Rapid evaporative ionization mass spectrometry: A survey through 15 years of applications

Abstract

Rapid evaporative ionization mass spectrometry (REIMS) is a relatively recent MS technique explored in many application fields, demonstrating high versatility in the detection of a wide range of chemicals, from small molecules (phenols, amino acids, di‐ and tripeptides, organic acids, and sugars) to larger biomolecules, that is, phospholipids and triacylglycerols. Different sampling devices were used depending on the analyzed matrix (liquid or solid), resulting in distinct performances in terms of automation, reproducibility, and sensitivity. The absence of laborious and time‐consuming sample preparation procedures and chromatographic separations was highlighted as a major advantage compared to chromatographic methods. REIMS was successfully used to achieve a comprehensive sample profiling according to a metabolomics untargeted analysis. Moreover, when a multitude of samples were available, the combination with chemometrics allowed rapid sample differentiation and the identification of discriminant features. The present review aims to provide a survey of literature reports based on the use of such analytical technology, highlighting its mode of operation in different application areas, ranging from clinical research, mostly focused on cancer diagnosis for the accurate identification of tumor margins, to the agri‐food sector aiming at the safeguard of food quality and security.

Article Related Abbreviations

CE

collision energy

DART

Direct analysis in real time

DESI

Desorption electrospray

DHA

Docosahexaenoic acid

DPA

Docosapentaenoic acid

EPA

Eicosapentaenoic acid

Iknife

Intelligent knife

LA‐REIMS

Laser REIMS

LC‐MS and UHPLC‐MS/MS

Liquid chromatographymass spectrometry, Ultra‐high performance liquid chromatography‐tandem mass spectrometry

MALDI

Matrix assisted laser desorption ionization

MSn and MS/MS

Tandem mass spectrometry

PCA‐LDA

Principal Component Analysis‐ Linear Discriminant Analysis

PDO

Protected designation of origin

PGI

Protected geographical indication

REIMS

Rapid evaporative ionization mass spectrometry

SPE

Solid phase extraction

1. INTRODUCTION

Rapid evaporative ionization mass spectrometry (REIMS) is based on the use of a recent ionization source, introduced in 2009 by a research team of the Imperial College of London, led by Prof. Zoltan Takats [1]. This ion source has been used exclusively in ambient MS approaches, in which, according to the definition of ambient techniques, the sample is analyzed in its native form, at atmospheric pressure and at room temperature [2]. In 2013, the same research group coined the term iknife, an abbreviation of “Intelligent Knife”, to indicate the coupling between an electrosurgical knife with REIMS [3].

The most widely accepted classification system among ambient MS techniques considers three main groups based on desorption/ionization mechanism: I) liquid extraction techniques that use solvent to extract/desorb molecules from the sample, while the ionization occurs in a second step (e.g., desorption electrospray, DESI); II) plasma‐based techniques that use plasma to desorb and ionize molecules from the sample (e.g., direct analysis in real‐time, DART); III) laser ablation in which a laser source assists the desorption and ionization mechanism [4]. Benefits and drawbacks of each category were extensively discussed, such as restrictions related to sample state and analyte polarity (e.g., only liquid samples and polar compounds can be analyzed by category I), the need for dedicated sample preparation (e.g., Matrix assisted laser desorption ionization, MALDI) or additional gas requirements for plasma‐based techniques [4, 5]. Alternative methods were developed to simplify the instrumental setup without the need for external chemicals (solvent, reagents, or gas) to promote desorption/ionization [4]. These methods involve the use of temperature, ultrasounds, or vibrations to favor analyte desorption. REIMS falls within thermal‐based ambient techniques, revealing to be particularly appropriate for in‐vivo analysis of biological tissues in clinical facilities since thermal evaporation is what happens during the surgical resection of tissues and no sample pretreatment is required. Differently from the solvent‐based desorption technique, the obtained molecular vapors are a mixture of both polar and non‐polar analytes.

Schafer et al. [1] proposed in the early beginnings two distinct mechanisms to explain the ion formation from the sample, one of which involved a desorption phase of neutral molecules followed by an ionization step in the gas phase through proton reaction with ionized water molecules (two‐step desorption‐ionization mechanism). The second mechanism currently shared and widely accepted, is based on a rapid thermal evaporation of the sample material during which the water molecules naturally present in the samples immediately react with chemical compounds to create an aqueous solution of molecular and ionic species (simultaneous desorption/ionization mechanism). The initially proposed REIMS setup was configured with an additional fluid line for the direct aspiration of the aerosol into the mass spectrometer without any postionization. [1] Subsequently, a Venturi air jet pump was added to improve the transfer of the vapor. Moreover, the instrumental setup was implemented with an orthogonal stream of nitrogen and a coaxial flow of 2‐propanol to reduce contamination effects [3]. Finally, the ionization is refined via the droplet‐surface collision phenomenon, which led to the introduction of a heated collision surface in the commercialized setup for a so‐called post‐ionization (Waters Corporation, www.waters.com).

Figure 1A displays the scheme of the REIMS setup. Briefly, the electrosurgical knife causes, by the Joule effect, the local heating of the sample and, as a consequence, the thermal evaporation/ionization of its molecular components. The vapors reach the REIMS source by the Venturi effect and are driven by a flow of nitrogen and a coaxial flow of 2‐propanol. Then, a post‐ionization takes place via thermionic effect by contact with a surface heated to over 800°C and the generated ions reach the analyzer and the mass detector.

FIGURE 1.

Scheme of a typical experimental workflow for a rapid evaporative ionization mass spectrometry (REIMS) application. Modified by www.waters.com.

The combination of described instrumental setups with the chemometric analysis of the data allows for the construction of spectral databases and statistical models. In this regard, raw data file processing is essential for mass peak alignment and lock‐mass correction in order to increase the identification reliability of single compounds and allow for direct sample comparison. Most chemometric software packages use mass binning options to reduce the number of unmatched peaks, for example, peaks likely corresponding to the same analyte but detected at slightly different m/z values. The bin size is set according to MS calibration results. Taking into account that REIMS applications were carried out solely through a high‐resolution MS system (i.e., quadrupole‐time of flight), a small bin size (0.1 or lower) was usually selected in order to avoid loss of informative data, which could be responsible for sample differentiation. When a mass binning approach is adopted the entire mass spectrum is processed to maximize the number of deductible information. An alternative or complementary approach consists of typical metabolomic analysis functions, named peak peaking, that is the selection and identification of a set of mass peaks in the spectrum to find candidate biomarkers [6]. Afterward, processed data files are collected in spectral libraries to be used for the building of classification models according to the concept of machine learning. The chemometric analysis allows for clustering the samples by minimizing the differences between samples belonging to the same class and maximizing those between different samples (generation of a statistical model); the tool is able to identify unknown samples through match with the built statistical model (real‐time identification of unknown samples). The entire experimental workflow is presented in Figure 1B.

Since electrocautery is normally employed in surgery, REIMS thas been used in principle in the clinical field for diagnostic purposes for the analysis of biological samples and the differentiation between healthy cells and cancer cells [7, 8, 9, 10, 11, 12, 13, 14]. It has immediately raised a lot of clamor for the possibility of identifying the tissues in real‐time, posing itself as an alternative to the longer and more laborious techniques of histopathology, or to temporary histology that showed a high probability of error. Furthermore, in‐situ and real‐time results could guide the surgeon during the surgery to accurately define the tumor margins and reduce the risk of relapses. However, these potentials have not yet been fully explored in vivo due to the intrinsic incompatibility between such an imposing instrumentation (not only in terms of weight and size, but also for electricity consumption, costs, ordinary and extraordinary maintenance operations), and the sterile environment of an operating room. Then, a certain number of applications were successfully developed through ex‐vivo analysis of biological samples, transferred from the surgery room to the analytical laboratory immediately after the resection, or properly stored at –80°C.

The aim of the present review is to provide the state‐of‐the‐art REIMS technique through a survey of 15‐year applications, highlighting the sampling device, the application fields, and the aim of different research works (e.g., lipidomics, metabolomics, food authenticity, diagnostics, etc.).

2. SAMPLING DEVICE FOR REIMS

Following the first REIMS applications with the hand‐held monopolar probe (iknife), different sampling devices have been developed to support various REIMS applications, as reported in Table 1 along with their benefits and drawbacks. In 2013, Strittmatter et al. used for the first time bipolar forceps coupled to REIMS for the analysis of intact bacteria [15]. Better sensitivity, simpler handling, complete elimination of memory effects, and less frequent blocking of transfer devices were all benefits of the forceps‐based configuration. This sampling device produces comparatively less aerosol since it requires less sample biomass for analysis and has a lower current density because of the electrodes' parallel geometry. Subsequently, this setup configuration was used also in clinic and food applications [6, 16, 17]. One of the main drawbacks of the bipolar probe consists in the lack of automation, different from the monopolar probe which was successfully automated through the use of a commercially available colony picker robot equipped with an autosampler for sample positioning on suitable plates. Such a robotic machine (TECAN) was employed for the characterization of clinically important bacteria and fungi colonies with high throughput and achieving uniform and localized heating during sampling [16]. The developed high‐throughput REIMS platform was compared in different works with the hand‐held bipolar forceps in terms of sensitivity and accuracy for the identification of microbiological samples, demonstrating the higher performances of hand‐held bipolar probe [16, 18, 19], despite the scares possibility for automation.

TABLE 1.

Typical sampling device for rapid evaporative ionization mass spectrometry (REIMS) analysis.

Sampling device	Applicability	Benefits	Drawbacks
Monopolar probe	Solid and conductive samples	Automatable through the use of the TECAN robotic platform (high throughput and reproducibility) Large sampling capability (extensive contact with sample material even in depth)	High memory effect: need for frequent cleaning operation Destructive
Bipolar forceps	Solid and conductive samples	Reduced memory effect and improved sensitivity as a result of the minor contamination High identification accuracy Low sample biomass required	Low reproducibility Not automatable Destructive
Laser	Solid and liquid samples	Not destructive Reduced memory effect and improved sensitivity as a result of the minor contamination High identification accuracy Low sample biomass required Automatable	Elevated initial cost investment High maintenance cost and energy consumption
Soldering ion	Solid and liquid samples	High ionic current intensity without the need for signal enhancer.	Not automatable Destructive High memory effect: need for frequent cleaning operation

Moreover, the use of an electrical current requires contact between the probe and the sample, necessitating either cleaning or changing the monopolar probe between experiments. To overcome this problem, while creating an automated system, Cameron et al., for the first time, coupled the laser desorption with REIMS (LA‐REIMS), reducing the need for direct contact between the sample and the electrical probe, allowing automation and then improving analytical throughput [18]. Additional advantages of a LASER as a sampling device are the higher versatility in analyzing both liquid and solid samples and the non‐destructiveness. In their work, Cameron and colleagues [18] carried out a three‐term comparison between hand‐held bipolar forceps, LA‐REIMS, and high‐throughput monopolar probes, proving a high degree of similarity between the first two sampling devices. On the other hand, LASER surely implies an elevated initial cost investment, as well as high maintenance cost and energy consumption.

A valid alternative to investigate liquid and poorly conductive samples was also presented by Wang et al., that combined the REIMS source with an electric soldering iron as a contact heating device [20]. This instrumental setup has been mostly used in the food field for the analysis of honey, milk, sorghum, olive oil, and cream [20, 21, 22, 23, 24] but was also employed for textile and clinical applications for the analysis of lather products and plasmalogen‐loaded zein nanoparticles, respectively [25, 26]. Liu and colleagues performed a clear comparison between iknife and soldering iron sampling devices in terms of signal intensity for the ambient ionization of sorghum powders, in order to assess that any ion enhancer is necessary when using the soldering iron even in the case of not conductive samples, while an insufficient aerosol is generated by iknife [22].

3. POTENTIAL AND LIMITS

In this section, the pros and cons of REIMS, as summarized in Table 2, are discussed in order to address potential users on the selection of this analytical methodology for the purposes of their research. One of the primary advantages of this setup is the high‐speed of the analysis. In fact, results are obtained within seconds, making it suitable for applications in which real‐time or high‐throughput analyses are required. In clinical and surgical applications, REIMS provides real‐time feedback, allowing for on‐the‐spot decision‐making and adjustments. Moreover, this technique is suitable for metabolomics investigation through the detection of a wide range of compounds, including both polar and non‐polar molecules. Compared to other analytical methods, such as chromatographic approaches, the REIMS setup requires minimal or no sample preparation, reducing analysis time and the risk of sample loss and contamination.

TABLE 2.

Benefits and drawbacks of rapid evaporative ionization mass spectrometry (REIMS) technique.

Benefits	Drawbacks
Rapid	Sophisticated and expensive system
Suitable for characterization of whole metabolome (untargeted approach)	Scarce automation (except for laser as sampling device or use of TECAN platform)
No sample preparation	No quantitative analysis
Easy combination with statistical methods for sample discrimination	Need of a large amount of samples to obtain a reliable database
Mass accuracy data	Poor discrimination of isomers or isobars

Finally, since the REIMS source was normally coupled with a high‐resolution tandem MS (MS/MS) analyzer, a high identification power was achieved by exploiting both mass accuracy and MSⁿ data.

At the same time, the REIMS setup has some drawbacks. Since it relies on MS databases for the identification of unknown samples, the accuracy of results is highly dependent on the quality and comprehensiveness of these databases. Intra‐sample variability represents a key‐aspect, making it difficult to establish universal reference profiles. In particular, biological samples can exhibit natural variations in composition, while the chemical profile of vegetable products can largely change during different harvest years. This can lead to false positives or misidentifications. A possible solution is the inclusion in the database of a huge number of samples, for example, obtained over a long period of time (more harvest years and pedoclimatic conditions) or, in biological terms, coming from a heterogeneous population (different ages, genders, and nationalities).

Another limitation can be represented by the difficulty of discriminating between isomers or even isobars, especially in highly complex matrices, such as biological tissues or certain food products. MS/MS experiments could be helpful in this context, but if more isomers coexist in the same matrix, the determination of the most abundant one is quite challenging. Furthermore, REIMS represents a not well‐suited technique for quantitative analysis, being able to provide only a relative quantification within each chemical class.

Besides, given that the initial and maintenance costs of REIMS are quite elevated, its accessibility for smaller research labs or institutions was limited.

Finally, the not automation of the setup (except when the laser is used as a sampling device) constituted a sort of barrier for users.

4. APPLICATIONS

Figure 2 shows the number of applications of the technique divided by year (Figure 2A) and by application field (Figure 2B). The union of this information generates a histogram that reports the evolution of REIMS applications in the various fields over the years (Figure 2C). From its inception until 2015, this instrument has been used only for diagnostic purposes, with the exception of a few microbiological applications [6, 15, 27]. The first application in the food field appeared in 2016 and deals with the identification of the species of origin of meat, including minced and cooked products [28]. The study aimed at the discrimination between equine and bovine meat, which had become necessary because of the horse meat scandal three years earlier, which alerted organizations active in the preservation of food security and traceability. In this context, the REIMS technique has proved to be a useful tool for the detection of food fraud by generating mass spectral profiles characteristic for each type of food (e.g. beef vs horse meat) and determining typical markers that can reliably identify authentic products [28]. The promising results of this first application paved the way for significant use of REIMS in the food sector, reaching nine applications in 2019 and 2021, 10 applications in 2022, and seven in 2023, definitely exceeding the number of applications in the clinical field.

FIGURE 2.

Applications of rapid evaporative ionization mass spectrometry (REIMS) technique grouped by (A) year, (B) application field, and (C) main areas of application divided by years.

Since 2019, applications have emerged in different areas, spanning from botany and veterinary medicine to entomology.

In all cases, this innovative technique combines two complementary analytical approaches. The first deals with a complete and accurate characterization of the sample from both qualitative and semi‐quantitative points of view. An illustrative instance is found in the study conducted by Arena et al. [29], wherein REIMS was utilized to obtain a metabolomic profile of Sterculia setigera bark. Moreover, MS/MS experiments were carried out to achieve a reliable identification of molecular constituents. As an example, Figure 3A illustrates the MS/MS spectrum acquired by the researchers for m/z 289.07, corresponding to catechin, along with structure elucidation.

FIGURE 3.

(A) Rapid evaporative ionization mass spectrometry (REIMS) spectrum of the bark of sterculia setigera (on the left) and tandem mass spectrometry (MS/MS) spectrum of m/z 289.07 for structure elucidation (on the right), reproduced [29]; (B) Prediction of external validation sets of partial least‐squares discriminant analysis (PLS‐DA) trained models of 32 beers according to the different brands (on the left) and types (on the right), reproduced with permission [30]; (C) Statistical model for pistachio differentiation from five different geographical regions (on the left), real‐time recognition of a Bronte pistachio sample (in the middle), MS/MS spectrum of the discriminant feature at 835.53 ion (on the right), corresponding to phosphatidylinositol PI (C16:0/C18:1), reproduced [31].

The second is a fingerprinting approach that aims to obtain characteristic profiles of various types of samples. In other words, the REIMS technique has been employed both to get extremely informative mass spectra from the matrix for the elucidation of the chemical composition and for the building of spectral databases to be used for the real‐time identification of unknown samples. In the latter case study, the identification and quantification of individual molecular constituents is not strictly required. Cardoso et al. [30] adopted this fingerprinting approach to discriminate between 32 beers according to the different brands and types (Figure 3B).

However, in most of the applications present in the literature, the two approaches have been combined to allow the differentiation of samples based on their unique spectral profiles and the identification of so‐called discriminant features. This combined approach has often been used both in the clinical field, as it is often necessary to identify specific biomarkers responsible for tissue discrimination, and in the food sector for the identification of authenticity markers. Figure 3C shows an example of this approach were the authors exploited the REIMS for the differentiation of Pistachio [31]. They achieved sample clustering based on the distinct fingerprinting of various Pistachio varieties and also identified specific discriminant features.

The next sections deal with the major application fields of the REIMS technique, which are the clinical and the food areas (Figure 2). A further section is dedicated to highlighting the great versatility of the REIMS technology in other fields.

Noteworthily, some works focused on the development and testing of the technique, by using biological tissues or food matrices as probe samples. In the present review, such applications were grouped under the analytical chemistry field since the aim of these works was to test the analytical technologies through the analysis of probe samples in order to highlight the potential and limits of the technique [32, 33, 34, 35, 36].

The entire list of applications is reported in Table 3.

TABLE 3.

Selected application of rapid evaporative ionization mass spectrometry (REIMS) technique.

Application field	Samples	Sampling device	Chemometric analysis	Molecular identification	References	MS analysis
Clinical	Human and canine tumors (breast, testicles, gastrointestinal tract, lung, and thyroid)	Electrosurgical knife	√	√	[1, 39]	MS and MS/MS
Clinical	Human tumors (brain, lung, breast, and gastrointestinal)	Electrosurgical knife and bipolar forceps	√	√	[3]	MS and MS/MS
Clinical	Cancer cells (breast, brain, gastrointestinal, skin, lung, gynecological, prostate, and renal)	Bipolar forceps	√	√	[7]	MS and MS/MS
Clinical	Gastrointestinal tumors	Electrosurgical knife	√	√	[12, 40, 54, 56]	MS and MS/MS
Clinical	Gastrointestinal tumors	Polypectomy snare	√	√	[41]	MS
Clinical	Gastrointestinal tumors	Electrosurgical knife	√	√	[47, 53]	MS
Clinical	Gastrointestinal tumors	Modified handheld diathermy pencil	√	√	[55]	MS and MS/MS
Clinical	Skin tumors	Electrosurgical knife	√	×	[8, 10, 14, 51]	MS
Clinical	Gynecological tumors	LASER	√	√	[11]	MS
Clinical	Gynecological tumors	Electrosurgical knife	√	√	[13, 43, 52]	MS and MS/MS
Clinical	Breast tumors	Bipolar forceps	√	√	[9]	MS and MS/MS
Clinical	Breast tumors	Electrosurgical knife	√	√	[42]	MS and MS/MS
Clinical	Breast tumors	Electrosurgical knife	√	×	[44, 51]	MS
Clinical	Brain tumors	Bipolar forceps	√	√	[57]	MS and MS/MS
Clinical	Brain tumors	Electrosurgical knife	√	√	[58]	MS and MS/MS
Clinical	Plasmalogen‐loaded zein nanoparticles	Soldering iron	√	√	[26]	MS and MS/MS
Clinical	Human feces	Robotic monopolar probe	√	√	[59]	MS
Clinical	Human feces	LASER	√	√	[49]	MS and MS/MS
Clinical	Saliva	LASER	√	√	[50]	MS
Clinical	Skin tissues	Electrosurgical knife	√	√	[61]	MS
Clinical	Aortic tissue	Electrosurgical knife	√	×	[45, 46]	MS
Veterinary clinical	Animal feces	Electrosurgical knife	√	×	[37]	MS
Veterinary clinical	Canine mammary tumors	Electrosurgical knife	√	√	[38]	MS
Clinical/ Microbiology	Gastrointestinal tumors and bacteria	REIMS Imaging Platform	√	√	[40]	MS and MS/MS
Clinical/ Microbiology	Bacteria isolates from cystic fibrosis patients	Bipolar forceps	√	√	[60]	MS and MS/MS
Microbiology	Bacteria	Electrosurgical knife and bipolar forceps	√	√	[15]	MS and MS/MS
Microbiology	Bacteria	Bipolar forceps	√	√	[6]	MS and MS/MS
Microbiology	Bacteria	LASER	√	√	[18]	MS
Microbiology	Bacteria and fungi	Bipolar forceps and robotic monopolar probe	√	×	[16]	MS
Microbiology	Fungi	Bipolar forceps and automatic monopolar probe	√	√	[19]	MS
Food	Fish	Electrosurgical knife	√	√	[62, 64, 66, 67, 69, 71]	MS and MS/MS
Food	Fish	Electric heating probe	√	√	[73]	MS and MS/MS
Food	Fish	Electrosurgical knife	√	√	[63, 70, 72]	MS
Food	Fish	Electrosurgical knife	√	×	[68]	MS
Food	Meat	Electrosurgical knife	√	√	[28, 76, 79, 84‐88]	MS and MS/MS
Food	Meat	Electrosurgical knife	√	√	[83]	MS
Food	Meat	Electrosurgical knife	√	×	[75, 80]	MS
Food	Meat	Electrosurgical probe	√	×	[77, 78, 81]	MS
Food	Pistachio	Electrosurgical knife	√	√	[31]	MS and MS/MS
Food	Extra virgin olive oil	Electrosurgical knife	√	√	[90]	MS
Food	Beer	Electrosurgical knife	√	×	[30]	MS
Food	Fish phospholipids	Electric heating probe	√	√	[74]	MS
Food	Honey	Soldering iron	√	×	[20]	MS
Food	Milk	Soldering iron	√	√	[21]	MS and MS/MS
Food	Sorghum	Soldering iron	√	√	[22]	MS
Food	Virgin olive oil adulteration	Soldering iron	√		[23]	MS
Food	Cream	Soldering iron	√	√	[24]	MS
Analytical	Meat	Electrosurgical knife and LASER	√	√	[32]	MS and MS/MS
Analytical	Porcine tissues	Ultrasonic harmonic scalpel	√	√	[33]	MS and MS/MS
Analytical	Food‐grade meat samples	Electrosurgical knife	√	√	[34]	MS and MS/MS
Analytical	Apples, animal, and human tissues	LASER	√	√	[35]	MS and MS/MS
Analytical	Fish	Electrosurgical knife	√	√	[36]	MS
Botanics	Sterculia setigera bark	Electrosurgical knife	×	√	[29]	MS and MS/MS
Botanics	Kigelia africana fruit	Electrosurgical knife	×	√	[94]	MS and MS/MS
Entomology	Drosophila species	Electrosurgical knife	√	√	[91]	MS
Entomology	Mosquito species	Electrosurgical knife	√	×	[93]	MS
Entomology	Aedes aegypti larvae	Electrosurgical knife	√	×	[92]	MS and MS/MS
Forensic	Drugs	Electrosurgical knife	√	√	[95]	MS and MS/MS
Pharmaceutics	Tissues and cell cultures	Electrosurgical knife and bipolar forceps	√	×	[96]	MS and MS/MS
Textile	Leather	Soldering iron	√	√	[25]	MS

Abbreviation: MS/MS, tandem mass spctrometry.

4.1. Clinical analysis

More than 30 applications of REIMS fall within the clinical field mainly aimed at the differentiation of cancerous tissues or pathological conditions and covering both veterinary [37, 38] and human medicine [39, 40, 41, 42, 43, 44, 45, 46] areas. In these cases, the first step was the collection of MS spectra of as many possible samples to build a database to be promptly used for identification purposes.

4.1.1. Cancer diagnosis

One of the largest databases of biological tissues was built in 2013 by Balog et al. [3], who acquired around 3000 tissue‐specific mass spectra from about 400 patients. The spectral comparison between different tissues and, specifically, between healthy and tumor tissues, highlighted differences in the lipid composition. More in detail, different ratios of lipids, such as phosphatidylcholine and phosphatidylinositol, were used as markers to determine the origin of metastatic cancers.

Some years later, the same research group explored REIMS as a valuable tool in endoscopic procedures [41, 47]. Indeed, endoscopy is routinely adopted in national screening programs, and endoscopic resection by using diathermic tools is becoming a valuable option for the treatment of early‐stage cancer and premalignant conditions, due to its less invasive nature. However, re‐intervention is needed for a high percentage of patients because of incomplete excision [48]. Moreover, the incorrect definition of margins and mucosal layers can lead to lesions and perforation risk. Within this context, the REIMS endoscopic method appeared as a valid methodology for the real‐time differentiation of normal and altered tissues. Even in this case, the lipidomic profile was responsible for the differentiation with many phospholipids over‐expressed in pathological states compared to normal conditions.

Meanwhile, the robustness of the REIMS spectral profiles was carefully evaluated in order to explore the capability of REIMS for obtaining a detailed and reliable shotgun lipid profile, for example, enabling a comprehensive characterization of lipid species within a cell or tissue [7]. Shotgun lipidomics is a high‐throughput approach offering valuable insights into the molecular composition of complex biological samples. The proof of principal study carried out by Strittmatter et al. [7] paved the way for interesting correlations between the cell lipid composition and cancer phenotypes, as well as for extending the knowledge about changes in lipid metabolism in response to alteration in gene and protein expressions.

Subsequently, the majority of works dealing with the use of REIMS in the clinical field have focused not only on the real‐time prediction of tumors but also on their metabolic fingerprints. The study of the metabolome can be very helpful for the determination of biomarkers to be used in the early diagnosis of cancer, as well as to drive personalized therapeutic approaches [49, 50], involving both pharmaceutical therapy and the use of surgery.

In the latter case, the intraoperative definition of tumor margins is crucial in determining patient outcomes, ensuring progression‐free survival, and avoiding post‐operative deficits. Within this context, most clinical REIMS applications regard the analysis of breast [1, 9, 42, 44, 51], gynecological [11, 13, 43, 52], skin [8, 10, 14, 51], and gastrointestinal tumors [12, 40, 41, 47, 53, 54, 55, 56], which are quite common and frequently subjected to surgical resection. On the other hand, the building of a large database represents a challenge for brain tumors, due to the special attention paid to the preservation of healthy tissue during surgery. Consequently, only a few works reported the analysis of brain tumors [57, 58]. Ma et al. used the lipidomic profile to distinguish glioblastoma, which is the most common and aggressive tumor occurring in the central nervous system, from control normal samples [57]. The high REIMS accuracy for glioblastoma classification was further demonstrated by Van Hese et al., who indeed analyzed different types and grades of gliomas [58]. An overall correctness score of 87.90% was achieved due to a certain number of failures in the identification of low‐grade gliomas. Such a result was ascribed to the slow evolution of this kind of tumor so that only subtle differences exist between them and the surrounding healthy tissues. The authors pointed out that the limitation of this study was represented by the poor sample availability and concluded that a larger number of samples would be necessary to increase the accuracy of the method, thus ensuring the correct identification of margins and maximizing tumor resection.

4.1.2. Metabolomic analysis of biological fluids

Beyond the analysis of cancer tissues, REIMS was successfully employed for metabolomics analysis of biological samples such as feces and saliva, for the diagnosis of specific diseases [49, 50]. More in detail, such research highlighted the complementarity between REIMS and liquid chromatography‐MS (LC‐MS) (or ultra‐high‐performance LC [UHPLC‐MS/MS]) approaches, being the first untargeted method able to provide a rapid fingerprinting of the sample and the latter a targeted profiling technique capable to reveal information on specific metabolites already known to be involved in a certain pathological condition. Figure 4 clearly summarizes the comprehensive metabolomics approach proposed by the authors for precision medicine strategies, which aimed at the identification of biomarkers and the definition of therapy tailored toward each individual depending on the metabolic fingerprint. Based on the scheme displayed in Figure 4, is understandable that REIMS can be conveniently used for a rapid discrimination of samples depending on qualitative and semi‐quantitative data (ratio between the intensity of different ions). On the other hand, LC‐MS data provide more reliable identification of metabolites exploiting the combination of mass accuracy data and retention behavior, as well as robust quantitative results. Then, such quali‐quantitive information could be reasonably used for in‐depth biological interpretations, viz. for the elucidation of a biological pathway involved in the disease progression. The authors of these papers also pinpointed that REIMS offers unique opportunities for the early diagnosis of the disease and therapy selection, as well as for the acquisition of information on the pathophysiological status during consultation (when the patient is still on‐site). However, identification based solely on accurate mass measurements poses limits for the discrimination of isobaric compounds and MS/MS experiments could not be sufficient to achieve a unique identification. Furthermore, the bulk analysis of samples without dedicated sample preparation procedures and/or chromatographic separation may typically lead to matrix effects, which also hinder an accurate quantification. Therefore, REIMS may be used first to obtain a full metabolome profile, while absolute quantitation and high‐confidence identification of “suspected” metabolites may be affordable through LC‐MS analyses. In this regard, it was also pointed out that the use of REIMS as an untargeted fingerprinting approach was intended to cover an increased number of chemical classes compared to currently available salivary profiling strategies carried out by more conventional LC‐MS methods.

FIGURE 4.

Laser desorption with rapid evaporative ionization mass spectrometry (LA‐REIMS) and ultra‐high‐performance liquid chromatography‐high‐resolution MS (UHPLC‐HRMS) metabolomics approaches as complementary strategies delivering rapid discriminative fingerprinting and in‐depth metabolite characterization for biological interpretation, reproduced with permission [49].

Cameron and colleagues also pinpointed such greater coverage of the metabolome following REIMS analysis of human feces [59]. In that work, a total of 371 metabolites was tentatively identified in a single analysis by exploiting mass accuracy data and by matching against the Human Metabolome Database (www.hmdb.ca) and/or LIPID MAPS database (www.lipidmaps.org), both freely online available. Such metabolites belonged to a wide range of chemical classes, namely lipids, amino acids, and peptides. Finally, the fatty acid composition of high‐intensity phospholipids was confirmed through MS/MS experiments. In this specific application, the authors detected many bacterial metabolites, including lipids containing odd‐chain fatty acids, bile acids derived from bacterial metabolism, and other small metabolites that could serve as biomarkers for the identification of clinically relevant microorganisms, composing or infecting the gut microbiome [59]. This and other applications were reported in Table 3 as related to both clinical and microbiological fields since regards the discrimination of species and sub‐species of pathogens through the investigation of their metabolic phenotypes. Another interesting example dealing with the elucidation of bacterial metabolites involved in the development and progression of infection was described by Bardin et al. [60], who utilized REIMS to perform metabolic phenotyping and strain characterization of Pseudomonas aeruginosa derived from patients with cystic fibrosis. The authors demonstrated the potential of REIMS methodology for understanding strain variations, which greatly impact disease progression and treatment. The research highlighted the clinical significance of this kind of study, based on the monitoring of the evolution of bacterial strains in cystic fibrosis patients to provide insights into treatment strategies and infection control.

More recently, a multi‐platform approach consisting of both REIMS and LC‐MS/MS analytical techniques was adopted by Yau et al. for the differentiation of damaged skin and healthy control tissues [61]. Specifically, the authors stated that the low amount of available tissue was only sufficient for direct‐MS analysis, while MS/MS acquisitions were not possible, thus hampering the univocal identification of selected compounds (biomarkers). Hence, targeted LC‐MS/MS analyses were used complementarily to REIMS acquisition, in order to achieve a reliable identification of lipid species, selected as potential biomarkers of excised skin. However, the unique characteristics of the REIMS method coupled with the iknife sampling device allowed the generation of real‐time MS spectra in surgery and the detection of spectral differences through chemometric data analysis, despite single molecular species were not identified.

It is possible to conclude that the use of complementary platforms can represent a powerful tool in metabolomics investigations, especially in clinical research where a rapid diagnosis achievable by REIMS is necessary to address early therapeutic actions, while an accurate determination of selected biomarkers through LC‐MS/MS is also useful to predict possible evolutions of a given pathology (e.g., the stage of cancer), thus making possible the implementation of preventive treatments.

4.2. Food analysis

The use of sophisticated and expensive technology in the food application field is justified by the growing awareness of the importance of healthy and sustainable nutrition. This means providing consumers with quality products that are free from risks for both human and ecosystem health.

Within this context, the REIMS technique has been successfully exploited to both guarantee food quality and ensure food safety. In the first case, it refers to the determination of the chemical composition of the food products, that is related to different factors, such as geographical origin/pedoclimatic conditions, cultivated varieties or bred species, and production technologies. In other words, the obtained metabolomic profiles have been used to confirm the authenticity of the product compared to label declaration, including the designation of origin as in the case of the protected designation of origin and protected geographical indication trademarks thus allowing the distinction between authentic and counterfeit food.

Food safety regards the total absence of risks for human health, which could arise from the incorrect handling and storage of foodstuffs, and lack of sensitive analytical controls for the determination of contaminants, pesticides, and veterinary drugs.

Finally, food security incorporates all these concepts but also includes the delicate matter of food availability and accessibility to the entire population. Thus, food security can be defined as the condition in which people have access to adequate quantities of safe and nutritious food. Such a condition strictly depends on social, economic, and institutional factors, as well as cultural and religious aspects. Surely, the preservation of biodiversity and the establishment of national and international rules aiming at the full exploitation of internal resources (preferring local products, minimizing waste, and encouraging recycling) could enhance food security, while being respectful of environmental health.

4.2.1. Preservation of food security

Starting from the safeguard of the entire ecosystem, many REIMS works dealt with the identification of animal species to monitor the meat and fish supply chains. Specifically, following the first case‐study proposed by Balog and colleagues [28] for the identification of meat specimens, Black et al. applied the same methodology to detect fish fraud [62]. In that work, the taxonomic specificity was confirmed through the building of a database of five different white fish species for which a total of almost 500 spectra were collected. The authors pointed out that fish fraud investigations normally utilize genomic profiling, which requires complex sample preparations and long analysis times, which are incompatible with the rapid management of fraudulent activities along the fast‐moving supply chain. Conversely, REIMS enabled the real‐time analysis (within seconds) of fish without sample preparation and delivered a 100% correct classification score for unknowns. In the same work, a statistical model discriminating two different fishery practices, namely line and trawl catch, was built and validated. Differences in the metabolomic profiles of line versus trawl‐caught fishes were presumably related to secondary metabolites, also referred to as stress markers. All these successful results paved the way for demonstrating the REIMS capabilities in multiple aspects of fish fraud.

Indeed, Rigano et al. built a larger spectral database of 18 Mediterranean Sea species, trying to create a robust statistical model for the unambiguous identification of each species [63]. The model, shown in Figure 5, displayed a clear clusterization according to taxonomic classification, for example, classes and order. Specifically, the chemometric analysis generated a sort of “X plot” with four well‐defined regions: (1) Cephalopoda and (2) Malacostraca classes, (3) Perciformes, and (4) Clupeiformes orders. In addition, sub‐models were built to enhance the differentiation of very similar species, belonging to the same order or family, or even discriminate species according to their age or fishing season. In this regard, the preservation of juvenile species or the reproductive period of some endangered species is fundamental to protecting the marine ecosystem. Figure 5B displayed the sub‐model built to discriminate species of the Carangidae family, including Seriola dumerili at two different maturation stages, known as juvenile and adult greater amberjack.

FIGURE 5.

(A) Statistical model for the differentiation of 18 fish species typical of the Mediterranean Sea according to classes and orders, (B) statistical sub‐model to discriminate four species belonging to the Carangidae family, and (C) statistical sub‐model of swordfish caught in two different seasons (autumn and summer), reproduced with permission [63].

Finally, in the same study, MS spectra of swordfish caught in two different seasonality were compared and collected to create a sub‐model able to identify the fishing season. The results reported in Figure 5C are quite promising to establish a method useful for detecting illegal fishing activities in prohibited periods, which are recommended to avoid overfishing and protect juvenile species.

Other valuable works can be mentioned in the context of fish speciation. Song et al. generated a statistical model for the authentication of Thunnus species, the most important commercial fish species consumed and traded worldwide [64]. Among them, bluefin tuna is the most valuable species, whose annual output is quite limited to hamper the risk of extinction [65]. At the same time, fraudulent practices at its expense, such as mislabelling and substitution, have been detected by the institutional authorities of the Food and Agriculture Organization of the United Nations and the European Commission through the Rapid Alert System for Food and Feed (https://ec.europa.eu/food/safety/rasff_en). Song and co‐authors developed a chemometric model for the authentication of four different tuna species (bluefin tuna, bigeye tuna, yellowfin tuna, and albacore tuna) also providing semi‐quantitative results about their fatty acyl and phospholipid composition in order to justify their differentiation (Figure 6) [64]. Bluefin tuna proved to possess the highest percentage of omega‐3 fatty acids, while saturated compounds were most concentrated in albacore tuna, confirming its lowest economic value.

FIGURE 6.

(A) Statistical model for the authentication of four different Thunnus species (bluefin tuna [BFT], bigeye tuna [BET], yellowfin tuna [YFT], and albacore tuna [AT]); (B) relative quantitative results about fatty acids in the four species, reproduced with permission [64].

Shen and co‐workers applied the same instrumental setup to discriminate basa catfish and sole fish, which have similar morphological features but different commercial values [66]. Even in this work, besides achieving a 100% correct classification score, the aim of the authors was to identify differences in the lipid profile of the investigated fishes. For this reason, a relative quantification based on the signal abundances in the low (200–400 m/z) and high (600–1000 m/z) mass regions was carried out to compare the fatty acyl and phospholipid composition, respectively. The obtained results revealed a higher level of highly unsaturated fatty acids (i.e., eicosapentaenoic acid [EPA, C20:5], docosahexaenoic acid [DHA, C22:6], and docosapentaenoic acid [DPA, C22:5]) in sole fish, while basa fish resulted richest in oleic and linoleic acid.

A similar strategy was adopted by the same authors to distinguish between salmon and rainbow trout, being aware of the beneficial nutritional properties of salmon and the controversial nature of rainbow trout, linked to potential contamination by parasites [67]. It is evident as in this case the food security area is strictly interconnected with food safety implications, thus highlighting the role of a sensitive, fast, and accurate analytical technology to fight fraudulent activities.

The last application related to fish speciation and aiming to contrast overfishing consists of a large model built by De Graeve et al. to differentiate between 17 commercially valuable fish species, as a tool to detect and prevent fish fraud at the industrial level [68].

In the future, the validated models built in the fish sector by many research groups could be combined to generate a large database of fish species to be used internationally for consumer and ecosystem protection, as well as to preserve industrial interests.

In all the above‐mentioned work, the authors concluded that the iknife/REIMS technology is an accurate tool for the real‐time and single‐step (without sample preparation) differentiation of marine species and could be implemented to preserve food safety and quality in the entire food supply chain.

4.2.2. Ensuring food safety and quality

With regard to food safety, the fish supply chain needs to be urgently monitored. Indeed, the conservation of seafood products is extremely important to prevent oxidation processes and inhibit enzymes naturally present in fish specimens and responsible for the generation of toxic molecules. In two valuable works, iknife was employed to monitor the storage conditions through the elucidation of lipid changes [69, 70]. Specifically, such research consisted of the quality assessment of sea cucumber during its storage in the dried form [69] and in the tuna meat characterization during the wet‐aging process [70]. These kinds of studies could be useful in order to define the shelf‐life of the product based on the detection of lipid oxidation derivatives. The iknife methodology was capable of providing a robust and accurate relative quantification of phospholipids and fatty acids, thus revealing the potential causes of lipid alteration during storage, mainly consisting of hydrolysis and oxidation processes.

Food treatment processes such as cooking techniques can also affect food quality because of the production of undesired and often unhealthy substances. Within this context, iknife was successfully exploited for the real‐time evaluation of lipid oxidation of prawn during air frying [71] and bigeye tuna under three different cooking processes [72]. The rapidity of such a direct‐MS approach was especially pinpointed since the real‐time output made the employed setup a front‐line fast detection tool to guarantee the quality of the cooked products. The results of these works showed that, due to the fact that polyunsaturated fatty acids are major targets of oxidation reactions, the percentage of saturated fatty acids increases significantly during cooking, especially at temperatures above 180°C. Among cooking methods, the air‐frying process provoked a slower decrease of unsaturated phospholipids and fatty acids, contributing to an increase in the awareness of the influence of the cooking process on the nutritional value of foodstuffs.

Furthermore, as a branch of metabolomics, lipidomics has opened new insights into the relationship between food and nutrition. In this context, REIMS was employed for the investigation of lipid metabolism following in vitro multiple‐stage digestion of phospholipids extracted from a grass carp, selected for its economic importance in China [73]. REIMS analyses confirmed the increase of the hydrolysis rate of phospholipids during in vitro digestion as the degree of unsaturation of the acyl chains increased. Later on, the same authors applied the same approach to demonstrate the inhibitory effect of dietary fiber on the digestion of fish phospholipids [74]. The spectral comparison revealed that the phospholipid signals were drastically reduced in the digested unfortified sample, while the digestion rate slowed down for the fortified samples. The findings of this study could contribute to the design of novel fortified foods for weight control diets.

Similar works were performed on meat, aiming at the assessment of quality attributes through the reliable identification of breed [28, 75, 76], production background (conventional vs organic; different feeding) [76, 77, 78], and cut type (neck, rump, shin, etc.) [78], as well as the sensitive detection of offals illegally added to more precious meat products [79] or other adulterants added in meat to falsify the nitrogen content or to improve the texture [80]. The detectability of the method was evaluated in some of those works by mixing authentic meat samples with different percentages of adulterants. For instance, Black et al. detected beef adulteration with beef brain, heart, kidney, large intestine, and liver tissues at levels as low as 1%, demonstrating that a mixed sample was recognized as an outlier against a model built by collecting spectra of pure beef or pure offal tissues [79]. A different strategy was adopted by Kosek et al. to achieve the same goal [80]. The authors built different two‐classes models in order to compare pure and adulterated meat at 0.5%–2.5% levels, demonstrating that REIMS can detect up to 2.5% of protein‐based additives and 1% of carrageenan. Such findings were very promising to guarantee high‐quality products and increase consumer satisfaction.

The latter also depends on sensory properties determining the palatability of meat products, related to the combination of tenderness, juiciness, and flavor. Hernandez‐Sintharakao et al. proposed three statistical models to distinguish tender from tough meat, dry from juicy meat, and positive from off‐flavors, obtaining prediction rates in the range of 74%–95%, depending on the employed statistical approach [81]. The relatively low accuracy was ascribed to the inability of REIMS to detect intact proteins, which affect these physical attributes. The authors stated that a large sample size and a more balanced class size could enhance the robustness of the model, being confident that REIMS is able to detect metabolites known as possible indicators of meat tenderness, such as malic acid, glucose, glucose‐6‐phosphate and compounds related to intramuscular fat [82].

Another issue in the meat industry is related to the use of banned or controversial veterinary drugs. Guitton et al. explored the use of REIMS in combination with chemometrics for high‐throughput screening of growth promoters in meat‐producing animals [83]. A correctness score higher than 95% was achieved for the classification of meat coming from treated and controlled animals. Lipid molecules, belonging to different classes, were identified as discriminant features through complementary statistical methods and software, such as Progenesis QI, also employed for the tentative identification of discriminant lipids.

Despite beef fraud being mostly carried out, interesting applications were developed on ovine meat. Wang et al. used REIMS to discriminate lamb from mutton meat, both derived from sheep species, but ad different ages so that lamb (<12 months) has a higher price on the market [84]. From an analytical point of view, an untargeted lipidomics UHPLC‐MS/MS approach combined with multivariate analysis was used in parallel to the REIMS workflow in order to achieve a reliable identification of authenticity markers. Thus, REIMS delivered near‐instantaneous results without sample preparation or chromatographic separation and is recommended for fast suspect screening analysis, while the UHPLC‐MS/MS method provided more detailed information at molecular levels.

Similarly to the seafood sector, aspects related to storage conditions, aging processes, and the impact of cooking processes on the chemical composition were investigated. He et al. highlighted the advantages of REIMS over enzymatic, spectroscopic, sensory, chromatographic, and DNA‐based methodologies for the evaluation of freshness degree, given that REIMS is capable of identifying the differential compounds/metabolites ions responsible for the discrimination, and associable to quality and safety attributes [85].

The metabolic fingerprints were used by Zhang et al. [86, 87, 88] to distinguish between different aging processes applied to lamb [87, 88] and beef [86] meats, based on differences in the involved biochemical mechanisms, which included lipid oxidation, proteolysis and dehydration. Differently from most REIMS applications in which lipids were the main identified metabolites, amino acids, peptides, cyclopeptides, volatiles, amines, and oxidized fatty acids were identified among discriminant features. The results suggested a comprehensive understanding of the biochemical transformations affecting lipid and protein components, as well as the development of flavor‐associated volatile compounds. These studies contribute to the ongoing efforts to optimize dry aging processes and enhance the overall quality and flavor of meat, in order to meet consumer preferences and culinary expectations [86, 87, 88]. Chromatographic analyses of free amino acids were also carried out, whereas, for many of the tentatively identified metabolites, the authors recommend conventional LC‐MS‐based lipidomics and metabolomics strategies to validate findings based on REIMS measurements [87].

Vegetable matrices were also selected for REIMS applications. As emphasized by Birse and Elliot in their recent review focused on food applications [89], the first works dealing with “non‐fish or meat” samples appeared in 2019. The works referred to the identification of the botanical origin of honey [20] and the geographical provenance of pistachio [31]. Moreover, a not‐fatty matrix was analyzed for the first time, being honey mainly composed of sugars. In this specific study, Wang and colleagues simulated the adulteration of high‐value honey with different percentages of low‐value honey or less expensive sweeteners and pointed out the high specificity of REIMS combined with chemometrics in discriminating up to 5% of adulteration levels [20].

Finally, the timeline reported by Birse and Elliot [89] indicated the year 2021 as a key‐date for the implementation of REIMS methodology, due to the publication of the first application on liquid food samples. Particularly, Mangraviti et al. developed a strategy for the valorization of Italian extra virgin olive oils by using the electrosurgical knife as a sampling device [90]. The authors applied a minimal sample preparation procedure to make their matrix a solid and conductive sample. The oil was mixed with a saline solution and the mixture was frozen to obtain ice cubes easy to cut.

As already mentioned in the introduction section, the coupling of REIMS with an electric soldering iron as a contact heating probe facilitated the sampling of liquid matrices, as demonstrated for the analysis of milk [21] and virgin olive oil [23].

All these applications demonstrated the potential of REIMS for the analysis of food products, independently from their physical state and nutritional composition.

4.3. Others

The elevate versatility of REIMS to ionize both polar and non‐polar compounds, such as lipids, phenols, sugars, and amino acids, was exploited in a wide range of application fields, providing a rapid and holistic profile of complex samples.

As already mentioned, one of the first application fields explored by REIMS was microbiological, aimed at the rapid identification of bacteria [6, 15, 16] and fungi [16, 19], based on their unique metabolic fingerprints. This can aid in the quick and accurate classification of bacterial isolates, streamlining the identification process in microbiology, with relevant implications in clinical diagnostics.

Similarly, REIMS found applications in entomology, offering a rapid and direct analytical approach to studying the chemical composition of insects [91, 92, 93]. REIMS can be used for rapid species identification by analyzing the unique chemical fingerprints of insect cuticles or other body parts. This aids in taxonomic studies and biodiversity assessments.

Other uses of REIMS technology fall into the botany field, where Arena et al. investigated the profile of two different botanical species such as the fruit of Kigelia Africana and the bark of Sterculia Setigera [29, 94], both largely used in the traditional African medicine. In this case, REIMS allowed us to identify also some polar molecules, such as phenolic compounds including flavonoids, phenolic acids, and phenol glucosides.

Finally, some applications are in the fields of forensics [95], pharmaceutics [96] and textiles [25]. Abu‐Rabie and colleagues collected both MS and MS/MS spectra at two different collision energies (CEs) for the building of three statistical models, demonstrating that MS/MS experiments at high CE enhance the accuracy of the prediction model [96]. As an exercise, the author built a combined model, from which was possible to notice that a clearer clusterization is obtained through MS/MS experiments at 80 eV CE. Figure 7 shows the hybrid plot deriving from supervised (principal component analysis and linear discriminant analysis) multivariate analysis carried out on spectra acquired under the three MS operating modes. MS/MS experiments were carried out by Van Hese et al. to confirm the identity of drugs responsible for the discrimination of human muscle tissues soaked in a solution of different drugs (methadone, morphine, cocaine, and diazepam) at different concentrations [95]. The method was partially validated through the comparison with a “gold standard” based on solid‐phase extraction (SPE) of opioid and benzodiazepine molecules from muscle tissues followed by LC‐MS/MS analyses. Specifically, the detection limit of REIMS was defined as the lowest concentration still correctly classified by the classification model and previously determined through conventional SPE‐LC‐MS/MS. Moreover, the ability to distinguish between different concentration levels sets out the basis for future implementation of quantitative methods.

FIGURE 7.

Principal component analysis and linear discriminant analysis (PCA‐LDA) statistical model of rat liver samples built by using mass spectrometry (MS), low collision energy tandem MS (MS/MS), and high collision energy MS/MS modes, reproduced with permission [96].

5. CONCLUSIONS AND FUTURE PERSPECTIVES

This review underscored the versatility and transformative impact of REIMS across a multitude of scientific disciplines. This innovative technology has revolutionized the approaching way of analytical challenges, offering a rapid and direct means of interrogation for complex samples. This has been possible thanks to the integration of a machine learning approach and the application of chemometrics to the REIMS data sets, for uncovering deeper insights and correlations within complex datasets.

An overview of REIMS's potentiality in several application fields was provided, with a special focus on its role in surgery and in the preservation of the agri‐food sector. In the medical area, real‐time results can guide surgeons during excision, improving margin definition and minimizing the risk of recurrence.

In the food field, applications for the safeguarding of both food quality and food security were reported, demonstrating its applicability against food fraud, thus protecting producers and consumers, as well as ecosystems.

As the technology continues to evolve, REIMS could open new frontiers in fields such as personalized medicine, drug development, and biotechnology.

Furthermore, the instrumental setup could be also modified to overcome the main limitations of such an ambient MS technique. For instance, the use of an ion mobility analyzer could increase the identification power and enable the discrimination of isomers and isobars, at the expense of analysis cost and the required level of expertise. On the other hand, less sophisticated and compact MS analyzers (single quadrupole) could be coupled with REIMS sources to make the technique cost‐effective and user‐friendly.

Finally, the complementarity between targeted and untargeted approaches is expected to be more and more explored in the near future to maximize the number of qualitative and quantitative data in metabolomics. Specifically, REIMS should be employed as an untargeted method for real‐time data generation, sample identification, and biomarker identification, while LC‐MS could be exploited as a confirmation technique to obtain accurate quantification of biomarkers.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Cafarella C, Mangraviti D, Rigano F, Dugo P, Mondello L. Rapid evaporative ionization mass spectrometry: A survey through 15 years of applications. J Sep Sci. 2024;47:2400155. 10.1002/jssc.202400155

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.