Discrimination of seized drug positional isomers based on statistical comparison of electron-ionization mass spectra

Positional isomers of novel psychoactive substances (NPS) present challenges in seized drug identification given the high degree of similarity in the resulting electron-ionization (EI) mass spectra. In this work, a previously developed method, which is based on the unequal variance t-test, is applied to statistically compare mass spectra with a specific focus on distinction of positional isomers. Two sets of positional isomers were statistically compared: ortho-, meta-, and para-isomers of ethylmethcathinone (EMC) and ortho-, meta-, and para-isomers of fluoromethamphetamine (FMA). Spectra were also subjected to principal components analysis (PCA) to identify ions contributing to the variance in each data set. For both the EMC5 and FMA5, spectra of corresponding isomers were statistically associated at the 99.9% confidence level. Random-match probabilities for associations, which estimate the probability that the mass spectral fragmentation pattern occurs by random chance alone, ranged from 6.5 x 10(-35)- 5.9 x 10(-27). For the EMC isomers, discrimination of the ortho-, meta-, and para-isomers was achieved at the 99.9% confidence level, with up to 16 ions responsible for discrimination. For the FMA isomers, discrimination of isomers was also achieved, albeit at the 99% confidence level due to the higher degree of spectral similarity among these isomers. For FMA comparisons, up to 10 discriminating ions were identified, with the lower number of ions a result of the higher spectral similarity. For both sets of isomers, there was correspondence in the discriminating ions identified by the statistical comparison method and by PCA.