Predicting Fibrosis Progression in Renal Transplant Recipients Using Laser-Based Infrared Spectroscopic Imaging

Point-of-care diagnosis of tissue fibrosis: a review of advances in vibrational spectroscopy with machine learning

Histopathology is the gold standard for diagnosing fibrosis, but its routine use is constrained by the need for additional stains, time, personnel and resources. Vibrational spectroscopy is a novel technique that offers an alternative atraumatic approach, with short scan times, while providing metabolic and morphological data. This review evaluates vibrational spectroscopy for the assessment of fibrosis, with a focus on point-of-care capabilities. OVID Medline, Embase and Cochrane databases were systematically searched using PRISMA guidelines for search terms including vibrational spectroscopy, human tissue and fibrosis. Studies were stratified based on imaging modality and tissue type. Outcomes recorded included tissue type, machine learning technique, metrics for accuracy and author conclusions. Systematic review yielded 420 articles, of which 14 were relevant. Ten of these articles considered mid-infrared spectroscopy, three dealt with Raman spectroscopy and one with near-infrared spectroscopy. The metrics for detecting fibrosis were Pearson correlation coefficients ranging from 0.65-0.98; sensitivity from 76-100%; specificity from 90-99%; area under receiver operator curves from 0.83-0.98; and accuracy of 86-99%. Vibrational spectroscopy identified fibrosis in myeloproliferative neoplasms in bone, cirrhotic and hepatocellular carcinoma in liver, end-stage heart failure in cardiac tissue and following laser ablation for acne in skin. It also identified interstitial fibrosis as a predictor of early renal transplant rejection in renal tissue. Vibrational spectroscopic techniques can therefore accurately identify fibrosis in a range of human tissues. Emerging data show that it can be used to quantify, classify and provide data about the nature of fibrosis with a high degree of accuracy with potential scope for point-of-care use.

Non-Perturbative Identification and Subtyping of Amyloidosis in Human Kidney Tissue with Raman Spectroscopy and Machine Learning

Amyloids are proteins with characteristic beta-sheet secondary structures that display fibrillary ultrastructural configurations. They can result in pathologic lesions when deposited in human organs. Various types of amyloid protein can be routinely identified in human tissue specimens by special stains, immunolabeling, and electron microscopy, and, for certain forms of amyloidosis, mass spectrometry is required. In this study, we applied Raman spectroscopy to identify immunoglobulin light chain and amyloid A amyloidosis in human renal tissue biopsies and compared the results with a normal kidney biopsy as a control case. Raman spectra of amyloid fibrils within unstained, frozen, human kidney tissue demonstrated changes in conformation of protein secondary structures. By using t-distributed stochastic neighbor embedding (t-SNE) and density-based spatial clustering of applications with noise (DBSCAN), Raman spectroscopic data were accurately classified with respect to each amyloid type and deposition site. To the best of our knowledge, this is the first time Raman spectroscopy has been used for amyloid characterization of ex vivo human kidney tissue samples. Our approach, using Raman spectroscopy with machine learning algorithms, shows the potential for the identification of amyloid in pathologic lesions.

Infrared spectral microscopy as a tool to monitor lung fibrosis development in a model system

Tissue fibrosis is a progressive and destructive disease process that can occur in many different organs including the liver, kidney, skin, and lungs. Fibrosis is typically initiated by inflammation as a result of chronic insults such as infection, chemicals and autoimmune diseases. Current approaches to examine organ fibrosis are limited to radiological and histological analyses. Infrared spectroscopic imaging offers a potential alternative approach to gain insight into biochemical changes associated with fibrosis progression. In this study, we demonstrate that IR imaging of a mouse model of pulmonary fibrosis can identify biochemical changes observed with fibrosis progression and the beginning of resolution using K-means analysis, spectral ratios and multivariate data analysis. This study demonstrates that IR imaging may be a useful approach to understand the biochemical events associated with fibrosis initiation, progression and resolution for both the clinical setting and for assessing novel anti-fibrotic drugs in a model system.

Tracking Extracellular Matrix Remodeling in Lungs Induced by Breast Cancer Metastasis. Fourier Transform Infrared Spectroscopic Studies

This work focused on a detailed assessment of lung tissue affected by metastasis of breast cancer. We used large-area chemical scanning implemented in Fourier transform infrared (FTIR) spectroscopic imaging supported with classical histological and morphological characterization. For the first time, we differentiated and defined biochemical changes due to metastasis observed in the lung parenchyma, atelectasis, fibrous, and muscle cells, as well as bronchi ciliate cells, in a qualitative and semi-quantitative manner based on spectral features. The results suggested that systematic extracellular matrix remodeling with the progress of the metastasis process evoked a decrease in the fraction of the total protein in atelectasis, fibrous, and muscle cells, as well as an increase of fibrillar proteins in the parenchyma. We also detected alterations in the secondary conformations of proteins in parenchyma and atelectasis and changes in the level of hydroxyproline residues and carbohydrate moieties in the parenchyma. The results indicate the usability of FTIR spectroscopy as a tool for the detection of extracellular matrix remodeling, thereby enabling the prediction of pre-metastatic niche formation.

Towards minimally-invasive, quantitative assessment of chronic kidney disease using optical spectroscopy

The universal pathologic features implicated in the progression of chronic kidney disease (CKD) are interstitial fibrosis and tubular atrophy (IFTA). Current methods of estimating IFTA are slow, labor-intensive and fraught with variability and sampling error, and are not quantitative. As such, there is pressing clinical need for a less-invasive and faster method that can quantitatively assess the degree of IFTA. We propose a minimally-invasive optical method to assess the macro-architecture of kidney tissue, as an objective, quantitative assessment of IFTA, as an indicator of the degree of kidney disease. The method of elastic-scattering spectroscopy (ESS) measures backscattered light over the spectral range 320-900 nm and is highly sensitive to micromorphological changes in tissues. Using two discrete mouse models of CKD, we observed spectral trends of increased scattering intensity in the near-UV to short-visible region (350-450 nm), relative to longer wavelengths, for fibrotic kidneys compared to normal kidney, with a quasi-linear correlation between the ESS changes and the histopathology-determined degree of IFTA. These results suggest the potential of ESS as an objective, quantitative and faster assessment of IFTA for the management of CKD patients and in the allocation of organs for kidney transplantation.

Hyperspectral Tissue Image Segmentation Using Semi-Supervised NMF and Hierarchical Clustering

Hyperspectral imaging (HSI) of tissue samples in the mid-infrared (mid-IR) range provides spectro-chemical and tissue structure information at sub-cellular spatial resolution. Disease states can be directly assessed by analyzing the mid-IR spectra of different cell types (e.g., epithelial cells) and sub-cellular components (e.g., nuclei), provided that we can accurately classify the pixels belonging to these components. The challenge is to extract information from hundreds of noisy mid-IR bands at each pixel, where each band is not very informative in itself, making annotations of unstained tissue HSI images particularly tricky. Because the tissue structure is not necessarily identical between the two sections, only a few regions in unstained HSI image can be annotated with high confidence, even when serial (or adjacent) hematoxylin and eosin stained section is used as a visual guide. In order to completely use both labeled and unlabeled pixels in training images, we have developed an HSI pixel classification method that uses semi-supervised learning for both spectral dimension reduction and hierarchical pixel clustering. Compared to the supervised classifiers, the proposed method was able to account for the vast differences in the spectra of sub-cellular components of the same cell type and to achieve an F1 score of 71.18% on twofold cross-validation across 20 tissue images. To generate further interest in this promising modality, we have released our source code and also showed that disease classification is straightforward after HSI image segmentation.

Digital Staining of High-Definition Fourier Transform Infrared (FT-IR) Images Using Deep Learning

Histological stains, such as hematoxylin and eosin (H&E), are routinely used in clinical diagnosis and research. While these labels offer a high degree of specificity, throughput is limited by the need for multiple samples. Traditional histology stains, such as immunohistochemical labels, also rely only on protein expression and cannot quantify small molecules and metabolites that may aid in diagnosis. Finally, chemical stains and dyes permanently alter the tissue, making downstream analysis impossible. Fourier transform infrared (FT-IR) spectroscopic imaging has shown promise for label-free characterization of important tissue phenotypes and can bypass the need for many chemical labels. Fourier transform infrared classification commonly leverages supervised learning, requiring human annotation that is tedious and prone to errors. One alternative is digital staining, which leverages machine learning to map IR spectra to a corresponding chemical stain. This replaces human annotation with computer-aided alignment. Previous work relies on alignment of adjacent serial tissue sections. Since the tissue samples are not identical at the cellular level, this technique cannot be applied to high-definition FT-IR images. In this paper, we demonstrate that cellular-level mapping can be accomplished using identical samples for both FT-IR and chemical labels. In addition, higher-resolution results can be achieved using a deep convolutional neural network that integrates spatial and spectral features.

Label-free Identification of Antibody-mediated Rejection in Cardiac Allograft Biopsies Using Infrared Spectroscopic Imaging

BACKGROUND

Antibody-mediated rejection (AMR) in cardiac allograft recipients remains less well-understood than acute cellular rejection, is associated with worse outcomes, and portends a greater risk of developing chronic allograft vasculopathy. Diffuse immunohistochemical C4d staining of capillary endothelia in formalin-fixed, paraffin-embedded right ventricular endomyocardial biopsies is diagnostic of immunopathologic AMR but serves more as a late-stage marker. Infrared (IR) spectroscopy may be a useful tool in earlier detection of rejection. We performed mid-IR spectroscopy to identify a unique biochemical signature for AMR.

METHODS

A total of 30 posttransplant formalin-fixed paraffin-embedded right ventricular tissue biopsies (14 positive for C4d and 16 negative for C4d) and 14 native heart biopsies were sectioned for IR analysis. Infrared images of entire sections were acquired and regions of interest from cardiomyocytes were identified. Extracted spectra were averaged across many pixels within each region of interest. Principal component analysis coupled with linear discriminant analysis and predictive classifiers were applied to the data.

RESULTS

Comparison of averaged mid-IR spectra revealed unique features among C4d-positive, C4d-negative, and native heart biopsies. Principal component analysis coupled with linear discriminant analysis and classification models demonstrated that spectral features from the mid-IR fingerprint region of these 3 groups permitted accurate automated classification into each group.

CONCLUSIONS

In cardiac allograft biopsies with immunopathologic AMR, IR spectroscopy reveals a biochemical signature unique to AMR compared with that of nonrejecting cardiac allografts and native hearts. Future study will focus on the predictive capabilities of this IR signature.