Difficult Spinal-Arachnoid Puncture (DSP) Score: Development and Performance Analysis.

Area of Science:

• Anesthesiology and critical care medicine
• Medical informatics and artificial intelligence

Background:

• Neuraxial blocks can be difficult and traumatic, impacting patient care.
• Previous score-based prediction methods have had limited practical application.
• Artificial Neural Network (ANN) analysis identified strong predictors of failed spinal-arachnoid punctures.

Purpose of the Study:

• To develop a clinical scoring system (DSP Score) for predicting difficult spinal-arachnoid puncture procedures.
• To validate the performance of the DSP Score on an independent cohort.
• To leverage ANN analysis for identifying key predictors of procedural difficulty.

Main Methods:

• Developed the Difficult Spinal-Arachnoid Puncture (DSP) Score using coefficient estimates from an ANN model (Pr(>|z|) <0.01).
• Analyzed 300 spinal-arachnoid punctures (index cohort) performed in an academic institute in India.
• Applied the DSP Score to the index cohort for ROC analysis and Youden's J point determination.

Main Results:

• The DSP Score incorporates spine grades, performer experience, and positioning difficulty (score range 0-7).
• The DSP Score achieved an excellent Area Under the ROC Curve of 0.858 (95% CI: 0.811-0.905).
• A cut-off value of 2 yielded a specificity of 98.15% and sensitivity of 56.5%.

Conclusions:

• The ANN model-based DSP Score demonstrates excellent predictive performance for difficult spinal-arachnoid punctures.
• With a cut-off of 2, the DSP Score offers a combined sensitivity and specificity of approximately 155%.
• The DSP Score is a valuable diagnostic tool for enhancing clinical practice in neuraxial procedures.