This PDF file contains the front matter associated with SPIE Proceedings Volume 12369, including the Title Page, Copyright information, Table of Contents and Conference Committee lists.

Whispering Gallery Mode (WGM) microresonators are a class of optical sensors with the ability to trap and confine light under optical resonance conditions. Typically, this resonance is excited inside a WGM resonator using expensive and bulky tunable diode lasers, which can be a limiting factor in low-resource settings and in developing economies. In the manuscript, we describe a method of “reverse tuning” to modify the resonance conditions, paving the way for lower cost WGM excitation and ultimately lower cost sensing. We demonstrate three different methods of reverse tuning the WGM using temperature, pressure, and refractive index in a microbubble resonator (MBR), a subclass of WGM sensors that is particularly well-suited for reverse tuning using the three aforementioned methods. By reducing the cost of the MBR platform through reverse tuning, we can make these ultra-sensitive devices more practical and accessible in low-resource settings.

Protein-based therapeutics are used to treat or prevent a range of diseases, but a challenge for the expanded use of these products is the need for cold storage that makes distribution difficult in low-resource settings. Lyophilization is a common method used to stabilize protein-based products. However, this process remains expensive, and many freeze-dried proteins require cold-chain storage. Anhydrous preservation in an amorphous trehalose matrix has been successfully used as an alternative to lyophilization. A new processing technique called light assisted drying (LAD) has been used to successfully dry proteins in preparation for anhydrous storage. Water is selectively heated via near-infrared (1064 nm) illumination, rapidly removing water from a sample, and forming an amorphous matrix that can be stored at supra-zero temperatures. In previous work, large volume samples (0.25 ml) were successfully LAD processed on glass coverslips, but this substrate is not typically used in industry. In this study, large volume samples are LAD processed in vials that are commonly used to lyophilize vaccines. After LAD processing, the samples are stored at room temperature (20◦C) or refrigerated (4◦C) for one month. The end moisture content of samples was determined immediately after processing/storage to evaluate the effectiveness of water removal via LAD. The trehalose matrix was characterized using polarized light imaging to determine if crystallization occurred during storage, potentially damaging embedded proteins. These preliminary studies indicate that LAD can effectively stabilize large volume samples in glass lyophilization vials and demonstrates the potential use of LAD to stabilize products such as vaccines.

Undernutrition is associated with approximately 45% of deaths among children under the age of 5. Furthermore, in 2020, around 149 million children suffered impaired physical/cognitive development due to lack of adequate nutrition. Environmental enteropathy (EE) is associated with undernutrition and is characterized by a multifaceted breakdown in gut function, including an increase in intestinal permeability that can lead to inflammatory responses. However, the role and mechanisms associated with EE (particularly gut permeability) are not well understood. This is partly because current techniques to assess changes in gut permeability, such as endoscopic biopsies, histopathology and chemical tests such as Lactulose:Mannitol assays, are either highly invasive, unreliable or difficult to perform on specific groups of patients (such as infants and patients with urine retention problems). Therefore, low-cost, non-invasive and reliable diagnostic tools are urgently needed for better evaluation of intestinal permeability. Here, we present a compact transcutaneous fluorescence spectroscopy sensor for non-invasive evaluation of gut permeability and report the first in vivo data collected from volunteers in an undernutrition trial. Using this technique and device, fluorescence signals are detected transcutaneously after oral ingestion of a fluorescent solution. Preliminary results demonstrate the potential use of the presented sensor for clinical assessment of gut permeability in low-income settings.

Oral squamous cell carcinoma (OSCC) is the most fatal cancer and has poor survival rate among all cancers. Regular screening of patients can reduce fatality rate of OSCC. Usually, oral cancer screening is accomplished by visual examinations and conventional endoscopes. These regular conventional endoscopes are not able to provide microstructure of altered OSCC tissue so use of micro-endoscopes are favorable for cancer screening, since it can give morphology of altered OSCC tissue. Similarly, the fluorescence (FL) spectroscopy of OSCC tissue can probe the molecular levels of tissue giving information about the progression of oral cancer as it is biophysical and biochemical sensitive. The motivation behind this study is to analyze OSCC by simultaneous micro-endoscopy and spectroscopy with the help of in-house-fabricated low-cost micro-endoscopic system. Simultaneous FL imaging with spectroscopy can provide the morphological changes in altered tissue and record the corresponding intensity counts, wavelength shift and area under the curve (AUC) of the sample. These spectroscopic parameters are directly linked with molecular transitions within the tissue and can be used in real-time cancer screening. FL micro-endoscopic images gives information about fluorescence contrast distribution within the tissue and differ in mean intensity, skewness, kurtosis and energy for OSCC and normal tissue. Present study included 10 OSCC patients and 8 control patients. The value of mean intensity, skewness, kurtosis, energy, peak wavelength, AUC and max counts for OSCC is 163.3±30, -0.567±0.33, 7.82±5.2, ~4.24×105 , 533.28±1.73, ~4.5×106 , ~4.05×104 , respectively and for normal tissue is 63.5±22.4, 0.148±0.05, 4.89±3.89, ~1.65×105 , 528.89±0.81, ~3.3×106 , ~4.63×103 , respectively. The observations from the present study are in-line with screening requirements of OSCC in regular basis screening and suitable for low-resource setting area.

Mobile microscopes, which are cost-effective and field-portable, are rapidly gaining popularity for a variety of different applications, including disease diagnostics. Different imaging modalities, based on transmission, absorption, scattering, phase change, and fluorescence have been developed, depending on the specific application. Mobile microscopes are typically designed to have single magnification, and the maximum achievable resolution is limited by the numerical aperture (N.A.) of the objective lens, due to diffraction of light. Here we present a low-cost imaging substrate containing a two-dimensional (2D) microlens array that enables multi-modal imaging (phase-contrast, dark-field, and fluorescence) with resolution beyond the diffraction limit. The substrate is placed in contact with the sample to form a sandwich structure and the image magnification is attributed to the formation of virtual images by the individual microspheres. This reusable substrate can be easily attached to and detached from the sample, with minimal to no sample damage. A variety of different sizes of glass microspheres (40- 500 𝜇m) were explored and their performance was characterized using fluorescent particles (200 nm – 1 μm). The substrate enabled imaging with sub-400nm resolution, using a low magnification objective lens with N.A.~0.25. Proof-of-concept experiments with this substrate were performed by imaging mammalian tissues, red blood cells, including sickle cells. This technique provides a low-cost, easy- to-use method of improving the resolution of multi-modal imaging systems, which is particularly useful for mobile microscopy applications.

A pH sensor can help understand chemical conditions of solutions, such as precise cell culture medium monitoring in real time. High-quality whispering-gallery-mode (WGM) microresonators have been utilized for surface sensing and are mainly based on the tracking of refractive index changes occurring within a wavelength range from their wall surface. This high sensitivity, reaching up to 10-5 RIU (~2.5 nm/RIU and measured at a femtometer resolution) leads to a broad range of applications, especially for biosensing purposes through the monitoring of molecular binding events. Here, we study the deposition of thin layers of poly(vinyl alcohol)/poly(acrylic acid) (PVA/PAA) hydrogels inside a whispering gallery mode (WGM) microbubble resonator (MBR), fabricated inline with a silica capillary. The generation of such layers is achieved by withdrawing a liquid solution of 25% PVA/PAA in pure water into the MBR and locally heating the microbubble region, resulting in hydrogel formation only in the cavity. The capillary is then rinsed and tested with varying pH solutions. The swelling ability of these gels is directly proportional to the pH of samples brought into contact with the cavity, leading to physical modifications of the WGM coupling properties. We show the preliminary results obtained for the polymerization and characterization of these gels in microbubbles and present the related signal shifts observed for several pH values. We also discuss the gel kinetics over time and investigate practical uses such as reversible and tunable detection of small pH changes.

Mozambique has one of the highest burdens of cervical cancer in the world. However, access to gynecology and pathology services to diagnose cervical precancer in Mozambique is limited. To enable effective cervical dysplasia detection at the point of care, we developed a rugged multi-modal imaging system to acquire colposcopic and high-resolution images of the cervix and identify regions with greater likelihood of high grade cervical precancer. Representative data collected from patients referred for colposcopy at Mavalane Hospital in Maputo, Mozambique illustrate the ability of the multimodal system to identify high grade cervical precancer at the point of care.

Cervical precancer – a major threat in low- and middle-income countries (LMICs) – is often screened through Visual Inspection with acetic Acid (VIA) but suffers from limited reproducibility and inaccuracy. Automatic Visual Evaluation (AVE) utilizes deep learning algorithms based on cervigram or digital images to detect cervical precancer accurately and efficiently. However, important patient data and test results that can provide additional context to the images have not been integrated into the AVE workflow. The NCI ALTS dataset provides over 30,000 cervigram images with hundreds of histopathology-verified positive cases, paired with HPV, cytology, colposcopy, and histology results. To take advantage of both the images and other data, we designed a deep learning image classification algorithm, then modified its architecture to integrate age, HPV status, and cytology result with two methods – concatenation, and multiplication via an added “MetaNet”. We then proposed a two-stage training process that maximized the predictive power of both the images and meta-information while allowing flexibility for missing data. This enabled us to run concurrent inference on a patient's images, test results, and age, and return a binary cervical precancer prediction. The experiments demonstrated that in certain scenarios, the MetaNet model produced a synergistic effect between the images and meta-information by outperformingmodels that only use each component individually, increasing specificity and maintaining high sensitivity. This work provides a direction for the integration of patient meta-information into an end-to-end AVE prediction model, and generally for other medical imaging data, to potentially increase prediction power andmore precisely calibrate patient risk.

India has over 74 million people currently diagnosed with diabetes today with ~35% at risk for diabetic foot ulcers (DFUs). Most patients with DFUs do not require hospitalization unless they have a severe infection with possible sepsis or require surgical intervention. Therefore, DFU care can remain remote for low-risk cases. However, high-risk DFU cases need to be identified and triaged to prevent worsening and hospitalization. These patients are catered to by nurses often performing home visits for wound dressing but with the least wound expertise. Hence, there is a need for point-of-care technologies to triage high-risk DFUs requiring clinical visitation vs stable low-risk DFUs. Recently, a smartphone device or SmartPhone Oxygenation Tool (SPOT) was developed as an add-on optical module to estimate 2D oxygen saturation maps. The hypothesis is that DFUs which are highly infectious or necrotic have poor oxygenation distribution around the wound and can be assessed using our SPOT device as high-risk ulcerations that would require clinical follow-up or visitation. A pilot study was conducted on 11 subjects (43-72 years and 9 male) at Dr. Mohan’s Diabetes Specialties Center (India) to observe oxygen distributions in DFUs and determine if SPOT can be established to triage high-risk DFU cases. Oxygenation patterns in complicated (or high-risk) DFUs, as determined by the clinician, were notably different from those that were stable or low-risk DFUs.

Some medical diagnostic tools, such as urinalysis dipsticks, rely on reading colors accurately for making clinical decisions. Reading color visually can be subject to 1) perceptual differences (inter-observer), 2) environmental factors such as illumination, and 3) target coloring (metamerism), especially among users with limited training and experience. Mobile phone cameras and compact camera modules offer potential low-cost platforms for automated, objective color readouts. However, image colors are a function of camera sensor and illumination characteristics. To restore color fidelity, color correction techniques must be applied to account for systematic deviation. This work aims to provide a quantitative assessment of color correction techniques and reduce variability in color interpretation for urinalysis dipstick results using a low-cost imager. Three color correction methods – linear, polynomial, and root-polynomial regression – were compared for performance in color difference reduction. A standard color checker card was used as reference to compute color correction matrices. A custom imaging system with a low-cost camera module was developed to capture images under controlled illumination. Reference values of the color checker card were obtained with a CM-26d handheld spectrophotometer. The CIE2000 ∆E was used to quantify the color difference between the camera image and the spectrometer to evaluate 3 color correction algorithms. The derived color correction matrices were applied to urinalysis dipstick images and compared to the spectrometer readings. Results indicated that polynomial fitting showed the lowest ∆E during calibration but failed to properly correct urine dipstick colors. Root polynomial offered the best performance in reducing color differences to be below 3 to 4 ∆E. Utilizing L*a*b values for classifying a given dipstick result according to reference concentration levels, it was found that quadratic discriminant analysis (QDA) and k Nearest Neighbor (kNN) classifiers achieved an 82.9% and 97.1% accuracy, respectively.

Helminth infections affect around 1.5 billion people worldwide but have historically been neglected by major healthcare initiatives. Manual microscopic examination to identify parasite eggs in urine or faeces remains the gold-standard diagnostic, but the technique is time consuming and requires bright-field microscopes which can be expensive to transport and maintain. We present a low-cost device which uses deep learning to automate helminth diagnosis from Kato-Katz (KK) faecal smears. The device comprises a 3D-printed microscope, which connects wirelessly to an Android smartphone. Egg detection is accomplished with a ResNet-50 object detection algorithm, trained on a dataset of over 6,000 images of eggs from six common helminth species. The model is exported to TensorFlow Lite and hosted locally in the app, enabling edge computing and removing the need for external internet connection.

Research for miniaturization and portability of optical coherence tomography, which has received considerable attention as one of the pre-diagnosis methods, has been conducted for several years to further expand the utility of optical coherence tomography. In this study, we introduce a method that can dramatically reduce the size of a system and resources using a Raspberry Pi miniature computer and the proposed small spectrometer. The optical systems of the sample stage and the reference stage were configured as half-inch optical components to reduce the size of the system. The size of the sample stage was minimized by using a MEMS scanning mirror. We designed a board that converts the unipolar drive signal into a bipolar signal to drive the MEMS scanning mirror with Raspberry Pi. The MEMS mirror was controlled by a commercial AD/DA conversion board and a developed board that can be controlled via the general-purpose input-output (GPIO) pin of Raspberry Pi. Furthermore, we also designed the spectrometer to fit the 1-inch optical system. The camera was selected as a product that can supply power and transmit data through the USB terminal to operate all other components, including the camera, through a portable charger. Due to camera performance limitations, A-scan 5 kHz was the maximum speed, but the resolution was axial 8.5 μm (Air) and lateral 17.54 μm, showing similar performance to a commercial system. Although the operating speed is slow, it is expected to be used in various fields due to its portability advantage.

By taking advantage of the inherent flexibility of low-cost 3D printing materials, we achieved an optical focus tuning accuracy of ~5 micron with a novel structure design. It shrinks the mechanical displacement by a factor of ~11 through a seesaw-like component. Combing with the built-in flashlight illumination and an off-the-shelf smartphone lens, the total manufacturing cost of our smartphone-based microscope is less than 4 USD. We demonstrated the capability of this design in imaging thick biological specimens. We further applied this device in the cell culture monitoring of VX2 tumor cells because of its portability and flexibility.