Search

Introduction Tamoxifen, a selective estrogen antagonist, is part of standard adjuvant therapy for hormone receptor-positive breast cancer. In addition to known systemic toxicities, well-described retinotoxic effects include peripheral and macular crystalline deposits, pseudocystic foveal cavitations, as well as an increased risk of developing punctate pigmentary changes. However, the true prevalence of retinal toxicity is unclear, with the literature reporting rates between 1.5% and 11%, with significant geographic and temporal discrepancies depending on choice of screening modality. Advances in retinal imaging, specifically spectral domain and swept-source optical coherence tomography (OCT), have led to more sensitive diagnosis of toxicity. However, the field of view of commercially available OCT is limited 30 degrees, whereas ultra-widefield (UWF) imaging can image up to 200 degrees of the retina, albeit with a superficial rather than a cross-sectional view. Therefore, UWF imaging may visualize previously underreported signs of toxicity in the peripheral retina such as pigmentary changes better than macula-focused imaging alone like OCT. As such, the goals of the present study were to: 1) determine the prevalence of tamoxifen retinopathy in our cohort, a major northeastern metropolitan area, based on multimodal retinal imaging including macular OCT, fundus photos, and autofluorescence; 2) determine whether the additional field of view captured in UWF imaging contributes to the diagnosis of tamoxifen retinopathy by assessing for peripheral findings, most specifically pigmentary retinopathy. Methods This is a retrospective study of patients who visited the retina service of a single academic tertiary referral center between September 2012 and November 2021. Female patients initiated on tamoxifen for at least 6 months before their first retinal exam were identified. Patients with poor image quality, prior vitreoretinal surgery, retinal laser, or confounding pathology such as severe diabetic retinopathy were excluded. Two independent graders performed blinded review of OCT images for evidence of macular toxicity and UWF images for signs of central and peripheral toxicity, and the prevalence of the latter was compared to age- and gender-matched controls. A two-tailed t-test was used to compare ages of patients in the two cohorts. A one-tailed two proportion Z-test was used to determine if eyes in the tamoxifen cohort exhibited a greater rate of peripheral pigmentary changes. Results 252 eyes from 128 patients were included in the tamoxifen cohort, and 244 eyes from 125 patients in the control cohort. The average age at first retinal imaging in the tamoxifen cohort (61.1 ± 1.1, n=128) versus the control cohort (61.0 ± 1.3, n = 125) was not significantly different (p=0.95). 4 eyes (1.6%) from 2 patients (1.6%) had evidence of definitive tamoxifen retinopathy. One patient had crystalline maculopathy bilaterally visible on OCT and UWF imaging and another had macular pseudocystic cavitations bilaterally on OCT. Neither patient had peripheral findings on UWF. 31 eyes (12.4%) among 16 (12.5%) patients in the tamoxifen cohort displayed peripheral reticular pigmentary changes, compared to 59 eyes (24.2%) among 30 patients (24.0%) in the control group. The rate of peripheral pigmentary change in the tamoxifen cohort was not significantly greater than in the control group (p=0.99). Conclusion Only 2 patients had definitive tamoxifen retinopathy findings, which were both concentrated in the macula and clearly visible on OCT imaging. Given the absence of meaningful UWF changes, macular OCT may be the most valuable tool in diagnosing tamoxifen retinopathy. Our prevalence (1.6%) diverges from higher prevalences reported by studies screening with advanced retinal imaging (10-12%), though notably reflects a very different ethnic and geographic population compared with other work. Further large population studies are needed. Citation Format: Ethan Zhao, Kyle Kovacs. Prevalence and Spectrum of Manifestations of Tamoxifen Retinopathy as Assessed by Multimodal and Ultra-widefield Retinal Imaging [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P1-12-06.

Background Olfactory Sensory Neuron (OSN) axons project from the zebrafish olfactory epithelium to reproducible intermediate target locations in the olfactory bulb called protoglomeruli at early stages in development. Two classes of OSNs expressing either OMP or TRPC2 exclusively target distinct, complementary protoglomeruli. Using RNAseq, we identified axon guidance receptors nrp2a and nrp2b, and their ligand sema3fa, as potential guidance factors that are differentially expressed between these two classes of OSNs. Methods To investigate their role in OSN axon guidance, we assessed the protoglomerular targeting fidelity of OSNs labeled by OMP:RFP and TRPC2:Venus transgenes in nrp2a, nrp2b, or sema3fa mutants. We used double mutant and genetic interaction experiments to interrogate the relationship between the three genes. We used live time-lapse imaging to compare the dynamic behaviors of OSN growth cones during protoglomerular targeting in heterozygous and mutant larvae. Results The fidelity of protoglomerular targeting of TRPC2-class OSNs is degraded in nrp2a, nrp2b, or sema3fa mutants, as axons misproject into OMP-specific protoglomeruli and other ectopic locations in the bulb. These misprojections are further enhanced in nrp2a;nrp2b double mutants suggesting that nrp2s work at least partially in parallel in the same guidance process. Results from genetic interaction experiments are consistent with sema3fa acting in the same biological pathway as both nrp2a and nrp2b. Live time-lapse imaging was used to examine the dynamic behavior of TRPC2-class growth cones in nrp2a mutants compared to heterozygous siblings. Some TRPC2-class growth cones ectopically enter the dorsal-medial region of the bulb in both groups, but in fully mutant embryos, they are less likely to correct the error through retraction. The same result was observed when TRPC2-class growth cone behavior was compared between sema3fa heterozygous and sema3fa mutant larvae. Conclusions Our results suggest that nrp2a and nrp2b expressed in TRPC2-class OSNs help prevent their mixing with axon projections in OMP-specific protoglomeruli, and further, that sema3fa helps to exclude TRPC2-class axons by repulsion from the dorsal-medial bulb.

Singer recognition aims to automatically recognize the singer of a given recording. Compared to spoken voices, singing voice is characterized by a much higher degree of vocal style. The task becomes more challenging when it operates on numerous singers. This paper explores different strategies in a deep metric learning framework, with special focus on their performance in a large-scale dataset consisting of audio samples from 5057 singers. We conduct thorough experiments to compare loss functions, including triplet loss, generalized end-to-end (GE2E) loss, and prototypical network (PN) loss. Effects of vocal source separation is also investigated. Using audio inputs with separated vocals, our model trained with PN loss outperforms other evaluated methods in the identification task. While in the verification task with one-on-one comparison of two single embeddings, triplet loss achieves the best results. However, verification using PN loss shows superior performance to methods with triplet loss when using the centroid of 5 embed dings to represent the singer embedding. Using longer segments for a singer representation consistently improves the performance for all evaluated tasks.

Audio super-resolution is a challenging task of recovering the missing high-resolution features from a low-resolution signal. To address this, generative adversarial networks (GAN) have been used to achieve promising results by training the mappings between magnitudes of the low and high-frequency components. However, phase information is not well-considered for waveform reconstruction in conventional methods. In this paper, we tackle the problem of music super-resolution and conduct a thorough investigation on the importance of phase for this task. We use GAN to predict the magnitudes of the high-frequency components. The corresponding phase information can be extracted using either a GAN-based waveform synthesis system or a modified Griffin-Lim algorithm. Experimental results show that phase information plays an important role in the improvement of the reconstructed music quality. Moreover, our proposed method significantly outperforms other state-of-the-art methods in terms of objective evaluations., This paper has been accepted to Interspeech 2020 (http://www.interspeech2020.org/)

We report monolithic aerogel foams as solid materials with hierarchical porosity created by a foam-like structure embedded in the skeletal framework of a regular aerogel. The foam-like structure is prepared without chemical foaming agents or templates, resulting in a less expensive, more efficient, and more readily adaptable process. Specifically, pressurized air (7 bar) is injected into a suitable sol, which is allowed to gel under pressure, followed by slow depressurization. Voids are created from the air bubbles formed during depressurization. The model material used for validation of the technique is based on poly(isocyanurate-urethane) aerogels (PIR-PUR) and selected material properties of the resulted aerogel foams are compared with those of their pristine aerogel counterparts. With an eye on scalability, all wet-gels were dried under ambient conditions. Aerogel foams exhibit lower bulk densities by about 25%, and higher porosities by about 10% in comparison with their pristine PIR-PUR aerogel counterparts. Interestingly, the thermal conductivities of aerogel foams were found reduced significantly (by 25%) from 0.104 to 0.077 Wm−1K−1 compared to the corresponding pristine aerogels. In addition, aerogel foams absorb 36% w/w more oil and show better oil retention in comparison with regular PIR-PUR aerogel samples made from the same sols. As this technique does not alter the chemical composition of the aerogel, it is anticipated that it can be used for a variety of different types of aerogels and formulations in order to lower their bulk density and improve desired physical properties such as thermal conductivity.

Integrins modulate chemically-induced nociception in a variety of inflammatory and neuropathic pain models. Yet, the role of integrins in mechanically-induced pain remains undefined, despite its well-known involvement in cell adhesion and mechanotransduction. Excessive spinal facet capsular ligament stretch is a common injury that induces morphological and functional changes in its innervating afferent neurons and can lead to pain. However, the local mechanisms underlying the translation from tissue deformation to pain signaling are unclear, impeding effective treatment. Therefore, the involvement of the integrin subunit β1 in pain signaling from facet injury was investigated in complementary in vivo and in vitro studies. An anatomical study in the rat identified expression of the integrin subunit β1 in dorsal root ganglion (DRG) neurons innervating the facet, with greater expression in peptidergic than non-peptidergic DRG neurons. Painful facet capsule stretch in the rat upregulated the integrin subunit β1 in small- and medium-diameter DRG neurons at day 7. Inhibiting the α2β1 integrin in a DRG-collagen culture prior to its stretch injury prevented strain-induced increases in axonal substance P (SP) in a dose-dependent manner. Together, these findings suggest that integrin subunit β1-dependent pathways may contribute to SP-mediated pain from mechanical injury of the facet capsular ligament.

Serotonergic (5-HT) and noradrenergic (NA) input to spinal motoneurons is essential for generating plateau potentials and self-sustained discharges. Extensor motoneurons are densely innervated by 5-HT and NA synapses and have robust plateau potentials and self-sustained discharges. Conversely, plateau potentials and self-sustained discharges are very rare in flexor motoneurons. The most likely reasons for this difference are that flexor motoneurons have few 5-HT and NA synapses and/or they are distributed distant to the channels responsible for plateau potentials and self-sustained discharges. However, the distribution of 5-HT and NA synapses on flexor motoneurons is unknown. Here we describe the distribution and density of 5-HT and NA synapses on motoneurons that innervate the flexor neck muscle, rectus capitis anterior (RCA), in the adult cat. Using a combination of intracellular staining, fluorescent immunohistochemistry, and 3D reconstruction techniques, we found that 5-HT and NA synapses are widely distributed throughout the dendritic trees of RCA motoneurons, albeit with a strong bias to small-diameter dendrites and to medial dendrites in the case of NA contacts. The number of 5-HT and NA contacts per motoneuron ranged, respectively, from 381 to 1,430 and from 642 to 1,382, which is 2.3- and 1.4-fold less than neck extensor motoneurons (Montague et al., J Comp Neurol 2013;521:638-656). These results suggest that 5-HT and NA synapses on flexor motoneurons may provide a powerful means of amplifying synaptic currents without incurring plateau potentials or self-sustained discharges. This feature is well suited to meet the biomechanical demands imposed on flexor muscles during different motor tasks.