7 January 2019 in Applied Optics
In breast cancer, 20%–30% of cases require a second surgery because of incomplete excision of malignant tissues. Therefore, to avoid the risk of recurrence, accurate detection of the cancer margin by the clinician or surgeons needs some assistance. In this paper, an automated volumetric analysis of normal and breast cancer tissue is done by a machine learning algorithm to separate them into two classes. The proposed method is based on a support-vector-machine-based classifier by dissociating 10 features extracted from the A-line, texture, and phase map by the swept-source optical coherence tomographic intensity and phase images. A set of 88 freshly excised breast tissue [44 normal and 44 cancers (invasive ductal carcinoma tissues)] samples from 22 patients was used in our study. The algorithm successfully classifies the cancerous tissue with sensitivity, specificity, and accuracy of 91.56%, 93.86%, and 92.71% respectively. The present computational technique is fast, simple, and sensitive, and extracts features from the whole volume of the tissue, which does not require any special tissue preparation nor an expert to analyze the breast cancer as required in histopathology. Diagnosis of breast cancer by extracting quantitative features from optical coherence tomographic images could be a potentially powerful method for cancer detection and would be a valuable tool for a fine-needle-guided biopsy.
8 February 2019 in Optics Express
Multi-spectral quantitative phase imaging (QPI) is an emerging imaging modality for wavelength dependent studies of several biological and industrial specimens. Simultaneous multi-spectral QPI is generally performed with color CCD cameras. Here, we present a new approach for accurately measuring the color crosstalk of 2D area detectors, without needing prior information about camera specifications. Color crosstalk is systematically studied and compared using compact interference microscopy on two different cameras commonly used in QPI, single chip CCD (1-CCD) and three chip CCD (3-CCD). The influence of color crosstalk on the fringe width and the visibility of the monochromatic constituents corresponding to three color channels of white light interferogram are studied both through simulations and experiments. It is observed that presence of color crosstalk changes the fringe width and visibility over the imaging field of view. This leads to an unwanted non-uniform background error in the multi-spectral phase imaging of the specimens. The color crosstalk of the detector is observed to be the limiting factor for phase measurement accuracy of simultaneous multi-spectral QPI systems.
1 April 2019 in Optics Letters
In the present Letter, a synthesized pseudothermal light source having high temporal coherence (TC) and low spatial coherence (SC) properties is used. The longitudinal coherence (LC) properties of the spatially extended monochromatic light source are systematically studied. The pseudothermal light source is generated from two different monochromatic laser sources: He–Ne (at 632 nm) and DPSS (at 532 nm). It was found that the LC length of such a light source becomes independent of the parent laser’s TC length for a large source size. For the chosen lasers, the LC length becomes constant to about 30 μm for a laser source size of . Thus, by appropriately choosing the source size, any monochromatic laser light source depending on the biological window can be utilized to obtain high axial resolution in an optical coherence tomography (OCT) system irrespective of its TC length. The axial resolution of 650 nm was obtained using a 1.2 numerical aperture objective lens at a 632 nm wavelength. These findings pave the path for widespread penetration of pseudothermal light into existing OCT systems with enhanced performance. A pseudothermal light source with high TC and low SC properties could be an attractive alternative light source for achieving high axial resolution without needing dispersion compensation as compared to a broadband light source.
13 August 2020 in Biomedical Optics Express
Optical coherence tomography (OCT) is being increasingly adopted as a label-free and non-invasive technique for biomedical applications such as cancer and ocular disease diagnosis. Diagnostic information for these tissues is manifest in textural and geometric features of the OCT images, which are used by human expertise to interpret and triage. However, it suffers delays due to the long process of the conventional diagnostic procedure and shortage of human expertise. Here, a custom deep learning architecture, LightOCT, is proposed for the classification of OCT images into diagnostically relevant classes. LightOCT is a convolutional neural network with only two convolutional layers and a fully connected layer, but it is shown to provide excellent training and test results for diverse OCT image datasets. We show that LightOCT provides 98.9% accuracy in classifying 44 normal and 44 malignant (invasive ductal carcinoma) breast tissue volumetric OCT images. Also, >96% accuracy in classifying public datasets of ocular OCT images as normal, age-related macular degeneration and diabetic macular edema. Additionally, we show ∼96% test accuracy for classifying retinal images as belonging to choroidal neovascularization, diabetic macular edema, drusen, and normal samples on a large public dataset of more than 100,000 images. The performance of the architecture is compared with transfer learning based deep neural networks. Through this, we show that LightOCT can provide significant diagnostic support for a variety of OCT images with sufficient training and minimal hyper-parameter tuning. The trained LightOCT networks for the three-classification problem will be released online to support transfer learning on other datasets.
15 November 2021 in Proceedings of the National Academy of Sciences
Significance: We propose the integration of chip-based optical nanoscopy with high spatially sensitive quantitative phase microscopy to obtain three-dimensional (3D) morphology of liver sinusoidal endothelial cells (LSEC). LSEC contain large numbers of transcellular nanopores —“fenestrations”—in the plasma membrane, typically clustered in groups of 10 to 50 within areas called sieve plates. Determining the diameter and the height of fenestrated regions is an important indicator of a cell’s functionality, and these dimensions can be influenced by agents such as drugs. Our proposed multimodal microscope offers a solution for 3D nanoscale characterization of fenestration diameter and measurement of the optical thickness of the sieve plates.
16 November 2020 in Optics Express
Quantitative phase microscopy (QPM) is a label-free technique that enables monitoring of morphological changes at the subcellular level. The performance of the QPM system in terms of spatial sensitivity and resolution depends on the coherence properties of the light source and the numerical aperture (NA) of objective lenses. Here, we propose high space-bandwidth quantitative phase imaging using partially spatially coherent digital holographic microscopy (PSC-DHM) assisted with a deep neural network. The PSC source synthesized to improve the spatial sensitivity of the reconstructed phase map from the interferometric images. Further, compatible generative adversarial network (GAN) is used and trained with paired low-resolution (LR) and high-resolution (HR) datasets acquired from the PSC-DHM system. The training of the network is performed on two different types of samples, i.e. mostly homogenous human red blood cells (RBC), and on highly heterogeneous macrophages. The performance is evaluated by predicting the HR images from the datasets captured with a low NA lens and compared with the actual HR phase images. An improvement of 9× in the space-bandwidth product is demonstrated for both RBC and macrophages datasets. We believe that the PSC-DHM + GAN approach would be applicable in single-shot label free tissue imaging, disease classification and other high-resolution tomography applications by utilizing the longitudinal spatial coherence properties of the light source.