Drug detection in oral fluid has become popular over the last few years with oral fluid being a non-invasive matrix. Handheld surface enhanced Raman spectroscopy (SERS) is a rapid technique that detects drugs at low concentration and in any field. This work proposes using handheld SERS with machine learning algorithms for detecting drugs in oral fluid. Three drugs were evaluated being cocaine, its metabolite (benzoylecgonine) and paracetamol. Raman spectra of drug solutions mixed with gold or silver nanoparticles were collected through glass vials. These spectra showed high signal to noise ratios and each spectrum of a drug solution showed characteristic drugs bands for each drug. Then three machine learning algorithms were applied and were correlation in wavenumber space, principal component analysis and partial least square regression. The correlation in wavenumber space and principal component analysis classified the different drugs in oral fluid with accuracy up to 99% depending on the model. The partial least square regression informed about the levels of drugs in the oral fluid solution with high accuracy and precision. In conclusion, handheld SERS could instantly detect and predict concentrations of drugs in oral fluid.

Drugs of abuse are often cut by pharmacological and non-pharmacological impurities that impact their adverse events. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) offers a rapid technique that requires minimal sample preparation and can be carried to the field in portable form. This study utilized ATR-FTIR spectroscopy for detecting street drugs of abuse (n = 291). Few milligrams of powder or powdered tablets were measured directly on the instrument. Reference analysis was made using immunoassays and gas chromatography-mass spectrometry. The results showed that ATR-FTIR spectroscopy could identify more than 50% of the drugs measured. The main drugs found contained cocaine, ketamine, benzodiazepines, piperazines, and amfetamine derivatives. Accurate drug identification depended on the number of impurities, types of impurities and concentration of the drug in the product. Hence, drugs could be identified via their IR signatures that were able to differentiate a base from a salt. When correlation method (CM) was applied to the spectra, drugs were correctly identified if they did not have multiple impurities. On the other hand, the presence of impurities affected the accuracy of identification and showed mismatches. Mismatches were addressed by applying principal component analysis (PCA) to the IR spectra of the products. PCA in this case four key clusters with few overlaps. corresponding to cocaine, piperazines, amfetamines and ketamine. There were yet some overlap in the PCA scores due to presence of common impurities between products example benzocaine, cocaine and lactose. In summary, ATR-FTIR combined with chemometrics was accurate in identifying street drugs on the spot.

Cocaine is often cut by multiple diluents and adulterants that increase the volume or alter the pharmacological activity. Handheld near-infrared (NIR) and Raman spectroscopy are rapid and can be carried on the site for detection of drugs of abuse. Both techniques are complementary and give a chemical and physical fingerprint of the drugs measured. This research combined NIR and Raman spectroscopy for quantifying cocaine in mixtures non-destructively. Mixtures of cocaine with adulterants and/or diluents were prepared and stored in glass vials. Then each mixture was measured three times through the glass vials. Likewise, pure substances were measured through glass vials. NIR and Raman spectra were exported in Matlab 2024b where partial least square regression (PLSR) was applied. PLSR models were evaluated for accuracy and precision considering the correlation coefficient (r2) and root mean square errors of calibration and prediction (RMSEC and RMSEP) values. In this respect, models based on NIR spectral data showed higher accuracy and precision than those based on Raman spectral data. Thus, the r2values of calibration and prediction sets for NIR spectral models were in the range of 0.9815-0.9925 and 0.9604-0.9907 respectively. Yet, the r2values of calibration and prediction sets for Raman spectral models were in the range of 0.8891-0.9905 and 0.6882-0.9394 respectively. Moreover, the RMSEC and RMSEP values for models based on NIR spectral data were in the range of 2.77-6.04 and 3.13-8.34% m/m respectively. In addition, the RMSEC and RMSEP values for models based on Raman spectral data were in the range of 3.01-8.61 and 6.13-12.7% m/m respectively. The difference in accuracy could be related to NIR spectral data showing more information regarding the measured powders being collected in diffuse reflectance mode. On the other hand, the Raman measurement mode comprised surface reflection. However, the difference in accuracy and precision was not major and both techniques proved accurate and precise in predicting cocaine in mixtures of drug and food products.

In recent years, saliva has emerged as an alternative biological matrix, offering several advantages such as its accumulative nature and non-invasive, non-intrusive sampling. Moreover, paired with novel vibrational spectroscopic techniques such as attenuated total reflectanceFourier transform infrared (ATR-FTIR) and Raman spectroscopy, saliva can be utilized for the detection and monitoring of drug use. Previous detection methods required extensive sample preparation and do not allow for rapid and portable analysis. Therefore, this work employed two vibrational spectroscopic instruments, those being: the Agilent 4500a ATR-FTIR spectrometer and the Agilent Resolve Raman spectrometer for the detection of drugs and their metabolites in artificial saliva. Both instruments were successful in the detection of illicit and over-the-counter drugs including benzoylecgonine, cocaine hydrochloride (HCl), diazepam, delta-9-tetrahydrocannabinol (THC) and paracetamol. A surface enhanced Raman spectroscopy (SERS) method was also developed for enhanced Raman signals. The chosen vibrational spectroscopic techniques utilized demonstrated the ability to detect key drug-related bands at concentrations as low as 0.05 mg/mL and over a four-week period. However, particularly in the case of SERS, the activity of drug-related bands decreased over this time.

As an alternative biological matrix, fingernails act as a non-invasive window into systemic diseases such as cardiovascular diseases (CVDs) or diabetes mellitus (DM). The presence of such diseases manifests physically, altering the fingernail plate's topography, as well as its elemental composition. Scanning electron microscopy (SEM) equipped with energy dispersive X-ray spectrometry (EDX) and scanning transmission electron microscopy detector (STEM), allows for detailed composition analysis, elemental mapping and spectroscopy. As complementary detectors, the presence of key elements such as calcium, oxygen, potassium and silicon can be quantified and investigated in relation to the presence of disease. Furthermore, topographical STEM images provide insight into the prevalence of tissue damage to the fingernail, which is often attributed to hyperglycemia or poor circulation. The EDX weight% values of identified elements within the fingernails were exported into Matlab R2024a, where a machine learning algorithm, principal component analysis (PCA), was applied to identify patterns between the elemental composition of healthy and diseased fingernails. Analysis of variance (ANOVA) results demonstrated relationships between confounding variables such as biological sex and diet and elemental composition. Female participants demonstrated a significantly higher weight% value of titanium than males due to the exposure of titanium in personal-care products and sun creams. The weight% value of calcium was also significantly different in fingernails of individuals who practiced extremely healthy, healthy and unhealthy diets. In the presence of disease, calcium and sodium showed significantly different weight% values in healthy versus diseased fingernails. STEM images further confirmed the presence of disease, with CVD and/or diabetic fingernail clippings showing rough, topographical textures. In contrast, fingernails taken from healthy participants displayed smooth areas, with small amounts of rough texture. Therefore, demonstrating the ability of SEM to detect the presence and severity of disease.

Identifying cosmetics on substrates is essential in any crime scene processing. It is important in such cases to warrant the sample integrity and continuity. Spectroscopic techniques offer the advantage of analyzing samples non-destructively thus addressing the aforementioned requirements. When used in handheld modality, spectroscopic techniques offer rapid and on-site analysis with the ability to collect numerous datasets in minimal time. |Each spectrum is a unique fingerprint of the sample measured and that urges the need to collect large datasets and apply machine learning algorithms to make meaningful conclusions from analysis. Therefore, this work involved applying machine learning algorithms to infrared, near-infrared, and Raman spectroscopic data of cosmetics applied to different substrates. The cosmetics included powder, creams, lipsticks and nail polish; and the substrates included paper, fabric and glass. Machine learning algorithms were correlation method and principal component analysis. Both algorithms showed to be complementary in identifying cosmetics on different substrates, Thus, the approach proved accurate and precise for identifying substrates encountered at crime scenes.

Covid-19 is a novel coronavirus that emerged in 2019 and spread across the globe, establishing a worldwide pandemic. Vaccination was presented as the most effective solution against the virulence of Covid-19. Accelerated vaccination programmes pushed several nucleic acid-based vaccines into production. Global desperation and limited vaccine supply allowed substandard and falsified (SF) Covid-19 vaccines to enter the supply chain. Conventional analytical methods can be cumbersome, costly and sophisticated to operate. Thus, this study presented a comparison of handheld surface-enhanced Raman spectroscopy (SERS) and Raman microscopy with machine learning algorithms (MLAs) for the rapid authentication of Covid-19 vaccines. Measurements were taken using the Metrohm MIRA XTR DS handheld Raman spectrometer and the Horiba XploRA Plus Raman microscope. Raman spectroscopy showed strong potential as a vaccine authentication method, allowing identification of nucleic acid-specific bands in spectra. SERS showed enhancement of up to $498 \%$ when applied to vaccines of sufficient concentration. Clustering based on principal component analysis (PCA) showed some accuracy but indicated poor repeatability for SERS, although, multiple classification models obtained $100 \%$ accuracy and area under the curve (AUC) for vaccine brand prediction based on spectral characteristics. Raman microscopy produced variable results with improved spectral quality over Raman spectroscopy for a number of samples. However, significant fluorescence was observed in numerous vaccine spectra, limiting the identification potential of the method. Clustering based on PCA showed accuracy in distinguishing between vaccine samples, but showed limited performance in vaccine brand identification. Therefore, this paper presents proof of concept for the use of both handheld Raman spectroscopy and confocal Raman microscopy alongside MLAs for the rapid, on-site authentication of Covid-19 vaccines, with further method optimisation required to combat fluorescence interference in vaccine spectra and expansion of sample size to address the potential of overfitting in the MLAs.

The prevalence of cardiovascular diseases (CVDs) and diabetes mellitus (DM) has become a global concern with figures as high as $\mathbf{1 7. 9}$ and $\mathbf{1. 5}$ million lives lost annually [1, 2]. Global figures also suggested that the majority of CVDs and DM are present within low- and middle-income countries (LMICs), where medical equipment, staff and training is limited. As a result, many patients go underdiagnosed or undertreated and instead are left to manifest into further complications such as heart failure or diabetic ketoacidosis, respectively. Therefore, this study aimed to investigate the use of nearinfrared (NIR) spectroscopy paired with machine learning algorithms (MLAs) for detection of CVDs and DM in fingernails. The findings showed key NIR bands related to the glycation of proteins within the fingernails and indicated the presence of disease. Furthermore, binary and multi-class classification models were explored for the classification of healthy, unhealthy, CVD and diabetic fingernails.

COVID-19 is a novel coronavirus first emerging in Wuhan, China in December 2019 and has since spread rapidly across the globe escalating into a worldwide pandemic causing millions of fatalities. Emergency response to the pandemic included social distancing and isolation measures as well as the escalation of vaccination programmes. The most popular COVID-19 vaccines are nucleic acid-based. The vast spread and struggles in containment of the virus has allowed a gap in the market to emerge for counterfeit vaccines. This study investigates the use of handheld Raman spectroscopy as a method for nucleic acid-based vaccine authentication and utilises machine learning analytics to assess the efficacy of the method. Conventional Raman spectroscopy requires a large workspace, is cumbersome and energy consuming, and handheld Raman systems show limitations with regards to sensitivity and sample detection. Surface Enhanced Raman spectroscopy (SERS) however, shows potential as an authentication technique for vaccines, allowing identification of characteristic nucleic acid bands in spectra. SERS showed strong identification potential through Correlation in Wavelength Space (CWS) with all vaccine samples obtaining an r value of approximately 1 when plotted against themselves. Variance was observed between some excipients and a selected number of DNA-based vaccines, possibly attributed to the stability of the SERS colloid where the colloid-vaccine complex had been measured over different time intervals. Further development of the technique would include optimisation of the SERS method, stability studies and more comprehensive analysis and interpretation of a greater sample size.

Near infrared (NIR) spectroscopy offers portable and rapid analysis of endogenous constituents and drugs within fingernails. Fingernails are a useful alternative biological matrix to blood and urine specimen as they provide the advantage of being non-invasive and require minimal sample size (1–3 mm). This work utilised NIR spectroscopy for the detection of (1) drugs in fingernails including benzocaine, calcium carbonate, cocaine hydrochloride (HCl), levamisole HCl, lidocaine HCl and procaine HCl; and (2) endogenous constituents such as carbohydrates, lipids, proteins and water. Fingernails were analysed initially ‘as received’ to identify the aforementioned endogenous constituents. Seven sets of fingernails were then spiked with one the identified drugs and measured over a six-week period. Spectra were exported into Matlab 2019a for spectral interpretation and machine learning analytics (MLAs). MLAs included correlation wavenumber space (CWS), principal component analysis (PCA) and Artificial Neural Networks Self-Organising Maps (SOM). The results showed that NIR spectra of spiked nails showed key characteristic features at specific wavelengths that corresponded to their spiked drug (1). When combined with CWS and PCA, NIR spectroscopy was able to differentiate between spiked and un-spiked nails and distinguish between the drugs that did not share similar chemical structures. CWS values (r values) and PCA loading scores highlighted spectra/spectral features that were significant. In addition, SOM showed further classes beyond PCA that corresponded to changes in physical properties of the fingernails. Thus, finding confirmed that NIR spectroscopy combined with MLAs possessed the ability to characterise fingernails based on their endogenous constituents and to detect the presence of drugs within fingernails.

The use of handheld Raman spectroscopy has increased in popularity based on its ability to provide portable and rapid analysis of endogenous compounds and diagnostic biomarkers in alternative biological matrices, such as fingernails. The application of fingernails as a diagnostic matrix allows for non-invasive, non-intrusive sampling, which can be carried out in the comfort of the patient’s home. This study aimed to identify diagnostic biomarkers in fingernails related to cardiovascular diseases (CVDs) and diabetes mellitus (DM) using Raman spectroscopy and machine learning algorithms (MLAs). The findings showed that Raman spectroscopy successfully identified the presence of disease specific biomarkers in CVD and diabetic fingernails. Furthermore, when used in combination with MLAs, Raman spectroscopy was able to differentiate between healthy, CVD and diabetic fingernails. Further investigation will look at applying additional MLAs for determining the prognosis of disease.