Conectar
To describe the structured process of threshold optimisation for a commercially available multiclass chest X-ray (CXR) deep learning model, to evaluate its diagnostic performance across different operating thresholds, and to estimate its potential operational impact within an artificial intelligence (AI)-enabled triage workflow in a primary care setting.
Retrospective diagnostic performance evaluation with threshold-based analysis.
Primary care radiography services in Singapore, using data derived from two primary care clinics and a tertiary hospital.
A total of 816 adult frontal chest radiographs were included (multiethnic Asian, 464 males, 352 females; mean age 60.8 years). Images were selected to represent the spectrum of findings often encountered in primary care. Exclusion criteria included paediatric studies, lateral or oblique radiographs, and findings not supported by the AI model (eg, bony abnormalities and medical devices).
Primary outcome measures were sensitivity, specificity, and negative and positive predictive value (NPV and PPV). Secondary outcomes included estimated potential operational improvement, which is calculated by dividing the number of true negatives by the total number of CXRs.
At the default threshold of 0.15, the AI model achieved a sensitivity of 87.3% (95% CI 83.9% to 90.4%) and an NPV of 87.0% (95% CI 83.6% to 90.2%). Lowering the threshold to 0.10 increased sensitivity to 93.2% (95% CI 90.7% to 95.5%) and NPV to 91.3% (95% CI 88.2% to 94.3%), with specificity of 71.7% (95% CI 67.3% to 76.1%). These trade-offs were considered acceptable for a safety-focused co-triage workflow prioritising minimisation of false negatives.
Threshold optimisation is critical for adapting AI models to context-specific clinical workflows. Our study shows that adjusting the operating threshold enabled prioritisation of sensitivity and NPV, supporting safe AI-assisted triage in primary care. This is a deeply collaborative process that must involve radiology and clinical teams: selecting appropriate thresholds aligned with clinical objectives for safe and effective implementation. Future work will assess real-world operational impact and user acceptance following prospective deployment.
by Sunghoon Jeon, Keunho Kim, Cheolwon Choe, Juil Choi, Gun Lee, Chung-Do Lee, Hyeon-Jeong Moon, Jun-Gyu Park, Jin-kyung Kim, Namsoon Lee, Dongwoo Chang
Quick-soluble gelatin microparticles (QS-GMP) are emerging embolic agents under investigation for temporary vascular occlusion, offering reduced ischemic risk compared to permanent materials. The aim of this preclinical study was to evaluate the safety and efficacy of QS-GMP for transarterial embolization in a rabbit model of urinary bladder embolization. Twelve male New Zealand White rabbits underwent bilateral umbilical artery embolization using QS-GMP. Animals were assigned to four time-points (immediately, 3, 7, and 14 days post-embolization), with comprehensive assessments including clinical observations, hematologic and serum biochemical analysis, angiography, and histopathology. The procedure was technically feasible in all animals without intraoperative complications. Temporary hematuria and a transient decrease in body weight were observed post-procedure, both of which resolved spontaneously. Complete occlusion of the cranial vesical artery and absence of bladder wall perfusion were achieved immediately after embolization, followed by full recanalization at 3 days. Angiographic imaging at 7 and 14 days revealed transient hypervascularization of the bladder wall. Histopathological analysis showed marked edema, epithelial necrosis, and inflammatory infiltration at 3 and 7 days, with full urothelial regeneration observed at 14 days. No signs of ureteral or renal injury, or adverse systemic responses were detected. These findings suggest that QS-GMP may serve as a feasible option for temporary arterial occlusion in future veterinary lower urinary tract applications, although further long-term evaluation is warranted.Diabetic foot disease (DFD) requires proactive healthcare management to minimise the risk of complications. Healthcare delivery has been shown to present significant challenges in rural and regional settings with Central Queensland being a large geographic area with limited local healthcare resources. Our objectives are to describe the distribution of DFD in a regional setting and assess the impact of remoteness on healthcare accessibility for this cohort.
A retrospective analysis of hospital admissions for DFD data between January 2017 and December 2023.
All recorded episodes of care for DFD patients provided by Central Queensland Hospital and Health Service (CQHHS) were included.
Primary outcome was the number of DFD episodes. Secondary outcomes included the number of hospital bed days used by DFD cases, the number of lower extremity amputations and the number of interhospital transfers (IHTs) for DFD cases.
1597 DFD episodes, 15 528 bed days, 340 lower extremity amputations and 452 IHTs were recorded. Population-adjusted outcomes showed 1054 DFD episodes per 100 000 males compared with 383 per 100 000 females (incidence rate ratio (IRR) 2.75, 95% CI: 2.46 to 3.07), and 1384 per 100 000 Indigenous Australians compared with 669 per 100 000 non-Indigenous Australians (IRR 2.07, 95% CI: 1.80 to 2.38). Remoteness was associated with lower DFD treatment completion rates at local hospitals, with only 9.3% of patients from the most remote regions completing treatment locally, compared with 76.3% from the least remote regions (p
Males and Indigenous Australians were disproportionately affected by DFD in Central Queensland. Remoteness poses a significant risk factor to completing treatment for DFD at a local facility. Further research into key drivers leading to the disproportionate outcomes demonstrated in this study between specific cohorts of people with DFD is suggested to help design future interventions to improve accessibility and outcomes.
FreshRSS 