Latent diffusion augmentation enhances deep learning analysis of neuro-morphology in limbal stem cell deficiency

Introduction

Limbal Stem Cell Deficiency (LSCD) is a blinding corneal disease characterized by the loss of function or deficiency in adult stem cells located at the junction between the cornea and the sclera (i.e., the limbus), namely the limbal stem cells (LSCs). Recent advances in in vivo imaging technology have improved disease diagnosis and staging to quantify several biomarkers of in vivo LSC function including epithelial thickness measured by anterior segment optical coherence tomography, and basal epithelial cell density and subbasal nerve plexus by in vivo confocal microscopy. A decrease in central corneal sub-basal nerve density and nerve fiber and branching number has been shown to correlate with the severity of the disease in parallel with increased nerve tortuosity. Yet, image acquisition and manual quantification require a high level of expertise and are time-consuming. Manual quantification presents inevitable interobserver variability.

Methods

The current study employs a novel deep learning approach to classify neuron morphology in various LSCD stages and healthy controls, by integrating images created through latent diffusion augmentation. The proposed model, a residual U-Net, is based in part on the InceptionResNetV2 transfer learning model.

Results

Deep learning was able to determine fiber number, branching, and fiber length with high accuracy (R2 of 0.63, 0.63, and 0.80, respectively). The model trained on images generated through latent diffusion on average outperformed the same model when trained on solely original images. The model was also able to detect LSCD with an AUC of 0.867, which showed slightly higher performance compared to classification using manually assessed metrics.

Discussion

The results suggest that utilizing latent diffusion to supplement training data may be effective in bolstering model performance. The results of the model emphasize the ability as well as the shortcomings of this novel deep learning approach to predict various nerve morphology metrics as well as LSCD disease severity.

Late Urinary Toxicity and QoL with Curative Radiotherapy for High-Risk Prostate Cancer: Dose-Effect Relations in the POP-RT Randomized Phase III Trial

PURPOSE/OBJECTIVE(S)

Whole pelvic radiotherapy (WPRT) showed better biochemical failure-free survival and metastasis-free survival than prostate-only radiotherapy (PORT) in the phase III randomized POP-RT trial for high and very high-risk prostate cancer, albeit with higher RTOG grade 2 late urinary toxicity. We report updated long term, symptom-wise comparison and dose-effect relations from this trial.

MATERIALS/METHODS

Late urinary toxicity, and cumulative severity of each symptom over the follow-up period was graded using CTCAE v5.0. Grade 2+ toxicities were compared between the trial arms by chi square test. Bladder dosimetry in 5-Gy increments (V5, V10, V15...V65 Gy, V68 Gy) from the trial database of approved radiotherapy plans, was compared for each urinary symptom and overall late gr2+ toxicity by student t-test. Observed differences in dosimetric parameters were tested using multivariable logistic regression analysis, including age at diagnosis, known diabetes, tumor stage, trial arm, and prior transurethral resection of prostate (TURP). Urinary QOL scores were compared between arms using generalised linear mixed model.

RESULTS

Combined late symptom-wise toxicity and dose-volume data were available for analysis for 193/224 patients. At a median follow-up of 75 months, cumulative CTCAE gr2+ late urinary toxicity remained higher with WPRT than PORT, though not statistically significant (36.5% vs 26.8%, p = 0.15). Grade 3 toxicity was low and similar in both arms. Symptom-wise cumulative rates showed no significant difference between arms (Table 1). Dosimetric comparison showed significantly higher bladder V5-V15 in patients with gr2+ toxicity over those with

CONCLUSION

Compared to prostate-only radiotherapy, whole pelvic radiotherapy resulted similar Grade 3 urinary toxicity of about 5% with about 10% higher cumulative grade 2+ urinary toxicity over long term follow up. This difference was not reflected in patient-reported QOL. WPRT particularly increased urgency and hematuria. Larger bladder volume being irradiated with 5Gy to 15Gy dose range could contribute to increase in urinary symptoms.

Collapse

Key Words Collapse

MESH Headings Collapse

Grants Collapse

Manipulating Relative Permittivity for High-Performance Wearable Triboelectric Nanogenerators

As the world marches into the era of the Internet of Things (IoT), the practice of human health care is on the cusp of a revolution, driven by an unprecedented level of personalization enabled by a variety of wearable bioelectronics. A sustainable and wearable energy solution is highly desired , but challenges still remain in its development. Here, we report a high-performance wearable electricity generation approach by manipulating the relative permittivity of a triboelectric nanogenerator (TENG). A compatible active carbon (AC)-doped polyvinylidene fluoride (AC@PVDF) composite film was invented with high relative permittivity and a specific surface area for wearable biomechanical energy harvesting. Compared with the pure PVDF, the 0.8% AC@PVDF film-based TENG obtained an enhancement in voltage, current, and power by 2.5, 3.5, and 9.8 times, respectively. This work reports a stable, cost-effective, and scalable approach to improve the performance of the triboelectric nanogenerator for wearable biomechanical energy harvesting, thus rendering a sustainable and pervasive energy solution for on-body electronics.