Anal Bioanal Chem DOI 10.1007/s00216-015-8947-0 RESEARCH PAPER A novel direct spray-from-tissue ionization method for mass spectrometric analysis of human brain tumors Alexey Kononikhin 1,3,5 & Evgeny Zhvansky 1,3 & Vsevolod Shurkhay 4 & Igor Popov 1,2,3 & Denis Bormotov 1,3 & Yury Kostyukevich 1,3,7 & Sofiia Karchugina 1 & Maria Indeykina 1,2 & Anna Bugrova 2,5 & Natalia Starodubtseva 3,5 & Alexander Potapov 4 & Eugene Nikolaev 1,2,6,7 Received: 17 June 2015 / Revised: 28 July 2015 / Accepted: 30 July 2015 # Springer-Verlag Berlin Heidelberg 2015 Abstract Real-time feedback about dissected tissue during the neurosurgical procedure is strongly requested. A novel direct ionization mass spectrometric method for identifying pathological differences in tissues is proposed. The method is based on simultaneous extraction of tissue lipids and electrospray ionization which allows mass spectrometric data to be obtained directly from soft tissues. The advantage of this method is the stable flow of solvent, which leads to stable time-dependent spectra. The tissues included necrotized tissues and tumor tissues in different combinations. Capability for direct analysis of samples of dissected tissues during the neurosurgical procedure is demonstrated. Data validation is Electronic supplementary material The online version of this article (doi:10.1007/s00216-015-8947-0) contains supplementary material, which is available to authorized users. * Eugene Nikolaev ennikolaev@rambler.ru 1 Institute for Energy Problems of Chemical Physics of the Russian Academy of Sciences, Leninskij pr. 38 k.2, 119334 Moscow, Russia 2 Emanuel Institute for Biochemical Physics of the Russian Academy of Sciences, Kosygina st. 4, 119334 Moscow, Russia 3 Moscow Institute of Physics and Technology, Institutskij bystr. 9, 141700 Dolgoprudnyi, Moscow Region, Russia 4 Burdenko Neurosurgical Institute of the Russian Academy of Medical Sciences, 4-th Tverskaya-Yamskaya st. 16, 125047 Moscow, Russia 5 Research Center for Obstetrics, Gynecology and Perinatology, Oparina st. 4, 117997 Moscow, Russia 6 Orekhovich Institute of Biomedical Chemistry of the Russian Academy of Medical Sciences, Pogodinskaya st. 10, 119121 Moscow, Russia 7 Skolkovo Institute of Science and Technology, Novaya St., 100, 143025 Skolkovo, Russian Federation conducted by compound identification using precise masses from the MS profile, MS/MS, and isotopic distribution structure analysis. The method can be upgraded and applied for real-time identification of tissues during surgery. This paper describes the technique and its application perspective. For these purposes, other methods were compared with the investigated one and the results were shown to be reproducible. Differences in lipid profiles were observed even in tissues from one patient where distinctions between different samples could be poor. The paper presents a proof of concept for the method to be applied in neurosurgery particularly and in tissue analysis generically. The paper also contains preliminary results proving the possibility of observing differences in mass spectra of different tumors. Keywords Ambient mass spectrometry . Brain . Tumor . Diagnostics . Electrospray . FT ICR Introduction One of the greatest barriers to achieving optimal surgical results for infiltrative tumors in the human brain is the difficulty to distinguish the tumor from intact tissues [1]. Tumors can closely resemble the healthy brain, often infiltrate brain tissue, and may be located near to the eloquent areas of the brain. Although gross total tumor resection leads to an increase in the overall and progression-free survival, in many cases, tumor margins cannot be clearly defined. Inadequate or very aggressive surgery leads to higher risks of postoperative complications and a possible decreased quality of life for the patient. A variety of mapping techniques has been developed to be used in brain tumor surgery; however, in most cases in clinical practice, determination of tumor borders is not done in a realtime mode. As for now, magnetic resonance imaging E. Zhvansky et al. techniques are of both diagnostic [2] and operative navigational [3] use. These techniques are routinely used at the preoperational stage, and only with limited success in intraoperational variants [4], mostly because real-time feedback is still not fully developed [5]. Other clinically relevant methods include fluorescent labeling [6–9], whose limited use is mostly restricted to certain application areas, caused by specificity and sensitivity issues [10], and intraoperative ultrasound [5], which is not applicable if tumor density is close to the density of the brain. Standard histopathological methods are the gold standard for detecting tumors and can provide diagnostic information during surgery within half an hour, but are usually limited to the evaluation of one or a very small number of segregated samples from each operation [1]. There are many ideas for solving the problem of the rapid distinguishing of healthy brain tissue from a brain tumor, some of which were tested on rodents [11], but none of them are currently clinically applied. One of the approaches for such differentiation is by comparing their lipid profiles [12]. For example, gangliosides are used to differentiate astrocytoma tissue from the surrounding tissues and normal control brain tissues [13, 14], or lipid profiles are used for the determination of tumor borders [15, 16]. An operative profiling method for molecular profiling can be created based on direct ionization mass spectrometry methods. Analysis of tissues during tumor removal is one of the challenging problems for modern MS, but despite the fact that many technical difficulties hinder the success of molecular profiling in surgery, many mass spectrometrybased methods have been recently suggested for intraoperative molecular diagnostics and differentiation of human brain tumors. Desorption electrospray ionization mass spectrometry (DESI MS) [17], matrix-assisted laser desorption ionization (MALDI) [18], rapid evaporative ionization mass spectrometry (REIMS) [19], and sonic spray ionization (SSI) [20] all yield informative lipid and/or protein profiles for differentiating tumors, while potentially allowing real-time feedback during the operation. However, these methods include a rather long sample preparation procedure, which makes the creation of a real-time feedback system rather difficult. With direct ionization from tissue, the period between tissue dissection and sample analysis is significantly reduced. Besides this, some compounds of the given tissue cannot be detected by standard mass spectrometric methods such as DESI, ESI, or leaf spray but can be detected by direct from tissue ionization [21]. Also, brain tumor tissue cannot be shaped into a required form and then fixed in these conditions; thus, it cannot form a tip, which is necessary for direct ionization by the leaf spray method. In this paper, a modified spray-from-tissue ionization method, basing on an idea initially proposed by Liu et al. [22], is discussed. In this method, a small bit of the tissue is placed on a standard medical injection needle through which the solvent is delivered. Application of a high voltage to the needle creates an electric field strong enough to initiate an electrospray directly from the sample. The technical novelty of the proposed method is in the absence of an irregular external solvent flow (as it was in the Liu et al. paper [22]), which is now continuously pumped to the sample from the needle inside. Thus, in the suggested method, the problem of sample drying is solved. This ionization source is very similar to the Bpaper spray^ [23] and Bleaf spray^ [24]. The spray-from-tissue ionization method supposes the presence of a Btip^ of tissue, as it was in the Bleaf^ [24] spray or a tip of something containing the adsorbed sample, as it was in the Bpaper^ [23] spray. So there existed a problem of obtaining a stable spray from soft tissues, which cannot form a tip, which is solved in this work. Described, for example, in the paper of Hu et al. [21], direct ionization from soft tissue has some complications with spectrum registration time due to the sample’s sharp end or lack of stable solvent flow and corresponding sample drying. Also, brain tumor tissue cannot be formed into a tip of a fixed shape. Methods Materials All tissues were provided by the Burdenko Neurosurgical Institute. Pathological tissues were dissected during the neurosurgical procedure and divided into two parts: the first one was sent for routine examination by a pathologist (including the immunohistochemical staining), and the second was frozen in normal saline and sent to the laboratory for mass spectrometric analysis. The experimental tissues were dissected during only one neurosurgery (dissection of glioblastoma WHO grade IV). Tissues included necrotic tissue with necrotized vessels, necrotic tissue with tumor stain, tumor with necrosis (with a predominance of tumor tissue), tumor, necrotized tumor (with a predominance of necrotic masses), and parts of tumor cells. Other samples of tissue from other neurosurgeries were histochemically identified as astrocytoma, meningioma, and neurinoma. The research has been presented to the ethical board of the Burdenko Neurosurgical Institute and was approved. Tissue electrospray ionization A lab-built prototype, based on electrospray with simultaneous lipid extraction, was used as the ion source: an injection needle (30 mm in length, 0.6 mm inner diameter KD-FINE disposable hypodermic needle (KD Medical GmbH Products, Germany)) was mounted on a holder platform adjacent to the inlet interface. A freshly thawed biopsy tissue was cut into approximately 2-mm3 homogeneous samples, and these slices were imposed on the needle point for analysis. As schematically shown in Fig. 1b, the sample was placed on the tip of a A novel direct spray-from-tissue ionization method Fig. 1 (a) The proposed ionization source interface: a sample of human brain tissue is placed on a regular medical needle. (b) Scheme of the ion source used for substance extraction in direct electrospray ionization standard medical injection needle. No modifications of the needle are required. The tip of a standard medical needle has a cutoff plane on the inner side onto which a bit of the sample can be put using forceps. Methanol (HPLC grade), purchased from Merck, was used as the spray solvent for MS acquisition. The solvent was pumped through a fused silica glass capillary into the needle by a syringe pump. The solvent flow was 4±1 μl/min; the precise value depended on the shape of the sample; higher flows resulted in a non-stationary Taylor cone and periodical break-off of large droplets of solvent, while lower flows resulted in the disappearance of a visible Taylor cone and spray instabilities. Before turning the electrospray high voltage (HV; 5.1 kV in positive mode and 6.0 kV in negative mode; lower voltages resulted in spray instabilities, while higher ones produced a corona discharge and significant changes in the mass spectrum) on, the solvent inlet system was used to wet the sample and to ensure that it does not move from the needle due to surface charges. The tissues contacted with the solvent only right before and during MS measurements. The solvent flowed around the tissue for enough time for surface molecule extraction to take place, and formed an electrospraying Taylor cone directed at the mass spectrometer inlet capillary (Fig. 1b). High voltage was applied to the solvent flow and, subsequently, to the whole needle and its tip. The spray stability was kept by means of precise needle position adjustment and was observed with a PULNiX zooming camera. Spray stability was evaluated by visual estimation of a stable Taylor cone formation as well as by recording the total ion current (TIC). Reproducibility of each mass spectrum continuously from scan to scan was observed by TIC variability, which was in range of no more than 15 %. Solvent flow was constant during spectrum registration. Depending on the sample shape, fine adjustments of the flow rate and voltage were required. The spray appeared on the end of the injection needle tip, where the intensity of the electrical field was high enough to make ionized droplets. Thus, the method does not use a manually created tissue tip (as well as Liu et al. [22]), which would be E. Zhvansky et al. Fig. 2 CID spectrum of PC(34:1)+Na+. The m/z=723.38 peak corresponds to a (N(CH3)3) neutral loss, while m/z=599.39 and 577.49 peaks correspond to choline phosphate neutral loss with the sodium substitute being either on the product ion or neutral loss, respectively A novel direct spray-from-tissue ionization method very difficult to make from a soft sample; rather, the solvent with the desorbed tissue itself forms a tip as it moistens the needle. An ion cyclotron resonance (ICR)-based high-resolution mass spectrometer Thermo Finnigan LTQ FT Ultra equipped with a 7T superconducting magnet and controlled by the XCalibur 2.0.7 software (Thermo Fisher Scientific, Waltham, MA USA) was used to obtain tandem mass spectra. The samples were analyzed in both positive and negative modes (m/z 200–2000, resolution 56,000 at m/z 800). To prove the stability of the proposed ionization method, a series of experiments was carried out and the major compounds identified (as described below) in the spectra of each tissue sample were compared. Collision-induced dissociation Experiments were performed and recorded in an automatic two-step data-dependent manner. The first step was to obtain FT ICR mass spectra in the whole mass range (m/z=100– 1000). Then the peaks were identified by MS/MS. On this second step, the five most intense peaks found in the full mass spectrum were isolated (isolation width = 2.0), fragmented by collision-induced dissociation (CID) (normalized collision energy=35.0), and then dynamically excluded from fragmentation candidates for the next 30 s. Mass spectra of the fragments formed by CID during this second stage were measured in the LTQ. Peaks identification The most intensive peaks were identified by their exact mass with the Lipid MS Predict tool, by CID MS/MS (see example on Fig. 2), and by their isotopic distribution with XCalibur 2.0.7 QualBrowser isotopic simulation (mostly to distinguish the [M+Cl]− ions). Results and discussions MS lipid profile analysis A fast head on analysis of the data allowed to identify 17 lipids in the negative mode and 14 lipids in the positive mode (see Tables 1 and 2). The most abundant ions observed in different samples belong to the following lipid classes: ceramides (Cer), plasmenyl phosphatidylethanolamines (PE-O), plasmenyl phosphatidic acid (PA-O), phosphatidylcholines (PC), phosphatidylserines (PS), galactosylceramides (GalCer), glycerolipids (GL), and phosphatidylinositols (PI) (Figs. 3 and 4). A strong coincidence between major peaks in spectra from different bits of one tissue was observed (Fig. 5). A limited coincidence between the lipid profile of Liu et al. and ours also took place, Table 1 Negative ion description and comparison with the DESI method. B+^ means that the lipid was denoted in the article [17] by Eberlin et al., while B−^ means it was not Mass Molecular species DESI [17] 916.6652 ST(26:1) + 888.6335 ST(24:1) + 885.5599 PI(38:4) + 844.5999 834.5369 810.5364 PS[40:1] PS(40:6) PS(38:4) − + + 794.5521 788.5502 PC(34:1)+Cl PS(36:1) − + 768.5384 PC(32:0)+Cl − 750.5497 PE-O(38:5) + 737.5422 726.5502 PA-O(40:4) PE-O(36:3) − + 682.5950 Cer[42:2]+Cl − 629.4952 600.5157 598.5006 GL(34:1) Cer[36:1]+Cl Cer[36:2]+Cl − − − 572.4846 Cer[34:1]+Cl − although, unfortunately, the results of the presented paper cannot be directly compared with the results of Liu et al. because the investigated tissues were different. The positive ions were observed in their protonated or sodiated forms, and the negative ions were either deprotonated or chlorinated. It should be mentioned that some lipids (specifically PC(32:0) and PC(34:1)) were present (and abundant) in both positive and negative modes, which gave an Table 2 Positive ion description and comparison with the DESI method. B+^ means that the lipid was denoted in the article [17] by Eberlin et al., while B−^ means it was not Mass Molecular species DESI [17] 832.6788 832.6016 818.6132 810.5914 808.6016 804.5700 790.5948 782.5803 780.5690 768.6064 760.5977 756.5638 754.5489 750.5549 GalCer(d42:2)+Na PC(38:4)+Na PC-O(38:4)+Na PC(36:1)+Na PC(36:2)+Na PC(36:4)+Na PC-O(36:4)+Na PC(34:1)+Na PC(34:2)+Na PC-O(34:1)+Na PC(34:1) PC(32:0)+Na PC(32:1)+Na PE-O(36:3)+Na + − − + + − − + − − − + + + E. Zhvansky et al. Fig. 3 A typical negative ion spectrum of brain tumor particles in the m/z range from 500 to 1040. Some of the most abundant peaks are annotated as the corresponding lipid species. A Negative ion spectrum, B positive ion spectrum opportunity to confirm the relevancy of the results—there was no contradiction between the peak intensities measured both in positive and negative modes. About 190 peak identifications are presented in Table S3 in the Electronic Supplementary Material (ESM). Discussion Fluorescent labeling is not applicable in some cases due to method limitations [10]. Ultrasound also has limitations caused by the specificity of tumor tissue, for example density [5]. And only mass spectrometry can give specific information on the molecular composition that can not only describe the differences in molecular profiles of normal and tumor tissue but also provide information about tumor borders by the gradient of tumor markers. The whole process of sample preparation and analysis takes about 1 min. The sample preparation takes about 30 s (including sample cutting, placing it on the injection needle, and also mass spectrum obtaining). Another 30 s is used for manual spectrum decoding from the raw file and getting a decoded spectrum for automatic processing. This method can actually be used in neurosurgery as a rather fast method, since the common method for brain tumor analysis during neurosurgery is immunohistochemistry, which is much more time consuming. The proposed method presents systematical advantages for tissue analysis in comparison with other methods. BPaper spray^ [23] can only be used for absorbed in paper substances; thus, it has long-term sample preparation. BLeaf spray^ [24] requires a tip to be formed from the investigated tissue, but it is impossible to form a fixed tip out of brain tumor tissue, and this would also lead to long-term preparation. The advantage of the investigated method of ionization in comparison to the method suggested by Liu et al. [22] is in the constant flow of solvent that prevents the sample from drying and in the long duration of a stable ionization process. Liu’s A novel direct spray-from-tissue ionization method Fig. 4 Examples of lipid classes detected in brain tumor samples [22] method allows stable ionization for only 30 s, while the method described in this paper allows to obtain spectra for more than 30 min without significant changes in the registered signal. For this moment, the suggested method has not been used for determining tumor margins; only the possibility to distinguish the different pathological tissue types, listed in Table 3 and ESM Table S1, has been shown. Also, the presence of lipids in different tumor types in negative mode is reflected in Table S2 in the ESM. Distinguishing different tumor types by their lipid profile was not the main purpose of the work, but this somehow reflects the potential of the method. Some lipids despite their affiliation to different lipid classes can produce ions that are very close on the m/z scale. Therefore, FT ICR high-resolution mass spectrometry is more suitable and provides more information for tissue profiling than low-resolution approaches, like triple quadrupole or quadrupole ion traps, typically used for mass spectrometric analysis of lipids [25, 26]. For example, positive ions of GalCer(d42:2)+Na+ (m=832.6788) and PC(38:4)+Na+ (m= 832.6016) have a sufficient mass difference to be distinguished in our experimental setup (see Table 2). The proposed method allowed to identify more than just the major components of the mass spectra but also lowintensity peaks of the minor components (just a brief analysis gave us information on about 30 compounds in the m/z range 100–1000). The results on the lipids observed in our experiments were compared with other works, for example Eberlin et al. [17], who used DESI MS, and Balog et al. [19], who used REIMS, for experiments with brain tumor tissue. Only about half of the identified lipids (plasmenylPE, PS, GalCer, PI) were previously mentioned by Eberlin et al. in their research (see Tables 1 and 2), and only one major peak at m/z=885.56 was found to be the same as the results of Balog et al. [19], which may be due to the completely different ionization methods used in these studies. There is no opportunity to compare the presented results with those obtained by different methods of direct ionization mass spectrometry. Thus, methods which are most close to direct ionization were selected for comparison. The proposed method of ionization gives reproducible results: spectra of the different tissue samples consisted of the same major peaks with similar isotopic distributions during E. Zhvansky et al. Fig. 5 A Negative ion spectra of two necrosis (a, b) and two tumor (c, d) samples. B Positive ion spectra of two necrosis (a, b) and two tumor (c, d) samples the whole measuring time (see Table 3, ESM Table S1). TIC fluctuations never exceeded 15 % of the average TIC value during the experiment. What is more important, there were no significant changes in the spectra during the registration of one sample. As shown, for example, by Köhler et al. [12], gliomas and healthy brain tissues are characterized by different lipid profiles. Thus, some of these common major components may be associated with the human brain tumor (Fig. 5). Ion Table 3 chromatograms for each identified lipid were obtained and are shown in Figures S1 to S31 in the ESM. The average number of analyzed samples of each tissue was about 7 and was limited by the size of the initial biopsy material, since it was cut into samples of about 2 mm3 each. The samples analyzed in the proposed experimental setup are rather small, which means that there is no need to take large samples of yet unknown tissues for analysis, thus providing Tissue composition Tissue no. Tissue histology description Peaks in tissue Reproducibility (number of observing of lipid classes in different tissue samples/ number of tissue samples) 1 2 3 4 5 6 Necrotic tissue with necrotized vessels Necrotic tissue with tumor stain Tumor with necrosis Tumor Necrotized tumor Parts of tumor cells PE-O, PA-O, PC, PS, Cer, PI, ST PE-O, PA-O, PC, PS, Cer, PI PE-O, PA-O, PC, PS, Cer, PI, ST PE-O, PA-O, PC, PS, Cer, GL, PI, ST PE-O, PA-O, PC, PS, Cer, PI PE-O, PA-O, PC, PS, Cer, PI, ST 100 100 100 100 100 100 % (9/9) % (3/3) % (6/6) % (11/11) % (9/9) % (8/8) A novel direct spray-from-tissue ionization method high spatial resolution for clinical use and reducing the risk of dissecting non-pathological tissue and operative trauma. The presented method has some advantages over the Bpaper^ and Bleaf sprays.^ Since the injection needle has a tip, the problem of high local electrical field tensity for spray production is solved by applying high voltage to the needle. This method allows to measure soft tissues, while Wang et al. [23] and Liu et al. [24] in their experiments measured only relatively hard tissues, which were simply reshaped to a sharp form. The advantage of the presented method is the simplicity of gaining a stable spray directly from soft tissue. 3. 4. 5. 6. 7. 8. 9. Conclusion A new method for soft tissue ionization is presented. It is rather close to DESI, and other electrospray methods, but has several considerable advantages such as ionization of soft samples and a directed spray flow. Stable ionization and consequently stable mass spectra were obtained. Partial compliance between our results and those from the literary data was demonstrated. Most of the major peaks and corresponding compounds are consistent with those obtained by a close direct ionization mass spectrometric method. This gives us the reason to believe that this method may be used as a convenient way of ionization in some cases, since according to the latest publications there is a great interest and potential in the use of direct spray-from-tissue ionization methods for intraoperative monitoring of the human brain tumor resection operations. Differences in lipid profiles were observed even in tissues from one patient where distinctions between different samples could be miserable. These results were obtained from one patient and have no control of healthy brain tissue, but will have one in the future. Therefore, obtained results are a proof of concept for the tissue express-analysis method. The paper also contains preliminary results proving the possibility of observing differences in mass spectra of different tumors. Acknowledgments The work was supported by the Russian Science Foundation grant no. 14-24-00114. Author contributions The manuscript was written through the contribution of all the authors. All the authors have given their approval to the final version of the manuscript. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Competing financial interest The authors declare no competing financial interest. 23. References 1. 2. Santagata S et al (2014) Intraoperative mass spectrometry mapping of an onco-metabolite to guide brain tumor surgery. Proc Natl Acad Sci U S A 111(30):11121–11126 Singhal S et al (2010) Nanotechnology applications in surgical oncology. Annu Rev Med 61:359–373 24. 25. 26. Machi J et al (2004) Intraoperative ultrasound. Surg Clin North Am 84(4):1085–1111 Schulder M et al (2003) Intraoperative magnetic resonance imaging: impact on brain tumor surgery. Cancer Control 10(2):115–124 Prada F et al (2014) Fusion imaging for intra-operative ultrasoundbased navigation in neurosurgery. J Ultrasound 17(3):243–251 Goriainov SA et al (2014) Intraoperative fluorescence diagnostics upon recurrent operations for brain gliomas. Zh Vopr Neirokhir Im N N Burdenko 78(2):22–31 Goriainov SA et al (2012) Photodynamic diagnosis and laser biospectroscopy as an intraoperative guidance in neurosurgery. Zh Vopr Neirokhir Im N N Burdenko 76(6):57–65 Potapov AA et al (2013) Intraoperative combined spectroscopy (optical biopsy) of cerebral gliomas. Zh Vopr Neirokhir Im N N Burdenko 77(2):3–10 Potapov AA et al (2012) Intraoperative fluorescent visualization and laser spectroscopy in intrinsic brain tumor surgery. Zh Vopr Neirokhir Im N N Burdenko 76(5):3–11 Jacquesson T et al (2013) Surgery of high-grade gliomas guided by fluorescence: a retrospective study of 22 patients. Neurochirurgie 59(1):9–16 Beljebbar A et al (2009) Monitoring of biochemical changes through the c6 gliomas progression and invasion by Fourier transform infrared (FTIR) imaging. Anal Chem 81(22):9247–9256 Köhler M et al (2009) Characterization of lipid extracts from brain tissue and tumors using Raman spectroscopy and mass spectrometry. Anal Bioanal Chem 393(5):1513–1520 Zamfir AD et al (2013) Profiling and sequence analysis of gangliosides in human astrocytoma by high-resolution mass spectrometry. Anal Bioanal Chem 405(23):7321–7335 Zamfir AD (2014) Neurological analyses: focus on gangliosides and mass spectrometry. Adv Exp Med Biol 806:153–204 Eberlin LS (2012) Classifying human brain tumors by lipid imaging with mass spectrometry. Cancer Res 72(3):645–654 Eberlin LS (2013) Ambient mass spectrometry for the intraoperative molecular diagnosis of human brain tumors. Proc Natl Acad Sci U S A 110(5):1611–1616 Eberlin LS et al (2010) Discrimination of human astrocytoma subtypes by lipid analysis using desorption electrospray ionization imaging mass spectrometry. Angew Chem Int Ed Engl 49(34):5953– 5956 Schwartz SA et al (2005) Proteomic-based prognosis of brain tumor patients using direct-tissue matrix-assisted laser desorption ionization mass spectrometry. Cancer Res 65(17):7674–7681 Balog J et al (2013) Intraoperative tissue identification using rapid evaporative ionization mass spectrometry. Sci Transl Med 5: 194ra93 Schäfer KC et al (2011) Real time analysis of brain tissue by direct combination of ultrasonic surgical aspiration and sonic spray mass spectrometry. Anal Chem 83(20):7729–7735 Hu B et al (2012) Direct ionization of biological tissue for mass spectrometric analysis. Analyst 137(16):3613–3619 Liu J et al (2011) Biological tissue diagnostics using needle biopsy and spray ionization mass spectrometry. Anal Chem 83(24):9221– 9225 Wang H et al (2010) Paper spray for direct analysis of complex mixtures using mass spectrometry. Angew Chem 122(5):889–892 Liu J et al (2011) Leaf spray: direct chemical analysis of plant material and living plants by mass spectrometry. Anal Chem 83(20):7608–7613 Isaac G et al (2007) New mass-spectrometry-based strategies for lipids. Genet Eng (N Y) 28:129–157 Shaner RL et al (2009) Quantitative analysis of sphingolipids for lipidomics using triple quadrupole and quadrupole linear ion trap mass spectrometers. J Lipid Res 50(8):1692–1707