**Background:** Cannabis sativa L. is a chemically complex plant containing over 565 identified compounds, including more than 120 cannabinoids, terpenoids, flavonoids, and other metabolites. While Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD) are the most studied for pharmacological activity, the 'entourage effect' suggests that therapeutic efficacy arises from synergistic interactions among multiple compounds. Current regulatory frameworks for cannabis safety testing—enforced by agencies such as Health Canada—focus on limited targeted analyses (cannabinoids, microbials, pesticides, heavy metals, residual solvents). However, these systems fail to capture the full chemical complexity, leading to risks from adulteration, degradation products, and inconsistent product quality. The authors argue that the 'authentomics' paradigm, already pioneered in food science (e.g., wine, olive oil, honey), should be applied to cannabis to provide comprehensive, non-targeted chemical fingerprinting combined with big data analysis.
**Methods:** This is a narrative review synthesizing literature on metabolomic technologies applied to cannabis analysis. The authors describe four major analytical categories: (1) Nuclear magnetic resonance (NMR) spectroscopy—used for metabolic fingerprinting of cultivars, with 1H NMR at 400–500 MHz enabling discrimination based on cannabinoid profiles (THC, THCA, CBD, CBDA, CBN) and primary metabolites (carbohydrates, amino acids). Studies cited include Choi et al. (2004) analyzing 12 C. sativa cultivars using PCA, and work by Happyana and Kayser (2014) combining 1H NMR with RT-PCR to track THCA synthase expression. (2) Gas chromatography coupled with flame ionization detection or mass spectrometry (GC-FID/MS)—used to quantify neutral cannabinoids and terpenes. Hazekamp’s group identified 8 major cannabinoids and 36 terpenes; a separate GC-MS seed study identified 236 untargeted metabolites, with 43 significantly different between accessions. Two-dimensional GC (GC×GC) identified 754 metabolites across chemical classes. (3) Liquid chromatography-mass spectrometry (LC-MS)—Berman et al. used ESI-LC/MS to identify 94 cannabinoids from 10 subclasses across 36 samples. LC-HRMS/MS untargeted approaches quantified CBD, CBDA, THC, THCA, CBGA, and CBN, along with trigonelline, proline, arginine, and choline. (4) Other techniques including thermal desorption ion mobility spectrometry (TD-IMS), thin-layer chromatography (TLC), hyperspectral coherent anti-stokes Raman scattering (HCARS), and sorptive tape-like extraction coupled with laser desorption ionization mass spectrometry (STELDI-MS).
**Key Results:** The review identifies several critical quality issues in the current cannabis system. Cannabinoid and terpenoid stability is problematic: THCA oxidizes to cannabinolic acid (CBN-A) and then to CBN upon exposure to air and light; terpenoids degrade via oxidation, isomerization, and polymerization. Lipid oxidation products (e.g., verbenol, linalool, alpha-terpineol, terpinen-4-ol, aldehydes, ketones) form in cannabis oils, especially at room temperature versus refrigerated conditions. Structural and stereoisomer complexity is noted: THC has seven possible structural isomers (Δ6a,10a-THC through Δ9,11-THC) and four stereoisomers, with only (−)-trans-Δ9-THC occurring naturally. Adulteration cases include mixing with synthetic cannabimimetics (Spice/K2), pine rosin (identified by NMR and ESI-MS via abietic and other resin acids), and vitamin E acetate as a diluent in illicit vape cartridges (detected by GC-MS and LC-MS/MS). The authors report that linear discriminant analysis (LDA) of 1H NMR spectra provided 99.8 ± 0.4% prediction accuracy for cultivar discrimination, and support vector classification machine trees (SVMTree) offered robust classification performance.
**Clinical Implications:** The application of authentomics to cannabis would enable comprehensive, non-targeted chemical profiling that captures both regulated components and unknown or unexpected adulterants—analogous to food fraud detection (e.g., melamine in milk, methanol in spirits). The authors emphasize that current consumer reliance on THC content as a quality indicator is misleading; a study of 121 participants showed similar neurobehavioral patterns between very high THC extracts (70% or 90% THC) and cannabis flower (16% or 24% THC). Establishing a cannabis metabolome database (analogous to the Human Metabolome Database with 114,265 metabolites or the Food Database with >28,000 metabolites) is essential for biomarker discovery and inter-laboratory standardization. The authors call for standardized analytical procedures (e.g., USP and AOAC methods for cannabinoids), open-access structure databases, authentic standards (13 USP quality standards currently available), and public–private partnerships such as the Emerald Test™ proficiency program (ISO/IEC 17043 accredited). Future directions include studying minor cannabinoids (CBDA, Δ9-THCV, CBDV, CBG, CBC) for their non-intoxicating therapeutic potential, and developing personalized medicine approaches based on endocannabinoid system (ECS) genetic variants and pharmacometabolomics. The review concludes that combining targeted and untargeted analysis within an authentomics platform—meeting ISO standards and supported by integrated database design—is essential to ensure cannabis product safety, efficacy, and authenticity for medical and health promotion uses.