Imaging assessment of conjunctival goblet cells in dry eye disease.

Dry eye disease (DED) is a widespread, multifactorial, and chronic disorder of the ocular surface with disruption of tear film homeostasis as its core trait. Conjunctival goblet cells (CGCs) are specialised secretory cells found in the conjunctival epithelium that participate in tear film formation by secreting mucin. Changes in both the structure and function of CGCs are hallmarks of DED, and imaging assessment of CGCs is important for the diagnosis, classification, and severity evaluation of DED. Existing imaging methods include conjunctival biopsy, conjunctival impression cytology and in vivo confocal microscopy, which can be used to assess the morphology, distribution, and density of the CGCs. Recently, moxifloxacin-based fluorescence microscopy has emerged as a novel technique that enables efficient, non-invasive and in vivo imaging of CGCs. This article presents a comprehensive overview of both the structure and function of CGCs and their alterations in the context of DED, as well as current methods of CGCs imaging assessment. Additionally, potential directions for the visual evaluation of CGCs are discussed.

Non-invasive assessment of response to transcatheter arterial chemoembolization for hepatocellular carcinoma with the deep neural networks-based radiomics nomogram.

BACKGROUND: Transcatheter arterial chemoembolization (TACE) is a mainstay treatment for intermediate and advanced hepatocellular carcinoma (HCC), with the potential to enhance patient survival. Preoperative prediction of postoperative response to TACE in patients with HCC is crucial. PURPOSE: To develop a deep neural network (DNN)-based nomogram for the non-invasive and precise prediction of TACE response in patients with HCC. MATERIAL AND METHODS: We retrospectively collected clinical and imaging data from 110 patients with HCC who underwent TACE surgery. Radiomics features were extracted from specific imaging methods. We employed conventional machine-learning algorithms and a DNN-based model to construct predictive probabilities (RScore). Logistic regression helped identify independent clinical risk factors, which were integrated with RScore to create a nomogram. We evaluated diagnostic performance using various metrics. RESULTS: Among the radiomics models, the DNN_LASSO-based one demonstrated the highest predictive accuracy (area under the curve [AUC] = 0.847, sensitivity = 0.892, specificity = 0.791). Peritumoral enhancement and alkaline phosphatase were identified as independent risk factors. Combining RScore with these clinical factors, a DNN-based nomogram exhibited superior predictive performance (AUC = 0.871, sensitivity = 0.844, specificity = 0.873). CONCLUSION: In this study, we successfully developed a deep learning-based nomogram that can noninvasively and accurately predict TACE response in patients with HCC, offering significant potential for improving the clinical management of HCC.

Quantitative model for assessment of lower-extremity perfusion in patients with diabetes.

BACKGROUND: Although diabetic and atherosclerotic vascular diseases have different pathophysiological mechanisms, the screening methods currently used for diabetic lower-extremity vascular diseases are mainly based on the evaluation methods used for atherosclerotic vascular diseases. Thus, assessment of microvascular perfusion is of great importance in early detection of lower-extremity ischemia in diabetes. PURPOSE: This cross-sectional study aimed to develop a quantitative model for evaluating lower-extremity perfusion. METHODS: We recruited 57 participants (14 healthy participants and 43 diabetes patients, of which 16 had lower-extremity arterial disease [LEAD]). All participants underwent technetium-99 m sestamibi (99mTc-MIBI) scintigraphy and ankle-brachial index (ABI) examination. We derived two key perfusion kinetics indices named activity perfusion index (API) and basal perfusion index (BPI). This study was registered in ClinicalTrials.gov (URL: https://www. CLINICALTRIALS: gov, NCT02752100). RESULTS: The estimated limb perfusion values in our lower-extremity perfusion assessment (LEPA) model showed excellent consistency with the actual measured data. Diabetes patients showed reduced lower-extremity perfusion in comparison with the control group (BPI: 106.21 ± 11.99 vs. 141.56 ± 17.38, p < 0.05; API: 12.34 ± 3.27 vs. 14.56 ± 3.12, p < 0.05). Using our model, the reductions in lower-extremity perfusion could be detected early in approximately 96.30% of diabetes patients. Patients with LEAD showed more severe reductions in lower-extremity perfusion than diabetes patients without LEAD (BPI: 47.85 ± 20.30 vs. 106.21 ± 11.99, p < 0.05; API: 7.06 ± 1.70 vs. 12.34 ± 3.27, p < 0.05). Discriminant analysis using API and BPI could successfully screen all diabetes patients with LEAD with a sensitivity of 100% and specificity of 80.77%. CONCLUSIONS: We established a LEPA model that could successfully assess lower-extremity microvascular perfusion in diabetes patients. This model has important application value for the recognition of early-stage LEAD in patients with diabetes.