Desorption electrospray ionization mass spectrometry imaging (DESI-MSI) in disease diagnosis: an overview

Abstract

Tissue analysis, which is essential to histology and is considered the benchmark for the diagnosis and prognosis of many illnesses, including cancer, is significant. During surgery, the surgical margin of the tumor is assessed using the labor-intensive, challenging, and commonly subjective technique known as frozen section histopathology. In the biopsy section, large numbers of molecules can now be visualized at once (ion images) following recent developments in [MSI] mass spectrometry imaging under atmospheric conditions. This is vastly superior to and different from the single optical tissue image processing used in traditional histopathology. This review article will focus on the advancement of desorption electrospray ionization mass spectrometry imaging [DESI-MSI] technique, which is label-free and requires little to no sample preparation. Since the proportion of molecular species in normal and abnormal tissues is different, DESI-MSI can capture ion images of the distributions of lipids and metabolites on biopsy sections, which can provide rich diagnostic information. This is not a systematic review but a summary of well-known, cutting-edge and recent DESI-MSI applications in cancer research between 2018 and 2023.

---

Introduction

To encourage clinical decision-making and enhance human healthcare, accurate and expeditious performance of medical diagnostic tests is highly desired. There has been a steady advancement of new analytical techniques and methods, which has had a significant impact on the early diagnosis of diseases and the development of effective therapeutic regimens. Although most in vitro diagnostics use immunoassays to identify the target analyte, some drawbacks of this approach (cross-reactivity, low sensitivity, limited dynamic range, and frequently expensive and time-consuming processing) force us to focus on better solutions with alternative strategies, which should be extremely sensitive, specific, and offer high throughput. Even though incredibly useful information is obtained, tissue discrimination with histopathology is frequently challenging and unreliable because of frozen specimens, replicates, and variability in tissue morphology. The histopathology-based categorization of cancer from benign tissue or determination of the grade of cancer, despite being the gold standard, severely limits active monitoring. By leaving tumor cells at the edge of the specimen that was removed, a questionable biopsy report commonly leads to a failed cancer surgery; this has been associated with a variety of cancer types with increased local relapses and reduced overall survival rates.

There has been a steady advancement in medical imaging techniques with the introduction of new modalities and sophistication, which has had a significant impact on image-guided diagnosis and therapy.¹ Since the invention of MRI, PET, CT, and US, patient care has advanced at a never-before-seen rate. The pathological characteristics (e.g., molecular or cellular fingerprints) related to the disease subtype and/or aggressiveness cannot be directly determined using any of these techniques on tissue specimens at a sufficiently high resolution.²

The development of several cutting-edge analytical methods for tissue discrimination in disease diagnosis has taken place over the years.^3–11 Due to a number of issues, including inconvenience, sophistication, limited responsiveness and availability, and incongruence with operating space workflow, none of these methods have been used in a clinical setting.² Over the past decade, MSI has developed into a significant instrument and is now beginning to show prospects for providing new information in many disciplines. This is due to its unique ability to acquire molecularly specific images and to provide multiplexed data without the need for labeling or staining.² In MSI, the chemical composition of molecules on a surface is examined in relation to their spatial distribution. MSI includes ambient ionization. MSI, where the samples are analyzed in their natural state while ionization occurs at atmospheric pressure. In ambient ionization, ions rather than the entire specimen are presented into the instrument, ionization happens outside of the mass spectrometer [MS], and the surface is analyzed with little to no pre-treatment.¹² A typical MSI workflow is shown in Fig. 1.

Fig. 1 MSI workflow. A thin microscopic section of the frozen tissue is created, and it is then exposed to a laser beam or a stream of charged microdroplets to ionize the molecular species present there. A mass spectrometer (MS) is then used to detect the ionic species. By scanning the section in two dimensions (2D) and desorbing and detecting charged biomolecules from the tissue surface spot by spot, an imaging experiment is carried out. As a result, a sizable collection of mass spectral data is stored alongside the spots' spatial coordinates (x_i, y_j) where the measurements were made. Each detected species' intensities can be visualized as ion images (shown in various colours), which allow for a two-dimensional (2D) mapping of their distribution on the tissue surface.

A method for visualizing mostly small molecules (metabolites, lipids, drugs, etc.) in the resected¹² or unresected¹³ tissues of a living system is called desorption electrospray ionization-mass spectrometry imaging, or DESI-MSI. DESI was developed in Professor Graham Cooks' lab in 2004, and the team reported its initial application in a tissue imaging experiment in 2006.^12–14 Since it can operate in ambient conditions (room temperature and atmospheric pressure) and requires little to no sample preparation, DESI in MSI has an advantage over all its counterpart ionization techniques (such as MALDI, SIMS, etc.). In contrast to MALDI or SIMS, the technique is also simple to use because it only needs an inexpensive solvent spray and doesn't need a vacuum chamber or an enclosure for ionization. The DESI probe can instantly capture molecular fingerprints from live organs, skin, tissues, biological fluids, and volatiles, which holds enormous promise for its performance in point-of-care diagnostics.¹⁵

The DESI-MSI (Fig. 2) procedure is usually carried out on a tissue section with a thickness of about 10–15 μm. This tissue section is then bombarded with a stream of high-speed charged microdroplets produced by electrospraying a solvent (such as water, methanol, acetonitrile, dimethylformamide, etc.) at high voltage while using nitrogen as the nebulizing gas.^16,17 The droplet solvent wets the tissue surface as a result, dissolving the biochemical species that are present in the tissue (mostly metabolites and lipids). This liquid film splashes when primary droplets later arrive, releasing secondary microdroplets that dissolve the biochemical species and then transform into gaseous ions for mass spectrometric detection. Like the electrospray ionization process, the desolvation mechanism for producing analyte ions in the gas phase involves repeated solvent evaporation and Coulomb fission of the secondary microdroplets that contain the analyte species.¹⁶ To carry out the bioimaging experiment, the tissue surface is scanned in the x and y directions while being impinged upon by a spray of charged microdroplets. These droplets are used to intercept molecular data from the mass spectra gathered across the sample. It is possible to map the tissue's biochemical composition using the pixel-to-pixel mass spectra as a two-dimensional image. The spray spot size on the tissue determines the typical spatial resolution of DESI-MSI, which also depends on the motorized moving stage step size and the MS scan rate. This resolution is typically between 50 and 200 μm.¹⁸ The distribution of specific classes of metabolites and lipids in the tissue can be seen with the help of the DESI spray solvent of choice.¹⁹ The metabolite/lipid ion map is superimposed on the histopathological H&E image of the specimen section to inform the DESI-MSI of the metabolic signature of normal and unhealthy tissue. Using tandem MS, high mass exactness, and isotopic distribution, individual species that appeared in the mass spectra are identified.¹⁵ Although DESI-MSI tissue diagnosis does not require ion characterization of the individual ions (metabolites or lipids), ion characterization can identify significant biomolecules and their relationship to disease biochemistry.

Fig. 2 Schematic illustration of DESI source.

This review concentrates on empirical studies reported in peer-reviewed papers in recent years (2018–2023), with an emphasis on the newest developments and applications of DESI-MSI in disease diagnosis. This is an overview of popular and cutting-edge applications rather than a systematic review. We created criteria that would enable us to conduct our research and sort through the various studies published. To be considered for inclusion, the article must fulfill the following criteria: (a) it needs to be published in a journal with peer review; (b) it must be written in English; (c) the research must be empirical; and (e) it must discuss either an advancement in or application of DESI-MSI in disease diagnosis. The review included a total of 22 studies. Table 1 lists key empirical studies, cancer types, metabolites, metabolic pathways, clinical relevance, and MS set up.

Table 1 DESI-MSI studies on disease diagnosis^a

Author	Diagnosis	Key metabolites	Representative ions (m/z)	Metabolic pathway	Clinical relevance	MS
a NR – not reported in the study; Q-TOF: quadrupole time of flight; FT-Fourier transform; TOF-time of flight.
Margulis et al. (2018)^87	Basal cell carcinoma	Fatty acids (FA), fatty acid clusters & glycerophospholipids	Arachidonic acid (303.3), FA dimers (oleic acid + palmitic acid) (537.4), glycerophosphoglycerol (747.6)	Krebs cycle	Early diagnosis of cancer	Linear ion trap
Morse et. al. (2019)^118	Prostrate cancer	Fatty acids (FA) & phospholipids	LysoPE (16 : 0) (452.7), LysoPE (18 : 1) (478.2), PI (37 : 6) (867.5), PCh (O-40 : 2), PC (P-40 : 1) (826.7), FA22 : 4 (331.2)	Krebs cycle	Prostate cancer metabolism and surgical guidance	Q-TOF
Tamura et al. (2019)^123	Renal cell carcinoma	Lipids and fatty acids (FA)	Azelaic acid (187.1), [FA(16 : 1)] (253.2) (palmitoleic acid), [FA(18 : 2)] (279.2) (linoleic acid), [FA(18 : 1)] (281.2) (oleic acid)	Warburg effect	Therapeutic strategy for targeting cancer cell metabolism	Q-TOF
Santoro et al. (2020)^124	Breast cancer	Lipids	Taurine (124.0), uric acid (167.0), glycerophospholipids (PE (P-16 : 0/22 : 6, 746.5), PS (38 : 3), m/z 812.5)	De novo lipogenesis	Breast cancer pathogenesis	Quadru-pole orbitrap
Ghadi et al. (2020)^73	Esophageal adeno carcinoma	Glycerol phospholipids (PG)	PG 36 : 4 (769.5), PG 38 : 6 (793.5), PG 40 : 8 (817.5), PI 34 : 1 (835.5)	De novo lipogenesis	Early diagnosis of cancer	FT-orbitrap
Bensussan et al. (2020)^96	Adenocarcinoma (ADC) and squamous cell carcinoma (SCC)	Fatty acids (FA) & Glycerol Phospholipids (PG)	FA (20 : 4) (303.2), PG (36 : 2) (773.5), PI (34 : 1) (835.5), PS (36 : 1) (885.5)	NR	Classification of ADC and SCC subtypes with tissues and FNA samples	LTQ-orbitrap
Zhang et al. (2020)^125	Renal oncocytoma vs. renal cell carcinoma	Free fatty acids (FFA) and monoradylglycerolipids (MG)	Ascorbic acid (175.0), hexose (m/z 215.0), FA(18 : 1) (281.2), FA (20 : 4) (303.2), cardiolipin (CL) (723.4 & 737.4)	NR	Discrimination of renal tumors	LTQ-orbitrap
Vijayalakshmi et al. (2020)^126	Renal cell carcinoma	Small metabolites, fatty acids and lipids	N-Acetyl glutamate (187.0), 2-hydroxy butyrate, (103.0), creatinine (112.9)	Metabolic alterations	Surgical margin assessment	LTQ-orbitrap
Fala et al. (2021)^127	Murine lymphoma	Metabolites	Pyruvate and lactate	Pyruvate metabolism	Investigate pyruvate delivery and lactate labeling	Quadrupole-orbitrap
Yang et al. (2021)^28	Oral cancer	Amino acids, carbohydrates, glycerolipids, glycerophospholipids, sphingolipids	Putrescine (89.1), betaine (156.0), phosphocholine (259.9), linoleic acid (317.1), hypoxanthine (175.0)	Arginine/proline metabolism, histidine metabolism	Satisfy the need for point-of-care testing	LTQ-orbitrap
Song et al. (2022)^128	TNBC	Metabolites	Arginine (175.1), lysine (147.1), N,N-dimethyl arginine (203.1), N,N,N-trimethyl lysine (189.1)	NR	Rapid diagnosis of TNBC	LTQ-orbitrap
Strittmatter et al. (2022)^73	Pancreatic cancer	Gemcitabine	Ceralasertib (451.3), 2′,2′-difluorodeoxycytidine (302.0), 2′,2′-difluorodeoxyuridine (299.0)	Gemcitabine metabolism	Visualizing drug metabolites	Quadrupole- orbitrap
Kauffmann et al. (2022)^89	Colorectal cancer (CRC)	Fatty acids	Phosphatidylserines (810.5, 812.535, 766.536), phosphatidylethanolamines (PE) (726.5)	De novo lipogenesis	CRC diagnosis and prognosis	TOF
Vaysse et al. (2022)^36	Oral cavity cancer	Metabolites	Ether-phosphatidylethanolamine PE(O-16 : 1/18 : 2) (698.5)	OSCC keratinization	In vivo tissue recognition	Q-TOF
Mondal et al. (2023)^111	Breast cancer	Lipids, fatty acids	Phosphatidylcholine (PC), phosphatidylethanolamine (PE), ceramide (Cer), sphingomyelin (SM), diacylglycerol (DG)	De novo lipogenesis	Therapeutic and diagnostic developments	LTQ-orbitrap
Zhan et al. (2023)^129	Brain tumor	Small metabolites	Glutamine, hexcer (d42 : 2)	Endogenous metabolic pathways	Tumor research	LTQ-orbitrap
Aramaki et al. (2023)^110	Luminal breast cancer	Lipids and fatty acids	Phosphatidylcholine (PC), triglycerides (TG), phosphatidylethanolamine, sphingomyelin, and ceramide	Cancer metabolism	Describe heterogeneity of cancer	Q-TOF
Seubnooch et al. (2023)^103	Liver zonation	Fatty acids (FA), phospholipids, triacylglycerols, diacylglycerols, ceramides, and sphingolipids	Phosphatidylinositols [PI(36 : 2), PI(36 : 3), PI(36 : 4), PI(38 : 5), and PI(40 : 6)]	De novo triacylglycerol biosynthesis	Lipid homeostasis	Q-TOF

---

DESI-MSI in disease diagnosis

Globally, the incidence rate of oral cancer is 3.9 per 100 000 people when adjusted for age.²⁰ Alcohol consumption, tobacco use, and the rising incidence of oropharyngeal cancer (more specifically, tonsil and base of tongue cancer) linked to human papillomavirus (HPV) infection are all common risk factors for developing oral cancer.²⁰ Patients with oral cancer are typically identified when their invasive squamous cell carcinoma (SCC) has reached an advanced stage.²¹ Roughly half of patients survive five years after initial disease diagnosis because of delayed disease presentation and a lack of appropriate screening techniques to detect early oral cancer.²² Despite improvements in radiotherapy, chemotherapy, and surgery, the mortality rate from oral cancer at five years has not changed drastically in more than 50 years.²¹ Local recurrence of the patient's primary tumor is the most frequent reason for treatment failure and death in oral cancer patients.²³ Oral cancer can be found in the oral cavity, where the oral surface epithelium can be easily examined visually and physically. A definitive determination of the presence of cancer or precancerous lesions does not appear to be possible with the current methods for oral cavity cancer screening.²⁴ Currently, a biopsy and pathological examination of the removed tissue serve as the only reliable methods of detection.^25,26 However, even in this situation, the pathologist's evaluation is still inherently subjective despite being guided by several diagnostic criteria.

Oral SCC [OSCC] has been associated with lipid-based biochemical changes, including overexpression of FAS.²⁷ FAS is essential for producing long chain fatty acids from acetyl- and malonyl-CoA in the body.²⁷ Endogenous fatty acid biosynthesis is active in several human carcinomas, including OSCC, which expresses high levels of FAS constitutively. FAS levels in OSCC are higher than in the nearby morphologically normal epithelium.²⁷

To create the testing and validation datasets, all pixels within the diagnosed tumor and mucosal margin areas were obtained by projecting MS data with hematoxylin–eosin staining from 18 enrolled OSCC participants. It was determined how well the test worked using a Lasso regression model.²⁸ Using leave-one-out validation, the model distinguished tumors from normal regions with an accuracy of 88.6% (Fig. 3). A group of lipid ions (n = 14) that significantly decreased from tumor to healthy tissues were given the highest weight coefficients in the model in order to ascertain the margin status and safe surgical excision distance of OSCC (Yang et al., 2021).²⁸ The safe surgical distance of the OSCC was determined using the 14-lipid ion molecular diagnostic model developed for clinical reference. Tumors, positive and negative margins could all be predicted with an overall exactness of 92.6%. The diagnostic model's spatial segmentation results not only distinguished the various surgical margin statuses but also distinguished between the tumor and normal tissue with great clarity (Yang et al., 2021).²⁸

Fig. 3 Different surgical margin status determined by 14 characteristic lipid ions' image OSCC tissue sections and their merged image. (A) H&E staining of frozen section, we redefined different margin status in the frozen section: positive margin (M1 region, 0–2 mm), close margin (M2 region, 2–10 mm), and negative margin (M3 region, >10 mm) according the shrinkage rate between fresh frozen and paraffin OSCC tissue sections. (B) The merge of the 14 characteristic lipid ion images in the case. (C–P) Separate image of the characteristic lipid ions. (Q–T) Prediction results given by the DESI-MSI for tumor, M1, M2, M3 region (n = 18 cases) compared with pathological diagnosis. Circle: the DESI-MSI prediction is consistent with the pathological results; triangle: the DESI-MSI prediction is not consistent with the pathological results. Pathological diagnosis include tumor, dysplasia, normal epithelial; DESI-MSI prediction include positive (red bright spot) and negative (non-red bright spot). Reproduced with permission from [ref. 28] @ ELSEIVER (2021).

Yang et al. (2021) used DESI-MSI, a lipidomic tier molecular diagnostic technique, to determine OSCC's safe surgical resection margin.²⁸ The morphological changes of the cells are the only changes that the histopathological analysis of the frozen section reflects; it does not reflect the underlying molecular alterations. In order to complement conventional pathological techniques, the idea of molecular margins has been introduced.^29–34 For surgeons, it is particularly critical to establish the safe margin distance during the procedure. Recent advancements in clinical research on the surgical margin of OSCC have shifted the focus to margin distance. The NCCN guidelines state that the safe margin distance for OSCC should be 5 mm, but because the postsurgical paraffin segment determines the 5 mm cut-off, real-time reconstruction is not possible. In a prior study, Yang et al. (2019) discovered that molecular markers were different at various surgical margin distances of OSCC, confirming the existence of metabolic molecular differences at various surgical margins.³⁵

In a recent study, Vaysse et al. (2022) examined the sensitivity of rapid evaporative ionization mass spectrometry (REIMS) to identify the smallest number of tumor cells detectable during oral cavity cancer surgery.³⁶ Electrocautery is a commonly used surgical technique for resecting the oral cavity because of its hemostatic function, which is necessary to cauterize vascularized tissues of the oral cavity. The byproducts of this modality include electrosurgical vapors. Recent studies have demonstrated significant promise for intraoperative tumor and healthy tissue recognition using direct analysis of these vapors by rapid evaporative ionization mass spectrometry [REIMS].^37,38 Molecules in tissue-derived vapors are ionized by REIMS to produce metabolic profiles that are almost real-time. These REIMS metabolic characteristics are reported to be tissue-specific for different cancer and normal tissue types.^39–41 In order to build libraries of tissue-specific metabolic profiles, tissue samples were cauterized ex vivo and examined by REIMS. When the surgeon is resecting the tumor, the objective of these ex vivo repositories is to convincingly categorize healthy and cancer tissues in vivo.^42,43

To support oral cavity surgery, REIMS's capacity to identify even small numbers of tumor cells is essential.^44–47 This is crucial considering the difficulties related to OSCC. This study included a total of 11 OSCC patients.³⁶ The tissue was categorized using 185 REIMS ex vivo metabolic characteristics from patients (n = 5), and the tissue categorization was compared to histopathology results using multivariate statistical analysis and cross-validation.³⁶ During four glossectomies, vapors were examined in vivo by REIMS. Six sections of the oral cavity were subjected to DESI-MSI, which was used to map tissue variability and support REIMS findings.³⁶ A novel cell-based assay made up of different cell lines (keratinocytes, myoblasts, and tumors) was used to evaluate the REIMS' sensitivity.³⁶ With 96.8% accuracy, REIMS classified soft tissues and tumors. Strong mass spectrometric signals were produced in vivo by REIMS. With 83% sensitivity and 82% specificity, REIMS was able to detect 10% of tumor cells among 90% of myoblasts.³⁶ The phosphatidylethanolamine PE(O-16 : 1/18 : 2)/cholesterol sulfate metabolic shift common to both mucosal maturation and OSCC differentiation was highlighted by DESI-MSI, as were the distinct metabolic characteristics of nerve attributes (Fig. 4). When performing oral cancer surgeries, the appraisal of tissue variability using DESI-MSI and REIMS specificity with cell compositions characterizes sensitive metabolic characteristics toward in vivo tissue recognition.³⁶

Fig. 4 Distinct nerve metabolic profiles in oral cavity tissues by DESI-MSI. (A) Histology surrounding tissue defects of needle electrode-sampling for REIMS analysis surrounded by nerve features delineated in yellow on one resected specimen. (B) Segmentation analysis discriminating nervous tissue from the rest of the imaged areas based on DESI-MS profiles. (C) Principal component analysis score plot of DESI-MS profiles (55 nerve, 54 muscle, and 53 tumor) from tissue provided by six patients on the mass range m/z 600–1000 (PC1, which explains 80.3% of the variance of the data; PC2, 7.8%). Reproduced with permission from [ref. 36] @ American Chemical Society (2022).

The most prevalent type of epilepsy, known as temporal lobe epilepsy with hippocampal sclerosis (TLE-HS), affects nearly 20% of all epilepsy cases.⁴⁸ Unprovoked recurrent focal seizures occur frequently in the temporal lobe, which causes a variety of clinical manifestations.⁴⁹ With distinct hippocampal morphological changes like neuronal loss, gliosis, dendritic alterations, and mossy fiber sprouting, a significant portion (30%) of TLE patients are intractable or drug-resistant.⁵⁰ The development and progression of various epilepsies may be attributed to a series of biochemical events, according to a recent trend in epilepsy research. The expression of excitatory amino acids rises, and this is followed by an overactivation of NMDA receptors, an increase in calcium entry into neurons, and the involvement of pro-inflammatory pathways.^49,51,52 Lipids, which make up 60% of the human brain's composition, are significant biomolecules involved in the construction of numerous biological membranes.⁵³ Phospholipids (PLs), which make up the majority of membrane bilayers, are known to play a variety of roles in cells, including modulating membrane fluidity/structure, permeability barriers, cell compartmentalization, neurotransmitter release, membrane trafficking, signal transduction, etc.^54,55 Additionally, they act as messengers in signaling pathways for downstream lipid signals involved in normal physiology and pathology.⁵⁵ Numerous studies using animal models of epilepsy have demonstrated the crucial role that PL metabolism plays in the pathological development of epilepsy.^56–58 When compared to patients with other temporal lobe space-occupying lesions, a recent study found a distinct lipid profile with the human hippocampal triglyceride level dropping in TLE-HS patients.⁵⁹ Therefore, identifying lipids and their associated metabolic pathways in the onset and progression of various neurological disorders provides new opportunities for developing novel therapeutic strategies and biomarkers for improved outcomes and early disease diagnosis.⁶⁰

By analyzing the spatially resolved tissue metabolome, DESI-MSI has been widely used to identify various tumor types. Its clinical applications for lipidomics profiling have been supported by numerous studies over the last ten years.^61,62 In a recent study, Banerjee et al. (2021) used 39 fresh frozen surgical specimens of the human hippocampus to perform DESI-MSI in order to look at the lipid profiles in control groups (n = 25) and experimental groups with TLE and hippocampal sclerosis (n = 14) (Fig. 5).⁶³ The human TLE hippocampus showed lower expression of several important lipids, most notably [PC] phosphatidylcholine and [PE] phosphatidylethanolamine, compared to several previous research on animal studies of epilepsy. Furthermore, metabolic pathway analyses revealed that TLE might have downregulated the Kennedy pathway, which would result in a sharp decline in PC and PE thresholds. This discovery enables more comprehensive studies of the associated molecular pathways and promising clinical targets for TLE. The study's authors noted that in order to expand the molecular coverage and detect additional classes of PLs in the future, a negative ion mode method must be used.⁶³

Fig. 5 Representative positive ion mode DESI-MS images showing the spatial distribution of 27 lipid species in the human hippocampal section obtained from a TLE (shown on the left-hand side) and a non-TLE (shown on the right-hand side) patient. The upper left box shows optical images of the corresponding H&E-stained tissues. Image data are total ion current (TIC) normalized. Reproduced with permission from [ref. 63] @ American Chemical Society (2021).

The eighth most frequent cancer in the world is esophageal cancer, accounting for about 500 000 fatalities annually.⁶⁴ Esophageal adenocarcinoma (EA), a subtype that is most common in Western nations and is associated with esophageal acid reflux, is the most prevalent variety of the illness. About 30% of patients are amenable to curative treatment when it is discovered at an advanced stage.⁶⁵ Five-year survival after treatment with curative intent is still 30–35%, and all-stage 5 year survival is 14%,⁶⁵ despite advancements in multimodal therapy. Therefore, new tools are required to encourage early diagnosis and efficient treatment. Although the relationship between metabolism and cancer has been extensively studied^66–68 lipids, which make up 70% of the human metabolome, have received relatively little attention.⁶⁹ Most lipids are locked away as glycerophospholipids (GPLs) in bilayer membranes. They are a desirable subject for biomarker studies due to their chemical and physical stability. Functionally, membrane properties and cellular signaling are impacted by GPL diversity. Phosphatidylglycerols (PGs) have powerful and opposing effects on the proliferation of squamous cells,⁷⁰ whereas phosphatidylinositols (PIs) mediate phosphatidylinositol-3 kinase (PI3K) signaling, one of the most frequently deregulated pathways in EA.⁷¹ As a result, measuring GPLs may offer diagnostic biomarkers with potential for therapeutic action; however, precise species composition must be established.

The advantages of DESI-MSI over other MSI techniques include the ability to detect a wider range of lipids, reduced sample preparation requirements, operation in atmospheric pressure conditions, and non-destructive tissue analysis that enables comparative histological analysis.⁷² With the aid of paired samples from surgical tissue samples, Abbassi-Ghadi et al. (2020) conducted an empirical study using DESI-MSI to contrast EA GPL and normal squamous and EA GPL profiles.⁷³ Furthermore, the authors examined the following: (a) the evolution of the EA GPL characteristics by contrasting healthy, inflammatory, metaplastic, dysplastic, and neoplastic cell kinds, which also functioned as an objective evaluation group; (b) the mechanistic premise for these GPL markers by defining the associated fatty acid group and genetic framework; and (c) the relationship between de novo lipogenesis and EA GPL attributes by suppressing a down regulated lipogenic gene. In both the discovery (area-under-curve = 0.97) and validation cohorts (AUC = 1), multivariable models created from 117 patients' phospholipid profiles were extremely selective for esophageal adenocarcinoma. The change is one of many, EA specimens exhibited a markedly enhanced presence of polyunsaturated PIs with longer acyl chains, with systematic enhancement in pre-malignant tissues. FA and GPL synthesis genes were significantly more expressed, and the properties of FA and GPL acyls were comparable. De novo lipogenesis is connected to the phospholipidome through the mechanism of the carbon switch ACLY being silenced in esophageal adenocarcinoma cells.⁷³

Globally, there were 935 000 fatal cases and an estimated 1.9 million new cases of CRC in 2022.⁷⁴ Individual CRC risk, tumor aggressiveness, and treatment outcomes may be influenced by both non-modifiable and modifiable factors, such as age, genetics, biological sex, environmental factors, and lifestyle factors, such as obesity.^75,76 Adenocarcinomas, which develop from the epithelial cells of the colorectal mucosa, make up most colorectal tumors.⁷⁷ Decisions regarding adjuvant oncologic treatment are increasingly based on molecular biomarkers.^78–80 There are currently four consensus molecular subtypes (CMS) of CRC that can affect a patient's prognosis and available treatments.^81,82 These subtypes are CMS1, CMS2, CMS3, and CMS4. The CMSs are determined by several variables, such as mutational status, immune activation, and metabolic dysregulation, all of which are discussed in detail elsewhere.⁷⁹ While the genetic and molecular causes of CRC have received extensive research, metabolomic analysis of CRC tissue may provide new opportunities to find biomarkers that are diagnostic, prognostic, and predictive. Methods based on mass spectrometry enable the characterization of metabolic pathways, such as those that initiate or support malignant transformation, as well as cell phenotypes and metabolic activity. Several mass spectrometry modalities have been used to examine the plasma, serum, and tissues of CRC patients. One investigation found that the metabolism of triacylglycerides (TAG) and amino acids was dysregulated in CRC.^83,84 Shen et al. also discovered that CRC tissues have increased glutathione metabolism to combat growing oxidative stress.⁸⁵ Studies investigating the connection between the metabolome and microbiota in CRC have also been published.^74,86 As a nondestructive metabolomic method DESI-MSI enables the profiling of small molecules from the surfaces of freshly frozen tissue sections without the need for any additional tissue processing beyond cryosectioning, as previously discussed.^87,88 Since the spatial location of the detected m/z ratios is preserved as a result, the same slide can undergo “gold standard” histological validation after DESI analysis and hematoxylin and eosin (H and E) staining.

Kaufmann et al. (2022) conducted a feasibility study on human CRC patient samples that contained malignant and nearby non-neoplastic tissue as well as the samples' simulated biopsies.⁸⁹ The goal of the study was to identify metabolite profiles linked to adenocarcinoma and other tissue types in the colon and rectum. It also aimed to investigate the reproducibility and clinical applicability of these profiles across patient populations. DESI examined CRC samples from patients (n = 10) undergoing surgery for the study.⁸⁹ MS features were contrasted to histopathological annotations and diagnostic biomarkers (Fig. 6). In fresh, frozen segments of colorectal cross sections, simulated endoscopic biopsy specimens containing tumor and non-neoplastic mucosa were made for each patient and blindly evaluated by DESI.⁸⁹ The sections were then hematoxylin and eosin stained, evaluated, and annotated by two different pathologists. Cross-sectional and biopsy DESI profiles with PCA/LDA-based models had 97% and 75% accuracy in detecting adenocarcinoma, respectively. According to molecular and targeted metabolomics signs of de novo lipogenesis in CRC tissue, very-long-chain FAs showed the highest differential presence in adenocarcinoma.⁸⁹ A higher concentration of oxidized phospholipids, indicative of pro-apoptotic pathways, was found in LVI-negative patients in comparison to LVI-positive patients in the samples that were stratified based on the existence of lymphovascular invasion (LVI), a weak CRC predictive sign. The research shows that spatially resolved DESI profiles may be clinically useful in enhancing clinicians' knowledge for CRC treatment.⁸⁹

Fig. 6 Histopathology annotation and corresponding mass spectrometry images of colorectal tissue. Hematoxylin and eosin (E and H) micrographs for representative colorectal cross sections (A and E) and biopsies (I and M), including annotation of adenocarcinoma (AdC) and other non-neoplastic tissue regions: BM—benign mucosa, Sub—submucosa, SM—smooth muscle, Ser: serosa, IC—inflammatory cells. Composite image of single-ion heatmaps featuring m/z ratios that are specifically abundant in the indicated region (B, F, J and N). Single-ion heatmap focusing on the relative abundance of m/z 309.279 which is spatially correlated with adenocarcinoma (C, G, K and O). Composite multivariate image based on the first three components of TIC-normalized pixels subjected to dimension reduction by PCA (D, H, L and P). Reproduced with permission from [ref. 89] @ MDPI (2023).

Lung cancer is the second most common cancer, with over 100 000 cases diagnosed each year.⁹⁰ Despite improvements in early cancer detection⁹¹ most lung cancer patients are diagnosed when the disease has progressed to the point where surgical resection is no longer curative.^92,93 The ability to distinguish between the adenocarcinoma (ADC) and squamous cell carcinoma [SCC] subtypes is crucial for the treatment of non-small cell lung cancer patients. Mass spectrometry (MS) imaging techniques for the diagnosis of cancer are becoming more popular due to their capacity to directly and untargetedly analyze tissues with high chemical specificity and analytical sensitivity.^94,95 Bensussan et al. (2020) evaluated the effectiveness of DESI-MSI to identify and distinguish lung cancer subtypes in tissues and fine needle aspiration (FNA) samples.⁹⁶ DESI-MSI was used to examine 22 normal lung tissues, 26 adenocarcinomas, and 25 squamous cell carcinomas (Fig. 7). The development and validation of statistical classifiers for the prognosis and subtyping of lung cancer used MS data collected from tissue sections.⁹⁶ DESI-MSI data gathered from 16 clinical FNA samples from patients undergoing interventional radiology-guided FNA were then used to evaluate classifiers. The MS data obtained from the lung tissues contained many metabolites and lipid species.⁹⁶ The classifiers developed from tissue sections produced 100% accuracy, sensitivity, and selectivity for diagnosing lung cancer for the training set of tissues per patient, and 73.5% accuracy for subtyping lung cancer.⁹⁶ On the validation set of tissues, 94.1% accuracy for subtyping lung cancer and 100% accuracy for diagnosing lung cancer were achieved.⁹⁶

Fig. 7 Optical microscopy images of three H&E stained lung tissue sections with corresponding DESI-MS ion images. Black dotted lines delineate regions of ADC or SCC. Red dotted lines delineate regions of normal lung cells and/or normal stroma. Blue dotted lines delineate regions of necrosis. Solid black squares indicate regions selected to show enlarged histology. (x : y) Indicate the number of carbon atoms : double bonds in each lipid. Scale bar ¼ 4 mm. Reproduced with permission from [ref. 96] @ cancer diagnostics (2020).

The liver is made up of tens of thousands of hexagonal structures called hepatic lobules, each with a central vein and hexagonal portal tracts at each corner. The liver lobule exhibits a gradient of nutrient and oxygen levels across the hepatocyte along the sinusoid from the periportal site to the pericentral site, leading to asymmetrically distributed metabolic processes and producing a pattern known as liver zonation.^97,98 The metabolic and biochemical pathways for xenobiotic, amino acid, carbohydrate, and lipid metabolism vary depending on the liver zonation. Even though the variations in metabolic zonation have previously been described, little is known about the lipid metabolism of each zone in the liver. Previous research has shown that hepatocytes isolated from various liver zones exhibit heterogeneous fatty acid (FA) metabolism. The periportal hepatocytes are primarily responsible for FA oxidation, whereas the pericentral hepatocytes demonstrated higher rates of lipogenesis.⁹⁹ A recent study found that periportal hepatocytes had differentially expressed genes involved in FA degradation, whereas pericentral hepatocytes had genes involved in cholesterol and bile acid metabolism.^100,101 A higher rate of FA uptake and FA synthesis was also found in periportal hepatocytes, according to zone-specific proteomics data from isolated hepatocytes.¹⁰² The zone-specific hepatic lipid signature, however, has only been the subject of a small number of studies.

Finally, spatial DESI-MSI was used to characterize the hepatic lipid metabolism throughout the zonation of the liver.¹⁰³ To comprehend the metabolic changes during the disease stage, it is essential to comprehend the role of zone-specific hepatic lipid metabolism in the healthy liver.¹⁰⁴ Desorption electrospray ionization mass spectrometry imaging was used to examine fresh frozen livers from healthy mice fed a control diet. Imaging was done with pixels measuring 50 μm by 50 μm. By manually co-registering with histological data, regions of interest (ROIs) were created to ascertain the spatial distribution of hepatic lipids across the zonation of the liver. With the help of double immunofluorescence, the ROIs were verified. To find statistically significant lipids throughout liver zonation, a mass list of ROIs was automatically generated, and univariate and multivariate statistical analysis was carried out. Fatty acids, phospholipids, triacylglycerols, diacylglycerols, ceramides, and sphingolipids were among the numerous lipid species that were discovered.¹⁰³ The periportal zone, midzone, and pericentral zones of the liver were all studied, and the authors were able to identify the hepatic lipid signatures in each of these regions. They also verified the method's ability to measure a variety of lipids consistently.¹⁰³ While phospholipids were found throughout the periportal and pericentral zones, fatty acids were primarily found in the periportal region. Unexpectedly, the midzone (zone 2) was where phosphatidylinositols, PI(36 : 2), PI(36 : 3), PI(36 : 4), PI(38 : 5), and PI(40 : 6) were primarily found (Fig. 8).¹⁰³ Most frequently found in the pericentral region were triacylglycerols and diacylglycerols. Over the three zones, the de novo triacylglycerol biosynthesis pathway seemed to be the most affected one. Accurately determining the distribution of hepatic lipids in each zone of the liver may help us comprehend how lipid metabolism changes as liver disease progresses.¹⁰³

Fig. 8 Comparison of the liver zonation pattern using liver histology, immunofluorescence imaging, and molecular imaging with DESI. (A) H&E staining of liver tissue after DESI-MSI analysis. (B) Double immunofluorescence staining shows the proto-central axis along with the liver zonation. GS-6, pericentral hepatocytes (blue); E-Cad, periportal hepatocytes (red); and DAPI (yellow). (C) Red, green, and blue overlay ion image from DESI-MSI : PI (38 : 3) located in the pericentral area (blue); PI(36 : 3) mainly expressed in the midzone (green); and PE(38 : 6) predominantly presented in the periportal region (red). CV, central vein; DESI, desorption electrospray ionisation; DESI-MSI, desorption electrospray ionisation mass spectrometry imaging; E-Cad, E-cadherin; GS-6, glutamine synthetase; MZ, midzone; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PV, portal vein. Reproduced with permission from [ref. 103] @ JHEP (2023).

Breast cancer heterogeneity can be distinguished into genotype and phenotype.¹⁰⁵ Utilizing next-generation sequencing, genotype heterogeneity has been studied.¹⁰⁶ Although gene mutations in breast cancer are less frequent than in other cancers,¹⁰⁷ several genetic tests have been used in this disease. Regardless of how sophisticated genomic methods become, the pathological approach to evaluating cancer phenotypes won't be rendered unnecessary but rather more crucial because phenotypes can reflect the effects of genomic abnormalities. While maintaining the position data of the molecules, MSI can recognize and visualize molecules on histological sections. MSI can be used to calculate the molar mass of lipids and fatty acids. Additionally, MSI research has demonstrated that, in contrast to normal tissues, cancer tissues express more specific lipids and fatty acids.^108,109

Aramaki et al. (2023) used cluster assessment of MSI data depending on the characteristics of lipid molecules and the levels of their expression to investigate variability in luminal breast cancer tissue sections.¹¹⁰ The clusters were composed of phosphatidylethanolamine, sphingomyelin, (PC) phosphatidylcholine, (TG) triglycerides, and ceramide. It was discovered that the percentage of TG and PC mainly correlated with the percentage of stroma and cancer on HE images. This group of lipids also differed from cluster to cluster in their carbon composition.¹¹⁰ This was in line with the notion that cancer metabolism results in an increase in the number of enzymes that synthesize long-chain fatty acids. Clusters of PCs could indicate high malignancy if they had high carbon counts. These results imply that cancers for which genetic testing alone is inadequate may be classified using phenotype variability depending on lipidomics.¹¹⁰

In a different study, Mondal et al. (2023) imaged fresh-frozen excision samples from 73 cancer patients who experienced lumpectomies using DESI-MSI. This included cancer and paired adjacent normal tissue.¹¹¹ The results demonstrate a marked metabolic upregulation of diacylglycerol, a lipid second carrier that activates protein kinase C to promote cancer growth. The study pinpointed four distinct sn-1,2-diacylglycerols.¹¹¹ Diacylglycerols outperformed other lipid classes in accurately predicting breast cancer, yielding 100% validity in the validation set, according to supervised machine learning on the entire dataset. When diacylglycerol signals were not included in the machine learning process, this accuracy significantly decreased. Therefore, DESI-MSI should support the intraoperative surgical pathology in lumpectomy by targeting diacylglycerol as the candidate oncomarker.¹¹¹

It can be difficult to ascertain the best surgical and therapeutic course of treatment for early-stage prostate cancer by physical examination or image analysis.¹¹² In contrast to breast¹¹³ and colon cancer,¹¹⁴ prostate cancer cannot be segmented into a appropriate sequence supported by specific genomic occurrences. Additionally, because prostate cancer grows in tiny cell clusters, it requires analytical techniques with high enough resolution to differentiate cancer cells from noncancerous tissue.¹¹⁵ Prostate cancer-related metabolic characteristics were observed in earlier research studies^116,117 using DESI-MSI. Across both studies, classification was performed for each sample, and each region of interest within a sample was approximated to create a single classification. This low-resolution technique has very little statistical influence and offers relatively limited information on sample variability. The classifiers from these experiments, for instance, lack the positional accuracy and validation required to be beneficial for intraoperative clinical guidelines.¹¹⁸

Morse et al. (2019) created an analytical work process to resolve these issues by evaluating DESI metabolomic characteristics to distinguish between noncancerous and cancerous tissue at a spatial resolution of 150 μm.¹¹⁸ A precise and highly-resolved metabolic characteristics of prostate cancer was created using over 900 spatially distributed DESI-MSI spectra. The study found 25 metabolites that were differentially prevalent, with cancer tissue exhibiting elevated levels of fatty acids, phospholipids, and benign tissue exhibiting elevated levels of lyso-phosphatidylethanolamine (PE).¹¹⁸ The study also discovered 2 lyso-PEs whose availability declined with cancer degree and phosphatidylcholines whose concentration increased with cancer degree for the very first time. From healthy to cancerous tissues, lysoPE(18 : 0) & lyso-PE(16 : 0) concentration gradually declined. According to the observation, prostate cancers deplete lysophospholipids in proportion to their rates of growth.¹¹⁷ Overall, it was observed that there was a difference in the quantity of the metabolites between cancer and noncancerous tissue, and those metabolites that varied with cancer degree represent defined modified metabolic processes.¹¹⁸

Limitations

Although DESI-MSI (a) has minimal prerequisites for sample pre-treatment and (b) does not tend to have high risks of spatial assignment inaccuracy brought on by specimen movement, enhancing the sensitivity of DESI-MSI has been a significant challenge for research studies employing the technique. According to a recent study¹¹⁹ it may be possible to enhance the responsiveness and specificity of the DESI-MSI by doping the solvent with silver [Ag] ions. Besides, due to variables like solvent composition, capillary dimensions, and air flow velocity, DESI-MSI has lesser spatial resolution than non-ambient techniques like MALDI-MSI (∼30 m).¹²⁰ Recent studies have used nano-DESI-MSI to overcome constraints and enhance spatial resolution (10 m), illustrating the potential future improvements of DESI-MSI.^121,122

Conclusions

As already mentioned, the DESI-MSI technique allows multiple assessments of thousands of molecules in a histological segment in a short period of time without the use of labels. This may offer a higher level of intricate molecular data to link the specialized fields of pathogenesis and cellular metabolism. Thus, each of the studies (Table 1) opened a new door in the field of clinical MS and diagnostic molecular pathology. As a result, the field of clinical MS forecasts a transition from a time-consuming traditional chromatographic methodology to a rapid and spatially resolved in situ MS imaging modality that is incredibly effective, insightful, and works in real time. As a result, ambient MSI is more significant than ever for enhancing our understanding of cellular biology and for accomplishing fast and effective disease diagnosis. Upwards of 100 different cancers have been linked to human occurrence. It is envisioned that the molecular characteristics of various cancers will differ. The metabolome associated with cancer is also dynamic and strongly influenced by dietary patterns, drugs, and heredity in comparison to the genome. The potential for unique results from the DESI-MSI research on each cancer model motivates further investigation into this developing field.

When it comes to performance, robustness, and accessibility, MS imaging has made significant strides in recent years. As a result, the technique is being used in biomedical and clinical research more and more frequently. This implementation has been primarily driven by ongoing advancements in instrumental design, reliable sample preparation, and cutting-edge data analysis, which have facilitated the implementation and thorough interrogation of multimodal imaging datasets. The various MSI modalities can be employed to target different chemical species and clarify their function in associated biomechanisms because of their complementary application profiles. MSI can be a potent tool to investigate biological mechanisms and advance our mechanistic understanding of diseases when used appropriately and in conjunction with a well-defined and controlled study design. Additionally, finding and validating mechanistically implied novel molecular species can help with the development of pathology-associated biofluid- and imaging-based biomarkers for neurodegenerative diseases. These biomarkers are extremely important for both routine clinical pathology monitoring and the development of pharmacotherapy strategies. The identification of metabolic biological markers with spatially defined biochemical data linked to different ailments is a crucial component for the success of this technique. To improve the diagnostic capabilities of this ambient ionization MSI, new statistical methods for quickly extracting thorough molecular data from ion images are required. Before applying this method to standard pathological practice, it must first be tried on a sizable cohort of patients with various cancers (or other diseases) in order to validate the technique.

Future studies may evaluate additional information and implement new methodologies. For instance, when more spatial metabolomics information on numerous cancer specimens becomes available, it may be beneficial to bridge metabolite distribution throughout the stages of carcinogenesis.

Author contributions

Each author contributed to the article's design and conceptualization. All authors read the final draft before giving their approval.

Conflicts of interest

“There are no conflicts to declare”.

Acknowledgements

The authors would like to thank the professors who provided helpful criticism and feedback.

References

1. J. Rego and K. M. Tan, Advances in imaging—the changing environment for the imaging specialist, Perm. J., 2006, 10(1), 26.
2. S. Banerjee, Ambient ionization mass spectrometry imaging for disease diagnosis: Excitements and challenges, J. BioSci., 2018, 43(4), 731–738.
3. A. Becker, C. Hessenius, K. Licha, B. Ebert, U. Sukowski, W. Semmler, B. Wiedenmann and C. Grötzinger, Receptor-targeted optical imaging of tumors with near-infrared fluorescent ligands, Nat. Biotechnol., 2001, 19(4), 327–331.
4. G. Paganelli, A. Luini and U. Veronesi, Radioguided occult lesion localization (ROLL) in breast cancer: maximizing efficacy, minimizing mutilation, Ann. Oncol., 2002, 13(12), 1839–1840.
5. B. Erguvan-Dogan, G. J. Whitman, V. A. Nguyen, M. J. Dryden, R. J. Stafford, J. Hazle, K. R. McAlee, M. J. Phelps, M. F. Ice, H. M. Kuerer and L. P. Middleton, Specimen radiography in confirmation of MRI-guided needle localization and surgical excision of breast lesions, Am. J. Roentgenol., 2006, 187(2), 339–344.
6. A. S. Haka, Z. Volynskaya, J. A. Gardecki, J. Nazemi, J. Lyons, D. Hicks, M. Fitzmaurice, R. R. Dasari, J. P. Crowe and M. S. Feld, Cancer Res., 2006, 66, 3317–3322.
7. F. T. Nguyen, A. M. Zysk, E. J. Chaney, J. G. Kotynek, U. J. Oliphant, F. J. Bellafiore, K. M. Rowland, P. A. Johnson and S. A. Boppart, Intraoperative evaluation of breast tumor margins with optical coherence tomography, Cancer Res., 2009, 69(22), 8790–8796.
8. T. E. Doyle, R. E. Factor, C. L. Ellefson, K. M. Sorensen, B. J. Ambrose, J. B. Goodrich, V. P. Hart, S. C. Jensen, H. Patel and L. A. Neumayer, High-frequency ultrasound for intraoperative margin assessments in breast conservation surgery: a feasibility study, BMC Cancer, 2011, 11, 1–5.
9. T. Shioyama, Y. Muragaki, T. Maruyama, T. Komori and H. Iseki, Intraoperative flow cytometry analysis of glioma tissue for rapid determination of tumor presence and its histopathological grade, J. Neurosurg., 2013, 118(6), 1232–1238.
10. L. Bellanova, T. Schubert, O. Cartiaux, F. Lecouvet, C. Galant, X. Banse and P. L. Docquier, MRI-based assessment of safe margins in tumor surgery, Sarcoma, 2014, 2014, 1–5.
11. R. P. Judy, J. J. Keating, E. M. DeJesus, J. X. Jiang, O. T. Okusanya, S. Nie, D. E. Holt, S. P. Arlauckas, P. S. Low, E. J. Delikatny and S. Singhal, Quantification of tumor fluorescence during intraoperative optical cancer imaging, Sci. Rep., 2015, 5(1), 1.
12. J. M. Wiseman, D. R. Ifa, Q. Song and R. G. Cooks, Tissue imaging at atmospheric pressure using desorption electrospray ionization (DESI) mass spectrometry, Angew. Chem., Int. Ed., 2006, 45(43), 7188–7192.
13. S. Santagata, L. S. Eberlin, I. Norton, D. Calligaris, D. R. Feldman, J. L. Ide, X. Liu, J. S. Wiley, M. L. Vestal, S. H. Ramkissoon and D. A. Orringer, Intraoperative mass spectrometry mapping of an onco-metabolite to guide brain tumor surgery, Proc. Natl. Acad. Sci. U. S. A., 2014, 111(30), 11121–11126.
14. Z. Takats, J. M. Wiseman, B. Gologan and R. G. Cooks, Mass spectrometry sampling under ambient conditions with desorption electrospray ionization, Science, 2004, 306(5695), 471–473.
15. A. Ajith, Y. Sthanikam and S. Banerjee, Chemical analysis of the human brain by imaging mass spectrometry, Analyst, 2021, 146(18), 5451–5473.
16. S. Banerjee and S. Mazumdar, Electrospray ionization mass spectrometry: a technique to access the information beyond the molecular weight of the analyte, Int. J. Anal. Chem., 2012, 2012, 282574.
17. J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong and C. M. Whitehouse, Electrospray ionization for mass spectrometry of large biomolecules, Science, 1989, 246(4926), 64–71.
18. A. Hollerbach, S. Ayrton, A. Jarmusch and R. G. Cooks, Encyclopedia of Spectroscopy and Spectrometry, Academic Press, 3rd edn, 2017, pp. 401–408.
19. L. S. Eberlin, C. R. Ferreira, A. L. Dill, D. R. Ifa and R. G. Cooks, Desorption electrospray ionization mass spectrometry for lipid characterization and biological tissue imaging, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids, 2011, 1811(11), 946–960.
20. F. D. Shah, R. Begum, B. N. Vajaria, K. R. Patel, J. B. Patel, S. N. Shukla and P. S. Patel, A review on salivary genomics and proteomics biomarkers in oral cancer, Indian J. Clin. Biochem., 2011, 26, 326–334.
21. M. W. Lingen, J. R. Kalmar, T. Karrison and P. M. Speight, Critical evaluation of diagnostic aids for the detection of oral cancer, Oral Oncol., 2008, 44(1), 10–22.
22. K. K. Dey, I. Pal, R. Bharti, G. Dey, B. P. Kumar, S. Rajput, A. Parekh, S. Parida, P. Halder, I. Kulavi and M. Mandal, Identification of RAB2A and PRDX1 as the potential biomarkers for oral squamous cell carcinoma using mass spectrometry-based comparative proteomic approach, Tumor Biol., 2015, 36, 9829–9837.
23. S. M. Lippman and W. K. Hong, Second malignant tumors in head and neck squamous cell carcinoma: the overshadowing threat for patients with early-stage disease, Int. J. Radiat. Oncol., Biol., Phys., 1989, 17(3), 691–694.
24. S. Fedele, Diagnostic aids in the screening of oral cancer, Head Neck Oncol., 2009, 1(1), 1–6.
25. K. S. Kerian, A. K. Jarmusch, V. Pirro, M. O. Koch, T. A. Masterson, L. Cheng and R. G. Cooks, Differentiation of prostate cancer from normal tissue in radical prostatectomy specimens by desorption electrospray ionization and touch spray ionization mass spectrometry, Analyst, 2015, 140(4), 1090–1098.
26. M. C. Pereira, D. T. Oliveira, G. Landman and L. P. Kowalski, Histologic subtypes of oral squamous cell carcinoma: prognostic relevance, J. Can. Dent. Assoc., 2007, 73(4), 339.
27. S. D. Silva, D. E. Perez, I. N. Nishimoto, F. A. Alves, C. A. Pinto, L. P. Kowalski and E. Graner, Fatty acid synthase expression in squamous cell carcinoma of the tongue: clinicopathological findings, Oral Dis., 2008, 14(4), 376–382.
28. X. Yang, X. Song, X. Zhang, V. Shankar, S. Wang, Y. Yang, S. Chen, L. Zhang, Y. Ni, R. N. Zare and Q. Hu, In situ DESI-MSI lipidomic profiles of mucosal margin of oral squamous cell carcinoma, EBioMedicine, 2021, 70, 103529.
29. D. P. Slaughter, H. W. Southwick and W. Smejkal, “Field cancerization” in oral stratified squamous epithelium. Clinical implications of multicentric origin, Cancer, 1953, 6(5), 963–968.
30. L. Mao and D. Clark, Molecular margin of surgical resections—Where do we go from here?, Cancer, 2015, 121(12), 1914–1916.
31. W. Y. Chiou, H. Y. Lin, F. C. Hsu, M. S. Lee, H. C. Ho, Y. C. Su, C. C. Lee, C. H. Hsieh, Y. C. Wang and S. K. Hung, Buccal mucosa carcinoma: surgical margin less than 3 mm, not 5 mm, predicts locoregional recurrence, Radiat. Oncol., 2010, 5, 1–8.
32. D. K. Zanoni, J. C. Migliacci, B. Xu, N. Katabi, P. H. Montero, I. Ganly, J. P. Shah, R. J. Wong, R. A. Ghossein and S. G. Patel, A proposal to redefine close surgical margins in squamous cell carcinoma of the oral tongue, JAMA Otolaryngol. Head Neck Surg., 2017, 143(6), 555–560.
33. E. A. Dik, N. A. Ipenburg, P. A. Kessler, R. J. van Es and S. M. Willems, The value of histological grading of biopsy and resection specimens in early stage oral squamous cell carcinomas, J Craniomaxillofac. Surg., 2018, 46(6), 1001–1006.
34. R. W. Nason, A. Binahmed, K. A. Pathak, A. A. Abdoh and G. K. Sándor, What is the adequate margin of surgical resection in oral cancer?, Oral Surg Oral Med Oral Pathol Oral Radiol Oral Endodontology, 2009, 107(5), 625–629.
35. X. H. Yang, X. X. Zhang, Y. Jing, L. Ding, Y. Fu, S. Wang, S. Q. Hu, L. Zhang, X. F. Huang, Y. H. Ni and Q. G. Hu, Amino acids signatures of distance-related surgical margins of oral squamous cell carcinoma, EBioMedicine, 2019, 48, 81–91.
36. P. M. Vaysse, I. Demers, M. F. van den Hout, W. van de Worp, I. G. Anthony, L. W. Baijens, B. I. Tan, M. Lacko, L. A. Vaassen, A. van Mierlo and R. C. Langen, Evaluation of the Sensitivity of Metabolic Profiling by Rapid Evaporative Ionization Mass Spectrometry: Toward More Radical Oral Cavity Cancer Resections, Anal. Chem., 2022, 94(19), 6939–6947.
37. J. Balog, J. Alexander, O. Golf, J. Huang, T. Wiggins, N. Abbassi-Ghadi, A. Enyedi, S. Kacska, J. Kinross, G. Hanna and J. K. Nicholson, In Vivo Endoscopic Tissue Identification Tool Utilising Rapid Evaporative Ionization Mass Spectrometry (REIMS).
38. J. Balog, T. Szaniszlo, K. C. Schaefer, J. Denes, A. Lopata, L. Godorhazy, D. Szalay, L. Balogh, L. Sasi-Szabo, M. Toth and Z. Takats, Identification of biological tissues by rapid evaporative ionization mass spectrometry, Anal. Chem., 2010, 82(17), 7343–7350.
39. P. M. Vaysse, I. Demers, M. F. van den Hout, W. van de Worp, I. G. Anthony, L. W. Baijens, B. I. Tan, M. Lacko, L. A. Vaassen, A. van Mierlo and R. C. Langen, Evaluation of the Sensitivity of Metabolic Profiling by Rapid Evaporative Ionization Mass Spectrometry: Toward More Radical Oral Cavity Cancer Resections, Anal. Chem., 2022, 94(19), 6939–6947.
40. E. R. St John, J. Balog, J. S. McKenzie, M. Rossi, A. Covington, L. Muirhead, Z. Bodai, F. Rosini, A. V. Speller, S. Shousha and R. Ramakrishnan, Rapid evaporative ionisation mass spectrometry of electrosurgical vapours for the identification of breast pathology: towards an intelligent knife for breast cancer surgery, Breast Cancer Res., 2017, 19(1), 1–4.
41. P. M. Vaysse, H. I. Grabsch, M. F. van den Hout, M. H. Bemelmans, R. M. Heeren, S. W. Olde Damink and T. Porta Siegel, Real-time lipid patterns to classify viable and necrotic liver tumors, Lab. Invest., 2021, 101(3), 381–395.
42. S. Mason, E. Manoli, L. Poynter, J. Alexander, P. Paizs, A. Adebesin, R. Goldin, A. Darzi, Z. Takats and J. Kinross, Mass spectrometry transanal minimally invasive surgery (MS-TAMIS) to promote organ preservation in rectal cancer, Surg. Endosc., 2020, 34, 3618–3625.
43. P. M. Vaysse, L. F. Kooreman, S. M. Engelen, B. Kremer, S. W. Olde Damink, R. M. Heeren, M. L. Smidt and T. Porta Siegel, Stromal vapors for real-time molecular guidance of breast-conserving surgery, Sci. Rep., 2020, 10(1), 20109.
44. Y. Li, S. Bai, W. Carroll, D. Dayan, J. C. Dort, K. Heller, G. Jour, H. Lau, C. Penner, M. Prystowsky, E. Rosenthal, N. F. Schlecht, R. V. Smith, M. Urken, M. Vered, B. Wang, B. Wenig, A. Negassa and M. Brandwein-Gensler, Head Neck Pathol., 2013, 7, 211–223.
45. M. Bryne, H. S. Koppang, R. Lilleng and Å. Kjærheim, Malignancy grading of the deep invasive margins of oral squamous cell carcinomas has high prognostic value, J. Pathol., 1992, 166(4), 375–381.
46. M. Brandwein-Gensler, M. S. Teixeira, C. M. Lewis, B. Lee, L. Rolnitzky, J. J. Hille, E. Genden, M. L. Urken and B. Y. Wang, Oral squamous cell carcinoma: histologic risk assessment, but not margin status, is strongly predictive of local disease-free and overall survival, Am. J. Surg. Pathol., 2005, 29(2), 167–178.
47. N. Rahman, M. MacNeill, W. Wallace and B. Conn, Reframing histological risk assessment of oral squamous cell carcinoma in the era of UICC 8th edition TNM staging, Head Neck Pathol., 2021, 15, 202–211.
48. I. Blümcke, E. Aronica, H. Miyata, H. B. Sarnat, M. Thom, K. Roessler, B. Rydenhag, L. Jehi, P. Krsek, S. Wiebe and R. Spreafico, International recommendation for a comprehensive neuropathologic workup of epilepsy surgery brain tissue: A consensus Task Force report from the ILAE Commission on Diagnostic Methods, Epilepsia, 2016, 57(3), 348–358.
49. C. E. Stafstrom and L. Carmant, Seizures and epilepsy: an overview for neuroscientists, Cold Spring Harbor Perspect. Med., 2015, 5(6), a022426.
50. C. F. Cavarsan, J. Malheiros, C. Hamani, I. Najm and L. Covolan, Is mossy fiber sprouting a potential therapeutic target for epilepsy?, Front. Neurol., 2018, 9, 1023.
51. Y. Luo, Q. Hu, Q. Zhang, S. Hong, X. Tang, L. Cheng and L. Jiang, Alterations in hippocampal myelin and oligodendrocyte precursor cells during epileptogenesis, Brain Res., 2015, 1627, 154–164.
52. C. M. O'Dell, A. Das, I. V. G. Wallace, S. K. Ray and N. L. Banik, Understanding the basic mechanisms underlying seizures in mesial temporal lobe epilepsy and possible therapeutic targets: a review, J. Neurosci. Res., 2012, 90(5), 913–924.
53. G. Dawson, Measuring brain lipids, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids, 2015, 1851(8), 1026–1039.
54. W. Dowhan, in Advances in Lipobiology, ed. R. W. Gross, JAI Press Inc., 1997, vol. 2, pp. 79−107.
55. J. A. Allen, R. A. Halverson-Tamboli and M. M. Rasenick, Lipid raft microdomains and neurotransmitter signalling, Nat. Rev. Neurosci., 2007, 8(2), 128–140.
56. R. Lerner, J. Post, S. Loch, B. Lutz and L. Bindila, Targeting brain and peripheral plasticity of the lipidome in acute kainic acid-induced epileptic seizures in mice via quantitative mass spectrometry, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids, 2017, 1862(2), 255–267.
57. X. L. Guan, X. He, W. Y. Ong, W. K. Yeo, G. Shui and M. R. Wenk, Non-targeted profiling of lipids during kainate-induced neuronal injury, FASEB J., 2006, 20(8), 1152–1161.
58. R. Lerner, J. M. Post, S. R. Ellis, D. N. Vos, R. M. Heeren, B. Lutz and L. Bindila, Simultaneous lipidomic and transcriptomic profiling in mouse brain punches of acute epileptic seizure model compared to controls, J. Lipid Res., 2018, 59(2), 283–297.
59. H. Zhang, P. Ren, Y. Huang, W. Zeng, K. Zhong, H. Gao, X. Cen and Q. Bu, Untargeted lipidomic analysis of human hippocampus for temporal lobe epilepsy with hippocampal sclerosis, Epilepsy Res., 2020, 161, 106299.
60. D. J. Stephenson, L. A. Hoeferlin and C. E. Chalfant, Lipidomics in translational research and the clinical significance of lipid-based biomarkers, Transl. Res., 2017, 189, 13–29.
61. L. S. Eberlin, C. R. Ferreira, A. L. Dill, D. R. Ifa and R. G. Cooks, Desorption electrospray ionization mass spectrometry for lipid characterization and biological tissue imaging, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids, 2011, 1811(11), 946–960.
62. M. Holzlechner, E. Eugenin and B. Prideaux, Mass spectrometry imaging to detect lipid biomarkers and disease signatures in cancer, Cancer Rep., 2019, 2(6), e1229.
63. A. Ajith, S. Mondal, S. Chattopadhyay, A. Kumar, Y. Sthanikam, A. G. Chacko, K. Prabhu, G. Chacko, H. A. Vanjare, R. V. Rajesh and S. Banerjee, Mass spectrometry imaging deciphers dysregulated lipid metabolism in the human hippocampus affected by temporal lobe epilepsy, ACS Chem. Neurosci., 2021, 12(21), 4187–4194.
64. L. A. Torre, F. Bray, R. L. Siegel, J. Ferlay, J. Lortet-Tieulent and A. Jemal, Global cancer statistics, 2012, Ca-Cancer J. Clin., 2015, 65(2), 87–108.
65. G. Chadwick, M. Varagunam, C. Brand, S. A. Riley, N. Maynard, T. Crosby, J. Michalowski and D. A. Cromwell, Coding of Barrett's oesophagus with high-grade dysplasia in national administrative databases: a population-based cohort study, BMJ Open, 2017, 7(6), e014281.
66. R. A. Weinberg and D. Hanahan, The hallmarks of cancer, Cell, 2000, 100(1), 57–70.
67. G. Kroemer and J. Pouyssegur, Tumor cell metabolism: cancer's Achilles' heel, Cancer Cell, 2008, 13(6), 472–482.
68. E. Holmes, I. D. Wilson and J. K. Nicholson, Metabolic phenotyping in health and disease, Cell, 2008, 134(5), 714–717.
69. D. S. Wishart, C. Knox, A. C. Guo, R. Eisner, N. Young, B. Gautam, D. D. Hau, N. Psychogios, E. Dong, S. Bouatra and R. Mandal, HMDB: a knowledgebase for the human metabolome, Nucleic Acids Res., 2009, 37(suppl_1), D603–D610.
70. D. Xie, M. Seremwe, J. G. Edwards, R. Podolsky and W. B. Bollag, Distinct effects of different phosphatidylglycerol species on mouse keratinocyte proliferation, PLoS One, 2014, 9(9), e107119.
71. A. M. Frankell, S. Jammula, X. Li, G. Contino, S. Killcoyne, S. Abbas, J. Perner, L. Bower, G. Devonshire, E. Ococks and N. Grehan, The landscape of selection in 551 esophageal adenocarcinomas defines genomic biomarkers for the clinic, Nat. Genet., 2019, 51(3), 506–516.
72. D. Gode and D. A. Volmer, Lipid imaging by mass spectrometry–a review, Analyst, 2013, 138(5), 1289–1315.
73. N. Abbassi-Ghadi, S. S. Antonowicz, J. S. McKenzie, S. Kumar, J. Huang, E. A. Jones, N. Strittmatter, G. Petts, H. Kudo, S. Court and J. M. Hoare, De novo lipogenesis alters the phospholipidome of esophageal adenocarcinoma, Cancer Res., 2020, 80(13), 2764–2774.
74. J. Yang, H. Wei, Y. Zhou, C. H. Szeto, C. Li, Y. Lin, O. O. Coker, H. C. Lau, A. W. Chan, J. J. Sung and J. Yu, High-fat diet promotes colorectal tumorigenesis through modulating gut microbiota and metabolites, Gastroenterology, 2022, 162(1), 135–149.
75. P. Rawla, T. Sunkara and A. Barsouk, Epidemiology of colorectal cancer: incidence, mortality, survival, and risk factors, Przegl. Gastroenterol., 2019, 14(2), 89–103.
76. R. L. Siegel, K. D. Miller, A. Goding Sauer, S. A. Fedewa, L. F. Butterly, J. C. Anderson, A. Cercek, R. A. Smith and A. Jemal, Colorectal cancer statistics, 2020, Ca-Cancer J. Clin., 2020, 70(3), 145–164.
77. M. Fleming, S. Ravula, S. F. Tatishchev and H. L. Wang, Colorectal carcinoma: Pathologic aspects, J. Gastrointest. Oncol., 2012, 3(3), 153.
78. S. Dayal, N. Battersby and T. Cecil, Evolution of surgical treatment for rectal cancer: a review, J. Gastrointest. Surg., 2017, 21, 1166–1173.
79. F. Ali, A. Keshinro and M. R. Weiser, Advances in the treatment of locally advanced rectal cancer, Ann. Gastroenterol. Surg., 2021, 5(1), 32–38.
80. G. Rospo, G. Corti, G. Crisafulli, L. Novara and A. Bardelli, Tracking colorectal cancer evolution in time and space, Ann. Oncol., 2017, 28(6), 1163–1165.
81. J. Guinney, R. Dienstmann, X. Wang, A. De Reynies, A. Schlicker, C. Soneson, L. Marisa, P. Roepman, G. Nyamundanda, P. Angelino and B. M. Bot, The consensus molecular subtypes of colorectal cancer, Nat. Med., 2015, 21(11), 1350–1356.
82. F. Hu, J. Wang, M. Zhang, S. Wang, L. Zhao, H. Yang, J. Wu and B. Cui, Comprehensive analysis of subtype-specific molecular characteristics of colon cancer: Specific genes, driver genes, signaling pathways, and immunotherapy responses, Front. Cell Dev. Biol., 2021, 9, 758776.
83. K. A. Lipinski, L. J. Barber, M. N. Davies, M. Ashenden, A. Sottoriva and M. Gerlinger, Cancer evolution and the limits of predictability in precision cancer medicine, Trends Cancer, 2016, 2(1), 49–63.
84. T. Liu, F. Peng, J. Yu, Z. Tan, T. Rao, Y. Chen and J. Peng, LC-MS-based lipid profile in colorectal cancer patients: TAGs are the main disturbed lipid markers of colorectal cancer progression, Anal. Bioanal. Chem., 2019, 411, 5079–5088.
85. Y. Shen, M. Sun, J. Zhu, M. Wei, H. Li, P. Zhao, J. Wang, R. Li, L. Tian, Y. Tao and P. Shen, Tissue metabolic profiling reveals major metabolic alteration in colorectal cancer, Mol. Omics, 2021, 17(3), 464–471.
86. F. Chen, X. Dai, C. C. Zhou, K. X. Li, Y. J. Zhang, X. Y. Lou and W. Cui, Integrated analysis of the faecal metagenome and serum metabolome reveals the role of gut microbiome-associated metabolites in the detection of colorectal cancer and adenoma, Gut, 2022, 71(7), 1315–1325.
87. T. Soudah, A. Zoabi and K. Margulis, Desorption electrospray ionization mass spectrometry imaging in discovery and development of novel therapies, Mass Spectrom. Rev., 2023, 42(2), 751–778.
88. Z. Takats, J. M. Wiseman and R. G. Cooks, Ambient mass spectrometry using desorption electrospray ionization (DESI): instrumentation, mechanisms and applications in forensics, chemistry, and biology, J. Mass Spectrom., 2005, 40(10), 1261–1275.
89. M. Kaufmann, N. Iaboni, A. Jamzad, D. Hurlbut, K. Y. Ren, J. F. Rudan, P. Mousavi, G. Fichtinger, S. Varma, A. Caycedo-Marulanda and C. J. Nicol, Metabolically Active Zones Involving Fatty Acid Elongation Delineated by DESI-MSI Correlate with Pathological and Prognostic Features of Colorectal Cancer, Metabolites, 2023, 13(4), 508.
90. R. L. Siegel, K. D. Miller and A. Jemal, Cancer statistics, Ca-Cancer J. Clin., 2019, 69(1), 7–34.
91. J. Shen, N. W. Todd, H. Zhang, L. Yu, X. Lingxiao, Y. Mei, M. Guarnera, J. Liao, A. Chou, C. L. Lu and Z. Jiang, Plasma microRNAs as potential biomarkers for non-small-cell lung cancer, Lab. Invest., 2011, 91(4), 579–587.
92. W. D. Travis, N. Rekhtman, G. J. Riley, K. R. Geisinger, H. Asamura, E. Brambilla, K. Garg, F. R. Hirsch, M. Noguchi, C. A. Powell and V. W. Rusch, Pathologic diagnosis of advanced lung cancer based on small biopsies and cytology: a paradigm shift, J. Thorac. Oncol., 2010, 5(4), 411–414.
93. C. Rolfo, F. Passiglia, M. Ostrowski, L. Farracho, T. Ondøichová, A. Dolcan, M. Castiglia, R. Remmen, K. Papadimitriou and P. Pauwels, Improvement in lung cancer outcomes with targeted therapies: an update for family physicians, J. Am. Board Fam. Med., 2015, 28(1), 124–133.
94. E. R. van Hove, D. F. Smith and R. M. Heeren, A concise review of mass spectrometry imaging, J. Chromatogr. A, 2010, 1217(25), 3946–3954.
95. C. L. Feider, A. Krieger, R. J. DeHoog and L. S. Eberlin, Ambient ionization mass spectrometry: recent developments and applications, Anal. Chem., 2019, 91(7), 4266–4290.
96. A. V. Bensussan, J. Lin, C. Guo, R. Katz, S. Krishnamurthy, E. Cressman and L. S. Eberlin, Distinguishing non-small cell lung cancer subtypes in fine needle aspiration biopsies by desorption electrospray ionization mass spectrometry imaging, Clin. Chem., 2020, 66(11), 1424–1433.
97. T. Kietzmann, Metabolic zonation of the liver: The oxygen gradient revisited, Redox Biol., 2017, 11, 622–630.
98. J. Paris and N. C. Henderson, Liver zonation, revisited, Hepatology, 2022, 76(4), 1219–1230.
99. M. Guzmán and J. Castro, Zonation of fatty acid metabolism in rat liver, Biochem. J., 1989, 264(1), 107–113.
100. A. Braeuning, C. Ittrich, C. Köhle, S. Hailfinger, M. Bonin, A. Buchmann and M. Schwarz, Differential gene expression in periportal and perivenous mouse hepatocytes, FEBS J., 2006, 273(22), 5051–5061.
101. K. B. Halpern, R. Shenhav, O. Matcovitch-Natan, B. Toth, D. Lemze, M. Golan, E. E. Massasa, S. Baydatch, S. Landen, A. E. Moor and A. Brandis, Single-cell spatial reconstruction reveals global division of labour in the mammalian liver, Nature, 2017, 542(7641), 352–356.
102. N. Berndt, E. Kolbe, R. Gajowski, J. Eckstein, F. Ott, D. Meierhofer, H. G. Holzhütter and M. Matz-Soja, Functional consequences of metabolic zonation in murine livers: insights for an old story, Hepatology, 2021, 73(2), 795–810.
103. P. Seubnooch, M. Montani, S. Tsouka, E. Claude, U. Rafiqi, A. Perren, J. F. Dufour and M. Masoodi, Characterization of hepatic lipid signature distributed across the liver zonation using mass spectrometry imaging, JHEP Rep., 2023, 100725.
104. A. L. Dill, D. R. Ifa, N. E. Manicke, Z. Ouyang and R. G. Cooks, Mass spectrometric imaging of lipids using desorption electrospray ionization, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2009, 877(26), 2883–2889.
105. F. Beca and K. Polyak, Intratumor heterogeneity in breast cancer, Novel Biomarkers in the Continuum of Breast Cancer, 2016, pp. 169–89.
106. L. R. Yates, M. Gerstung, S. Knappskog, C. Desmedt, G. Gundem, P. Van Loo, T. Aas, L. B. Alexandrov, D. Larsimont, H. Davies and Y. Li, Subclonal diversification of primary breast cancer revealed by multiregion sequencing, Nat. Med., 2015, 21(7), 751–759.
107. L. G. Morris, N. Riaz, A. Desrichard, Y. Şenbabaoğlu, A. A. Hakimi, V. Makarov, J. S. Reis-Filho and T. A. Chan, Pan-cancer analysis of intratumor heterogeneity as a prognostic determinant of survival, Oncotarget, 2016, 7(9), 10051.
108. Y. Hosokawa, N. Masaki, S. Takei, M. Horikawa, S. Matsushita, E. Sugiyama, H. Ogura, N. Shiiya and M. Setou, Recurrent triple-negative breast cancer (TNBC) tissues contain a higher amount of phosphatidylcholine (32: 1) than non-recurrent TNBC tissues, PLoS One, 2017, 12(8), e0183724.
109. Y. Ide, M. Waki, T. Hayasaka, T. Nishio, Y. Morita, H. Tanaka, T. Sasaki, K. Koizumi, R. Matsunuma, Y. Hosokawa and H. Ogura, Human breast cancer tissues contain abundant phosphatidylcholine (36:1) with high stearoyl-CoA desaturase-1 expression, PLoS One, 2013, 8(4), e61204.
110. S. Aramaki, S. Tsuge, A. Islam, F. Eto, T. Sakamoto, S. Oyama, W. Li, C. Zhang, S. Yamaguchi, D. Takatsuka and Y. Hosokawa, Lipidomics-based tissue heterogeneity in specimens of luminal breast cancer revealed by clustering analysis of mass spectrometry imaging: A preliminary study, PLoS One, 2023, 18(5), e0283155.
111. S. Mondal, Y. Sthanikam, A. Kumar, A. Nandy, S. Chattopadhyay, D. Koner, N. Rukmangadha, H. Narendra and S. Banerjee, Mass Spectrometry Imaging of Lumpectomy Specimens Deciphers Diacylglycerols as Potent Biomarkers for the Diagnosis of Breast Cancer, Anal. Chem., 2023, 8054–8062.
112. G. J. Kelloff, P. Choyke and D. S. Coffey, Challenges in clinical prostate cancer: role of imaging, AJR, Am. J. Roentgenol., 2009, 192(6), 1455.
113. A. Abeshouse, J. Ahn, R. Akbani, A. Ally, S. Amin, C. D. Andry, M. Annala, A. Aprikian, J. Armenia, A. Arora and J. T. Auman, The molecular taxonomy of primary prostate cancer, Cell, 2015, 163(4), 1011–1025.
114. M. Fleming, S. Ravula, S. F. Tatishchev and H. L. Wang, Colorectal carcinoma: pathologic aspects, J. Gastrointest. Oncol., 2012, 3(3), 153–173.
115. J. I. Epstein, L. Egevad, M. B. Amin, B. Delahunt, J. R. Srigley and P. A. Humphrey, The 2014 International Society of Urological Pathology (ISUP) consensus conference on Gleason grading of prostatic carcinoma, Am. J. Surg. Pathol., 2016, 40(2), 244–252.
116. K. S. Kerian, A. K. Jarmusch, V. Pirro, M. O. Koch, T. A. Masterson, L. Cheng and R. G. Cooks, Differentiation of prostate cancer from normal tissue in radical prostatectomy specimens by desorption electrospray ionization and touch spray ionization mass spectrometry, Analyst, 2015, 140(4), 1090–1098.
117. S. Banerjee, R. N. Zare, R. J. Tibshirani, C. A. Kunder, R. Nolley, R. Fan, J. D. Brooks and G. A. Sonn, Diagnosis of prostate cancer by desorption electrospray ionization mass spectrometric imaging of small metabolites and lipids, Proc. Natl. Acad. Sci. U. S. A., 2017, 114(13), 3334–3339.
118. N. Morse, T. Jamaspishvili, D. Simon, P. G. Patel, K. Y. Ren, J. Wang, R. Oleschuk, M. Kaufmann, R. J. Gooding and D. M. Berman, Reliable identification of prostate cancer using mass spectrometry metabolomic imaging in needle core biopsies, Lab. Invest., 2019, 99(10), 1561–1571.
119. J. Lillja and I. Lanekoff. Silver-doped nano-DESI MSI for increased specificity and sensitivity of alkenes, Mass Spectrometry Imaging of Small Molecules: Methods and Protocols, 2022, pp. 241–249.
120. K. Qi, L. Wu, C. Liu and Y. Pan, Recent advances of ambient mass spectrometry imaging and its applications in lipid and metabolite analysis, Metabolites, 2021, 11(11), 780.
121. R. Yin, J. Kyle, K. Burnum-Johnson, K. J. Bloodsworth, L. Sussel, C. Ansong and J. Laskin, High spatial resolution imaging of mouse pancreatic islets using nanospray desorption electrospray ionization mass spectrometry, Anal. Chem., 2018, 90(11), 6548–6555.
122. D. Unsihuay, R. Yin, D. M. Sanchez, M. Yang, Y. Li, X. Sun, S. K. Dey and J. Laskin, High-resolution imaging and identification of biomolecules using Nano-DESI coupled to ion mobility spectrometry, Anal. Chim. Acta, 2021, 1186, 339085.
123. K. Tamura, M. Horikawa, S. Sato, H. Miyake and M. Setou, Discovery of lipid biomarkers correlated with disease progression in clear cell renal cell carcinoma using desorption electrospray ionization imaging mass spectrometry, Oncotarget, 2019, 1688–1703.
124. A. L. Santoro, R. D. Drummond, I. T. Silva, S. S. Ferreira, L. Juliano, P. H. Vendramini, M. B. da Costa Lemos, M. N. Ebelin and V. P. Andrade, In Situ DESI-MSI Lipidomic Profiles of Breast Cancer Molecular Subtypes and Precursor Lesions, Cancer Res., 2020, 1246–1257.
125. J. Zhang, S. Q. Li, J. Q. Lin, W. Yu and L. S. Eberlin, Mass Spectrometry Imaging enables Discrimination of Renal Oncocytoma from Renal Cell Cancer Subtypes and Normal Kidney Tissues, Cancer Res., 2020, 689–698.
126. K. Vijayalakshmi, V. Shankar, R. M. Bain, R. Nolley, G. A. Sonn, C. S. Kao, H. Zhao, R. Tibshirani, R. N. Zare and J. D. Brooks, Identification of diagnostic metabolic signatures in clear cell renal cell carcinoma using mass spectrometry imaging, Int. J. Cancer, 2020, 256–265.
127. M. Fala, V. Somai, A. Dannhorn, G. Hamm, R. Hesketh, D. Laurent Couturier, K. Gibson, A. J. Wright, J. Takats, J. Bunch, S. T. Barry, R. J. A. Goodwin and K. M. Brindle, Comparison of 13C MRI of hyperpolarized [1-13C]pyruvate and lactate with the corresponding mass spectrometry images in a murine lymphoma model, Magn. Reson. Med., 2021, 3027–3035.
128. Y. Song, Y. Zhang, S. Xie and X. Song, Screening and diagnosis of triple negative breast cancer based on rapid metabolic fingerprinting by conductive polymer spray ionization mass spectrometry and machine learning, Front. Cell Dev. Biol., 2022, 1075810.
129. L. Zhan, C. Liu, K. Qi, L. Wu, Y. Xiong, X. Zhang, J. Zang and Y. Pan, Enhanced imaging of endogenous metabolites by negative ammonia assisted DESI/PI mass spectrometry, Talanta, 2023, 123864.

---