SignificanceMueller matrix (MM) microscopy has proven to be a powerful tool for probing microstructural characteristics of biological samples down to subwavelength scale. However, in clinical practice, doctors usually rely on bright-field microscopy images of stained tissue slides to identify characteristic features of specific diseases and make accurate diagnosis. Cross-modality translation based on polarization imaging helps to improve the efficiency and stability in analyzing sample properties from different modalities for pathologists.AimIn this work, we propose a computational image translation technique based on deep learning to enable bright-field microscopy contrast using snapshot Stokes images of stained pathological tissue slides. Taking Stokes images as input instead of MM images allows the translated bright-field images to be unaffected by variations of light source and samples.ApproachWe adopted CycleGAN as the translation model to avoid requirements on co-registered image pairs in the training. This method can generate images that are equivalent to the bright-field images with different staining styles on the same region.ResultsPathological slices of liver and breast tissues with hematoxylin and eosin staining and lung tissues with two types of immunohistochemistry staining, i.e., thyroid transcription factor-1 and Ki-67, were used to demonstrate the effectiveness of our method. The output results were evaluated by four image quality assessment methods.ConclusionsBy comparing the cross-modality translation performance with MM images, we found that the Stokes images, with the advantages of faster acquisition and independence from light intensity and image registration, can be well translated to bright-field images.
Advances in vectorial polarization-resolved imaging are bringing new capabilities to applications ranging from fundamental physics through to clinical diagnosis. Imaging polarimetry requires determination of the Mueller matrix (MM) at every point, providing a complete description of an object’s vectorial properties. Despite forming a comprehensive representation, the MM does not usually provide easily interpretable information about the object’s internal structure. Certain simpler vectorial metrics are derived from subsets of the MM elements. These metrics permit extraction of signatures that provide direct indicators of hidden optical properties of complex systems, while featuring an intriguing asymmetry about what information can or cannot be inferred via these metrics. We harness such characteristics to reveal the spin Hall effect of light, infer microscopic structure within laser-written photonic waveguides, and conduct rapid pathological diagnosis through analysis of healthy and cancerous tissue. This provides new insight for the broader usage of such asymmetric inferred vectorial information.
We propose a cross-modality method that translates polarimetric images into bright-field. In the lung tissue histological analysis, immunohistochemical (IHC) staining of tissues is widely used to specify particular cellular events especially in precision medicine. In this work, we measured hematoxylin and eosin (HE) stained slices by Mueller matrix (MM) microscopy and then fed polarimetric data into a well-designed generative adversarial network (GAN). The network can generate images that are equivalent to the IHC stained from bright-field microscopy. This will assist pathologists with the real IHC staining procedure and pathological diagnosis. Instead of preparing specimens from scratch, we collected already existing specimens, i.e., the adjacent HE and IHC stained slices from the same tissue volume. We adopted the CycleGAN to learn the translation between unaligned images from two domains. We used a U-Net based generator and a PixelGAN based discriminator in the model. The efficacy of this method was demonstrated on smooth muscle actin (SMA) staining in lung tissue. The results are evaluated by three image quality assessment methods by comparing the generated and real staining images.
Early diagnosis and fast screening of cervical cancer is the key to prognosis of treatment and patient survival. Polarimetry technique with high sensitivity to microstructures and low requirement for resolution is promising at facilitating the fast screening and quantitative diagnosis. In this study, we apply the Mueller matrix microscope and multichannel convolutional neural network for the detection of human cervical intraepithelial neoplasia (CIN) samples from normal samples. The Mueller matrix polar decomposition and transformation parameters, rotation invariant parameters, and Mueller matrix symmetry-related parameters of the cervical tissues in epithelial region and at different stages are calculated and analyzed. For detection of early cervical lesions, the selection method of polarimetry parameters based on statistical features and multichannel convolutional neural network (CNN) for classification are proposed. To illustrate, we select the input parameters of CNN models from all commonly used polarimetry parameters according to the amount of information which are evaluated by the mean value, standard deviation, and information entropy of all pixels in 2D parameters images of the training samples. In multichannel CNN classification, each selected parameter is treated as an input of a channel. The proper multichannel CNN models learn deep features from the selected polarimetry parameters of training samples and show good performance for detecting CIN samples under a low-resolution system.
Breast diseases with many distinct histopathological types are showing a rising trend in incidence for decades worldwide. The proliferation of cells and the remodeling of collagen fibers in breast carcinoma tissues may be used to predict breast disease diagnosis, prognosis of treatment, and patient survival. Pathologists can label related typical pathological features as cell nuclei, aligned collagen, and disorganized collagen in hematoxylin and eosin (HE) sections of breast tissues. In this study, we apply the Mueller matrix microscopic imaging to various breast pathological section samples, and calculate corresponding polarimetry basis parameters (PBPs). A pixel-based extraction approach of polarimetry feature parameters (PFPs) is proposed using a mutual information (MI) method and a linear discriminant analysis (LDA) classifier. The three PFPs derived by the proposed learning algorithm are the simplified linear combinations of PBPs with physical meanings, and provide quantitative characterization of the three pathological features in different breast tissues respectively. We present results of the three PFPs of tissue samples from a cohort of 32 clinical patients diagnosed as normal, breast fibroma, breast ductal carcinoma in situ, invasive ductal carcinoma, and breast mucinous carcinoma with analysis of 210 regions-of-interest (ROI). The results demonstrate that the three PFPs of each breast disease tissue have specific value ranges, which has a potential to quantitatively distinguish typical pathological features between different breast tissues. This technique has good prospects for automation of the microstructure identification and prediction of breast disease diagnosis, resulting in the reduction of pathologists’ workload.
As one of the most fundamental features of light, polarization can be used to develop imaging techniques which can provide insight into the optical and structural properties of tissues. Especially, the Mueller matrix polarimetry is suitable to detect the changes in collagen and elastic fibres, which are the main compositions of skin tissue. Here we demonstrate a novel quantitative, non-contact and in situ technique to monitor the microstructural variations of skin tissue during ultraviolet radiation (UVR) induced photoaging based on Mueller matrix polarimetry. Specifically, we measure the twodimensional (2D) backscattering Mueller matrices of nude mouse skin samples, then calculate and analyze the Mueller matrix derived parameters during the skin photoaging and self-repairing processes. To induce three-day skin photoaging, the back skin of each mouse is irradiated with UVR (0.05J/cm2) for five minutes per day. After UVR, the microstructures of the nude mouse skin are damaged. During the process of UV damage, we measure the backscattering Mueller matrices of the mouse skin samples and examine the relationship between the Mueller matrix parameters and the microstructural variations of skin tissue quantitatively. The comparisons between the UVR damaged groups with and without sunscreens show that the Mueller matrix derived parameters are potential indicators for fibrous microstructure variation in skin tissue. The pathological examinations and Monte Carlo simulations confirm the relationship between the values of Mueller matrix parameters and the changes of fibrous structures. Combined with smart phones or wearable devices, this technique may have a good application prospect in the fields of cosmetics and dermatological health.
It has been demonstrated in many biomedical applications that polarization imaging is capable of probing the characteristic microstructural features of complex biological specimens quantitatively and non-invasively. In a recent study, we carried on backscattering Muller matrix imaging on living nude mice skin using oblique illumination by a 633nm LED light source. We quantitatively measured how the anisotropy properties of the living mice skin changes as functions of the UV exposure time. The time course features provide vital clue for the mechanism of UV damage and the effectiveness of sunscreen for reducing such damage. In this work, we report an upgraded system with LED light sources of five different colors ranging from blue to red. The system is calibrated by taking multi-color Mueller matrix images using a single set of rotating achromatic quarter-wave plates. In both in situ applications on living nude mice skin and ex vivo imaging of thick fresh tissue samples, we demonstrated that the multi-color polarized light backscattering measurements are able to reveal more details on the microstructure of the sample, particularly helpful in separating different effects due to photon scattering and propagation.
Polarization imaging is regarded as a promising technique for probing the microstructures, especially the anisotropic fibrous components of tissues. Among the available polarimetric techniques, Mueller matrix imaging has many distinctive advantages. Recently, we have developed a Mueller matrix microscope by adding the polarization state generator and analyzer to a commercial transmission-light microscope, and applied it to differentiate human liver and cervical cancerous tissues with fibrosis. Here we apply the Mueller matrix microscope for quantitative detection of human breast ductal carcinoma, which is a primary form of breast cancers, at different stages. The Mueller matrix polar decomposition (MMPD) and Mueller matrix transformation (MMT) parameters of the breast ductal tissues in different regions at in situ and invasive stages are calculated and analyzed. For more comparisons, Monte Carlo simulations based on the sphere-birefringence model are also carried out. The experimental and simulated results indicate that the Mueller matrix microscope and the polarization parameters can facilitate the quantitative detection of breast ductal carcinoma tissues at different stages.
As one of the salient features of light, polarization contains abundant structural and optical information of media. Recently, as a comprehensive description of polarization property, the Mueller matrix polarimetry has been applied to various biomedical studies such as cancerous tissues detections. In previous works, it has been found that the structural information encoded in the 2D Mueller matrix images can be presented by other transformed parameters with more explicit relationship to certain microstructural features. In this paper, we present a statistical analyzing method to transform the 2D Mueller matrix images into frequency distribution histograms (FDHs) and their central moments to reveal the dominant structural features of samples quantitatively. The experimental results of porcine heart, intestine, stomach, and liver tissues demonstrate that the transformation parameters and central moments based on the statistical analysis of Mueller matrix elements have simple relationships to the dominant microstructural properties of biomedical samples, including the density and orientation of fibrous structures, the depolarization power, diattenuation and absorption abilities. It is shown in this paper that the statistical analysis of 2D images of Mueller matrix elements may provide quantitative or semi-quantitative criteria for biomedical diagnosis.
Mueller matrix polarimetry is a powerful tool for detecting microscopic structures, therefore can be used to monitor physiological changes of tissue samples. Meanwhile, spectral features of scattered light can also provide abundant microstructural information of tissues. In this paper, we take the 2D multispectral backscattering Mueller matrix images of bovine skeletal muscle tissues, and analyze their temporal variation behavior using multispectral Mueller matrix parameters. The 2D images of the Mueller matrix elements are reduced to the multispectral frequency distribution histograms (mFDHs) to reveal the dominant structural features of the muscle samples more clearly. For quantitative analysis, the multispectral Mueller matrix transformation (MMT) parameters are calculated to characterize the microstructural variations during the rigor mortis and proteolysis processes of the skeletal muscle tissue samples. The experimental results indicate that the multispectral MMT parameters can be used to judge different physiological stages for bovine skeletal muscle tissues in 24 hours, and combining with the multispectral technique, the Mueller matrix polarimetry and FDH analysis can monitor the microstructural variation features of skeletal muscle samples. The techniques may be used for quick assessment and quantitative monitoring of meat qualities in food industry.
Polarized light is sensitive to the microstructures of biological tissues and can be used to detect physiological changes. Meanwhile, spectral features of the scattered light can also provide abundant microstructural information of tissues. In this paper, we take the backscattering polarization Mueller matrix images of bovine skeletal muscle tissues during the 24-hour experimental time, and analyze their multispectral behavior using quantitative Mueller matrix parameters. In the processes of rigor mortis and proteolysis of muscle samples, multispectral frequency distribution histograms (FDHs) of the Mueller matrix elements can reveal rich qualitative structural information. In addition, we analyze the temporal variations of the sample using the multispectral Mueller matrix transformation (MMT) parameters. The experimental results indicate that the different stages of rigor mortis and proteolysis for bovine skeletal muscle samples can be judged by these MMT parameters. The results presented in this work show that combining with the multispectral technique, the FDHs and MMT parameters can characterize the microstructural variation features of skeletal muscle tissues. The techniques have the potential to be used as tools for quantitative assessment of meat qualities in food industry.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.