Carfentanil in the illicit drug supply: A machine learning approach for trace detection using surface-enhanced Raman...

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Machine Learning

Background:

• Carfentanil, a potent fentanyl analogue, is increasingly found in illicit drug supplies.
• Current drug checking technologies struggle to detect carfentanil at trace levels.
• This limitation hinders effective overdose prevention strategies.

Purpose of the Study:

• To develop and validate a novel method for reliable carfentanil detection.
• To address the limitations of existing drug checking technologies.
• To enhance harm reduction by improving drug identification capabilities.

Main Methods:

• Surface-enhanced Raman spectroscopy (SERS) using gold nanoparticles and a handheld spectrometer.
• Analysis of 250 illicit drug samples from a community drug checking service.
• Optimization of a random forest machine learning classifier with SERS data.

Main Results:

• The SERS-based model achieved 96% sensitivity and 84% specificity in predicting carfentanil.
• Successfully identified carfentanil at concentrations as low as 0.06% w/w.
• Demonstrated strong discriminatory power (AUC = 0.96) in complex mixtures.

Conclusions:

• SERS combined with machine learning reliably detects trace carfentanil in illicit drug mixtures.
• This portable and sensitive technology can enhance community-based drug checking.
• Findings support evidence-informed drug policy and harm reduction practices.