Multispectral Optoacoustic Tomography for Functional Imaging in Vascular Research

Summary

The present protocol describes the acquisition of multispectral optoacoustic images of in vivo human skin vasculature. These include the quantification of hemoglobin and melanin, regarded as chromophores of interest for functional analysis.

Abstract

Microcirculatory impairment has been recognized in various disease processes, underlying this growing theme within vascular research. In recent years, the development of live imaging systems has set the (analytical) pace in both basic and clinical research, with the objective of creating new instruments capable of providing real-time, quantifiable endpoints with clinical interest and application. Near-infrared spectroscopy (NIRS), positron emission tomography (PET), computed tomography (CT), and magnetic resonance imaging (MRI) are available, among other techniques, but cost, image resolution, and reduced contrast are recognized as common challenges. Optoacoustic tomography (OT) offers a new perspective on vascular functional imaging, combining state-of-the-art optical absorption and spatial resolution capacities (from micrometer optical to millimeter acoustic resolution) with tissue depth. In this study, we tested the applicability of multispectral optoacoustic tomography (MSOT) for functional imaging. The system uses a tunable optical parametric oscillator (OPO) pumped by an Nd: YAG laser, providing excitation pulses sensed by a 3D probe at wavelengths from 680 nm to 980 nm. Images obtained from the human forearm were reconstructed through a specific algorithm (supplied within the manufacturer's software) based on the response of specific chromophores. Maximal Oxygenated Hemoglobin (Max HbO₂) and Deoxygenated Hemoglobin (Max Hb), Total Hemoglobin (HbT), and mean Oxygen Saturation (mSO₂) to vascular density (µVu), inter-unit average distances (ζAd), and capillary blood volume (mm³) may be measured using this system. The applicability potential found with this OT system is relevant. Ongoing software developments will surely improve the utility of this imaging system.

Introduction

Cardiovascular diseases are recurrent top causes of death worldwide and represent a huge burden for any health system^1,2. Technology has been a major contributor to the expansion of our understanding of fundamental cardiac and vascular pathophysiology, providing more precise diagnostic tools and the possibility of early disease detection and more effective management. Imaging techniques offer the possibility to measure not only cardiac and major vessel performance but also, on a much smaller scale, to calculate the capillary density, local perfusion and volume, and endothelial dysfunction, among other characteristics. These technologies have offered the first quantitative insights into vascular biology with direct clinical application. Changes in capillary density, local perfusion reduction, or occlusion likely correspond to an ischemic condition, which helps to explain the growing role of imaging, becoming an indispensable tool in cardiovascular research and practice^3,4,5.

In recent years, functional imaging has consistently set the pace in technological innovation, with ultrasound (US) near-infrared spectroscopy (NIRS), positron emission tomography (PET), computed tomography (CT), and magnetic resonance imaging (MRI) as some well-known examples. However, multiple concerns limit their application, from cost and patient safety (as well as comfort) to image contrast and resolution^6,7. Optoacoustic tomography (OT) has recently emerged as a new direction in optical-based vascular research. This technology, centered on the detection of ultrasonic waves generated by thermoelastic expansion of the tissue impacted with ultrashort laser pulses, has been known for some time^6,8. This physical reaction of heat development and tissue expansion evokes an acoustic signal detected by an ultrasound transducer. The use of pulses of light from visible to near-infrared and the absence of an acoustic background signal benefit the resolution depth. The detected contrast results from the most important chromophores present (hemoglobin or melanin). Compared with other technologies, OT has the advantages of (1) needing no contrast (label-free imaging), (2) better contrast and resolution with fewer artifacts than ultrasonography, and (3) lower price, and faster acquisition and ease of operation^6,9,10,11.

Multispectral optoacoustic tomography (MSOT) is among the most recent generation of OT instruments. Built with a tunable optical parametric oscillator (OPO) pumped by an Nd:YAG laser providing excitation pulses, a 3D image is acquired by time-resolved signals detected from high-frequency ultrasonic excitation pulses at wavelengths from 680 nm to 980 nm with a repetition rate of up to 50 Hz¹². The optoacoustic imaging platform provides the quantification of different chromophores in-depth (as low as 15 mm). Variables such as HbO₂, Hb, and melanin are easily accessible. Other variables of interest, such as maximal Oxygenated Hemoglobin (Max HbO₂) and Deoxygenated Hemoglobin (Max Hb), are also available. Reconstruction algorithms from the manufacturer's software allow the calculation of other variables such as vascular density (µVu), inter-unit average distance (ζAd), and capillary volume (mm³).

The present study explores the essential operating aspects of this new system to understand better its practicalities and potential applications in cardiovascular preclinical research. 

Protocol

The experimental protocol was previously approved by the Ethics Committee of the University's School of Health Sciences (EC.ECTS/P10.21). Procedures fully respected the principles of good clinical practice defined for human research¹³. A convenient sample of six healthy participants of both sexes (n = 3 per sex) with a mean age of 32.8 ± 11.9 years old was chosen from the university community. Selected participants were required to be normotensive, non-smokers, and free of any medication or food supplementation. Blood pressure, cardiac frequency, and the Body Mass Index were also registered. All participants were previously informed of the objectives and duration of the study and provided informed written consent.

NOTE: This study was performed using the MSOTAcuity (see Table of Materials), henceforth referenced as the optoacoustic imaging platform.

1. Preparation for Acquisition

NOTE: In the experimental description that follows, screen commands are in Boldface type.

1. Loading subject information: Turn on the optoacoustic imaging equipment. While the equipment is warming up, introduce the participant information. The main welcome window of the software opens to the Scan overview. Introduce data (including name, the study denomination, personal data, and any relevant observations) after clicking on Patient ID, and finish the application by pressing Select.
2. Preset selection: Make sure the Laser is Ready message appears on the equipment screen. Following the warmup time, the laser status bar on the equipment screen must change from Laser Standby to Laser is Ready. For this protocol, the preset is designed for chromophores Hb, HbO₂, and melanin. After selecting the correct preset, the laser power will be tested.
3. At this point, ensure that there is a message on the screen that reminds every participant in the room to apply laser safety googles. Press the laser (power) switch foot pedal and wait for the laser power self-check. After a few seconds, a window appears with the current laser status with a check-up report. Release this window by pressing the available OK button.
   NOTE: The optoacoustic imaging platform uses a Nd:YAG laser, considered a class 4 laser particularly dangerous to the human eye. Thus, this laser must be handled with adequate care.
   ​CAUTION: No acquisition should take place without ensuring all safety procedures, including appropriate eye protection, are in place.

2. Positioning and Image Acquisition

1. Acclimatize the participant to the laboratory environment (21 °C ± 1 °C; 40%-60% relative humidity), choosing a comfortable position to minimize unnecessary movement. Ensure that the area to be scanned is previously cleaned.
   NOTE: The manufacturer's recommendation to clean the area to be imaged with a 70% ethanol/water solution is recommended. Additionally, for best image acquisition, removal of hair (when applicable) is suggested.
2. Probe holder and image stabilization
   1. Apply a thin layer of ultrasound gel to the 3D cup. Image stabilization requires holding the 3D cup in the desired imaging position. Position and stabilize a lockable arm for the area of interest. The arm used in this study was designed in-house and built with aluminum profile components (Figure 1).
   2. After placing the 3D cup on the area of interest, partially lock the stabilizing arm lock for image acquisition.
      NOTE: The quality and even application of the ultrasound gel is critical; the presence of air-bubbles may compromise the image definition.
3. Image acquisition for dynamic conditions, using the Post-Occlusive Reactive Hyperemia (PORH) maneuver, in the Examination menu tab.
   1. Acquire the baseline control scan. After finding a field of view for imaging, with the deflated blood pressure cuff in place, securely lock the 3D cup positioning arm.
   2. Apply minimal pressure to the imaging site, as higher pressures might compromise the readouts. Push the manufacturer's default preset Hb, HbO₂, and Melanin that simultaneously measures chromophores for Hb, HbO₂, and melanin.
      CAUTION: It is mandatory to protect the eyes with proper safety goggles during operation.
      NOTE: Skin phototypes IV to VI (dark skin) are prone to misreading, requiring a baseline control image for further processing. The use of safety goggles during image acquisition (when the laser is active) only allows the human eye to recognize yellow and blue colors. Colors can be edited during image processing.
   3. Select the anatomical area for baseline image acquisition. For exploratory purposes, the ventral forearm is recommended. Proceed by pressing the laser footswitch pedal.
      NOTE: The touch screen button labeled View (colored yellow), which shows the live image on the screen when pressed. The image stability status is shown as a grey bar in the middle of the touch screen, indicating the stability of the 3D probe.
      1. When the image stability is maximized, take (or capture) a snapshot of the area by pressing the Snapshot button on the touchscreen. Each scan will acquire 10-12 frames at an acoustic depth of 150 mm for every wavelength defined within the preset over an acquisition time of 2 s. This baseline acquisition scan will include a total of 30-36 frames.
         NOTE: 10 frames for each chromophore detected (Hb, HbO₂, and melanin) are collected with a maximal depth of 15 mm.
      2. Continue pressing the laser footswitch pedal for continuous video acquisition and pay attention to the View button (colored yellow) on the touch screen. The stabilized image will appear. Press Record (colored blue) to begin live image recording.
      3. Stop the recording by pressing the Stop button (colored black). The optoacoustic imaging platform will stop recording and automatically render the video to preview mode.
   4. Dynamic measurements (PORH illustration): Adjust the pressure cuff to the patient's arm above the elbow to illustrate this maneuver. Inflate the cuff with supra systolic pressure (~200 mmHg) and proceed according to steps 2.3.1 to 2.3.3.1 to acquire the imaged vasculature under pressure.
   5. To acquire a video to assess the impact of the pressure release on the imaged vasculature, open the pressure valve while acquiring the video as in 2.3.3.2. As before, follow the live image on the screen.
      NOTE: To execute this maneuver, supra systolic pressure should be maintained for 1-5 min; it is important to be aware that this pressure might induce different degrees of tolerance and discomfort in the patient. This aspect should be carefully managed during the experiments.

3. Image analysis protocol

1. Copy the recorded scans to a selected/dedicated folder for backup and further analysis on a separate computer workstation using the manufacturer's dedicated analytical software. Each scan is stored by acquisition time and ordered by the program in a study folder with a running code.
   NOTE: A backup copy is strongly recommended. Working directly with the recorded raw data is possible but strongly discouraged, as any potential hard drive crash might damage the raw data.
2. Open the analysis program on the workstation computer. Choose program Menu > Open Study to import files and access backup scans. Open the study and scroll to the bottom of the folder (with recorded scans) to find files with a .NOD extension. This is the only file type recognized by the software to open a study.
   NOTE: .NOD files are named automatically with a running number given to each study and carry no patient information in the file name.
3. For image reconstruction, open the image analysis module by accessing the software Menu > Advanced processing.
   1. Ensure that the program workflow tabs are visible (colored black) on the top menu bar (Supplementary Figure 1): Menu; Scan Overview; Reconstruction; Fluence Correction; Spectral Unmixing; Visualization & Analysis. During analysis, any activated workflow tab is colored blue.
      NOTE: If Advance Processing is not opened, the software shows only Scan Overview and Visualization & Analysis.
4. Reconstruct the image via the Reconstruction tab of the software. Select the scans to reconstruct from the left side of the main program menu. Loaded scans appear on the right side of the screen. Leave the default six optoacoustic emission wavelengths (700, 730, 760, 800, 850, and 900 nm), as they include the maximum optoacoustic signal for HbO₂ at 900 nm, for Hb at 760 nm, and melanin at 700 nm.
   1. Perform the scan reconstruction using the icon on the right side. Follow the program workflow by selecting the Scan preset and Field of View (resolution). Information is presented at the top left corner of the main screen. Adjust the speed of sound to adjust scan focus (Supplementary Figure 2). The Reconstruction panel also shows the number of frames of each acquired scan and allows the selection of the repetitions to be analyzed (if necessary).
      NOTE: Each scan is loaded with a default speed of sound of -90, which should be adjusted by the user. The speed of sound may also be adjusted automatically with an auto-focus function (AF).
5. Push the button Reconstruct Scans at the top of the screen to advance to scan reconstruction. A temporary dashboard will appear with the message Job Processing. This panel can also be accessed from the Menu > Processing Status. After finishing the reconstruction, the image post-processing analysis must advance to Fluence Correction.
6. Activate Fluence Correction of reconstructed images in the dashboard menu. Reconstructed images must be loaded for fluence correction. These appear with a flag next to each scan number. Loaded files will be immediately displayed on the right side of the screen as Selected Reconstructions. Activate Fluence Correction by interacting with the icon on the right side of the screen (Supplementary Figure 3). Push the Save Fluence Correction(s) to progress.
7. After saving the fluence correction, perform the spectral unmixing of the acquired preset (Hb, HbO₂, and melanin). Select the Spectral Unmixing tab to open the list of Selected Reconstructions for spectral unmixing. A list with each scan of the selected study will be displayed with the history of the previous image processing steps.
8. Load the previously saved fluence correction files. Loaded scans will be immediately displayed on the right side of the screen as Selected Reconstructions (Supplementary Figure 4). Activate spectral unmixing by pushing the icon on the right side of the screen.
   1. Observe the wavelengths to be unmixed. All six optoacoustic emission wavelengths (700, 730, 760, 800, 850, and 900 nm) taken into the reconstruction step (step 3.4) are automatically chosen for spectral unmixing. Edit the desired spectra to be processed (e.g., Spectra: Hb, HbO₂, and melanin) using the XYZ icon, if necessary.
   2. After confirming the adjusted parameters, click on Start Spectral Unmixing for the spectral unmixing to progress. A processing menu bar appears, displaying the operation progress.
      NOTE: Various parameter adjustments are possible during spectral unmixing, and several unmixing methods are available. In this protocol, the Linear Regression method is used as a standard to unmix Hb, HbO₂, and melanin.
9. Access the Visualization and Analysis tab. Click an activated scan to display all subject information and comments introduced in step 1.1 (Supplementary Figure 5).
   NOTE: Multiple scans may be visualized in parallel.
   1. Push the + button to create a multiple scan analysis. In this window, introduce a multiple scan view and press the Save button. After saving the view name, a new dashboard is displayed, including all the scans of the study being analyzed.
   2. Select each desired scan to Add (each) to the saved Analysis View. Add additional scans at the top left corner icon, and they will be automatically displayed in the Analysis View.
10. Within the analysis view, set proper color lookup tables to prepare the image for analysis. Click More Image Control Options on the top menu bar and activate the Max Intensity Projection icon. Attribute colors to layers by pressing the icon available in the bottom right corner of the image display, adjacent to the 2D+ image display.
11. Select More to edit the colors of all channels simultaneously. This menu shows all chromophores unmixed and allows the selection of multiple layers for display.
    NOTE: Moving the mouse over the software icons shows their name in gray as displayed in the protocol.
12. Adjust each layer's color intensity with the tools available at the bottom left of the screen.
    ​NOTE: Adjustment with min/max interpolation for each channel generally yields good results.

4. Region of Interest (ROI) analysis

NOTE: The selection of a Region of Interest (ROI) is mandatory for data analysis.

1. Identify the ROI to be analyzed. Surround the ROI with the shapes available (within the menu bar) in the XY image while tracing the same ROI within the orthogonal views available in the XZ and YZ axis (Supplementary Figure 6).
   NOTE: A polygon shape was used for the current ROI analysis.
   1. Follow the ROI shape in the remaining XZ and YZ axis (example in Figure 2) while placing multiple polygon layers with the Add Interpolate and Remove sub-regions function. The data can be plotted after defining/selecting the desired ROI.
   2. Press the icon Import Region of Interest to Quantification and observe the multi spectra component shown on the right side of the screen as a graphical detail of the selected ROI.
   3. Export ROI data by pressing the Excel icon at the bottom of the graphical view of the ROI data. The entire data package from all regions is exported as a bundle to a spreadsheet for subsequent analysis. Figure 3 shows data from one participant who submitted to a pressure cuff inflated to 200 mmHg and vasculature was analyzed in comparison to vasculature resting state at 0 mmHg.
2. Quantify multiple ROI objects simultaneously by following steps 4.1-4.1.3.
3. Export images from the same menu as TIFF files with all embedded data and built-in ROI outline (Figure 2).

Representative Results

Data provided by optoacoustic imaging can be analyzed in post-processed export images (Figure 2) and plotted data (Figure 3). The purpose here was to introduce the operation of optoacoustic functional imaging and to explore its application in more commonly known vascular research. For that, we compared images acquired during rest and after a 200 mmHg occlusion of a major supplying artery (Figure 2). These observations can be quantified after ROI analysis and export. In the XY plane, the higher signal of melanin compared to planes YZ and XZ can be observed, which indicates the epidermis limit. The occlusion of the brachial artery (arm) provokes some stasis in the vessels ahead of the OT probe placement (ventral forearm). In consequence, we detected an increase in the overall signals shown as an increase of blue (Hb) and red (HbO₂) at axes XY, YZ, and XZ. Stasis might be followed in the XY plane while holding the 200 mmHg pressure within the cuff. YZ and XZ axes depict increased blood volume due to the occlusion above compared to the normal perfusion conditions (no occlusion), highlighted by the magenta masked areas.

Exported ROI analysis of the same microvasculature area quantifies chromophores HbO₂ (red), Hb (blue) and HbT (pink), mSO₂ (deep red), and melanin (yellow) from stabilized images collected over 8.6 s. The pressure release is immediately detected; Figure 3 shows the post-occlusion evolution of Hb, HbO₂, and HbT recovery, while the optoacoustic data output follows the observations in Figure 1. The software calculates blood oxygen saturation (mSO₂) and HbT values from the addition of the Hb and HbO₂ arbitrary signals. Melanin concentration remains constant within the 200 mmHg occlusion and at the resting state within the time interval of image acquisition.

Figure 1: Schematic diagram representing the flexible arm designed to hold the measuring probe in stabilized contact with the participant's skin. Please click here to view a larger version of this figure.

Figure 2: Representative optoacoustic images highlighting changes of the vasculature at rest or under pressure of 200 mmHg. The image shown includes three colors, representing Hb (blue), HbO₂ (red), and melanin (yellow), as described within the image analysis in section 4. Each optoacoustic image represents a maximum intensity projection of all the planes associated with each scanned chromophore. (A) The XY plane of the optoacoustic acquisition. (B) The YZ orthogonal view of the same optoacoustic imaged site. (C) The XZ view of the scanned area. The magenta arrows point to the areas with increased stasis; the magenta masked area marks the increased volume of blood trapped inside vessels due to the occlusion of the brachial artery compared to the normal perfusion conditions (no occlusion). Please click here to view a larger version of this figure.

Figure 3: Representative data export of a quantified ROI. Natural chromophores of HbO₂ (Red), Hb (Blue) and HbT (Pink), mSO₂ (deep red), and melanin (yellow) are depicted from the data extracted from the stabilized images collected over 8.6 s. The graphs from Hb, HbO₂, and HbT show a recovery slope from occlusion towards the non-occluded resting state. The calculated blood oxygenation mSO₂ and the melanin concentration remain constant within the 200 mmHg occlusion and at the resting state within the time interval of image acquisition. Images extracted are data points depicted as mean ± sd of n = 10 images per frame. Please click here to view a larger version of this figure.

Supplementary Figure 1: Scan overview panel and Main Menu of the analysis software. Pushing the Menu button (in black), the main menu will drop down options to select the selected study. This action will select and load the ".nod" file recognised by the software. The Scan overview (in blue) shows all study's scans. Details (black) appear in the right. Please click here to download this File.

Supplementary Figure 2: Reconstruction analysis workflow. Panel 1 - Select the scan to be reconstructed and press the right-facing arrow on the right side of the display (purple arrow) to advance. Panel 2 - Observe the speed of sound and adjust the slider to the best focus (blue arrow); a) adjusted focus displayed in the window right hand side; b) select repetitions to be analysed (yellow arrow); c) press the Reconstruct scans button to proceed (green arrow). Please click here to download this File.

Supplementary Figure 3: Fluence correction panel Workflow. Panel 1 - Select scans to be corrected and press the right-facing arrow on the right side of the screen. Panel 2 - press Save Fluence Correction(s) to proceed (green arrow). Please click here to download this File.

Supplementary Figure 4: Spectra unmixing panel Workflow. Panel 1 - Select scans to unmix and press the right-facing arrow (purple arrow). Panel 2 a) select the scan to unmix (blue arrow) and a preview of the adjusted image will be displayed in the right hand side; b) Select the repetitions to unmix (yellow arrow); c) press Start spectral unmixing to proceed (green arrow). Please click here to download this File.

Supplementary Figure 5: Visualization panel and selection of chromophore colors. Panel 1) Select scans to display with a double click (purple arrow); Panel 2) Acquired image in axis XY (blue square), XZ (yellow square), and YZ (green square); 2a) Image analysis button showing acquired wavelengths; 2b) choose More Image Control Options in the top menu bar and activate the Max Intensity Projection icon; select More to edit channels' colors. Please click here to download this File.

Supplementary Figure 6: Selection of the Region of Interest (ROI). Select the lasso tool (yellow arrow) and define the boundaries of the ROI within the XY axis (magenta arrow). It is possible to define various shape areas (polygon, rectangle, square, circle, or elipse). Follow the ROI in XZ and YZ axis and add sub-regions (green arrow) to the initial selection. Multiple subregions are displayed (cyan arrow). To extract data from the selected ROI, press the icon Import Region of Interest to Quantification and proceed. Please click here to download this File.

Discussion

This protocol emphasizes the working steps regarded as practical requirements to operate this new optoacoustic imaging instrument, from the adequate positioning (participant, probe) needed for 3D cup probe stabilization to image acquisition, ROI selection, and image reconstruction and analysis.

The proposed experimental approach, using "instantaneous" acquisitions together with images obtained under dynamic conditions, illustrates the interest and utility of this instrument in accessing in vivo human vascular physiology. As shown, the 150 µm acoustic image resolution collected in a volume of up to 15 mm³ is unmatched by other tomography techniques.

Special attention is necessary regarding (i) the importance of the probe stabilization for image acquisition; the use of a flexible, secure probe holder clearly improves image acquisition; (ii) the correct identification of the vascular structures; sonographic references such as melanin in the epidermal-dermal transition might be used as a marker to identify the upper plexus vessels in the skin; and (iii) the functional image analysis performed through the manufacturer´s reconstruction software.

Advanced analysis of ROI data and image export requires a deeper comprehension of the dedicated software and the algorithms developed. The current optoacoustic imaging instrument is able to reconstruct a 3D volume of 15 mm³ of tissue with a resolution of 150 µm. This capacity should be potentiated to better quantify microvascular function(s) in depth. Nevertheless, the basic operation allows the direct observation of reference chromophores and the acquisition of multiple presets from the same area, providing fast scanning and live video recordings.

The applicability potential found with the optoacoustic imaging system is relevant. Ongoing software developments will surely improve the utility of this imaging system.

Materials

Name	Company	Catalog Number	Comments
Cuff	PIC	107001
Drapes	Pajunk	021151-1501
Ethanol 70%	Sigma Aldrich	EX0281
Gogless	Univet	559G.00.00.201
Kimwipes	Amoos	5601856202331.00
MSOT	iThera	MSOTAcuity
Stabilizing arm	ITEM		Self designed and assemble
Ultrasound gel	Parker Laboratories	308
Waxing cream	Veet	kkdg08hagd