Adaptive Design of Fluorescence Imaging Systems for Custom Resolution, Fields of View, and Geometries

Objective and Impact Statement:We developed a generalized computational approach to design uniform,high-intensity excitation light for low-cost,quantitative fluorescence imaging of in vitro,ex vivo,and in vivo samples with a single device.Introduction:Fluorescence imaging is a ubiquitous tool for biomedical applications.Researchers extensively modify existing systems for tissue imaging,increasing the time and effort needed for translational research and thick tissue imaging.These modifications are applicationspecific,requiring new designs to scale across sample types.Methods:We implemented a computational model to simulate light propagation from multiple sources.Using a global optimization algorithm and a custom cost function,we determined the spatial positioning of optical fibers to generate 2 illumination profiles.These results were implemented to image core needle biopsies,preclinical mammary tumors,or tumor-derived organoids.Samples were stained with molecular probes and imaged with uniform and nonuniform illumination.Results:Simulation results were faithfully translated to benchtop systems.We demonstrated that uniform illumination increased the reliability of intraimage analysis compared to nonuniform illumination and was concordant with traditional histological findings.The computational approach was used to optimize the illumination geometry for the purposes of imaging 3 different fluorophores through a mammary window chamber model.Illumination specifically designed for intravital tumor imaging generated higher image contrast compared to the case in which illumination originally optimized for biopsy images was used.Conclusion:We demonstrate the significance of using a computationally designed illumination for in vitro,ex vivo,and in vivo fluorescence imaging.Applicationspecific illumination increased the reliability of intraimage analysis and enhanced the local contrast of biological features.This approach is generalizable across light sources,biological applications,and detectors.

Multicontrast Pocket Colposcopy Cervical Cancer Diagnostic Algorithm for Referral Populations

Objective and Impact Statement.We use deep learning models to classify cervix images—collected with a low-cost,portable Pocket colposcope—with biopsy-confirmed high-grade precancer and cancer.We boost classification performance on a screened-positive population by using a class-balanced loss and incorporating green-light colposcopy image pairs,which come at no additional cost to the provider.Introduction.Because the majority of the 300,000 annual deaths due to cervical cancer occur in countries with low-or middle-Human Development Indices,an automated classification algorithm could overcome limitations caused by the low prevalence of trained professionals and diagnostic variability in provider visual interpretations.Methods.Our dataset consists of cervical images(n=1,760)from 880 patient visits.After optimizing the network architecture and incorporating a weighted loss function,we explore two methods of incorporating green light image pairs into the network to boost the classification performance and sensitivity of our model on a test set.Results.We achieve an area under the receiver-operator characteristic curve,sensitivity,and specificity of 0.87,75%,and 88%,respectively.The addition of the class-balanced loss and green light cervical contrast to a Resnet-18 backbone results in a 2.5 times improvement in sensitivity.Conclusion.Our methodology,which has already been tested on a prescreened population,can boost classification performance and,in the future,be coupled with Pap smear or HPV triaging,thereby broadening access to early detection of precursor lesions before they advance to cancer.