Bridging the Diagnostic Gap: Reviewing Current Endometriosis Screening Tools and Models

Area of Science:

• Reproductive Health
• Medical Diagnostics
• Gynecology

Background:

• Endometriosis impacts 10% of women globally, causing infertility and delayed diagnosis.
• The lack of noninvasive diagnostic tools contributes to disease progression and poor outcomes.
• Current diagnostic delays average up to 10 years, highlighting an urgent need for improved methods.

Purpose of the Study:

• To synthesize literature on existing noninvasive endometriosis diagnostic tools.
• To evaluate the use, applicability, and identify gaps in current diagnostic approaches.
• To inform the development of more effective and accessible diagnostic solutions.

Main Methods:

• A comprehensive literature search was performed following PRISMA-Scoping guidelines.
• Eighteen studies were included, categorizing screening tools into four types.
• Psychometric parameters and accuracy metrics were analyzed for identified tools.

Main Results:

• Screening tools included questionnaire-based, app-based, machine learning (ML) and artificial intelligence (AI)-driven, and subtype-focused models.
• 14 studies reported psychometric data, with Area Under the Curve (AUC) values from 0.77 to 0.95.
• ML and AI-driven models demonstrated the highest accuracy metrics in diagnosing endometriosis.

Conclusions:

• Significant gaps exist in the external validation of current diagnostic tools.
• Comprehensive psychometric evaluation and cultural adaptation are needed for broader applicability.
• Further research is required to refine and validate advanced ML/AI models for endometriosis diagnosis.