搜索结果：Proceedings of SPIE--the International Society for Optical Engineering

Low-dose computed tomography (CT) remains a popular research topic with the advent of an increasing number of algorithmic solutions to control noise. One such approach that enforces data consistency through a model-based data likelihood term but that also includes a deep learning generative prior is Diffusion Posterior Sampling (DPS). This technique is formulated within a probabilistic framework and is capable of generating high-quality reconstructions under noisy and/or undersampled conditions. However, one major unanswered question is, given the opportunity to design a low-dose protocol, how should low dose be achieved - through sparse sampling or reduced fluence per projection. In this work, we conducted a simulation study and systematically investigated the impact of acquisition parameters - the number of views ( n view ) and incident photons per view ( I 0 ) - on DPS-based CT reconstruction. We performed a 2D sweep over different combinations of the number of views ( n view ) and incident photons per view ( I 0 ) and compared reconstructions with an equivalent total incident photons (TIP). Reconstruction quality was evaluated in terms of PSNR (Peak Signal-to-Noise Ratio), bias, and posterior sample variability. We found that the number of views had a strong influence on image quality and that most performance curves showed a transition where too few views had a large negative impact on performance. We observed that there is an advantage to be gained by jointly optimizing both the fluence per view and the number of views, with a trend of an increasing number of views required for a higher total incident fluence. These findings provide a strategy for optimizing CT acquisition protocols that adapt both fluence per view and sparsity to optimally maintain image quality at reduced radiation doses.

Paired inspiratory-expiratory computed tomography (CT) scans enable quantification of gas trapping alterations due to small airway disease and emphysema through the motion of the lung tissue for people with chronic obstructive pulmonary disease (COPD). Deformable image registration of these paired CT scans is often used to assess the regional volumetric changes in the lung. However, variations in reconstruction protocols, particularly the reconstruction kernels between paired inspiratory-expiratory scans are often overlooked, and these variations introduce errors during quantitative image analysis. In this work, we propose a two-stage pipeline to harmonize reconstruction kernels between paired inspiratory-expiratory scans and perform deformable image registration for data acquired from the COPDGene study. We use a cycle generative adversarial network (GAN) for image synthesis to harmonize inspiratory scans reconstructed with a hard kernel (BONE) to match expiratory scans reconstructed with a soft kernel (STANDARD). We then perform deformable image registration to register the expiratory scans to the inspiratory scans. We validated harmonization by measuring emphysema using a publicly available segmentation algorithm, both before and after harmonization. Our results show that harmonization significantly reduces inconsistencies in emphysema measurement, decreasing the median emphysema scores from 10.479% to 3.039% with a reference median score of 1.305% from the STANDARD kernel as a harmonization target. We validate the registration accuracy by observing the Dice overlap between emphysema regions on the inspiratory, expiratory and deformed images. The Dice coefficient between the fixed inspiratory emphysema masks and deformably registered emphysema masks increases across different stages of registration with statistical significance (p<0.001). Additionally, we show that deformable registration is robust to kernel variation.

Autofluorescence describes light emitted from a naturally occurring substance when exposed to light of a shorter wavelength. It has been shown that cancer-related changes in tissue composition can be responsible for shifts in autofluorescence intensity of biological samples. Several prior studies have worked to characterize the spectral properties of autofluorescent molecules using spectrofluorometers and microscope systems. Here, we quantitatively characterized endogenous fluorophores using spectrofluorometry and fluorescence microscopy with the objective of establishing a reference spectral library for subsequent spectral unmixing of hyperspectral images of mouse tissues during colorectal cancer (CRC) progression. Endogenous fluorophores of interest include collagen, elastin, nicotinamide adenine dinucleotide (NADH), flavin adenine dinucleotide (FAD), protoporphyrin IX (PPIX), tryptophan, and tyrosine. In the tumor microenvironment (TME), collagen is overproduced and remodeled affecting immune cell infiltration and treatment resistance; elastin degradation generates fragments to promote or inhibit tumor development; NADH and FAD play essential roles in reduction-oxidation reactions; PPIX precedes heme in the heme biosynthesis pathway; and tryptophan and tyrosine are autofluorescent amino acids. Concentration fluctuations of these endogenous fluorophores are directly related to autofluorescent properties which contribute to shifts in bulk autofluorescence spectra concurrent with tissue restructuring caused by the TME. Current hyperspectral fluorescence microscopy techniques utilize emission-scanning hyperspectral imaging (Em-HSI). However, this method requires long acquisition times that would be incompatible with real-time endoscopic screening. Here, we utilize a novel excitation-scanning hyperspectral imaging (Ex-HSI) approach to establish a comprehensive spectral library of several biologically-relevant autofluorescent molecules for downstream spectral unmixing of mouse CRC tissue spectral image stacks.

Efficiency in screening mammogram interpretation is complex and involves the time to view the images and report the findings. The purpose of this study was to evaluate the factors impacting the interpretation time of screening mammograms by radiology residents. In this IRB-approved prospective study, radiology residents performed screening mammography simulations during their breast imaging rotations. The research simulation space replicated the clinical practice area without the interruptions. Residents interpreted 50 enriched screening mammograms and completed eye tracking on the first ten cases (six digital breast tomosynthesis (DBT) and four full-field digital mammogram (FFDM) cases). The amount of time to interpret (visual inspection) and report the 10 eye-tracking cases (electronic reporting) was evaluated and compared based on the breast imaging rotation (first, second, or third) and postgraduate year level (PGY). We present the time for each of the 10 eye tracking cases that were subdivided into visual inspection time of the mammogram and reporting time. Comparisons were performed with t-test and significance at p&#x2264;0.05. 25 residents performed the simulation with eye tracking: 14 residents during their first rotation, 7 in their second, and 4 in their third. There were 5 PGY2, 8 PGY3, 7 PGY4 and 5 PGY5 residents.The average overall interpretation time for 50 cases was 4.12 hours (SD=1.26), and 1.35 hours for the 10 eye tracking cases. For the eye-tracking cases, the average visual inspection time per case was 2.13 minutes with a reporting time of 3.08 minutes for an overall interpretation time of 5.21 minutes per case.For eye-tracking cases, the overall interpretation time significantly decreased from those completing their first (6.0 min) to third (4.2 min) breast rotation (p<0.001) and from PGY2 (6.6 min) to PGY5 (4.8 min) (p=0.004). The visual inspection time significantly decreased from PGY2 (2.6 min) compared to PGY5 (1.8 min) (p=0.001) level and from the first to the third rotation (p=0.021).No significant differences were observed for overall interpretation time between DBT (5.29 min) and FFDM (5.69 min) (p=0.319) or visual inspection time per case for DBT (2.20 min) and FFDM (2.03 min) (p=0.335). Interpretation time for screening mammography decreases with increasing PGY level and the number of breast imaging rotations, which is likely due to experience and confidence with the types of exams. No difference was noted between DBT and FFDM average times.

Mid-infrared (mid-IR) optical fiber sensors offer highly specific and sensitive detection and analysis of various chemical species due to many molecular vibrations and fundamental absorption bands in this range. In this paper, we present a novel optical fiber probe design allowing for controlled optical pathlength. The optical fiber probe was fabricated using a silver halide polycrystalline fiber, a gold-coated short fiber as a mirror, and a connector to align the two parts to face each other. The outer diameter of the connector, 1.59 mm, dictates the overall probe diameter. To demonstrate the sensing performance, a quantum cascade laser (QCL) was coupled to the optical fiber probe to measure glucose solutions at physiological concentration levels by monitoring the C-O stretching vibration at 1,035 cm-1. A detection limit of 8.91 mM for glucose was achieved. The results highlight the potential of the proposed optical fiber probe for molecular detection and analysis, offering a promising solution for chemical and biomedical applications.

Radiation-induced xerostomia remains a common and debilitating side effect in head-and-neck cancer radiotherapy, despite advances in volumetric modulated arc therapy (VMAT). Traditional dose-volume histogram (DVH) metrics capture only part of the variation in toxicity, motivating the use of multimodal imaging biomarkers such as dosiomics and radiomics to characterize dose distribution and tissue response better. In this pilot study, we present an integrated framework combining DVH metrics, 3D dosiomics features, baseline planning CT (pCT) radiomics, and novel longitudinal delta-radiomics derived from daily cone-beam CT-based synthetic CT (sCT) images to predict post-treatment xerostomia severity. In a cohort of ten high-risk oropharyngeal cancer patients treated with VMAT at the Cleveland Clinic, wrapper-based feature selection yielded a compact set of 15 predictors (5 DVH, 3 dosiomics, 4 pCT radiomics, 3 &#x394;-sCT radiomics). Using cross-validation, four classifiers, including support-vector machine (SVM), regularized logistic regression (GLMnet), Na&#xef;ve Bayes, and k-nearest neighbors, achieved consistently strong performance for discriminating grade I vs. grade II xerostomia, with AUC of 0.97-1.00, accuracy of 0.90-0.93, uniformly high sensitivity (1.00), specificity of 0.75-0.83, and F1 scores of 0.923-0.945. SVM and GLMnet showed the best overall balance of discrimination and robustness. These results demonstrate the potential of integrating dosiomics with multiphase radiomics, particularly time-resolved delta-radiomics, for individualized xerostomia risk prediction.

The cAMP signaling system coordinates a plethora of cellular responses, including contraction, movement, excitability, gene transcription and translation, and Ca2+ handling. However, the mechanisms by which cAMP signals differentially regulate specific physiological responses remain elusive. Recent studies from a variety of groups provide compelling evidence that the subcellular location of cAMP signals contributes to signal specificity. Our previous work using a spectral Nikon A1R confocal microscope system highlights the need to consider cAMP signals in three spatial distributions (x,y,z). However, this and other commercially available microscope systems lack sufficient sensitivity and speed to track the kinetics of cAMP signals in four dimensions (4D: x,y,z,t). Here, we propose a combination of excitation scan-based hyperspectral imaging and digital deconvolution approaches to overcome this limitation. Imaging was performed using a custom excitation-scanning hyperspectral microscope based upon an Eclipse Ti2 microscope platform (Nikon Instruments), Titan 300 Xenon arc lamp (Sunoptics), VersaChrome thin-film tunable filters (Semrock Inc.) in a VF-5 tiltable filter wheel (Sutter Instruments), and a Prime 95B sCMOS camera (Photometrics). HEK293 and LNCaP cells were transfected with caDDis cAMP probes (Montana Molecular), and subsequently labeled with MitoTracker and/or NucBlue. Cells were imaged at excitation wavelengths ranging from 360 to 525 nm in 5 nm increments and emission was detected using a 532 nm long pass dichroic filter (Semrock Inc.). Hyperspectral image stacks were acquired at 250-500 nm axial increments (z-stacks). NIS Elements analysis software was used to deconvolve image data at each acquisition wavelength. The contributions from each fluorophore were then assessed using nonnegative linear unmixing. Preliminary results indicate that digital deconvolution of hyperspectral image stacks offers an effective approach to assess the spatial distributions of multiple fluorophores including cAMP probes in 5D (x, y, z, &#x3bb;, and t). This work was supported by the University of South Alabama Center for Lung Biology, NIH R01HL169522 and S10OD028606, and NSF DBI-2408000.

Accurate detection and segmentation of cancerous lesions from computed tomography (CT) scans is essential for automated treatment planning and cancer treatment response assessment. Transformer-based models with self-supervised pretraining can produce reliably accurate segmentation from in-distribution (ID) data but degrade when applied to out-of-distribution (OOD) datasets. We address this challenge with RF-Deep, a random forest classifier that utilizes deep features from a pretrained transformer encoder of the segmentation model to detect OOD scans and enhance segmentation reliability. The segmentation model comprises a Swin Transformer encoder, pretrained with masked image modeling (SimMIM) on 10,432 unlabeled 3D CT scans covering cancerous and non-cancerous conditions, with a convolution decoder, trained to segment lung cancers in 317 3D scans. Independent testing was performed on 603 3D CT public datasets that included one ID dataset and four OOD datasets comprising chest CTs with pulmonary embolism (PE) and COVID-19, and abdominal CTs with kidney cancers and healthy volunteers. RF-Deep detected OOD cases with a FPR95 of 18.26%, 27.66%, and < 1.0% on PE, COVID-19, and abdominal CTs, consistently outperforming established OOD approaches. The RF-Deep classifier provides a simple and effective approach to enhance reliability of cancer segmentation in ID and OOD scenarios.

While much of the evaluation of artificial intelligence (AI) in healthcare has focused on technical performance metrics such as accuracy or area under the curve, real-world adoption critically depends on how AI reshapes clinical workflows, operations, and revenue streams. Simulation models provide a means to anticipate these impacts before implementation, allowing stakeholders to weigh benefits against potential harm. In this study, we used discrete-event simulation to evaluate an AI-assisted workflow for same-day diagnostic breast imaging following abnormal screening mammograms. The revised workflow captured an additional of 1.1% mammography screening patients who might otherwise be lost to follow-up. It also eliminated the need for a second visit for diagnostic workup for 11% of mammography screening patients. It also increased daily work relative value units by 4.8%, translating to an estimated $15,979 in additional annual gain, while extending clinic operating hours by 2.9%, amounting to 109.5 hours annually. These findings highlight how simulation modeling can inform the operational and financial implications of AI adoption in imaging workflows in clinical practice.

Latent Diffusion Models (LDMs) introduce exciting opportunities in medical imaging, from disease progression prediction to interpolation to generate entire datasets of rare data. The stochastic nature of generative models makes it challenging to validate their outputs and assess their robustness across diverse datasets. BrLP, a state-of-the-art LDM for T1-weighted (T1w) images that incorporates auxiliary brain volume information, has been evaluated on Alzheimer's Disease (AD) progression, and achieves structural similarity index (SSIM) of 0.91 &#xb1; 0.03. In this work, we conducted a pilot study of the BrLP model using the Baltimore Longitudinal Study of Aging (BLSA) dataset. Our objectives are to (1) evaluate the model performance on an external dataset using pretrained image-based and brain image-based metrics such as mean squared error (MSE), similarity index, and mean absolute error (MAE) between conditional and unconditional brain regions; and (2) determine if a harmonization step, in addition to the proposed model's preprocessing steps, is required to improve performance. We found that the BrLP is robust to T1w imaging scanner effects, and harmonization is not required. However, there exists a bias toward a younger population compared to the BLSA cohort (BrLP reported 0.91 &#xb1; 0.03 SSIM in a cohort with age range 74 &#xb1;7 years; SSIM in our experiment is 0.91 &#xb1; 0.03 in cognitively normal subjects with age range 79.47 years &#xb1; 7.35 and AD cohort 0.90 &#xb1; 0.012 with age range 83.73 years &#xb1; 6.01). Interestingly, when the model's input was changed to simulate AD progression instead of normal aging, a higher SSIM of 0.91 &#xb1; 0.0012 was achieved compared to a ground truth non-AD scan, suggesting a potential mismatch. However, the resulting lower conditional volume regions function as expected. The model's architecture shows promise for longitudinal T1w imaging studies.

Advances in artificial intelligence have increased the availability of auto-segmentation tools. However, conventional accuracy metrics cannot capture regional segmentation differences between AI models or with respect to reference segmentations, necessary to interpret contouring variations. To address this, we developed a novel distance metric based on topological data analysis (TDA) to evaluate 3D point cloud representations of segmentations applied to six organs-at-risk (OARs) and lung gross tumor volume (GTV). A total of 34 CTs and 54 CBCTs were analyzed to compare a bespoke AI segmentation method with reference clinical contours. TDA involved: (1) converting segmentations into 3D point clouds, (2) clustering them into regions via K-means with fixed seeds and cluster numbers determined by the Elbow method, (3) constructing directed graphs for AI and reference clusters using centroids as nodes, and (4) computing distances using unbalanced optimal mass transport. Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD95) were also calculated. TDA successfully identified local regions of high deviation in both OARs and GTVs of varying shapes. It correlated positively with HD95 and negatively with DSC based on Pearson's correlation coefficient. Computation was efficient, averaging 1.72 s, and TDA effectively highlighted regions of greatest mismatch, providing quantitative visualization of poor concordance. In conclusion, we developed a new TDA metric for comparing auto-segmentation of GTV and OARs. Importantly, it allows visualization of mismatching regions thus potentially allowing faster contour editing and evaluation.

Trachomatous trichiasis (TT) is an ocular condition in which the eyelid turns inward causing the eyelashes to scratch the eye, leading to blindness and affecting millions of people worldwide. The standard treatment for TT is surgery, where an incision is made to rotate the eyelid margin outward, repositioning the eyelashes to their original position. However, outcomes after surgery are often suboptimal, with a high risk of post-operative trichiasis (PTT). Studies have shown that the appearance of the immediate post-operative eyelid strongly correlates with the success of the procedure after six weeks, emphasizing the importance of early identification and correction of poor surgical results. We propose a mobile application that detects post-operative eyelids at higher risk of poor outcomes, enabling surgeons in the field to have immediate feedback to perform the necessary corrections and improve patient outcomes. The algorithm is based on the well-established Faster R-CNN model, which detects and classifies parts of the eyelid into three categories: under-correction, overcorrection, or appropriate correction. The model achieved 75.7% recall for under-correction and 75.6% recall for overcorrection, demonstrating strong sensitivity in identifying potential for adverse outcomes. The UI/UX of the application was designed with an intuitive interface that allows users to take a picture of an eyelid and evaluate the surgical result using the model. The algorithm runs in under 12 seconds and has been tested by TT surgeons in the field. This work has the potential to significantly improve post-operative trichiasis outcomes, reducing PTT rates, and improving life quality in resource-limited settings.

Deep learning has demonstrated an excellent capacity to capture prior information about image classes. This has driven advances in image formation to break through traditional limits of data fidelity in denoising, reconstruction, and processing of undersampled data. Diffusion posterior sampling (DPS) is one approach that combines a generative prior with an analytic model for the measurements to both enforce consistency with measurements and also integrate sophisticated prior information. In such approaches, performance gains can be limited by the quality of the deep learning prior, which, in turn, is limited by the data used to train that network model. In many cases, like tomographic reconstruction, training data is assembled from standard clinical protocols. Specialized, high-fidelity datasets (e.g. very high spatial resolution scans) are often limited in number and/or only target regional anatomy. We propose a novel DPS framework for image reconstruction that uses mixed prior models to enhance regional spatial resolution while maintaining global information for stability and consistency. Specifically, the method integrates a global diffusion model, trained on (untruncated) normal-resolution data, with a regional patch-based diffusion model, trained on high-resolution patches. The prior models are combined using frequency-domain methods, where low-frequency components are extracted from the global model and high-frequency components come from the patch-based model. To address boundary discontinuities inherent to patch-based diffusion model, we adopt a shifted patch division mechanism, which dynamically moves patch boundaries across sampling steps. This strategy removes the stitching artifacts by dispersing them as stochastic noise, while the diffusion prior and posterior constraints gradually eliminate residual inconsistencies. Furthermore, a resampling step is applied after each likelihood update, ensuring stability and preventing error accumulation across iterations. Finally, we introduce a regional sampling scheme, where a binary mask ensures the regional prior is applied within the appropriate anatomy, while the global prior is applied in the background. Experimental results demonstrate that the proposed framework achieves superior reconstruction quality by preserving fine-grained details in regions of interest without sacrificing global structural coherence. This work highlights the potential of combining multi-scale diffusion priors for high-fidelity and efficient posterior sampling in inverse imaging problems.

Curating, processing, and combining large-scale medical imaging datasets from national studies is a non-trivial task due to the intense computation and data throughput required, variability of acquired data, and associated financial overhead. Existing platforms or tools for large-scale data curation, processing, and storage have difficulty achieving a viable cost-to-scale ratio of computation speed for research purposes, either being too slow or too expensive. Additionally, management and consistency of processing large data in a team-driven manner is a non-trivial task. We design a BIDS-compliant method for an efficient and robust data processing pipeline of large-scale diffusion-weighted and T1-weighted MRI data compatible with low-cost, high-efficiency computing systems. Our method accomplishes automated querying of data available for processing and process running in a consistent and reproducible manner that has long-term stability, while using heterogenous low-cost computational resources and storage systems for efficient processing and data transfer. We demonstrate how our organizational structure permits efficiency in a semi-automated data processing pipeline and show how our method is comparable in processing time to cloud-based computation while being almost 20 times more cost-effective. Our design allows for fast data throughput speeds and low latency to reduce the time for data transfer between storage servers and computation servers, achieving an average of 0.60 Gb/s compared to 0.33 Gb/s for using cloud-based processing methods. The design of our workflow engine permits quick process running while maintaining flexibility to adapt to newly acquired data.

Patients with diffuse lung disease (DLD) undergo CT scans for diagnosis and evaluation. Attempts to characterize the radiographic appearance of these regions with quantitative features like radiomics have been hindered by variability in CT acquisition and reconstruction parameters. The purpose of this investigation is to characterize the effect of CT reconstruction kernel on radiomic features in normal and fibrotic regions in DLD patients. Raw CT projection data of DLD patients receiving a thoracic CT exam was collected from 3 CT scanners (Definition AS, Flash, and Force; Siemens Healthineers, Forchheim, Germany) and retrospectively reconstructed with 5 reconstruction kernels (smooth, medium-smooth, medium, medium-sharp, and sharp). The medium kernel is part of the clinical protocol at our institution and considered the reference kernel for this investigation. Regions of classic normal and classic fibrosis were annotated by 2 thoracic radiologists on images reconstructed with reference kernels. The annotations were copied across each reconstruction, and 72 radiomic features were extracted from each annotation using Pyradiomics. Agreement between features in reference and non-reference kernels was assessed with concordance correlation coefficient (CCC). Features were considered robust if average CCC was > 0.9. There were 66 patients included with 116 regions of normal and 208 regions of fibrosis annotated across all patients. Across kernels in normal tissue, 0/16 GLSZM, 1/16 GLRLM, 0/22 GLCM, and 5/18 first-order features were robust (average CCC > 0.9). Across kernels in fibrotic tissue, 1/16 GLSZM, 3/16 GLRLM, 3/22 GLCM, and 7/18 first-order features were robust, however the effect magnitude of sharper kernels was greater than in normal tissue. First-order features were more robust than other features; GLCM features were the least robust. There are more robust features across kernels in fibrotic tissue, but the magnitude of kernel effect is greater in fibrotic tissue than in normal tissue.

Deep learning image generation has been an active area of research in a number of applications. However, traditional generative models are not able to control specific properties of the image outputs. In this work, we propose a deep learning model that produces images according to user-specified texture feature values. We adopted a diffusion transformer architecture and used texture features to condition the reverse process. The model was trained on lung patches from a public lung CT database. Two texture features, autocorrelation and inverse difference derived from the Gray-Level Co-Occurrence Matrix were used as conditional inputs. We evaluated the ability of the model to produce samples with similar feature values as the conditional inputs. Both in-distribution and out-of-distribution conditions were evaluated. Results indicate that the model is able to generate image patches resembling lung parenchyma. The autocorrelation and inverse difference of generated images have good agreement with and exhibit low variability around the conditional inputs. The concordance correlation coefficient between real and generated samples is 0.9962 for autocorrelation and 0.9402 for inverse difference. Visual assessment of image samples reveals that real and generated images share similar features, consistent with their radiomic properties. Findings from this work indicate that the diffusion transformer model is able to generate images with texture features closely aligning with the conditional inputs, supporting its utility for highly controlled data generation for a variety of applications.

Dose prediction plays a key role in knowledge-based planning (KBP) by automatically generating patient-specific dose distribution. Recent advances in deep learning-based dose prediction methods necessitates collaboration among data contributors for improved performance. Federated learning (FL) has emerged as a solution, enabling medical centers to jointly train deep-learning models without compromising patient data privacy. We developed the FedKBP framework to evaluate the performances of centralized, federated, and individual (i.e. separated) training of dose prediction model on the 340 plans from OpenKBP dataset. To simulate FL and individual training, we divided the data into 8 training sites. To evaluate the effect of inter-site data variation on model training, we implemented two types of case distributions: 1) Independent and identically distributed (IID), where the training and validating cases were evenly divided among the 8 sites, and 2) non-IID, where some sites have more cases than others. The results show FL consistently outperforms individual training on both model optimization speed and out-of-sample testing scores, highlighting the advantage of FL over individual training. Under IID data division, FL shows comparable performance to centralized training, underscoring FL as a promising alternative to traditional pooled-data training. Under non-IID division, larger sites outperformed smaller sites by up to 19% on testing scores, confirming the need of collaboration among data owners to achieve better prediction accuracy. Meanwhile, non-IID FL showed reduced performance as compared to IID FL, posing the need for more sophisticated FL method beyond mere model averaging to handle data variation among participating sites.

Fluorescence Lifetime Imaging Microscopy (FLIM) is a popular imaging technique that provides users another dimension for investigating biomolecular states, interactions, and environments. It can, for instance, be used to monitor metabolism or discern different F&#xf6;rster resonance energy transfer (FRET) proximity relationships within cells. FLIM often reveals multiple lifetimes which are usually analyzed via multi-exponential fitting in the time domain and phasor localization in the frequency domain. Both methods yield maps of amplitudes (i.e., alphas/intensities/concentrations) associated with their lifetimes. The resulting "pixel histograms" or "phasor clouds" give the user pixel-weighted visualizations of the distribution of amplitudes vs. lifetime. Using this approach, however, pixel areas with low photon counts can improperly skew FLIM analysis. In contrast, a global decay-associated analysis of image data first optimizes the nonlinear parameters that are common across the image-namely, the lifetimes-along with their amplitudes, and then provides statistically rigorous confidence limits on both lifetimes and amplitudes. With these limits, one may assess cellular metabolism and FRET relationships with confidence. Additionally, global analysis is known to be capable of unmixing multiple components that are inseparable in conventional or graphical analyses. We present global Decay-Associated Image Software (gDAIS) for analyzing FLIM data which incorporates a fast regression method and yields images of the amplitudes and images of all associated uncertainties. These outputs can then be used to generate error limits for amplitude ratios (e.g., free:bound NADH).

Radiomics relies on quantitative features to discern the underlying biological signatures. However, feature dependence on the imaging systems themselves hampers the creation of reproducible and generalizable models. We have previously proposed a novel framework to remove the effects of system blur and image noise on radiomic calculations and performed validation in simulation studies. In this work, we extended the analysis and evaluated the method on CT data acquired of an anthropomorphic phantom with realistic lung textures. Data was acquired at five different dose levels and reconstructed using eight different reconstruction kernels. To test the generalizability of the method, we applied our proposed method to standardize from all possible starting and reference kernel pairs under all measured dose levels for a total of 320 cases (8&#xd7;8&#xd7;5). Standardization was performed for radiomics features from four classes, histogram, GLCM, gray-level run length matrix-(GLRLM), and wavelet transforms. Results indicate that standardized radiomics features are closer to the reference and on average, the average absolute percentage difference from reference over all features is improved by a factor of three compared with unstandardized features. In addition, we found that standardization from a smoother to a sharper kernel is a more challenging task and that performance is comparable across all dose levels. This work shows that the proposed standardization method is effective in standardizing radiomics feature values across a wide range of imaging conditions in clinical CT.