Correctly diagnosing and staging prostate cancer continues to be a significant clinical challenge. Currently, the standard of care consists of a pathologist’s visual assessment of hematoxylin-and-eosin-stained (HE) histological sections, and designation of a Gleason score based on the top two most common patterns. However, this process is subjective and thus prone to error. Further, lack of standard protocols for staining, makes quantitative analysis of stained tissues difficult. Therefore, there is a significant need to develop new quantitative methods that can provide robust, objective, and accurate information of the aggressiveness and stage of prostate cancer. In this work, we seek to address this challenge using multi-spectral deep-UV microscopy of unstained tissue sections. This method yields valuable insight into the aggressiveness and stage of the disease due to its subcellular spatial resolution and high sensitivity to many endogenous biomolecules, including nucleic acid and proteins. In our approach we use a simple and cost effective wide-field imaging configuration with sequential illumination at multiple wavelengths ranging from 220 nm to 450 nm. Spectral signatures are analyzed in conjunction with the morphology using a geometrical representation of principal component analysis and principles of mathematical morphology. Our results reveal distinct morphological and molecular alterations in the tissue as cancer becomes more aggressive. In this presentation we will detail the design of the multispectral, deep UV microscope; describe our quantitative image analysis; and show preliminary results.
Clinical hematological practice often relies on analysis of the peripheral blood based on microscopic evaluation of blood smears and complete blood count (CBC). Accurate examination of blood cell abnormalities using such methods necessitates complex, time-consuming, and expensive sample preparation as well as instruments which require a many reagents and intensive maintenance. Further, hematology analysis is performed at healthcare centers by trained personnel which significantly limits monitoring frequency for patients with severe conditions and can compromise the treatment outcome. Therefore, a portable, easy-to-use, and inexpensive hematology analysis device can potentially improve quality of life for patients with blood diseases and allow point-of-care monitoring and diagnosis. In this work, we demonstrate label-free blood cell assessment based on deep-ultraviolet (UV) microscopy. Our approach provides quantitative endogenous molecular information from live cells and enables assessment and differentiation of blood cell types based on their molecular and structural signatures. We show the ability of our method by performing classification of polymorphonuclear leukocyte (PMNL) subtypes based on features extracted from deep-UV images. In addition, we demonstrate a pseudo-colorization scheme which accurately mimics the colors produced by standard Giemsa staining and enable visual examination of blood smears. The results of our work paves the way for development of a low-cost and easy-to-use hematological analysis device that can be used for point-of-care applications.
The ultraviolet region of the spectrum offers unique capabilities for label-free molecular imaging of biological samples by providing highly-specific, quantitative information of many important endogenous biomolecules. However, the application of UV spectral imaging to biomedicine has been limited. To this end, we have recently introduced ultraviolet hyperspectral interferometric (UHI) microscopy, which applies interferometry to overcome significant challenges associated with UV spectroscopy when applied to molecular imaging. Here we present an alternative approach for UV multi-spectral microscopy which enables faster wide-field imaging at the expense of fewer spectral data points. Instead of line-scanning to recover high-resolution spectral information with an imaging spectrometer, we detect a wide field-of-view using a UV-sensitive camera and recover the spectral information using several (>5) UV-filters. Moreover, rather than using interferometry to recover the phase to correct for chromatic aberrations, we leverage the chromatic aberrations themselves to obtain a stack of through-focus intensity images (at various wavelengths) and then apply an iterative solution of the Transport of Intensity (TIE) equation to recover the phase and produce in-focus images at all wavelengths without moving the sample or objective. This configuration greatly simplifies the instrumentation, reducing its footprint and making it less expensive, while enabling fast, wide area imaging with better photon efficiency. We assess the capabilities of this technique through a series of simulations and experiments on red blood cells, which show good quantitative agreement with UHI and tabulated hemoglobin absorption properties. Potential biomedical applications are also discussed.
Polymorphonuclear leukocyte (PMNL) count is employed as an immune status indicator for diagnosis of numerous medical conditions. Currently, assessment of PMNLs (i.e., neutrophils, eosinophils, basophils) is a part of complete blood count (CBC) that is performed by trained technicians at healthcare centers and involves sample preparation which is costly and time consuming, both of which limits monitoring frequency. A prominent application of PMNL counting is in identification of neutropenia—a condition describing an abnormally low number of neutrophils in the bloodstream (<1500/μL)—common among cancer patients receiving chemotherapy. Susceptibility to infections in neutropenia patients puts them at an increased risk for medical emergencies, and thus requires constant monitoring of their neutrophil count. Therefore, a portable and easy-to-use, in-home device can potentially circumvent these requirements and enable neutropenia diagnosis. In this work, we demonstrate the feasibility of accurately identifying PMNL subtypes using deep-ultraviolet (UV) microscopy as label-free molecular imaging technique. Our approach benefits from quantitative endogenous molecular information provided by deep-UV imaging, to enable assessment of different cell types based on their molecular and structural signatures. We show the ability of our system to measure neutrophil count in samples containing a mixture of PMNL subtypes as well as whole blood samples by extracting various features from deep-UV images and performing classification to obtain cell count for each subtype. Finally, we will discuss the potential of this technology to empower cancer patients and improve their quality of life via a simple and relatively inexpensive device for point-of-care neutropenia assessment.
Early detection of the most prevalent oral disease worldwide, i.e., dental caries, still remains as one of the major challenges in dentistry. The current dental standard of care relies on caries detection methods, such as visual inspection and x-ray radiography, which lack the sufficient specificity and sensitivity to detect caries at early stages of formation when they can be healed. We report on the feasibility of early caries detection in a clinically and commercially viable thermophotonic imaging system. The system incorporates intensity-modulated laser light along with a low-cost long-wavelength infrared (LWIR; 8 to 14 μm) camera, providing diagnostic contrast based on the enhanced light absorption of early caries. The LWIR camera is highly suitable for integration into clinical platforms because of its low weight and cost. In addition, through theoretical modeling, we show that LWIR detection enhances the diagnostic contrast due to the minimal LWIR transmittance of enamel and suppression of the masking effect of the direct thermal Planck emission. Diagnostic performance of the system and its detection threshold are experimentally evaluated by monitoring the inception and progression of artificially induced occlusal and smooth surface caries. The results are suggestive of the suitability of the developed LWIR system for detecting early dental caries.
Dental caries is one of the most prevailing oral diseases which can be healed if detected in early stages of formation. In this paper, we present a clinically and commercially viable thermophotonic imaging technology for detection of early enamel caries using an inexpensive long-wavelength infrared (LWIR) camera. The efficacy of the system is verified through theoretical simulations as well as experiments carried out on extracted teeth with natural and artificially-induced caries.