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

Habitat Radiomics Based on MRI for Predicting Platinum Resistance in Patients with High-Grade Serous Ovarian Carcinoma: A Multicenter Study

RATIONALE AND OBJECTIVES

This study aims to explore the feasibility of MRI-based habitat radiomics for predicting response of platinum-based chemotherapy in patients with high-grade serous ovarian carcinoma (HGSOC), and compared to conventional radiomics and deep learning models.

MATERIALS AND METHODS

A retrospective study was conducted on HGSOC patients from three hospitals. K-means algorithm was used to perform clustering on T2-weighted images (T2WI), contrast-enhanced T1-weighted images (CE-T1WI), and apparent diffusion coefficient (ADC) maps. After feature extraction and selection, the radiomics model, habitat model, and deep learning model were constructed respectively to identify platinum-resistant and platinum-sensitive patients. A nomogram was developed by integrating the optimal model and clinical independent predictors. The model performance and benefit was assessed using the area under the receiver operating characteristic curve (AUC), net reclassification index (NRI), and integrated discrimination improvement (IDI).

RESULTS

A total of 394 eligible patients were incorporated. Three habitats were clustered, a significant difference in habitat 2 (weak enhancement, high ADC values, and moderate T2WI signal) was found between the platinum-resistant and platinum-sensitive groups (P < 0.05). Compared to the radiomics model (0.640) and deep learning model (0.603), the habitat model had a higher AUC (0.710). The nomogram, combining habitat signatures with a clinical independent predictor (neoadjuvant chemotherapy), yielded a highest AUC (0.721) among four models, with positive NRI and IDI.

CONCLUSION

MRI-based habitat radiomics had the potential to predict response of platinum-based chemotherapy in patients with HGSOC. The nomogram combining with habitat signature had a best performance and good model gains for identifying platinum-resistant patients.

Quantification of intratumoural heterogeneity in mice and patients via machine-learning models trained on PET-MRI data

In oncology, intratumoural heterogeneity is closely linked with the efficacy of therapy, and can be partially characterized via tumour biopsies. Here we show that intratumoural heterogeneity can be characterized spatially via phenotype-specific, multi-view learning classifiers trained with data from dynamic positron emission tomography (PET) and multiparametric magnetic resonance imaging (MRI). Classifiers trained with PET-MRI data from mice with subcutaneous colon cancer quantified phenotypic changes resulting from an apoptosis-inducing targeted therapeutic and provided biologically relevant probability maps of tumour-tissue subtypes. When applied to retrospective PET-MRI data of patients with liver metastases from colorectal cancer, the trained classifiers characterized intratumoural tissue subregions in agreement with tumour histology. The spatial characterization of intratumoural heterogeneity in mice and patients via multimodal, multiparametric imaging aided by machine-learning may facilitate applications in precision oncology.

Advances in PET imaging of cancer

Molecular imaging has experienced enormous advancements in the areas of imaging technology, imaging probe and contrast development, and data quality, as well as machine learning-based data analysis. Positron emission tomography (PET) and its combination with computed tomography (CT) or magnetic resonance imaging (MRI) as a multimodality PET-CT or PET-MRI system offer a wealth of molecular, functional and morphological data with a single patient scan. Despite the recent technical advances and the availability of dozens of disease-specific contrast and imaging probes, only a few parameters, such as tumour size or the mean tracer uptake, are used for the evaluation of images in clinical practice. Multiparametric in vivo imaging data not only are highly quantitative but also can provide invaluable information about pathophysiology, receptor expression, metabolism, or morphological and functional features of tumours, such as pH, oxygenation or tissue density, as well as pharmacodynamic properties of drugs, to measure drug response with a contrast agent. It can further quantitatively map and spatially resolve the intertumoural and intratumoural heterogeneity, providing insights into tumour vulnerabilities for target-specific therapeutic interventions. Failure to exploit and integrate the full potential of such powerful imaging data may lead to a lost opportunity in which patients do not receive the best possible care. With the desire to implement personalized medicine in the cancer clinic, the full comprehensive diagnostic power of multiplexed imaging should be utilized.

Quantifying Tumor Heterogeneity via MRI Habitats to Characterize Microenvironmental Alterations in HER2+ Breast Cancer

This study identifies physiological habitats using quantitative magnetic resonance imaging (MRI) to elucidate intertumoral differences and characterize microenvironmental response to targeted and cytotoxic therapy. BT-474 human epidermal growth factor receptor 2 (HER2+) breast tumors were imaged before and during treatment (trastuzumab, paclitaxel) with diffusion-weighted MRI and dynamic contrast-enhanced MRI to measure tumor cellularity and vascularity, respectively. Tumors were stained for anti-CD31, anti-ɑSMA, anti-CD45, anti-F4/80, anti-pimonidazole, and H&E. MRI data was clustered to identify and label each habitat in terms of vascularity and cellularity. Pre-treatment habitat composition was used stratify tumors into two "tumor imaging phenotypes" (Type 1, Type 2). Type 1 tumors showed significantly higher percent tumor volume of the high-vascularity high-cellularity (HV-HC) habitat compared to Type 2 tumors, and significantly lower volume of low-vascularity high-cellularity (LV-HC) and low-vascularity low-cellularity (LV-LC) habitats. Tumor phenotypes showed significant differences in treatment response, in both changes in tumor volume and physiological composition. Significant positive correlations were found between histological stains and tumor habitats. These findings suggest that the differential baseline imaging phenotypes can predict response to therapy. Specifically, the Type 1 phenotype indicates increased sensitivity to targeted or cytotoxic therapy compared to Type 2 tumors.

Using Variable Flip Angle (VFA) and Modified Look-Locker Inversion Recovery (MOLLI) T1 mapping in clinical OE-MRI

BACKGROUND AND PURPOSE

The imaging technique known as Oxygen-Enhanced MRI is under development as a noninvasive technique for imaging hypoxia in tumours and pulmonary diseases. While promising results have been shown in preclinical experiments, clinical studies have mentioned experiencing difficulties with patient motion, image registration, and the limitations of single-slice images compared to 3D volumes. As clinical studies begin to assess feasibility of using OE-MRI in patients, it is important for researchers to communicate about the practical challenges experienced when using OE-MRI on patients to help the technique advance.

MATERIALS AND METHODS

We report on our experience with using two types of T1 mapping (MOLLI and VFA) for a recently completed OE-MRI clinical study on oropharyngeal squamous cell carcinoma.

RESULTS

We report: (1) the artefacts and practical difficulties encountered in this study; (2) the difference in estimated T1 from each method used - the VFA T1 estimation was higher than the MOLLI estimation by 27% on average; (3) the standard deviation within the tumour ROIs - there was no significant difference in the standard deviation seen within the tumour ROIs from the VFA versus MOLLI; and (4) the OE-MRI response collected from either method. Lastly, we collated the MRI acquisition details from over 45 relevant manuscripts as a convenient reference for researchers planning future studies.

CONCLUSION

We have reported our practical experience from an OE-MRI clinical study, with the aim that sharing this is helpful to researchers planning future studies. In this study, VFA was a more useful technique for using OE-MRI in tumours than MOLLI T1 mapping.

Multiparametric Analysis of Longitudinal Quantitative MRI data to Identify Distinct Tumor Habitats in Preclinical Models of Breast Cancer

This study identifies physiological tumor habitats from quantitative magnetic resonance imaging (MRI) data and evaluates their alterations in response to therapy. Two models of breast cancer (BT-474 and MDA-MB-231) were imaged longitudinally with diffusion-weighted MRI and dynamic contrast-enhanced MRI to quantify tumor cellularity and vascularity, respectively, during treatment with trastuzumab or albumin-bound paclitaxel. Tumors were stained for anti-CD31, anti-Ki-67, and H&E. Imaging and histology data were clustered to identify tumor habitats and percent tumor volume (MRI) or area (histology) of each habitat was quantified. Histological habitats were correlated with MRI habitats. Clustering of both the MRI and histology data yielded three clusters: high-vascularity high-cellularity (HV-HC), low-vascularity high-cellularity (LV-HC), and low-vascularity low-cellularity (LV-LC). At day 4, BT-474 tumors treated with trastuzumab showed a decrease in LV-HC (p = 0.03) and increase in HV-HC (p = 0.03) percent tumor volume compared to control. MDA-MB-231 tumors treated with low-dose albumin-bound paclitaxel showed a longitudinal decrease in LV-HC percent tumor volume at day 3 (p = 0.01). Positive correlations were found between histological and imaging-derived habitats: HV-HC (BT-474: p = 0.03), LV-HC (MDA-MB-231: p = 0.04), LV-LC (BT-474: p = 0.04; MDA-MB-231: p < 0.01). Physiologically distinct tumor habitats associated with therapeutic response were identified with MRI and histology data in preclinical models of breast cancer.

An Automated Segmentation Pipeline for Intratumoural Regions in Animal Xenografts Using Machine Learning and Saturation Transfer MRI

Saturation transfer MRI can be useful in the characterization of different tumour types. It is sensitive to tumour metabolism, microstructure, and microenvironment. This study aimed to use saturation transfer to differentiate between intratumoural regions, demarcate tumour boundaries, and reduce data acquisition times by identifying the imaging scheme with the most impact on segmentation accuracy. Saturation transfer-weighted images were acquired over a wide range of saturation amplitudes and frequency offsets along with T₁ and T₂ maps for 34 tumour xenografts in mice. Independent component analysis and Gaussian mixture modelling were used to segment the images and identify intratumoural regions. Comparison between the segmented regions and histopathology indicated five distinct clusters: three corresponding to intratumoural regions (active tumour, necrosis/apoptosis, and blood/edema) and two extratumoural (muscle and a mix of muscle and connective tissue). The fraction of tumour voxels segmented as necrosis/apoptosis quantitatively matched those calculated from TUNEL histopathological assays. An optimal protocol was identified providing reasonable qualitative agreement between MRI and histopathology and consisting of T₁ and T₂ maps and 22 magnetization transfer (MT)-weighted images. A three-image subset was identified that resulted in a greater than 90% match in positive and negative predictive value of tumour voxels compared to those found using the entire 24-image dataset. The proposed algorithm can potentially be used to develop a robust intratumoural segmentation method.

Multiparametric MRI and Coregistered Histology Identify Tumor Habitats in Breast Cancer Mouse Models

It is well-recognized that solid tumors are genomically, anatomically, and physiologically heterogeneous. In general, more heterogeneous tumors have poorer outcomes, likely due to the increased probability of harboring therapy-resistant cells and regions. It is hypothesized that the genomic and physiologic heterogeneity are related, because physiologically distinct regions will exert variable selection pressures leading to the outgrowth of clones with variable genomic/proteomic profiles. To investigate this, methods must be in place to interrogate and define, at the microscopic scale, the cytotypes that exist within physiologically distinct subregions ("habitats") that are present at mesoscopic scales. MRI provides a noninvasive approach to interrogate physiologically distinct local environments, due to the biophysical principles that govern MRI signal generation. Here, we interrogate different physiologic parameters, such as perfusion, cell density, and edema, using multiparametric MRI (mpMRI). Signals from six different acquisition schema were combined voxel-by-voxel into four clusters identified using a Gaussian mixture model. These were compared with histologic and IHC characterizations of sections that were coregistered using MRI-guided 3D printed tumor molds. Specifically, we identified a specific set of MRI parameters to classify viable-normoxic, viable-hypoxic, nonviable-hypoxic, and nonviable-normoxic tissue types within orthotopic 4T1 and MDA-MB-231 breast tumors. This is the first coregistered study to show that mpMRI can be used to define physiologically distinct tumor habitats within breast tumor models. SIGNIFICANCE: This study demonstrates that noninvasive imaging metrics can be used to distinguish subregions within heterogeneous tumors with histopathologic correlation.

Data-driven identification of tumor subregions based on intravoxel incoherent motion reveals association with proliferative activity

PURPOSE

Intravoxel incoherent motion (IVIM) analysis gives information on tissue diffusion and perfusion and may thus have a potential for e.g. tumor tissue characterization. This work aims to study if clustering based on IVIM parameter maps can identify tumor subregions, and to assess the relevance of obtained subregions by histological analysis.

METHODS

Fourteen mice with human neuroendocrine tumors were examined with diffusion-weighted imaging to obtain IVIM parameter maps. Gaussian mixture models with IVIM maps from all tumors as input were used to partition voxels into k clusters, where k = 2 was chosen for further analysis based on goodness of fit. Clustering was performed with and without the perfusion-related IVIM parameter D * , and with and without including spatial information. The validity of the clustering was assessed by comparison with corresponding histologically stained tumor sections. A Ki-67-based index quantifying the degree of tumor proliferation was considered appropriate for the comparison based on the obtained cluster characteristics.

RESULTS

The clustering resulted in one class with low diffusion and high perfusion and another with slightly higher diffusion and low perfusion. Strong agreement was found between tumor subregions identified by clustering and subregions identified by histological analysis, both regarding size and spatial agreement. Neither D * nor spatial information had substantial effects on the clustering results.

CONCLUSIONS

The results of this study show that IVIM parameter maps can be used to identify tumor subregions using a data-driven framework based on Gaussian mixture models. In the studied tumor model, the obtained subregions showed agreement with proliferative activity.

Running in the wheel: Defining individual severity levels in mice

The fine-scale grading of the severity experienced by animals used in research constitutes a key element of the 3Rs (replace, reduce, and refine) principles and a legal requirement in the European Union Directive 2010/63/EU. Particularly, the exact assessment of all signs of pain, suffering, and distress experienced by laboratory animals represents a prerequisite to develop refinement strategies. However, minimal and noninvasive methods for an evidence-based severity assessment are scarce. Therefore, we investigated whether voluntary wheel running (VWR) provides an observer-independent behaviour-centred approach to grade severity experienced by C57BL/6J mice undergoing various treatments. In a mouse model of chemically induced acute colitis, VWR behaviour was directly related to colitis severity, whereas clinical scoring did not sensitively reflect severity but rather indicated marginal signs of compromised welfare. Unsupervised k-means algorithm–based cluster analysis of body weight and VWR data enabled the discrimination of cluster borders and distinct levels of severity. The validity of the cluster analysis was affirmed in a mouse model of acute restraint stress. This method was also applicable to uncover and grade the impact of serial blood sampling on the animal’s welfare, underlined by increased histological scores in the colitis model. To reflect the entirety of severity in a multidimensional model, the presented approach may have to be calibrated and validated in other animal models requiring the integration of further parameters. In this experimental set up, however, the automated assessment of an emotional/motivational driven behaviour and subsequent integration of the data into a mathematical model enabled unbiased individual severity grading in laboratory mice, thereby providing an essential contribution to the 3Rs principles.

Animal-based biomedical research is often accompanied by experience of discomfort or pain by the animal. Recognition of disturbed animal welfare is mandatory, and the classification and assessment of its severity is a crucial part of the legislative framework in the European Union (EU). In the present study, we analysed voluntary wheel running (VWR) behaviour as a measure of compromised welfare in a mouse colitis model. Unsupervised mathematical clustering of clinical and VWR data enabled us to allocate and classify severity levels. This cluster model was verified using VWR data from a restraint stress model and allowed us to uncover the impact of routine experimental procedures on these mice. We propose that clustering of VWR behaviour provides a useful method for assessing the severity level of experimental procedures conducted on mice.

Fully Automatic Lesion Localization and Characterization: Application to Brain Tumors Using Multiparametric Quantitative MRI Data

When analyzing brain tumors, two tasks are intrinsically linked, spatial localization, and physiological characterization of the lesioned tissues. Automated data-driven solutions exist, based on image segmentation techniques or physiological parameters analysis, but for each task separately, the other being performedmanually or with user tuning operations. In this paper, the availability of quantitative magnetic resonance (MR) parameters is combined with advancedmultivariate statistical tools to design a fully automated method that jointly performs both localization and characterization. Non trivial interactions between relevant physiologicalparameters are capturedthanks to recent generalized Student distributions that provide a larger variety of distributional shapes compared to the more standard Gaussian distributions. Probabilisticmixtures of the former distributions are then consideredto account for the different tissue types and potential heterogeneity of lesions. Discriminative multivariate features are extracted from this mixture modeling and turned into individual lesion signatures. The signatures are subsequently pooled together to build a statistical fingerprintmodel of the different lesion types that captures lesion characteristics while accounting for inter-subject variability. The potential of this generic procedure is demonstrated on a data set of 53 rats, with 36 rats bearing 4 different brain tumors, for which 5 quantitative MR parameters were acquired.

Data-driven mapping of hypoxia-related tumor heterogeneity using DCE-MRI and OE-MRI

PURPOSE

Previous work has shown that combining dynamic contrast-enhanced (DCE)-MRI and oxygen-enhanced (OE)-MRI binary enhancement maps can identify tumor hypoxia. The current work proposes a novel, data-driven method for mapping tissue oxygenation and perfusion heterogeneity, based on clustering DCE/OE-MRI data.

METHODS

DCE-MRI and OE-MRI were performed on nine U87 (glioblastoma) and seven Calu6 (non-small cell lung cancer) murine xenograft tumors. Area under the curve and principal component analysis features were calculated and clustered separately using Gaussian mixture modelling. Evaluation metrics were calculated to determine the optimum feature set and cluster number. Outputs were quantitatively compared with a previous non data-driven approach.

RESULTS

The optimum method located six robustly identifiable clusters in the data, yielding tumor region maps with spatially contiguous regions in a rim-core structure, suggesting a biological basis. Mean within-cluster enhancement curves showed physiologically distinct, intuitive kinetics of enhancement. Regions of DCE/OE-MRI enhancement mismatch were located, and voxel categorization agreed well with the previous non data-driven approach (Cohen's kappa = 0.61, proportional agreement = 0.75).

CONCLUSION

The proposed method locates similar regions to the previous published method of binarization of DCE/OE-MRI enhancement, but renders a finer segmentation of intra-tumoral oxygenation and perfusion. This could aid in understanding the tumor microenvironment and its heterogeneity. Magn Reson Med 79:2236-2245, 2018. © 2017 The Authors Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance in Medicine. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Linking imaging to omics utilizing image-guided tissue extraction

Imaging provides an insight into biological patho-mechanisms of diseases. However, the link between the imaging phenotype and the underlying molecular processes is often not well understood. Methods such as metabolomics and proteomics reveal detailed information about these processes. Unfortunately, they provide no spatial information and thus cannot be easily correlated with functional imaging. We have developed an image-guided milling machine and unique workflows to precisely isolate tissue samples based on imaging data. The tissue samples remain cooled during the entire procedure, preventing sample degradation. This enables us to correlate, at an unprecedented spatial precision, comprehensive imaging information with metabolomics and proteomics data, leading to a better understanding of diseases.

Phenotypic heterogeneity is commonly observed in diseased tissue, specifically in tumors. Multimodal imaging technologies can reveal tissue heterogeneity noninvasively in vivo, enabling imaging-based profiling of receptors, metabolism, morphology, or function on a macroscopic scale. In contrast, in vitro multiomics, immunohistochemistry, or histology techniques accurately characterize these heterogeneities in the cellular and subcellular scales in a more comprehensive but ex vivo manner. The complementary in vivo and ex vivo information would provide an enormous potential to better characterize a disease. However, this requires spatially accurate coregistration of these data by image-driven sampling as well as fast sample-preparation methods. Here, a unique image-guided milling machine and workflow for precise extraction of tissue samples from small laboratory animals or excised organs has been developed and evaluated. The samples can be delineated on tomographic images as volumes of interest and can be extracted with a spatial accuracy better than 0.25 mm. The samples remain cooled throughout the procedure to ensure metabolic stability, a precondition for accurate in vitro analysis.