SpotOn: A Tool to Enhance Echo Acoustic Standard Deposition for Quantitative Mass Spectrometry Imaging

Abstract

The quantitative evaluation of therapeutics in the drug discovery pipeline is essential for elucidating their pharmacokinetic and pharmacodynamic profiles within complex biological systems. Traditionally, autoradiography and liquid chromatography-mass spectrometry (LC-MS) have been employed for such analyses. However, matrix-assisted laser desorption ionization mass spectrometry imaging (MALDI-MSI) is emerging as a promising alternative, offering a more rapid untargeted workflow. Quantitative MALDI-MSI (qMSI) relies on accurate calibration to achieve robust and reproducible quantification of analytes. Here, we present an open-source software tool, SpotOn, designed to facilitate the discrete deposition of low-volume calibration standards onto glass slides or tissue sections. By enabling precise placement and significantly reducing the volume of standards required, our tool affords the generation of more calibration points across a wider dynamic range, thereby improving the robustness and accuracy of quantitative analyses. Software was validated using JQ1 on both glass slides and murine liver tissue. The resulting calibration curves demonstrated strong linearity, with coefficients of determination (R ²) exceeding 0.92. This tool streamlines the qMSI workflow, conserves valuable reagents, and enhances analytical reliability, making it a helpful resource.

Introduction

The drug discovery pipeline is a dynamic, intricate, and resource-intensive process. To identify novel therapeutics, both target-based and phenotypic drug discovery strategies are used, focusing on specific molecular targets or desired biological effects, respectively. , Understanding how drugs distribute and interact within tissues is critical to drug development. Innovative imaging technologies have provided new insights into the effects of therapeutics within tissues for pharmacokinetic-pharmacodynamic (PK/PD)-based studies. Mass spectrometry imaging (MSI) offers a unique opportunity to visualize drug distributions alongside additional molecular classes with high chemical specificity and spatial resolution.

Using mass spectrometry in this modality enables the simultaneous generation of thousands of annotated chemical heat maps across the two-dimensional space of many sample types. Different ionization techniques may be used for MSI; the most common, termed matrix-assisted laser desorption/ionization (MALDI), was first described in 1997 for the analysis of peptides and proteins within tissue samples. During MALDI-based imaging, a laser scans across the sample surface coated with a crystalline matrix, providing a mass spectrum for each pixel. Peaks in these spectra can be used to deduce a known molecular formula, and their intensities to generate a molecular heat map. Since its introduction, the technique has evolved significantly, offering improved sensitivity, spatial resolution, and data acquisition speeds. ,

Building on the capabilities of MALDI, quantitative MSI (qMSI) has emerged as a powerful approach for measuring the relative or absolute abundance of specific molecules, such as therapeutics, within complex samples. In drug discovery, profiling drug metabolism and pharmacokinetic characteristics represents a critical step in the early phases of development. While traditional methods, such as whole-body autoradiography, often coupled with chromatography-based mass spectrometry for validation, have been widely used, they are limited by the need for labor-intensive preparation of radiolabeled compounds and homogenized tissues. In contrast, qMSI offers a label-free, spatially resolved, and time-efficient alternative, enabling direct visualization and quantitative analysis of drug distribution in situ without the extensive preparation required by conventional techniques. − Although outside the scope of this technical note, additional references are included to highlight the importance of sample preparation, including factors such as the effects of MALDI crystal size, on the analysis of qMSI samples. −

Herein, we present the utilization of Echo acoustic liquid handlers for the discrete deposition of standards chosen using a Java-based tool integrated with ImageJ. We further demonstrate the use of the qMSI technique when applying standards to glass slides and preclinical tissue sample sections for intravenously dosed drugs to generate calibration curves and complete quantitation experiments.

Results and Discussion

SpotOn Software

SpotOn is a Java plugin designed for use with ImageJ. , Functionally, SpotOn aims to provide reasonable control over acoustic spotting at array or discrete positions: the user clicks a point in a suitable target image, and the software maps that point in the image to the corresponding coordinates in the grid used by the Echo acoustic liquid handler. The software adds the selected coordinates to a comma-separated values (CSV) file when the plugin run completes. Additional interface elements are included to allow the user to adjust the plate well source and comments as they add positions, but all that is required is a linear coordinate transformation to build the list of Echo points to be targeted. The challenge for the plugin is to determine the precise coefficients of that linear transformation.

For a successful conversion from image coordinates to dispenser coordinates, the plugin must determine the slide holder’s position in the original scanned image. This orientation step is accomplished through a series of image-processing steps, designed to be as flexible and robust as possible given the wide range of noise and error sources in the application. The algorithm is compatible with slides laser-etched with custom letter-fiducial patterns (Figure S1 in the Supporting Information) and with Bruker Intellislides. Additionally, it can be applied to unmarked slides, although this is not recommended and should be performed with caution. If the holder, scanner, and arrangement are sufficiently similar to the conditions under which the plugin was developed, the slide holder can be identified directly; alternatively, the more robustly identifiable slide markings can be used to infer the slide holder’s overall position. The plugin can be run with all slide slots occupied or a subset, including the extreme case of a single slide.

To ensure that relevant image elements, such as slides and slide holders, are reliably and robustly identified, the plugin uses multiple fiducials or image elements to estimate object locations, then employs an iterative consensus algorithm to down-weight noisy or inconsistent estimates. This allows the algorithm to robustly locate key elements (essential for subsequent calculations) even when one or more markers or fiducials are missing, obstructed, or obscured.

Evaluation of Standard Deposition to Glass Slides

To critically evaluate the deposition of standards onto glass slides, studies of the generated patterns and calibration curves were conducted. The first set of analyses used white paint on glass microscopy slides to ensure that specified patterns could be deposited as 2.5 nL droplets without merging. For this experiment, DMSO containing 2 mg/mL cresyl violet was used to visualize droplets. As shown in Figure , tightly spaced droplets can be successfully deposited onto slides at a volume of 2.5 nL. There is a shift of all patterns from the perceived to the actual location onto the slides. This alteration may be due to the scanned images being taken from the “back” of the glass slide while target deposition is completed on the “top” of the slide. Calculations were included in the final plugin version to mitigate refraction-related errors. However, manual inspection indicates that this systemic shift does not significantly affect the inherent variability of the deposition process.

1.

Evaluation of spatial patterns deposited on slides using SpotOn software in conjunction with Echo acoustic deposition of spots. (A) Red spots indicate the deposition locations chosen by users in the software. The yellow arrows on slide 1 indicate the white paint markings used for reference, which are replicated on subsequent slides. (B) Slide image taken after deposition of spots onto slides with white paint boundaries. (C) Slide image after removal of white paint marks to show deposition locations and patterns. Slide images (B) and (C) were inverted in color and white-balanced to enhance droplet visibility.

For MALDI MSI experiments, spots were deposited on indium tin oxide (ITO) glass slides, and this round of samples showed inconsistent droplet patterns (Figure S2). Simple washing of the ITO slides with a mixture of LC-MS-grade hexanes and methanol alleviated this consistency problem. After depositing JQ1 solutions at varied concentrations, which also contained cresyl violet, the samples underwent a typical MALDI sample preparation workflow. Since the choice of instrumentation for qMSI is not a critical issue, more attention was devoted to developing spectrometer methods targeting ions specific to JQ1, thereby maximizing sensitivity. A similar approach using other instrument configurations is recommended.

When not commercially available, molecules like JQ1 may be affordably labeled using one-step synthetic methods, taking advantage of readily exchangeable protons that are outcompeted in heavy water (D₂O) or solvent (CD₃OD) in the presence of an acid or base catalyst. Since JQ1 has a methylene adjacent to an ester, we leveraged this moiety to produce a doubly deuterated standard. Optimal qMSI data were then collected by including this heavy-atom-enriched D₂-JQ1 as an external standard in the workflow.

Using the protonated JQ1 signal, the SpotOn output image was coregistered with the MALDI MSI data to assess spotting accuracy (Figure ). Although deviations from the initially selected spotting location were observed, the nonuniform variations in spot placement suggest that these deviations were most likely attributable to the Echo acoustic deposition process.

2.

Calibration curve of JQ1 spotted onto a glass slide using SpotOn software. Coregistration of output locations with MSI data for protonated JQ1 (m/z 457.144 Da) was completed in MSiReader using the coregistration tool, fitting the points to the two locations indicated by arrows. Some deviations are observed between the chosen spotting locations and the actual images obtained by MALDI MSI.

Calibration Curve Generated on Glass Slides

Using the previously spotted samples to evaluate spotting locations, a calibration curve was generated (Figure ). The calibration curve for the protonated JQ1 signal was generated using 8 standard solutions ranging from 0.92 to 3.66 μg/μL (2006–8019 μM), with each measurement repeated 7 times. While slight variations in response were observed, no outliers were identified in the data set. Linear regression yielded an R-squared of 0.9430, indicating strong linearity within the calibration range (Figure A). The observed signal intensity for each spot is shown in Figure B.

3.

Calibration curve generated on ITO glass slides for protonated JQ1 species (457.144 Da) normalized to the internal standard of deuterated JQ1 (459.159 Da). (A) Generated calibration curve for JQ1 from 8 concentrations (n = 7). (B) MSI image generated for protonated JQ1 (XY units: mm).

Utilization of the Tool for Deposition to Tissues

To determine whether matrix effects from tissue changed the accuracy of both deposition and calibration curve generation, JQ1 standards were spotted onto sections of murine liver and brain tissue. In contrast to the spotting on glass slides, spots were shifted farther from the anticipated tissue than on plain ITO glass slides. After careful experimentation, this shift was found to result from taking images from the back of slides for spot determination, rather than from the top of the glass where samples are deposited. As shown in supplemental Figure S3, using the etch marks on the back of the slide to direct the laser based on the top of the slide resulted in a discrepancy between the actual and perceived laser locations. However, as the deposited spots were primarily resolved and all on top of the tissues, the experiments proceeded to evaluate calibration curves generated on the tissue.

For the calibration curve, three replicates of eight standards were spotted at various locations on the murine brain section (Figure ). As shown in Figure A, while the mean response of the protonated JQ1 peak shows some variation, strong linearity was observed over the range of 0.90–3.68 μg/μL (1978–8053 μM), with only a mild loss of goodness of fit compared to spotting onto ITO slides.

4.

Calibration curve for protonated JQ1 generated on control murine brain tissue section. (A) Calibration curve for eight concentrations (n = 3). (B) MSI results for protonated JQ1 species (457.146 Da). (C) Blended image of JQ1 (Red) and putative lipid species (Blue, 772.525 Da) to show drug distribution on tissue section. Figures (B) and (C) are normalized to the deuterated JQ1 signal (459.159 Da) and obtained at a 20 μm spatial resolution on the FT-ICR MS. (XY units: mm). Additional images of liver tissues from the experiments are shown in Figures S4 and S5.

To compare our Echo acoustic deposition spotting methods with a manual process, serial sections of brain tissue were prepared using a single calibrated 50 nL pin held by hand to deposit identical volumes of DMSO solutions via capillary action. Compared with the Echo deposition, which allowed three replicates of calibration curves (n = 24 spots) to be dropped reproducibly, it was more difficult to maintain precision and define more than one calibration curve per tissue using manual pipetting. While the same low dead volume plate used for the Echo experiment could be used to prepare manual pipetting standards on tissue, a higher volume of standard solutions was needed for a set of eight locations (20 nL as opposed to 400 nL), with additional wash solutions of DMSO required to clean the pen tool between runs. The relative sizes of spotted solutions from Echo and manual deposition were calculated using MSiReader. While the relative standard deviation between spots was similar in variation (7.9% for manual and 10.8% for Echo), the spot sizes for manual spotting with higher volumes were approximately twice the size of the Echo deposition, making it more difficult to obtain multiple calibration curves per tissue section (0.38 ± 0.03 mm² for manual as opposed to 0.21 ± 0.02 mm² for Echo). Further comparison of the calculated calibration curves for Echo deposition using SpotOn versus manual deposition revealed a higher degree of linearity and goodness of fit for Echo deposition (R ² = 0.93) compared with manual deposition (R ² = 0.72). This difference may be attributed to the increased statistical power afforded by averaging more data points in the calculations. Evaluation of reproducibility between spots in Echo deposition across triplicate runs showed an increasing relative standard deviation at lower concentrations, as the instrument approaches the limit of detection. MSI images of manual pipetting of standards, along with RSD values for concentration curves using the Echo and spot sizes for both methods, can be found in Tables S1 and S2 and Figure S6.

Conclusions

We have created a free software tool to enhance qMSI workflows by depositing standards at discrete locations. The preliminary studies presented demonstrate the software’s utility for generating calibration curves on both glass and tissue samples, which can be used to estimate the putative concentrations of analytes of interest. This tool provides an alternative approach to traditional qMSI that can be completed faster than generating tissue mimetics.

Previous qMSI studies have used strategies such as tissue-mimetic models and depositing standards onto tissue sections or glass slides with pipettes or syringes. In the case of tissue mimetics, this requires a larger amount of tissue and more time to generate these models. As the tissue mimetics homogenize the tissue with the standard of interest, this method is limited by the requirement to evaluate one analyte at a time. While spotting standards onto tissue sections reduces the amount of tissue required, most depositions from manual methods yield large volumes of material, resulting in only one calibration curve per tissue. Using SpotOn software, we can minimize the need for additional tissues and deposit multiple calibration curves using low-volume depositions. Increasing the number of replicates enhances the statistical robustness and reliability of calibration curves generated on the tissue of interest. Additionally, the ability to use nanoliter amounts of standards conserves laboratory resources.

Using the SpotOn application, we generated calibration curves on glass slides in specified patterns and on tissue sections, achieving high linearity (R ² > 0.92). Producing calibration curves in either experimental context provides distinct advantages and addresses different sources of error. While calibration curves on glass slides increase the number of analysis spots, they do not account for the background tissue matrix observed during tissue imaging. Yet, with the application of standards to glass, the higher number of spots reduces variability associated with hot spots and variations in crystal formation due to the MALDI matrix. , Conversely, applying calibration curves to control tissues demonstrates the technique’s ability to account for suppression by tissue matrix signals while reliably detecting drug-specific signals, thereby quantifying drug standard concentrations in tissue samples with lower spatial variability. Ultimately, while both methods are valuable for method development and validation, tissue-based calibration curves are more relevant for drug discovery applications, as they better reflect the conditions encountered in real-world sample analysis.

Looking ahead, the adoption of this software tool could facilitate broader applications in quantitative imaging, including studies involving diverse tissue types, analyte classes, and disease models. Furthermore, although the present study used MALDI as the ionization source, the workflow and software tool described herein are broadly applicable and may be readily adapted to alternative ionization modalities in MSI.

In summary, our workflow represents a unique advancement in qMSI methodology, offering an efficient, versatile, and scalable approach for quantitative imaging studies. By reducing experimental timelines, conserving resources, and improving analytical performance, this tool aims to accelerate research and enhance the impact of MSI in drug discovery and development.

Materials and Methods

Animal Models

Animal experiments and collection of materials in this study were conducted according to protocols approved by the Institutional Animal Care and Use Committee (IACUC) of St. Jude Children’s Research Hospital in accordance with Principles of Laboratory Animal Care. Nonperfused liver samples were collected from a female 17–18-week-old C57BL/6 mouse. Immediately after collection, samples were frozen using liquid nitrogen vapor, then immersed in liquid nitrogen and stored at −80 °C until sectioned on a cryostat.

Preparation for Internal Standard

JQ1 was obtained from a catalog (MedKoo Biosciences Inc., Cat. No: 40669), and deuterium exchange was performed under basic conditions to yield a highly enriched D₂-JQ1, followed by standard normal-phase preparative chromatography. Details of the synthetic procedure and NMR structural characterization are provided in the protocol by Sekera et al. Overall enrichment was calculated as 99.8% from the high-resolution mass spectra (R ≥ 100 K at 459 m/z) using the doubly deuterated monoisotopic peak intensity (D₂-JQ1, [M + H]⁺ calc. 459.15830 Da m/z, obs. 459.15801 Da m/z, Int. 262409440) compared to the residual peak of native JQ1 (JQ1, [M + H]⁺ calc. 457.14595 Da m/z, obs. 457.14520 Da m/z, Int. 686939).

Preparation of Slides

Slides are etched from a custom vector (SVG) file prepared using freely available software (Inkscape v1.3, GNU v3) (Figure S1) with a 30W CO₂ laser etcher (Universal Laser Systems Inc., VLS 2.30). After etching, slides are washed for 10 min in hexanes, then for 10 min in LC-MS-grade methanol. Slides are dried overnight at 37 °C. Liver tissue sections were cryosectioned at – 20 °C to a thickness of 10 μm and subsequently thaw-mounted onto ITO-coated slides and stored at −80 °C until used for analyses.

Development and Implementation of Code

SpotOn is written in Java and implemented as a plugin for the ImageJ image processing software. It is compatible with the standard ImageJ installation but is not yet supported in other distributions such as Fiji. The source code and plugin files are available in the SpotOn GitHub repository (https://github.com/nathanieltwarog/SpotOn_plugin).

To begin, the slide holder image is “normalized” to ensure consistent orientation and size. Though the plugin works best if the slide holder is scanned with the “short flat” side (the shorter side of the slide holder that is opposite the two truncated corners) against the edge of the scanner, all images are rotated and rescaled so that they are in a portrait orientation and 1200 dpi within ImageJ (Figure A). If the option to locate the holder directly is enabled, the plugin then uses a stored 128 × 128-pixel template to locate three circular impressions on the slide holder resulting from the production process (Figure B,C). A similar template is used to locate three other impressions at the bottom of the working image. These two sets of images are then averaged to estimate the locations of the slide-holder’s ends; from these ends, calculations are performed to estimate the slide-holder’s center and orientation.

5.

Slide holder markings and fiducials are used for the software target recognition. (A) Orientation of the slide holder that has been rotated and transformed by SpotOn software from an original 2400 DPI horizontal orientation to a 1200 DPI vertical orientation. (B) Fiducials are used by the SpotOn software to recognize the slide holder rather than etched slides. The red and blue circles indicate the markings on the top and bottom of the holder, respectively. (C) Magnified image of the markings on the slide holder utilized as a template for direct recognition of the holder without etched slides. (D) Slide marking examples for custom laser-etched slides. (E) One of two slide marking examples used for the recognition of Bruker Intellislides.

It is important to note that not all template matches are weighted equally in the average. If one of the circular impressions was obscured, damaged, or blocked by something on the surface of the scanner, a poorly matched template location could throw off the estimate of the location of the end of the holder, which could reduce the precision of the estimation of the holder location overall, making the resulting Echo coordinates less accurate. Thus, when a set of matched templates is “averaged”, each match is given a weight determined by how well it correlated with the stored template. Furthermore, if the estimate of a single template location match differs substantially from the others, its weight is further reduced through iterative reweighting, gradually silencing template matches that deviate significantly. This robust, consensus-based averaging is also used to estimate slide locations.

Templates are matched by selecting a subset of the image (e.g., all pixels in the top 10% and left 30% of the image), and running a standard, fast-Fourier-transform-based convolution to identify the point in that region that most closely correlates with the provided template. The match weight is determined by linearly scaling the template to account for the nature of the match. Some templates are more robust but produce a wider range of image intensities, whereas others are weaker but generate a more consistent appearance.

Each matched template in an image can be used to estimate the location of a specific point on the slide holder or the slide. Therefore, each match can be converted into a point with a weight that represents its confidence and the estimated location of the shared target point. Given a set of weighted points (x _i , y _i , w _i,0) (where w _i,0 is the initial weight given to each point from the template match), the weights are updated over a set of N iterations, where in each iteration:

$${\stackrel{-}{x}}_k=\frac{\sum _i{w}_{i,k-1}{x}_i}{\sum _i{w}_{i,k-1}},{\stackrel{-}{y}}_k=\frac{\sum _i{w}_{i,k-1}{y}_i}{\sum _i{w}_{i,k-1}}$$ 1

$$w_{i,k}=w_{i,0}\exp -\frac{(x_i-{\stackrel{-}{x}}_k{)}^2+(y_i-{\stackrel{-}{y}}_k{)}^2}{2{\sigma }^2}$$ 2

This iterative process further reduces the weight of potentially weak matches that represent an estimated location and are significantly farther from the consensus of other matches. Even with reduced initial weights, such points could still significantly pull the estimated location away from the correct one. The parameter σ represents the expected amount of variation due to image noise and scanner variability, and is set to 15 pixels (approximately 0.3 mm) by default. After N iterations of weight updates, the final center is determined as

$$\stackrel{-}{x}=\frac{\sum _i{w}_{i,N}{x}_i}{\sum _i{w}_{i,N}},\stackrel{-}{y}=\frac{\sum _i{w}_{i,N}{y}_i}{\sum _i{w}_{i,N}}$$ 3

The resulting value is used as a robust estimate of the target point’s position in the image.

After the slide-holder has been located, or if it was not located at all due to differing scanning conditions, the plugin then locates any slides in the holder’s four slots. For slides laser-etched with the custom pattern shown in the figure, the plugin uses another set of custom templates to locate the “A”, “C”, “F”, and “H” markings on each slide (Figure D). Again, the robust averaging described above allows accurate placement of each slide, even when one or two of the matched letter markings are obscured. For Intellislides, a template matching the “square-in-circle” medallion on the slides, as well as the company logo, is used in the same manner as the letter markings to place each slide (Figure E). Assuming slides are located, the plugin keeps track of which slots contain slides, so that only those slots are presented to the user.

Once slides are located, if the slide holder’s position is not directly identified, the known relative positions of the four slide slots are used to estimate the slide holder’s position from the slides’ positions. It is always recommended to estimate the slide holder’s position directly, as slides can shift within the holder, leading to imperfect alignment. However, because the slide markings are much more precise and less dependent on the scanner model, estimating the slide-holder position from the slides will produce better results across a broader range of experimental settings.

Finally, once all preprocessing and template matching are complete, any identified slides are presented in a new window, allowing the user to zoom in or out and place Echo-targeted spots at any desired location on the slide. These placed spots are added to an internal list tracked by the plugin, and once the user has annotated all (up to 4) slides, the plugin exports them to a table that can be saved for deposition.

Utilization of SpotOn Software for Echo and Manual Depositions

Either plain etched ITO slides or slides containing tissue sections were used. When using tissue sections, slides were brought to room temperature in a lyophilizer before scanning to select deposition locations. Images of slides placed in a microscope slide holder adapter (Thor C4SH01) were captured at 2400 DPI. Images were opened in ImageJ (v 1.8.0), and the SpotOn plugin was run to choose discrete locations on the slide for acoustic deposition by Echo.

Standard solutions of JQ1 were prepared in DMSO, and 1 μL of a 2 mg/mL cresyl violet solution in DMSO was added to aid droplet visualization. Samples were prepared in a 384-well low-dead-volume plate and briefly centrifuged at 2,000 rpm before using the Echo system. Custom labware definitions for the Thor microscope holder are provided in Table S3 of the Supporting Information. Region definitions generated by SpotOn and exported as CSV files were imported into the Echo Reformat Tool. Within the protocol menu, destination plate optimization was selected. Manual applications of solutions used a pin tool (VNP Scientific Inc., FP1S50), cleaned with three successive DMSO washes followed by dry blotting paper. Following completion of the spotting protocols, slides were further dried in a lyophilizer for 15 min before continuing with MALDI analysis preparations.

MALDI MSI Experiments and Data Analysis

A MALDI matrix solution comprised of 40 mg/mL 2,5-dihydroxybenzoic acid in 6 mL of 70% methanol with 6 μL of a 10 mM D₂-JQ1 solution in DMSO was applied to slides using an HTX M5 sprayer. The matrix was applied at a nozzle temperature of 75 °C for 9 passes over the samples with a heated bed temperature of 60 °C and a drying time of 20 s between passes. The flow rate of the matrix was 0.200 mL/min at a velocity of 1920 mm/min, a track spacing of 2 mm, a pressure of 10 psi, a gas flow rate of 3 L/min, and a nozzle height of 40 mm. Samples were analyzed immediately after the MALDI matrix was applied.

MALDI MSI images for calibration curves on glass were acquired on a Bruker timsTOF fleX at 20 μm, and liver calibration curves were acquired on a Bruker SolariX FT-ICR at 40 and 20 μm, both in positive ion mode. For on-tissue analyses, a continuous accumulation of selected ions (CASI) experiment was used to increase drug signals. Both data sets were normalized to the D₂-JQ1 signal. Data was converted to an imzML format and processed using MSiReader (v 3.10).

Briefly, post normalization to 459.16 Da, a region of interest is chosen on the tissue using the polygon tool for ROI to enable the quantification tool ‘Absolute – QMSI spatial calibration curve’. In this example, based on MALDI experiments using liver tissue, a tissue density of 1.06 g/cm³ was chosen, with a thickness (equivalent to the approximate amount of tissue ablated by MALDI) of 0.1 μm. The internal standard, JQ1, with an observed mass of 457.15 Da, was used as the spotted standard, with a spotted liquid volume of 0.0025 μL. For each spot, a region was chosen encompassing all pixels in that region under ‘Create LX’, and the concentration of JQ1 used was reported. Observed abundance means were exported from MSiReader and subsequently imported into GraphPad Prism (v 10.6.1) for further statistical analysis.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/cbmi.6c00010.

• images for use in laser etching systems; observed Echo deposition of DMSO drug standards after various cleaning methods for glass slides; labware definitions utilized in the Echo software for the Thor microscope adapter; calculated averages, standard deviations, and relative standard deviations for both observed signals and spot sizes; and additional mass spectrometry imaging experimental data (PDF)

The authors declare no competing financial interest.