Improved correlation of histological data with DCE MRI parameter maps by 3D reconstruction, reslicing and parameterization of the histological images

Additive Manufacturing of Resected Oral and Oropharyngeal Tissue: A Pilot Study

Better visualization of tumor structure and orientation are needed in the postoperative setting. We aimed to assess the feasibility of a system in which oral and oropharyngeal tumors are resected, photographed, 3D modeled, and printed using additive manufacturing techniques. Three patients diagnosed with oral/oropharyngeal cancer were included. All patients underwent preoperative magnetic resonance imaging followed by resection. In the operating room (OR), the resected tissue block was photographed using a smartphone. Digital photos were imported into Agisoft Photoscan to produce a digital 3D model of the resected tissue. Physical models were then printed using binder jetting techniques. The aforementioned process was applied in pilot cases including carcinomas of the tongue and larynx. The number of photographs taken for each case ranged from 63 to 195. The printing time for the physical models ranged from 2 to 9 h, costs ranging from 25 to 141 EUR (28 to 161 USD). Digital photography may be used to additively manufacture models of resected oral/oropharyngeal tumors in an easy, accessible and efficient fashion. The model may be used in interdisciplinary discussion regarding postoperative care to improve understanding and collaboration, but further investigation in prospective studies is required.

Development of a Patient-specific Tumor Mold Using Magnetic Resonance Imaging and 3-Dimensional Printing Technology for Targeted Tissue Procurement and Radiomics Analysis of Renal Masses

OBJECTIVE

To implement a platform for colocalization of in vivo quantitative multiparametric magnetic resonance imaging features with ex vivo surgical specimens of patients with renal masses using patient-specific 3-dimensional (3D)-printed tumor molds, which may aid in targeted tissue procurement and radiomics and radiogenomic analyses.

MATERIALS AND METHODS

Volumetric segmentation of 6 renal masses was performed with 3D Slicer (http://www.slicer.org) to create a 3D tumor model. A slicing guide template was created with specialized software, which included notches corresponding to the anatomic locations of the magnetic resonance images. The tumor model was subtracted from the slicing guide to create a depression in the slicing guide corresponding to the exact size and shape of the tumor. A customized, tumor-specific, slicing guide was then printed using a 3D printer. After partial nephrectomy, the surgical specimen was bivalved through the preselected magnetic resonance imaging (MRI) plane. A thick slab of the tumor was obtained, fixed, and processed as a whole-mount slide and was correlated to multiparametric MRI findings.

RESULTS

All patients successfully underwent partial nephrectomy and adequate fitting of the tumor specimens within the 3D mold was achieved in all tumors. Distinct in vivo MRI features corresponded to unique pathologic characteristics in the same tumor. The average cost of printing each mold was US$160.7 ± 111.1 (range: US$20.9-$350.7).

CONCLUSION

MRI-based preoperative 3D printing of tumor-specific molds allow for accurate sectioning of the tumor after surgical resection and colocalization of in vivo imaging features with tissue-based analysis in radiomics and radiogenomic studies.

3D prostate histology reconstruction: an evaluation of image-based and fiducial-based algorithms

PURPOSE

Evaluation of in vivo prostate imaging modalities for determining the spatial distribution and aggressiveness of prostate cancer ideally requires accurate registration of images to an accepted reference standard, such as histopathological examination of radical prostatectomy specimens. Three-dimensional (3D) reconstruction of prostate histology facilitates these registration-based evaluations by reintroducing 3D spatial information lost during histology processing. Because the reconstruction accuracy may constrain the clinical questions that can be answered with these data, it is important to assess the tradeoffs between minimally disruptive methods based on intrinsic image information and potentially more robust methods based on extrinsic fiducial markers.

METHODS

Ex vivo magnetic resonance (MR) images and digitized whole-mount histology images from 12 radical prostatectomy specimens were used to evaluate four 3D histology reconstruction algorithms. 3D reconstructions were computed by registering each histology image to the corresponding ex vivo MR image using one of two similarity metrics (mutual information or fiducial registration error) and one of two search domains (affine transformations or a constrained subset thereof). The algorithms were evaluated for accuracy using the mean target registration error (TRE) computed from homologous intrinsic point landmarks (3-16 per histology section; 232 total) identified on histology and MR images, and for the sensitivity of TRE to rotational, translational, and scaling initialization errors.

RESULTS

The algorithms using fiducial registration error and mutual information had mean ± standard deviation TREs of 0.7 ± 0.4 and 1.2 ± 0.7 mm, respectively, and one algorithm using fiducial registration error and affine transforms had negligible sensitivities to initialization errors. The postoptimization values of the mutual information-based metric showed evidence of errors due to both the optimizer and the similarity metric, and variation of parameters of the mutual information-based metric did not improve its performance.

CONCLUSIONS

The extrinsic fiducial-based algorithm had lower mean TRE and lower sensitivity to initialization than the intrinsic intensity-based algorithm using mutual information. A model relating statistical power to registration error for certain imaging validation study designs estimated that a reconstruction algorithm with a mean TRE of 0.7 mm would require 27% fewer subjects than the method used to initialize the algorithms (mean TRE 1.3 ± 0.7 mm), suggesting the choice of reconstruction technique can have a substantial impact on the design of imaging validation studies, and on their overall cost.

High-field small animal magnetic resonance oncology studies

This review focuses on the applications of high magnetic field magnetic resonance imaging (MRI) and spectroscopy (MRS) to cancer studies in small animals. High-field MRI can provide information about tumor physiology, the microenvironment, metabolism, vascularity and cellularity. Such studies are invaluable for understanding tumor growth and proliferation, response to treatment and drug development. The MR techniques reviewed here include (1)H, (31)P, chemical exchange saturation transfer imaging and hyperpolarized (13)C MRS as well as diffusion-weighted, blood oxygen level dependent contrast imaging and dynamic contrast-enhanced MRI. These methods have been proven effective in animal studies and are highly relevant to human clinical studies.

Characterisation of mobile lipid resonances in tissue biopsies from patients with cervical cancer and correlation with cytoplasmic lipid droplets

The aims of this study were to characterise the major saturated and unsaturated lipid peaks in histologically normal cervical epithelium and stroma, dysplastic epithelium (low-grade cervical intraepithelial neoplasia, CIN) and cancer-containing tissue samples from patients with cervical cancer using diffusion-weighted (1) H high-resolution magic angle spinning MRS, to determine whether mobile lipid resonances (MLRs) distinguish tissue types and to test for a correlation between MLRs and the number of cytoplasmic lipid droplets. Diffusion-weighted spectra of tissue biopsies were acquired using a stimulated echo sequence with bipolar gradients. Major saturated and unsaturated MLRs were identified and multivariate analysis of peak combinations was used to determine the best separation between tissue classes. Lipid droplets were visualised with Nile red staining and fluorescence microscopy. Correlations of saturated lipid resonances (0.9 and 1.3 ppm), polyunsaturated resonances (2.8 ppm), triglycerides (4.3 ppm) and unsaturated resonances (5.3 ppm) with average droplet number (per image) were investigated using a Spearman rank test. A large heterogeneity in lipid content among samples was observed, resulting in no significant differences in MLR intensities of individual peaks between the three tissue classes. Linear discriminant analysis separated 'no cancer' from 'cancer' based on the intensities at 0.9, 1.3, 2.2 and 2.8 ppm [area under the curve (AUC) = 0.939, p < 0.001], 'low-grade CIN' from 'cancer' based on the intensities at 0.9, 4.1, 4.3 and 5.3 ppm (AUC = 0.987, p < 0.001) and 'no cancer' from 'low-grade CIN' based on intensities at 0.9, 2.2 and 4.3 ppm (AUC = 0.984, p < 0.001). The distribution of cytoplasmic lipid droplets was nonuniform and was not related to the presence of epithelial or stromal components. On average, there were more droplets visible in low-grade CIN and cancer-containing tissues. Significant correlations between MLR peaks and lipid droplet number were seen for 0.9 (p = 0.002), 1.3 (p = 0.003) and 2.8 ppm (p = 0.018). MLR combinations indicative of average lipid structure efficiently separated tissue classes. Increased lipid resonances correlated with increased numbers of cytoplasmic lipid droplets.

Mouse model of lymph node metastasis via afferent lymphatic vessels for development of imaging modalities

Animal studies of lymph node metastasis are constrained by limitations in the techniques available for noninvasive monitoring of the progression of lymph node metastasis, as well as difficulties in the establishment of appropriate animal models. To overcome these challenges, this study has developed a mouse model of inter-lymph-node metastasis via afferent lymphatic vessels for use in the development of imaging modalities. We used 14- to 18-week-old MRL/MpJ−/lpr/lpr (MRL/lpr) mice exhibiting remarkable systemic lymphadenopathy, with proper axillary lymph nodes (proper-ALNs) and subiliac lymph nodes (SiLNs) that are 6 to 12 mm in diameter (similar in size to human lymph nodes). When KM-Luc/GFP malignant fibrous histiocytoma-like cells stably expressing the firefly luciferase gene were injected into the SiLN, metastasis could be detected in the proper-ALN within 3 to 9 days, using in vivo bioluminescence imaging. The metastasis route was found to be via the efferent lymphatic vessels of the SiLN, and metastasis incidence depended on the number of cells injected, the injection duration and the SiLN volume. Three-dimensional contrast-enhanced high-frequency ultrasound imaging showed that the blood vessel volume and density in the metastasized proper-ALN significantly increased at 14 days after tumor cell inoculation into the SiLN. The present metastasis model, with lymph nodes similar in size to those of humans, has potential use in the development of ultrasound imaging with high-precision and high-sensitivity as well as other imaging modalities for the detection of blood vessels in lymph nodes during the progression of metastasis.

Decision support system for localizing prostate cancer based on multiparametric magnetic resonance imaging

PURPOSE

There is a growing need to localize prostate cancers on magnetic resonance imaging (MRI) to facilitate the use of image guided biopsy, focal therapy, and active surveillance follow up. Our goal was to develop a decision support system (DSS) for detecting and localizing peripheral zone prostate cancers by using machine learning approach to calculate a cancer probability map from multiparametric MR images (MP-MRI).

METHODS

This IRB approved Health Insurance Portability and Accountability Act compliant retrospective study consisted of 31 patients (mean age and serum prostate specific antigen of 60.4 and 6.62 ng∕ml, respectively) who had MP-MRI at 3 T followed by radical prostatectomy. Seven patients were excluded due to technical issues with their MP-MRI (e.g., motion artifact, failure to perform all sequences). Cancer and normal regions were identified in the peripheral zone by correlating them to whole mount histology slides of the excised prostatectomy specimens. To facilitate the correlation, tissue blocks matching the MR slices were obtained using a MR-based patient-specific mold. Segmented regions on the MP-MRI were correlated to histopathology and used as training sets for the learning system that generated the cancer probability maps. Leave-one-patient-out cross-validation on the cancer and normal regions was performed to determine the learning system's efficacy, an evolutionary strategies approach (also known as a genetic algorithm) was used to find the optimal values for a set of parameters, and finally a cancer probability map was generated.

RESULTS

For the 24 patients that were used in the study, 225 cancer and 264 noncancerous regions were identified from the region maps. The efficacy of DSS was first determined without optimizing support vector machines (SVM) parameters, where a region having a cancer probability greater than or equal to 50% was considered as a correct classification. The nonoptimized system had an f-measure of 85% and the Kappa coefficient of 71% (Rater's agreement, where raters are DSS and ground truth histology). The efficacy of the DSS after optimizing SVM parameters using a genetic algorithm had an f-measure of 89% and a Kappa coefficient of 80%. Thus, after optimization of the DSS there was a 4% increase in the f-measure and a 9% increase in the Kappa coefficient.

CONCLUSIONS

This DSS provides a cancer probability map for peripheral zone prostate tumors based on endorectal MP-MRI. These cancer probability maps can potentially aid radiologists in accurately localizing peripheral zone prostate cancers for planning targeted biopsies, focal therapy, and follow up for active surveillance.

Use of patient-specific MRI-based prostate mold for validation of multiparametric MRI in localization of prostate cancer

OBJECTIVE

To demonstrate the use of a patient-specific magnetic resonance imaging (MRI)-based prostate mold to generate histologic sections that directly correlate to axial MRI slices in a patient with anteriorly located prostate cancer. Anteriorly located prostate cancer has traditionally been difficult to detect on digital rectal examination and transrectal ultrasound-guided biopsy. Multiparametric MRI has potential as a valuable tool for the diagnosis and focal treatment of prostate cancer. A significant difficulty to date has been accurate correlation between the magnetic resonance images and histopathologic specimens.

METHODS

A patient-specific mold from a preoperative T2-weighted MRI scan was created to hold and shape the prostate specimen. Slots for slicing were positioned at 6-mm increments coplanar to the axial MRI slices. After surgical excision, the specimen was inked to maintain the orientation and fixed in formalin. The seminal vesicles were excised, and the prostate was oriented in the mold such that the color-coding matched the anatomic labels on the mold. The specimen was sliced with a single blade and the resultant 6-mm tissue blocks were used for histologic analysis.

RESULTS

Preoperative multiparametric MRI revealed a lesion in the right anterior transition zone that was positive on T2-weighed MRI, apparent diffusion coefficient maps of diffusion-weighted MRI, magnetic resonance spectroscopy, and dynamic contrast-enhanced MRI. The histologic sections obtained using the mold demonstrated a similar Gleason score 6 (3+3) lesion in the right anterior transition zone, correlating with the MRI findings.

CONCLUSION

The use of patient-specific prostate molds to register the MRI findings with the histopathologic specimen in prostate cancer could offer several benefits compared with current specimen processing techniques. This technique might further validate MRI as an accurate tool for prostate cancer localization and staging.

Validation of preclinical multiparametric imaging for prediction of necrosis in hepatocellular carcinoma after embolization

BACKGROUND & AIMS

The hepatocellular carcinoma (HCC) exhibits varying degrees of vascularization with more poorly differentiated carcinoma commonly exhibiting high amounts of vascularization. Transcatheter arterial embolization (TAE) of HCC tumor nodules results in varying amounts of tumor necrosis. Reliable quantification of necrosis after TAE, would aid in treatment planning and testing of novel combinatorial treatment regimen. The aim of this work was to validate different imaging parameters as individual or combined predictors of tumor necrosis after TAE in an orthotopic rat HCC tumor model.

METHODS

Unifocal rat HCC was imaged by T(2)-weighted MRI, quantitative dynamic contrast enhanced (DCE) MRI, diffusion weighted MRI (DWI) and [(18)F]-FDG PET imaging before (day-1) and after (days 1 and 3) TAE. Univariate and multivariate regression analyses were carried out to analyze the ability of each imaging parameter to predict the percent residual vital tumor (vtu) and vital tissue (vti) as determined by quantitative histopathology.

RESULTS

TAE induced a wide range of tumor necrosis. Tumor volume was the only parameter showing a correlation with vti (r(2) = 0.63) before TAE. After TAE, moderate correlations were found for FDG tracer uptake (r(2) = 0.56) and plasma tissue transfer constant (r(2) = 0.55). Correlations were higher for the extravascular extracellular volume fraction (v(e), r(2) = 0.68) and highest for the apparent diffusion coefficient (ADC, r(2) = 0.86). Multivariate analyses confirmed highest correlation of ADC and v(e) with vtu and vti.

CONCLUSIONS

DWI and DCE-MRI with the respective parameters ADC (day 3) and v(e) (day 1) were identified as the most promising imaging techniques for the prediction of necrosis. This study validates a preclinical platform allowing for the improved tumor stratification after TAE and thus the testing of novel combinatorial therapy approaches in HCC.

Peptide and protein imaging mass spectrometry in cancer research

MALDI mass spectrometry is able to acquire protein profiles directly from tissue that can describe the levels of hundreds of distinct proteins. MALDI imaging MS can simultaneously reveal how each of these proteins varies in heterogeneous tissues. Numerous studies have now demonstrated how MALDI imaging MS can generate different protein profiles from the different cell types in a tumor, which can act as biomarker profiles or enable specific candidate protein biomarkers to be identified. MALDI imaging MS can be directly applied to patient samples where its utility is to accomplish untargeted multiplex analysis of the tissue's protein content, enabling the different regions of the tissue to be differentiated on the basis of previously unknown protein profiles/biomarkers. The technique continues to rapidly develop and is now approaching the cusp whereby its potential to provide new diagnostic/prognostic tools for cancer patients can be routinely investigated. Here the latest methodological developments are summarized and its application to a range of tumors is reported in detail. The prospects of MALDI imaging MS are then described from the perspectives of modern pathological practice and MS-based proteomics, to ensure the outlook addresses real clinical needs and reflects the real capabilities of MS-based proteomics of complex tissue samples.

A method for correlating in vivo prostate magnetic resonance imaging and histopathology using individualized magnetic resonance-based molds

A method for the design and rapid manufacture of a patient specific tissue slicing device based on in vivo images in order to facilitate the process of correlating the images with histopathology is presented. The method is applied to radical prostatectomy specimens where the customized mold is designed using magnetic resonance (MR) images of each patient obtained prior to surgery. In this case, the mold holds the prostate in place while a knife with a single blade or multiple blades is inserted in slots which are positioned to obtain tissue blocks corresponding to the slices in the MR images. The resulting histological specimens demonstrate good anatomical correlation with the MR images.

Molecular profiling of angiogenesis with targeted ultrasound imaging: early assessment of antiangiogenic therapy effects

Molecular ultrasound is capable of elucidating the expression of angiogenic markers in vivo. However, the capability of the method for volumetric "multitarget quantification" and for the assessment of antiangiogenic therapy response has rather been investigated. Therefore, we generated cyanoacrylate microbubbles linked to vascular endothelial growth factor receptor 2 (VEGFR2) and alphavbeta3 integrin binding ligands and quantified their accumulation in squamous cell carcinoma xenografts (HaCaT-ras-A-5RT3) in mice with the quantitative volumetric ultrasound scanning technique, sensitive particle acoustic quantification. Specificity of VEGFR2 and alphavbeta3 integrin binding microbubbles was shown, and changes in marker expression during matrix metalloproteinase inhibitor treatment were investigated. In tumors, accumulation of targeted microbubbles was significantly higher compared with nonspecific ones and could be inhibited competitively by addition of the free ligand in excess. Also, multimarker imaging could successfully be done during the same imaging session. Molecular ultrasound further indicated a significant increase of VEGFR2 and alphavbeta3 integrin expression during tumor growth and a considerable decrease in both marker densities after matrix metalloproteinase inhibitor treatment. Histologic data suggested that the increasing VEGFR2 and alphavbeta3 integrin concentrations in tumors during growth are related to an up-regulation of its expression by the endothelial cells, whereas its decrease under therapy is more related to the decreasing relative vessel density. In conclusion, targeted ultrasound appears feasible for the longitudinal molecular profiling of tumor angiogenesis and for the sensitive assessment of therapy effects in vivo.

Pixel-by-pixel analysis of DCE MRI curve patterns and an illustration of its application to the imaging of the musculoskeletal system

Dynamic contrast enhanced (DCE) MRI is a widespread method that has found broad application in the imaging of the musculoskeletal (MSK) system. A common way of analyzing DCE MRI images is to look at the shape of the time-intensity curve (TIC) in pixels selected after drawing an ROI in a highly enhanced area. Although often applied to a number of MSK affections, shape analysis has so far not led to a unanimous correlation between these TIC patterns and pathology. We hypothesize that this might be a result of the subjective ROI approach. To overcome the shortcomings of the ROI approach (sampling error and interuser variability, among others), we created a method for a fast and simple classification of DCE MRI where time-curve enhancement shapes are classified pixel by pixel according to their shape. The result of the analysis is rendered in multislice, 2D color-coded images. With this approach, we show not only that differences on a short distance range of the TIC patterns are significant and cannot be appreciated with a conventional ROI analysis but also that the information that shape maps and conventional standard DCE MRI parameter maps convey are substantially different.