Rapid, high-resolution, non-destructive assessments of metabolic and morphological homogeneity uniquely identify high-grade cervical precancerous lesions.

Purpose: Two-photon microscopy (2PM) is an emerging clinical imaging modality with the potential to non-invasively assess tissue metabolism and morphology in high-resolution. This study aimed to assess the translational potential of 2PM for improved detection of high-grade cervical precancerous lesions. Experimental Design: 2P images attributed to reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) and oxidized flavoproteins (FP) were acquired from the full epithelial thickness of freshly excised human cervical tissue biopsies (N = 62). Fifteen biopsies harbored high-grade squamous intraepithelial lesions (HSILs), 14 biopsies harbored low-grade SILs (LSILs), and 33 biopsies were benign. Quadratic discriminant analysis (QDA) leveraged morphological and metabolic functional metrics extracted from these images to predict the presence of HSILs. We performed gene set enrichment analysis (GSEA) using datasets available on the Gene Expression Omnibus (GEO) to validate the presence of metabolic reprogramming in HSILs. Results: Integrating metabolic and morphological 2P-derived metrics from finely sampled, full-thickness epithelia achieved a high 90.8 ± 6.1% sensitivity and 72.3 ± 11.3% specificity of HSIL detection. Notably, sensitivity (91.4 ± 12.0%) and specificity (77.5 ± 12.6%) were maintained when utilizing metrics from only two images at 12- and 72-µm from the tissue surface. Upregulation of glycolysis, fatty acid metabolism, and oxidative phosphorylation in HSIL tissues validated the metabolic reprogramming captured by 2P biomarkers. Conclusion: Label-free 2P images from as few as two epithelial depths enable rapid and robust HSIL detection through the quantitative characterization of metabolic and morphological reprogramming, underscoring the potential of this tool for clinical evaluation of cervical precancers.

Restoration of metabolic functional metrics from label-free, two-photon human tissue images using multiscale deep-learning-based denoising algorithms.

Significance: Label-free, two-photon excited fluorescence (TPEF) imaging captures morphological and functional metabolic tissue changes and enables enhanced understanding of numerous diseases. However, noise and other artifacts present in these images severely complicate the extraction of biologically useful information. Aim: We aim to employ deep neural architectures in the synthesis of a multiscale denoising algorithm optimized for restoring metrics of metabolic activity from low-signal-to-noise ratio (SNR), TPEF images. Approach: TPEF images of reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) and flavoproteins (FAD) from freshly excised human cervical tissues are used to assess the impact of various denoising models, preprocessing methods, and data on metrics of image quality and the recovery of six metrics of metabolic function from the images relative to ground truth images. Results: Optimized recovery of the redox ratio and mitochondrial organization is achieved using a novel algorithm based on deep denoising in the wavelet transform domain. This algorithm also leads to significant improvements in peak-SNR (PSNR) and structural similarity index measure (SSIM) for all images. Interestingly, other models yield even higher PSNR and SSIM improvements, but they are not optimal for recovery of metabolic function metrics. Conclusions: Denoising algorithms can recover diagnostically useful information from low SNR label-free TPEF images and will be useful for the clinical translation of such imaging.

Formalin fixation and paraffin embedding interfere with the preservation of optical metabolic assessments based on endogenous NAD(P)H and FAD two-photon excited fluorescence.

Endogenous NAD(P)H and FAD two-photon excited fluorescence (TPEF) images provide functional metabolic information with high spatial resolution for a wide range of living specimens. Preservation of metabolic function optical metrics upon fixation would facilitate studies which assess the impact of metabolic changes in the context of numerous diseases. However, robust assessments of the impact of formalin fixation, paraffin embedding, and sectioning on the preservation of optical metabolic readouts are lacking. Here, we evaluate intensity and lifetime images at excitation/emission settings optimized for NAD(P)H and FAD TPEF detection from freshly excised murine oral epithelia and corresponding bulk and sectioned fixed tissues. We find that fixation impacts the overall intensity as well as the intensity fluctuations of the images acquired. Accordingly, the depth-dependent variations of the optical redox ratio (defined as FAD/(NAD(P)H + FAD)) across squamous epithelia are not preserved following fixation. This is consistent with significant changes in the 755 nm excited spectra, which reveal broadening upon fixation and additional distortions upon paraffin embedding and sectioning. Analysis of fluorescence lifetime images acquired for excitation/emission settings optimized for NAD(P)H TPEF detection indicate that fixation alters the long lifetime of the observed fluorescence and the long lifetime intensity fraction. These parameters as well as the short TPEF lifetime are significantly modified upon embedding and sectioning. Thus, our studies highlight that the autofluorescence products formed during formalin fixation, paraffin embedding and sectioning overlap highly with NAD(P)H and FAD emission and limit the potential to utilize such tissues to assess metabolic activity.

Non-invasive Imaging Techniques for Monitoring Cellular Response to Treatment in Stable Vitiligo.

Punch grafting procedures, where small pieces of normal skin are transplanted into stable vitiligo patches, results in repigmentation in only half of patients treated, yet the factors that determine whether a patient responds to treatment or not are still unknown. Reflectance confocal microscopy (RCM) is adept at visualizing melanocyte migration and epidermal changes over large areas while multiphoton microscopy (MPM) can capture metabolic changes in keratinocytes. With the overall goal of identifying optical biomarkers for early treatment response, we followed 12 vitiligo lesions undergoing punch grafting. Dendritic melanocytes adjacent to the graft site were observed before clinical evidence of repigmentation in treatment responsive patients but not in treatment non-responsive patients, suggesting that the early visualization of melanocytes is indicative of a therapeutic response. Keratinocyte metabolic changes in vitiligo skin adjacent to the graft site also correlated with treatment response, indicating that a keratinocyte microenvironment that more closely resembles normal skin is more hospitable for migrating melanocytes. Taken together, these studies suggest that successful melanocyte transplantation requires both the introduction of new melanocytes and modulation of the local tissue microenvironment.

Formalin fixation and paraffin embedding interfere with preservation of optical metabolic assessments based on endogenous NAD(P)H and FAD two photon excited fluorescence.

Endogenous NAD(P)H and FAD two-photon excited fluorescence (TPEF) images provide functional metabolic information with high spatial resolution for a wide range of living specimens. Preservation of metabolic function optical metrics upon fixation would facilitate studies which assess the impact of metabolic changes in the context of numerous diseases. However, robust assessments of the impact of formalin fixation, paraffin embedding, and sectioning on the preservation of optical metabolic readouts are lacking. Here, we evaluate intensity and lifetime images at excitation/emission settings optimized for NAD(P)H and FAD TPEF detection from freshly excised murine oral epithelia and corresponding bulk and sectioned fixed tissues. We find that fixation impacts the overall intensity as well as the intensity fluctuations of the images acquired. Accordingly, the depth-dependent variations of the optical redox ratio (defined as FAD/(NAD(P)H + FAD)) across squamous epithelia are not preserved following fixation. This is consistent with significant changes in the 755 nm excited spectra, which reveal broadening upon fixation and additional distortions upon paraffin embedding and sectioning. Analysis of fluorescence lifetime images acquired for excitation/emission settings optimized for NAD(P)H TPEF detection indicate that fixation alters the long lifetime of the observed fluorescence and the long lifetime intensity fraction. These parameters as well as the short TPEF lifetime are significantly modified upon embedding and sectioning. Thus, our studies highlight that the autofluorescence products formed during formalin fixation, paraffin embedding and sectioning overlap highly with NAD(P)H and FAD emission and limit the potential to utilize such tissues to assess metabolic activity.

Restoration of metabolic functional metrics from label-free, two-photon cervical tissue images using multiscale deep-learning-based denoising algorithms.

Label-free, two-photon imaging captures morphological and functional metabolic tissue changes and enables enhanced understanding of numerous diseases. However, this modality suffers from low signal arising from limitations imposed by the maximum permissible dose of illumination and the need for rapid image acquisition to avoid motion artifacts. Recently, deep learning methods have been developed to facilitate the extraction of quantitative information from such images. Here, we employ deep neural architectures in the synthesis of a multiscale denoising algorithm optimized for restoring metrics of metabolic activity from low-SNR, two-photon images. Two-photon excited fluorescence (TPEF) images of reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) and flavoproteins (FAD) from freshly excised human cervical tissues are used. We assess the impact of the specific denoising model, loss function, data transformation, and training dataset on established metrics of image restoration when comparing denoised single frame images with corresponding six frame averages, considered as the ground truth. We further assess the restoration accuracy of six metrics of metabolic function from the denoised images relative to ground truth images. Using a novel algorithm based on deep denoising in the wavelet transform domain, we demonstrate optimal recovery of metabolic function metrics. Our results highlight the promise of denoising algorithms to recover diagnostically useful information from low SNR label-free two-photon images and their potential importance in the clinical translation of such imaging.