Clinical Scorecard: Retinal Biomarker from Deep Learning Linked to Amputation Risk in Type 2 Diabetes Patients with Diabetic Foot Complications

At a Glance

Key Highlights

• Study included 392 individuals with type 2 diabetes, 79 with DF-related amputation.
• AUC of the full model for predicting amputation was 0.791.
• Retinal biomarker showed incremental association with DF-related amputation beyond conventional diabetes variables.
• Sensitivity of 87.5% and specificity of 90.4% at specified cutoffs.
• Five-year mortality after DF-related amputation exceeds 50%.

Guideline-Based Recommendations

Diagnosis

• Utilize deep learning-derived retinal biomarkers to assess amputation risk in diabetic foot patients.

Management

• Implement timely limb-salvage strategies for high-risk individuals identified through retinal imaging.

Monitoring & Follow-up

• Regularly monitor retinal health and related biomarkers in patients with type 2 diabetes.

Risks

• Patients with diabetic foot complications have a high risk of amputation and subsequent mortality.

Patient & Prescribing Data

Individuals with type 2 diabetes and diabetic foot complications.

Early identification of high-risk patients can lead to improved management and prevention of amputations.

Clinical Best Practices

• Incorporate retinal imaging in routine assessments for patients with diabetic foot complications.
• Educate patients on the importance of glycemic control and foot care to prevent complications.

Related Resources & Content

• Rim et al. study on deep learning model for coronary artery calcium