Fentanyl-Hunter: Machine Learning–Driven Screening for Opioid Metabolites

Fentanyl, a widely used synthetic opioid, can be fatal even at low exposures, and in recent years, the illegal abuse of fentanyl and its analogs has become a serious global concern. Scientists headed by a team at the Shanghai Institute for Doping Analysis, Shanghai University of Sport, and the Department of Environmental Science & Engineering, Fudan University, have now reported on the development of a machine learning platform called Fentanyl-Hunter, which can screen for opioid metabolites in biological and environmental samples.

In tests, the platform identified 27 previously unknown metabolites of fentanyl compounds in in vitro samples of human liver cells, as well as two in patient urine samples. The screening platform also found biomarkers of the opioid, and their derivatives, in more than 250 human and environmental samples from eight countries.

Lead author Changzhi Shi, PhD, at the Shanghai University of Sport, and colleagues, reported on their development in Science Advances, in a paper titled “Machine learning– and multilayer molecular network–assisted screening hunts for fentanyl compounds,” reporting, “In this study we introduce a rapid, accurate, and comprehensive screening platform for fentanyl and its transformation products, named Fentanyl-Hunter, which outperforms previous analytical workflow.”

Fentanyl and its many derivatives have infiltrated communities around the world, and newer fentanyl-type compounds continue to crop up and circumvent detection. The authors cited figures indicating that in 2023, fentanyl abuse accounted for approximately 75,000 deaths in the United States, making up nearly 70% of all drug overdose fatalities. “The comprehensive monitoring of fentanyl metabolites is essential for assessing drug abuse, toxicity, and metabolism, with critical applications in preventing overdoses and providing key forensic and toxicological evidence,” the investigators further noted.

More than 1,400 fentanyl analogs with similar or even greater potency to fentanyl have also been synthesized, Shi and colleagues pointed out. “They are gradually becoming prevalent in the drug market.” These previously unidentified fentanyls are designed to avoid analytical detection, which is posing challenges for global regulation and control. The number of previously unknown fentanyl analogs is also constantly increasing,” the team acknowledged. “… however, most have a low detection rate. As structural modifications become more complex, rapid screening of fentanyls is becoming increasingly challenging.”

Over the last decade, high-resolution mass spectrometry (HRMS) and nontargeted analysis (NTA) have demonstrated capabilities in identifying fentanyl with high-throughput and confident identification, the team stated. However, they noted, “Current NTA platforms that rely on MS and spectral libraries are insufficient for detecting unique fentanyl compounds and their metabolites.”

Changzhi Shi and colleagues developed Fentanyl-Hunter to address challenges in detection, pinpointing new and emergent fentanyl analogs via metabolites. The screening platform uses a machine learning classifier and multilayer molecular network to select and annotate fentanyl compounds using mass spectrometry (MS), they explained. “It includes a machine learning model for screening fentanyl compounds (Fentanyl_Finder) and a multilayer molecular network–assisted structure annotation tool (Fentanyl_ID). The machine learning classification was trained on 772 fentanyl spectra and the multilayer network, based on spectral similarity and paired mass distances, covers more than 87% of known fentanyls.

The team used Fentanyl-Hunter to screen for fentanyl metabolites both in vitro and in vivo. They first examined in vitro biotransformations of four derivatives (fentanyl, remifentanil, sufentanil, and alfentanil) in human liver cell fragments. Fentanyl-Hunter found eight metabolites that were already documented and 27 previously unknown metabolites. In urine samples from fentanyl users, the platform also successfully detected metabolites, including two previously unreported ones only seen in the in vitro tests.

Shi et al. next had Fentanyl-Hunter look at wastewater from a treatment plant, which confirmed norfentanyl’s dominance as the main metabolite in wastewater. The platform also uncovered fentanyl, sufentanil, norfentanyl, or remifentanil in 250 samples from sewage sludge, surface water, seawater, and human samples from eight different countries: France, Luxembourg, the United States, China, Israel, Indonesia, Australia, and Argentina.

Their findings, they suggest, “… validate our methods and suggest that the risks of fentanyl abuse and environmental pollution should be closely scrutinized and carefully reevaluated.” The results, in addition, “… underscore the urgent need for the strengthened regulation of the fentanyl family, particularly regarding variants and metabolites … This platform could serve as a robust tool for the detection and regulation of fentanyl variants and metabolites in public health, forensic science, environmental monitoring, and law enforcement applications.”