found that T1-weighted, gadolinium-enhanced T1-weighted, T2-weighted and fluid-attenuated inversion recovery magnetic resonance imaging are used the most in various segmentation algorithms. However, there is limited use of perfusion-weighted and diffusion-weighted magnetic resonance imaging. Moreover, we found that the U-Net deep learning technology is cited the most, and has high accuracy (Dice score 0.9) for magnetic resonance imaging-based brain tumour segmentation.

Conclusion

U-Net is a promising deep learning technology for magnetic resonance imaging-based brain tumour segmentation. The community should be encouraged to contribute open-access datasets so training, testing and validation of deep learning algorithms can be improved, particularly for diffusion- and perfusion-weighted magnetic resonance imaging, where there are limited datasets available.

3D-PulCNN: Pulmonary cancer classification from hyperspectral images using convolution combination unit based CNN

Pulmonary cancer is one of the most common malignancies worldwide. Accurate classification of its subtypes is required in differential diagnosis. However, existing algorithms are mostly based on color images, and the improvement of accuracy is quite challenging. In this study, we propose a convolution combination unit (CCU)-based three-dimensional convolutional neural network (3D-PulCNN) for classifying pulmonary cancer presented in microscopic hyperspectral image with both spatial and spectral information. CCU is designed to fuse the features acquired by different convolution scales. Compared with VGGNet, only two fully connected layers are used in this model, reducing the network parameters and model complexity. Experimental results show that 3D-PulCNN achieves overall average (OA) of 0.962 and Precision, Recall, and Kappa of more than 0.920, superior to 2D-VGGNet. Then, 3D-UNet is leveraged to segment cancer cells, and their morphological characteristics are calculated to supply quantitative virtual analysis data for classification results explanation and prognosis assessment.

Evaluating potential of multi-parametric MRI using co-registered histology: Application to a mouse model of glioblastoma

BACKGROUND

Conventional MRI fails to detect regions of glioblastoma cell infiltration beyond the contrast-enhanced T1 solid tumor region, with infiltrating tumor cells often migrating along host blood vessels.

PURPOSE

MRI is capable of generating a range of image contrasts which are commonly assessed individually by qualitative visual inspection. It has long been hypothesized that better diagnoses could be achieved by combining these multiple images, so called multi-parametric or multi-spectral MRI. However, the lack of clinical histology and the difficulties of co-registration, has meant this hypothesis has never been rigorously tested. Here we test this hypothesis, using a previously published multi-dimensional dataset consisting of registered MR images and histology.

STUDY TYPE

Animal Model.

SUBJECTS

Mice bearing orthotopic glioblastoma xenografts generated from a patient-derived glioblastoma cell line.

FIELD STRENGTH/SEQUENCES

7 Tesla, T₁/T₂ weighted, T₂ mapping, contrast enhance T_1, diffusion-weighted, diffusion tensor imaging.

ASSESSMENT

Immunohistochemistry sections were stained for Human Leukocyte Antigen (probing human-derived tumor cells). To achieve quantitative MRI-tissue comparison, multiple histological slices cut in the MRI plane were stacked to produce tumor cell density maps acting as 'ground truth'.

STATISTICAL TESTS

Sensitivity, specificity, accuracy and Dice similarity indices were calculated. ANOVA, t-test, Bonferroni correction and Pearson coefficients were used for statistical analysis.

RESULTS

Correlation coefficient analysis with co-registered 'ground truth' histology showed interactive regression maps had higher correlation coefficients and sensitivity values than T2W, ADC, FA, and T2map. Further, the interaction regression maps showed statistical improved detection of tumor volume.

DATA CONCLUSION

Voxel-by-voxel analysis provided quantitative evidence confirming the hypothesis that mpMRI can, potentially, better distinguish between the tumor region and normal tissue.

Multi-parametric MR Imaging Biomarkers Associated to Clinical Outcomes in Gliomas: A Systematic Review

PURPOSE

To systematically review evidence regarding the association of multiparametric biomarkers with clinical outcomes and their capacity to explain relevant subcompartments of gliomas.

MATERIALS AND METHODS

Scopus database was searched for original journal papers from January 1st, 2007 to February 20th, 2017 according to PRISMA. Four hundred forty-nine abstracts of papers were reviewed and scored independently by two out of six authors. Based on those papers we analyzed associations between biomarkers, subcompartments within the tumor lesion, and clinical outcomes. From all the articles analyzed, the twenty-seven papers with the highest scores were highlighted to represent the evidence about MR imaging biomarkers associated with clinical outcomes. Similarly, eighteen studies defining subcompartments within the tumor region were also highlighted to represent the evidence of MR imaging biomarkers. Their reports were critically appraised according to the QUADAS-2 criteria.

RESULTS

It has been demonstrated that multi-parametric biomarkers are prepared for surrogating diagnosis, grading, segmentation, overall survival, progression-free survival, recurrence, molecular profiling and response to treatment in gliomas. Quantifications and radiomics features obtained from morphological exams (T1, T2, FLAIR, T1c), PWI (including DSC and DCE), diffusion (DWI, DTI) and chemical shift imaging (CSI) are the preferred MR biomarkers associated to clinical outcomes. Subcompartments relative to the peritumoral region, invasion, infiltration, proliferation, mass effect and pseudo flush, relapse compartments, gross tumor volumes, and highrisk regions have been defined to characterize the heterogeneity. For the majority of pairwise cooccurrences, we found no evidence to assert that observed co-occurrences were significantly different from their expected co-occurrences (Binomial test with False Discovery Rate correction, α=0.05). The co-occurrence among terms in the studied papers was found to be driven by their individual prevalence and trends in the literature.

CONCLUSION

Combinations of MR imaging biomarkers from morphological, PWI, DWI and CSI exams have demonstrated their capability to predict clinical outcomes in different management moments of gliomas. Whereas morphologic-derived compartments have been mostly studied during the last ten years, new multi-parametric MRI approaches have also been proposed to discover specific subcompartments of the tumors. MR biomarkers from those subcompartments show the local behavior within the heterogeneous tumor and may quantify the prognosis and response to treatment of gliomas.

Precision Digital Oncology: Emerging Role of Radiomics-based Biomarkers and Artificial Intelligence for Advanced Imaging and Characterization of Brain Tumors

Advances in computerized image analysis and the use of artificial intelligence-based approaches for image-based analysis and construction of prediction algorithms represent a new era for noninvasive biomarker discovery. In recent literature, it has become apparent that radiologic images can serve as mineable databases that contain large amounts of quantitative features with potential clinical significance. Extraction and analysis of these quantitative features is commonly referred to as texture or radiomic analysis. Numerous studies have demonstrated applications for texture and radiomic characterization methods for assessing brain tumors to improve noninvasive predictions of tumor histologic characteristics, molecular profile, distinction of treatment-related changes, and prediction of patient survival. In this review, the current use and future potential of texture or radiomic-based approaches with machine learning for brain tumor image analysis and prediction algorithm construction will be discussed. This technology has the potential to advance the value of diagnostic imaging by extracting currently unused information on medical scans that enables more precise, personalized therapy; however, significant barriers must be overcome if this technology is to be successfully implemented on a wide scale for routine use in the clinical setting. Keywords: Adults and Pediatrics, Brain/Brain Stem, CNS, Computer Aided Diagnosis (CAD), Computer Applications-General (Informatics), Image Postprocessing, Informatics, Neural Networks, Neuro-Oncology, Oncology, Treatment Effects, Tumor Response Supplemental material is available for this article. © RSNA, 2020.

Characterization of hardware-related spatial distortions for IR-PETRA pulse sequence using a brain specific phantom

OBJECTIVE

Inversion recovery-pointwise encoding time reduction with radial acquisition (IR-PETRA) is an effective magnetic resonance (MR) pulse sequence in generating pseudo-CTs. The hardware-related spatial-distortion (HRSD) in MR images potentially deteriorates the accuracy of pseudo-CTs. Thus, we aimed at characterizing HRSD for IR-PETRA.

MATERIALS AND METHODS

gross-HRSD_overall (Euclidean-sum of gross-HRSD_i (i = x, y, z)) for IR-PETRA was assessed using a brain-specific phantom for two MR scanners (1.5 T-Aera and 3.0 T-Prisma). Moreover, hardware imperfections were analyzed by determining gradient-nonlinearity spatial-distortion (GNSD) and B₀-inhomogeneity spatial-distortion (B₀ISD) for magnetization-prepared rapid acquisition gradient-echo (MP-RAGE) which has well-known distortion characteristics.

RESULTS

In 3.0 T, maximum of gross-GNSD_overall (Euclidean-sum of gross-GNSD_i) and gross-B₀ISD for MP-RAGE was 2.77 mm and 0.57 mm, respectively. For this scanner, the mean and maximum of gross-HRSD_overall for IR-PETRA were 0.63 ± 0.38 mm and 1.91 mm, respectively. In 1.5 T, maximum of gross-GNSD_overall and gross-B₀ISD for MP-RAGE was 3.41 mm and 0.78 mm, respectively. The mean and maximum of gross-HRSD_overall for IR-PETRA were 1.02 ± 0.50 mm and 3.12 mm, respectively.

DISCUSSION

The spatial accuracy of MR images, besides being impacted by hardware performance, scanner capabilities, and imaging parameters, is mainly affected by its imaging strategy and data acquisition scheme. In 3.0 T, even without applying vendor correction algorithms, spatial accuracy of IR-PETRA image is sufficient for generating pseudo-CTs. In 1.5 T, distortion-correction is required to provide this accuracy.

A potential field segmentation based method for tumor segmentation on multi-parametric MRI of glioma cancer patients

Background

Accurate segmentation of brain tumors is vital for the gross tumor volume (GTV) definition in radiotherapy. Functional MR images like apparent diffusion constant (ADC) and fractional anisotropy (FA) images can provide more comprehensive information for sensitive detection of the GTV. We synthesize anatomical and functional MRI for accurate and semi-automatic segmentation of GTVs and improvement of clinical efficiency.

Methods

Four MR image sets including T1-weighted contrast-enhanced (T1C), T2-weighted (T2), apparent diffusion constant (ADC) and fractional anisotropy (FA) images of 5 glioma patients were acquired and registered. A new potential field segmentation (PFS) method was proposed based on the concept of potential field in physics. For T1C, T2 and ADC images, global potential field segmentation (global-PFS) was used on user defined region of interest (ROI) for rough segmentation and then morphologically processed for accurate delineation of the GTV. For FA images, white matter (WM) was removed using local potential field segmentation (local-PFS), and then tumor extent was delineated with region growing and morphological methods. The individual segmentations of multi-parametric images were ensembled into a fused segmentation, considered as final GTV. GTVs were compared with manually delineated ground truth and evaluated with segmentation quality measure (Q), Dice’s similarity coefficient (DSC) and Sensitivity and Specificity.

Results

Experimental study with the five patients’ data and new method showed that, the mean values of Q, DSC, Sensitivity and Specificity were 0.80 (±0.07), 0.88 (±0.04), 0.92 (±0.01) and 0.88 (±0.05) respectively. The global-PFS used on ROIs of T1C, T2 and ADC images can avoid interferences from skull and other non-tumor areas. Similarity to local-PFS on FA images, it can also reduce the time complexity as compared with the global-PFS on whole image sets.

Conclusions

Efficient and semi-automatic segmentation of the GTV can be achieved with the new method. Combination of anatomical and functional MR images has the potential to provide new methods and ideas for target definition in radiotherapy.

Characterization of active and infiltrative tumorous subregions from normal tissue in brain gliomas using multiparametric MRI

BACKGROUND

Targeted localized biopsies and treatments for diffuse gliomas rely on accurate identification of tissue subregions, for which current MRI techniques lack specificity.

PURPOSE

To explore the complementary and competitive roles of a variety of conventional and quantitative MRI methods for distinguishing subregions of brain gliomas.

STUDY TYPE

Prospective.

POPULATION

Fifty-one tissue specimens were collected using image-guided localized biopsy surgery from 10 patients with newly diagnosed gliomas.

FIELD STRENGTH/SEQUENCE

Conventional and quantitative MR images consisting of pre- and postcontrast T1 w, T2 w, T2 -FLAIR, T2 -relaxometry, DWI, DTI, IVIM, and DSC-MRI were acquired preoperatively at 3T.

ASSESSMENT

Biopsy specimens were histopathologically attributed to glioma tissue subregion categories of active tumor (AT), infiltrative edema (IE), and normal tissue (NT) subregions. For each tissue sample, a feature vector comprising 15 MRI-based parameters was derived from preoperative images and assessed by a machine learning algorithm to determine the best multiparametric feature combination for characterizing the tissue subregions.

STATISTICAL TESTS

For discrimination of AT, IE, and NT subregions, a one-way analysis of variance (ANOVA) test and for pairwise tissue subregion differentiation, Tukey honest significant difference, and Games-Howell tests were applied (P < 0.05). Cross-validated feature selection and classification methods were implemented for identification of accurate multiparametric MRI parameter combination.

RESULTS

After exclusion of 17 tissue specimens, 34 samples (AT = 6, IE = 20, and NT = 8) were considered for analysis. Highest accuracies and statistically significant differences for discrimination of IE from NT and AT from NT were observed for diffusion-based parameters (AUCs >90%), and the perfusion-derived parameter as the most accurate feature in distinguishing IE from AT. A combination of "CBV, MD, T2 _ISO, FLAIR" parameters showed high diagnostic performance for identification of the three subregions (AUC ∼90%).

DATA CONCLUSION

Integration of a few quantitative along with conventional MRI parameters may provide a potential multiparametric imaging biomarker for predicting the histopathologically proven glioma tissue subregions.

LEVEL OF EVIDENCE

2 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2018;48:938-950.

Findings of DTI-p maps in comparison with T2/T2-FLAIR to assess postoperative hyper-signal abnormal regions in patients with glioblastoma

PURPOSE

The aim of this study was to compare diffusion tensor imaging (DTI) isotropic map (p-map) with current radiographically (T₂/T₂-FLAIR) methods based on abnormal hyper-signal size and location of glioblastoma tumor using a semi-automatic approach.

MATERIALS AND METHODS

Twenty-five patients with biopsy-proved diagnosis of glioblastoma participated in this study. T₂, T₂-FLAIR images and diffusion tensor imaging (DTI) were acquired 1 week before radiotherapy. Hyper-signal regions on T₂, T₂-FLAIR and DTI p-map were segmented by means of semi-automated segmentation. Manual segmentation was used as ground truth. Dice Scores (DS) were calculated for validation of semiautomatic method. Discordance Index (DI) and area difference percentage between the three above regions from the three modalities were calculated for each patient.

RESULTS

Area of abnormality in the p-map was smaller than the corresponding areas in the T₂ and T₂-FLAIR images in 17 patients; with mean difference percentage of 30 ± 0.15 and 35 ± 0.15, respectively. Abnormal region in the p-map was larger than the corresponding areas in the T₂-FLAIR and T₂ images in 4 patients; with mean difference percentage of 26 ± 0.17 and 29 ± 0.28, respectively. This region in the p-map was larger than the one in the T₂ image and smaller than the one in the T₂-FLAIR image in 3 patients; with mean difference percentage of 34 ± 0.08 and 27 ± 0.06, respectively. Lack of concordance was observed ranged from 0.214-0.772 for T₂-FLAIR/p-map (average: 0.462 ± 0.18), 0.266-0.794 for T₂ /p-map (average: 0.468 ± 0.13) and 0.123-0.776 for T₂/ T₂-FLAIR (average: 0.423 ± 0.2). These regions on three modalities were segmented using a semi-automatic segmentation method with over 86% sensitivity, 90% specificity and 89% dice score for three modalities.

CONCLUSION

It is noted that T₂, T₂-FLAIR and DTI p-maps represent different but complementary information for delineation of glioblastoma tumor margins. Therefore, this study suggests DTI p-map modality as a candidate to improve target volume delineation based on conventional modalities, which needs further investigations with follow-up data to be confirmed.

Supervised learning based multimodal MRI brain tumour segmentation using texture features from supervoxels

BACKGROUND

Accurate segmentation of brain tumour in magnetic resonance images (MRI) is a difficult task due to various tumour types. Using information and features from multimodal MRI including structural MRI and isotropic (p) and anisotropic (q) components derived from the diffusion tensor imaging (DTI) may result in a more accurate analysis of brain images.

METHODS

We propose a novel 3D supervoxel based learning method for segmentation of tumour in multimodal MRI brain images (conventional MRI and DTI). Supervoxels are generated using the information across the multimodal MRI dataset. For each supervoxel, a variety of features including histograms of texton descriptor, calculated using a set of Gabor filters with different sizes and orientations, and first order intensity statistical features are extracted. Those features are fed into a random forests (RF) classifier to classify each supervoxel into tumour core, oedema or healthy brain tissue.

RESULTS

The method is evaluated on two datasets: 1) Our clinical dataset: 11 multimodal images of patients and 2) BRATS 2013 clinical dataset: 30 multimodal images. For our clinical dataset, the average detection sensitivity of tumour (including tumour core and oedema) using multimodal MRI is 86% with balanced error rate (BER) 7%; while the Dice score for automatic tumour segmentation against ground truth is 0.84. The corresponding results of the BRATS 2013 dataset are 96%, 2% and 0.89, respectively.

CONCLUSION

The method demonstrates promising results in the segmentation of brain tumour. Adding features from multimodal MRI images can largely increase the segmentation accuracy. The method provides a close match to expert delineation across all tumour grades, leading to a faster and more reproducible method of brain tumour detection and delineation to aid patient management.

A Novel Unsupervised Segmentation Approach Quantifies Tumor Tissue Populations Using Multiparametric MRI: First Results with Histological Validation

Purpose

We aimed to precisely estimate intra-tumoral heterogeneity using spatially regularized spectral clustering (SRSC) on multiparametric MRI data and compare the efficacy of SRSC with the previously reported segmentation techniques in MRI studies.

Procedures

Six NMRI nu/nu mice bearing subcutaneous human glioblastoma U87 MG tumors were scanned using a dedicated small animal 7T magnetic resonance imaging (MRI) scanner. The data consisted of T2 weighted images, apparent diffusion coefficient maps, and pre- and post-contrast T2 and T2* maps. Following each scan, the tumors were excised into 2–3-mm thin slices parallel to the axial field of view and processed for histological staining. The MRI data were segmented using SRSC, K-means, fuzzy C-means, and Gaussian mixture modeling to estimate the fractional population of necrotic, peri-necrotic, and viable regions and validated with the fractional population obtained from histology.

Results

While the aforementioned methods overestimated peri-necrotic and underestimated viable fractions, SRSC accurately predicted the fractional population of all three tumor tissue types and exhibited strong correlations (r_necrotic = 0.92, r_{peri-necrotic} = 0.82 and r_viable = 0.98) with the histology.

Conclusions

The precise identification of necrotic, peri-necrotic and viable areas using SRSC may greatly assist in cancer treatment planning and add a new dimension to MRI-guided tumor biopsy procedures.

Electronic supplementary material

The online version of this article (doi:10.1007/s11307-016-1009-y) contains supplementary material, which is available to authorized users.

A fuzzy feature fusion method for auto-segmentation of gliomas with multi-modality diffusion and perfusion magnetic resonance images in radiotherapy

The diffusion and perfusion magnetic resonance (MR) images can provide functional information about tumour and enable more sensitive detection of the tumour extent. We aimed to develop a fuzzy feature fusion method for auto-segmentation of gliomas in radiotherapy planning using multi-parametric functional MR images including apparent diffusion coefficient (ADC), fractional anisotropy (FA) and relative cerebral blood volume (rCBV). For each functional modality, one histogram-based fuzzy model was created to transform image volume into a fuzzy feature space. Based on the fuzzy fusion result of the three fuzzy feature spaces, regions with high possibility belonging to tumour were generated automatically. The auto-segmentations of tumour in structural MR images were added in final auto-segmented gross tumour volume (GTV). For evaluation, one radiation oncologist delineated GTVs for nine patients with all modalities. Comparisons between manually delineated and auto-segmented GTVs showed that, the mean volume difference was 8.69% (±5.62%); the mean Dice’s similarity coefficient (DSC) was 0.88 (±0.02); the mean sensitivity and specificity of auto-segmentation was 0.87 (±0.04) and 0.98 (±0.01) respectively. High accuracy and efficiency can be achieved with the new method, which shows potential of utilizing functional multi-parametric MR images for target definition in precision radiation treatment planning for patients with gliomas.

Semi-automated brain tumor segmentation on multi-parametric MRI using regularized non-negative matrix factorization

Background

Segmentation of gliomas in multi-parametric (MP-)MR images is challenging due to their heterogeneous nature in terms of size, appearance and location. Manual tumor segmentation is a time-consuming task and clinical practice would benefit from (semi-) automated segmentation of the different tumor compartments.

Methods

We present a semi-automated framework for brain tumor segmentation based on non-negative matrix factorization (NMF) that does not require prior training of the method. L1-regularization is incorporated into the NMF objective function to promote spatial consistency and sparseness of the tissue abundance maps. The pathological sources are initialized through user-defined voxel selection. Knowledge about the spatial location of the selected voxels is combined with tissue adjacency constraints in a post-processing step to enhance segmentation quality. The method is applied to an MP-MRI dataset of 21 high-grade glioma patients, including conventional, perfusion-weighted and diffusion-weighted MRI. To assess the effect of using MP-MRI data and the L1-regularization term, analyses are also run using only conventional MRI and without L1-regularization. Robustness against user input variability is verified by considering the statistical distribution of the segmentation results when repeatedly analyzing each patient’s dataset with a different set of random seeding points.

Results

Using L1-regularized semi-automated NMF segmentation, mean Dice-scores of 65%, 74 and 80% are found for active tumor, the tumor core and the whole tumor region. Mean Hausdorff distances of 6.1 mm, 7.4 mm and 8.2 mm are found for active tumor, the tumor core and the whole tumor region. Lower Dice-scores and higher Hausdorff distances are found without L1-regularization and when only considering conventional MRI data.

Conclusions

Based on the mean Dice-scores and Hausdorff distances, segmentation results are competitive with state-of-the-art in literature. Robust results were found for most patients, although careful voxel selection is mandatory to avoid sub-optimal segmentation.

Comparison of unsupervised classification methods for brain tumor segmentation using multi-parametric MRI

Tumor segmentation is a particularly challenging task in high-grade gliomas (HGGs), as they are among the most heterogeneous tumors in oncology. An accurate delineation of the lesion and its main subcomponents contributes to optimal treatment planning, prognosis and follow-up. Conventional MRI (cMRI) is the imaging modality of choice for manual segmentation, and is also considered in the vast majority of automated segmentation studies. Advanced MRI modalities such as perfusion-weighted imaging (PWI), diffusion-weighted imaging (DWI) and magnetic resonance spectroscopic imaging (MRSI) have already shown their added value in tumor tissue characterization, hence there have been recent suggestions of combining different MRI modalities into a multi-parametric MRI (MP-MRI) approach for brain tumor segmentation. In this paper, we compare the performance of several unsupervised classification methods for HGG segmentation based on MP-MRI data including cMRI, DWI, MRSI and PWI. Two independent MP-MRI datasets with a different acquisition protocol were available from different hospitals. We demonstrate that a hierarchical non-negative matrix factorization variant which was previously introduced for MP-MRI tumor segmentation gives the best performance in terms of mean Dice-scores for the pathologic tissue classes on both datasets.

•

Unsupervised classification algorithms are applied for brain tumor segmentation on multi-parametric MRI datasets.

•

Reported mean Dice-scores are in the range of state-of-the-art segmentation algorithms.

•

Hierarchical NMF obtained the best segmentation results in terms of mean Dice-scores for most of the tissue classes.

Independent value of image fusion in unenhanced breast MRI using diffusion-weighted and morphological T2-weighted images for lesion characterization in patients with recently detected BI-RADS 4/5 x-ray mammography findings

OBJECTIVES

The aim of this study was to evaluate the accuracy and applicability of solitarily reading fused image series of T2-weighted and high-b-value diffusion-weighted sequences for lesion characterization as compared to sequential or combined image analysis of these unenhanced sequences and to contrast- enhanced breast MRI.

METHODS

This IRB-approved study included 50 female participants with suspicious breast lesions detected in screening X-ray mammograms, all of which provided written informed consent. Prior to biopsy, all women underwent MRI including diffusion-weighted imaging (DWIBS, b = 1500s/mm²). Images were analyzed as follows: prospective image fusion of DWIBS and T2-weighted images (FU), side-by-side analysis of DWIBS and T2-weighted series (CO), combination of the first two methods (CO+FU), and full contrast-enhanced diagnostic protocol (FDP). Diagnostic indices, confidence, and image quality of the protocols were compared by two blinded readers.

RESULTS

Reading the CO+FU (accuracy 0.92; NPV 96.1 %; PPV 87.6 %) and the CO series (0.90; 96.1 %; 83.7 %) provided a diagnostic performance similar to the FDP (0.95; 96.1 %; 91.3 %; p > 0.05). FU reading alone significantly reduced the diagnostic accuracy (0.82; 93.3 %; 73.4 %; p = 0.023).

CONCLUSIONS

MR evaluation of suspicious BI-RADS 4 and 5 lesions detected on mammography by using a non-contrast-enhanced T2-weighted and DWIBS sequence protocol is most accurate if MR images were read using the CO+FU protocol.

KEY POINTS

• Unenhanced breast MRI with additional DWIBS/T2w-image fusion allows reliable lesion characterization. • Abbreviated reading of fused DWIBS/T2w-images alone decreases diagnostic confidence and accuracy. • Reading fused DWIBS/T2w-images as the sole diagnostic method should be avoided.

Dynamic contrast-enhanced MR in the diagnosis of lympho-associated benign and malignant lesions in the parotid gland

OBJECTIVE

The aim of this study was to determine if dynamic contrast-enhanced (DCE)-MRI can differentiate mucosa-associated lymphoid tissue (MALT) lymphoma from benign lymphoepithelial lesion (BLEL) in the parotid gland.

METHODS

25 patients with tumour-like BLEL and 20 patients with MALT lymphoma in the parotid gland confirmed by pathology were examined pre-operatively using routine MR series and DCE-MRI with a 1.5-T MR unit. The time to peak (TTP), time to start (TTS), SIstart, SImax and SIending were measured and the initial slope of increase (ISI) and relative washout ratio (RWO) were calculated separately from the time-intensity curve (TIC), and the types of TIC were analysed.

RESULTS

There were significant differences in the TTP and ISI between the two lesions (p < 0.001). The sensitivity, specificity and accuracy of TTP were all more than 90%. TICs were divided into three types according to the threshold of TTP and ISI: tumour-like BLEL: gradual type (Type II) and late increase type (Type III); MALT lymphoma: rapid increase and gradual type (Type I).

CONCLUSIONS

DCE-MRI contributed greatly to the differential diagnosis between tumour-like BLEL and MALT lymphoma in the parotid gland.

Interobserver variability of selective region-of-interest measurement protocols for quantitative diffusion weighted imaging in soft tissue masses: Comparison with whole tumor volume measurements

BACKGROUND

To assess the interobserver reliability of three selective region-of-interest (ROI) measurement protocols for apparent diffusion coefficient (ADC) quantifications in soft tissue masses (STMs) compared with whole tumor volume (WTV) ADC measurements.

METHODS

Institutional review board approval was obtained and informed consent was waived. Three observers independently measured minimum and mean ADCs of 73 benign and malignant musculoskeletal STMs using three selective methods (single-slice [SS], predefined three slices [PD], observer-based [OB]) and WTV measurements at 3.0 Tesla. Minimum and mean ADC values derived from each method were compared with WTV measurements, and inter-reader variation was assessed using the intraclass correlation coefficient (ICC). The time required for each method of ADC measurement was recorded.

RESULTS

For the SS, PD, OB, and WTV methods, minimum ADC values ((×10(-3) mm2 /s)) were 0.97, 0.78, 0.73, and 0.67, respectively, and mean ADC values ((×10(-3) mm2 /s)) were 1.49, 1.49, 1.51, and 1.49, respectively. Interobserver agreement was good to excellent for the minimum and mean ADC values for the three readers using the SS, PD, OB, and WTV (ICC range 0.78-0.90). The SS, PD and OB methods required the least amount of measurement time (14 ± 5, 40 ± 17, and 38 ± 15 s, respectively) while the reference WTV method required the longest measurement time (111 ± 54 s) (P < 0.01).

CONCLUSION

While all selective and WTV measurements offer good to excellent interobserver agreement, the selective OB method of ADC measurement results in the closest values to WTV measurements and requires significantly less measurement time than that required for the WTV method.

Automatic segmentation of subcutaneous mouse tumors by multiparametric MR analysis based on endogenous contrast

OBJECT

Contrast-enhanced T1-weighted imaging is usually included in MRI procedures for automatic tumor segmentation. Use of an MR contrast agent may not be appropriate for some applications, however. We assessed the feasability of automatic tumor segmentation by multiparametric cluster analysis that uses intrinsic MRI contrast only.

MATERIALS AND METHODS

Multiparametric MRI consisting of quantitative T1, T2, and apparent diffusion coefficient (ADC) mapping was performed in mice bearing subcutaneous tumors (n = 21). k-means and fuzzy c-means clustering with all possible combinations of MRI parameters, i.e. feature vectors, and 2-7 clusters were performed on the multiparametric data. Clusters associated with tumor tissue were selected on the basis of the relative signal intensity of tumor tissue in T2-weighted images. The optimum segmentation method was determined by quantitative comparison of automatic segmentation with manual segmentation performed by three observers. In addition, the automatically segmented tumor volumes from seven separate tumor data sets were quantitatively compared with histology-derived tumor volumes.

RESULTS

The highest similarity index between manual and automatic segmentation (SI manual,automatic = 0.82 ± 0.06) was observed for k-means clustering with feature vector {T2, ADC} and four clusters. A strong linear correlation between automatically and manually segmented tumor volumes (R (2) = 0.99) was observed for this segmentation method. Automatically segmented tumor volumes also correlated strongly with histology-derived tumor volumes (R (2) = 0.96).

CONCLUSION

Automatic segmentation of mouse subcutaneous tumors can be achieved on the basis of endogenous MR contrast only.