Here, we present a protocol for high-resolution optical detection of chemical information around implanted medical devices with X-ray excited luminescence chemical imaging (XELCI). This novel imaging technique is developed in our lab which enables studying implant-associated infection biochemistry.
Microbial infections associated with implantable medical devices are a major concern in fracture fixation failure. Early diagnosis of such infection will allow successful eradication with antibiotics without an extra cost for a second surgery. Herein, we describe XELCI as a technique with high X-ray resolution, implant specificity, and chemical sensitivity to noninvasively image chemical concentrations near the surface of implanted medical devices. The devices are coated with chemically reporting surfaces. This chemically responsive surface consists of two layers coated on an implantable medical device; a pH-sensitive layer (bromothymol blue or bromocresol green incorporated hydrogel) which is coated over a red-light emitting scintillator (Gd2O2S: Eu) layer for monitoring. A focused X-ray beam irradiates a spot on the implant, and the red light generated by the scintillator (with 620 nm and 700 nm peaks) is transmitted through the sensing layer which alters the spectral ratio depending on the pH. An image is generated by scanning the X-ray beam across the implant and measuring the spectral ratio of light passing through the tissue point-by-point. We used this imaging technique for monitoring implant-associated infections previously on the bone surface of the femur with a modified implantable plate sensor. Now we are studying pH changes that occur from tibial intramedullary rod infections. Two different types of intramedullary rod designs are used in pre-pilot rabbit studies, and we learned that the XELCI technique could be used to monitor any chemical changes that occur not only on the bone surface but also inside the bone. Thus, this enables noninvasive, high spatial resolution, low background local pH imaging to study implant-associated infection biochemistry.
In the United States, about 2 million fracture fixation devices are inserted annually, and 5%-10% of them lead to implant-associated infections1. These infections are harder to treat with antibiotics at later stages due to the heterogeneity and antibiotic-resistant nature of the biofilms2,3. If they are diagnosed early, infections can be treated with antibiotics and surgical debridement to prevent extra medical costs for a second surgery to replace hardware at the treated fracture site. Plain radiography and other advanced radiographic techniques are applied in the diagnosis of orthopedic implant-associated infections, non-unions, and related complications. Although these techniques are used frequently to acquire structural information of the surrounding bone and tissue at the orthopedic implant, they are unable to provide biochemical information in the specific environment. Thus, we developed a novel X-ray excited luminescence chemical imaging (XELCI) technique for high-resolution imaging of biochemical information noninvasively at the implant site. Diagnosis of orthopedic implant-associated infections is commonly carried out by one or a combination of different means. Clinical observations (pain, swelling, redness, wound discharge, etc.) suggest the first signs of infection. Later, radiological and laboratory experiments are carried out to confirm the failure of bone healing progression and identify the pathogenic organism4,5. Nuclear medicinal techniques such as computerized tomography (CT), magnetic resonance imaging (MRI), and radionucleotide methods such as Single Photon Emission Computed Tomography (SPECT) and Positron Emission Tomography (PET) are in use for better visualization of the infected implant and the associated infection6,7. CT and MRI are advantageous in determining bone necrosis and soft tissue abnormalities, respectively but cause interferences at a close distance to the metal implants8. Different X-ray methodologies such as SPECT and PET in combination with radioisotope-labeled analytes as in vivo imaging contrast agents are widely utilized to diagnose implant-associated osteomyelitis2. Current applications combine both data from CT scanning and labeling data from either SPECT or PET to generate anatomical information9. Although one or more of these imaging modalities are used to aid infection diagnosis, they cannot detect the pH variations associated with infection early to initiate the treatments with antibiotics to avoid extra medical and surgical expenses.
The main advantage of utilizing the imaging system used in this study for monitoring implant-associated infections is its ability to reveal biochemical information about the biofilm microenvironment with a spectral reference. Although the main focus is on imaging and mapping pH at the infected site, this method can be altered to monitor other biomarkers specific to implant-associated infections. Thus, XELCI allows understanding the pathophysiology of the infection. The high spatial resolution imaging allows mapping heterogeneity as the infection grows. pH at the surface where the biofilm formation occurs is very important to understanding biochemical changes. Also, other microenvironment changes can occur due to antibiotic-related stress responses by bacteria10,11. Due to surface-specific and high spatial resolution imaging, the antibiotic effect on the biofilm microenvironment can be monitored. The technique can also be used to study the biofilm environment for targeted drug delivery experiments. We can study targeted low pH drug release or raising pH to make them more susceptible to work at higher pH.
Three specific characteristics of this imaging technique are X-ray resolution, Implant surface specificity, and chemical sensitivity (Figure 1A). These characteristics can be compared with the currently available imaging techniques for imaging orthopedic implant-related infections (Figure 1B). Once irradiated with X-rays, phosphor particles coated on the implant surface generate red and near-IR (NIR) light that can penetrate through a few centimeters of tissue (albeit with some attenuation)12,13. Table 1 shows some of the features of the developed imaging system compared to other ways that have been used to measure pH in biofilms or through tissue.
XELCI is a novel imaging technique to acquire high spatial resolution chemical information optically near implanted medical devices in combination with X-ray excitation, as shown in Figure 2. Here the selective excitation and optical detection of X-ray excitable phosphor particles is utilized. The implant is coated with two layers, a pH-sensitive dye incorporated polymer layer over a layer of scintillator particles. Once a sequence of focused X-ray beams irradiates the implant, the scintillator layer generates visible light (620 nm and 700 nm). This produced light passes through the pH-sensitive layer modulating the luminescence spectrum depending on the pH of the surrounding environment. Low pH is generally associated with infection and biofilm formation; as the infection progresses, the pH changes from physiological pH (pH 7.2) to acidic (less than pH 7), and the pH dye in the sensor changes color and thus absorbance. The variation of the luminescence spectrum is shown in Figure 2E for Bromocresol green pH dye at pH 7 and pH 4. The transmitted light through tissue and bone is collected and the spectral ratio determines pH. To generate a pH image, the focused X-ray beam irradiates a point at a time in the scintillator film and scans the beam point-by-point across the sample. Previously, this technique was applied to image pH variation on the surface of the orthopedic implants14,15 and have tested it to monitor pH variations in the intramedullary canal through bone and tissue.
Figure 3 below shows a schematic of the imaging system. Basic components of the imaging system are the X-ray excitation source with poly capillary optics, a one-piece acrylic light guide connecting to two photomultiplier tubes, the x, y, and z motorized stage (30 cm x 15 cm x 6 cm travel) and the computer connected for data acquisition. The X-ray source, x,y,z stage, and collection optics (elbow, light guide, photomultiplier tubes (PMTs)) are in the X-ray proof enclosure, while the X-ray controller, power source for PMTs, function generator connected to the data acquisition (DAQ) board and computer are kept outside. A push-button, normally open switch, placed between the enclosure and the front of the door serves as an interlock. If the door is not fully closed (the interlock switch is open), the X-ray source will not turn on, and it will automatically turn off the X-ray source if it is opened during operation. The motors can execute a continuous scan as well as they can be moved to any discrete location. The scan speed for y-axis is usually 1-5 mm/s, while the step size on the x-axis can be chosen typically from 150-2000 µm. The parameters can be chosen depending on the required spatial resolution. Even exposure times are confirmed by consistent speed throughout a continuous scan.
Once the focused X-ray beam is irradiated on the X-ray luminescence particles, the generated light will pass through the pH-sensitive film by modulating the light depending on the surrounding pH. The transmitted light will interact (scatter and absorb partially) with a tissue, while the light attenuation by scattering and absorption will increase as the tissue thickness increases. The collection optics includes a one-piece bifurcated acrylic light guide fitted with a reflective aluminum elbow (with a 90° bend and polished reflective interior surface) at the beginning. This is to ensure the light is collimated as soon as light reaches the light guide. These additions significantly improved the light collection efficiency. For further details, Figure 4 shows the machine drawings of the elbow and light guide. The 90° elbow was machined out of aluminum with the internal surface polished to a mirror finish and the light guide was machined with Acrylic. We have also attached a broad range long-pass blue light filter (blocking 350-450 nm light) at the beginning of the elbow to ensure that only red light will pass through. The end of the one-piece acrylic light guide bifurcates into two streams leading to two different PMTs. The PMTs are enclosed in a small light-tight metal box that is in contact with a thermoelectric cooler to cool down the PMTs to ~5 °C. At the beginning of one of the PMTs, a narrow range long-pass filter (blocking 570-640 nm light and passing 640-740 nm light) is attached to measure only the 700 nm light. Therefore, the 620 nm and 700 nm light can be calculated separately. The PMTs are set up in photon counting mode, and they generate transistor-transistor logic (TTL) pulses for each photon detected. A DAQ system counts the pulses (saturation point 20 million pulses per second) using USB communication. Two separate intensity maps are generated after processing the data, and a final image is created by considering the ratio of the signal wavelength intensity (620 nm) to the reference wavelength intensity (700 nm). This ratio accounts for differences in total light collection efficiency, which depend strongly on the position of collection optics, X-ray irradiation intensity, and tissue thickness. In addition, a spatially separated reference region without any pH indicator dye accounts for spectral distortion from wavelength-dependent tissue penetration. A graphics-based programming language is used for controlling the imaging system, and a basic flow chart of the operation is shown below. The imaging setup, except for the computer, X-ray controller, and DAQ unit, is enclosed in a safe X-ray enclosure to minimize radiation exposure.
This procedure follows the animal use protocols approved by the Clemson University Institutional Animal Care and Use Committee (IACUC). The experiments are carried out according to the Clemson University Biosafety Committee (IBC) and radiation safety committee (RSC) as well as following the relevant guidelines and regulations.
NOTE: A flow diagram of completing a XELCI scan is shown below in Figure 5 followed by a detailed step-by-step description of the imaging procedure.
1. Initialize the system and acquire a plain radiograph
2. Optionally, perform a background scan with the X-ray off
3. Perform a sample scan with the X-ray on
4. Forming the image
5. Culturing bacteria for imaging in sterile conditions (If bacterial grown sensors are imaged)
As a preliminary study, we imaged the intramedullary rod sensor in a reamed tibia of a rabbit cadaver14. The sensor has three distinct regions: the reference region, pH 8 region (basic pH), and pH 4 region (acidic pH). The reference region is the scintillator (Gd2O2S:Eu) particle incorporated in roughened epoxy film. The distinctive acidic and basic pH regions represent infected and non-infected situations inside the intramedullary canal (Figure 6A,B)14.
After completing the scan, the images at 620 nm, 700 nm, and ratio (Figure 7A–C), respectively, were generated in MATLAB. The changes in color are indicative of the changes in the pH. As the bromothymol blue at the basic pH region significantly absorbs emitted light than the acidic pH region, the lower pH region appears as a brighter signal at 620 nm. The scintillator emission at 700 nm functions as a spectral reference for inconsistencies in the scintillator film, tissue composition changes, and any changes that occur in the position of the detection optics from scan to scan.
Figure 1: Specific features of the XELCI imaging compared to the currently available techniques (A) Key features; (B) comparison with the currently available imaging techniques. Please click here to view a larger version of this figure.
Figure 2: Working principle of the technique and sensor behavior with and without infection. (A) Schematic showing implant irradiated with focused X-ray beam and luminescence collected for detection; (B) zoomed-in view of the scintillator and pH-sensitive film-coated intramedullary rod sensor; (C) at low pH caused by infection, intramedullary rod sensor turns from blue to yellow, while dye-free reference region is unchanged; (D) zoomed-in view of the intramedullary rod during an infection; (E) spectra showing of epoxy-PEG pH sensor film (10% PEG hydrogel containing Bromocresol Green pH dye coated on top of epoxy film containing Gd2O2S:Eu scintillator particles) at pH 7 and pH 4 and epoxy-scintillator layer without pH indicator PEG hydrogel. This figure has been reproduced with permission from Uzair et al.14. Please click here to view a larger version of this figure.
Figure 3: Imaging system. (A) Schematic diagram (the red arrows show the direction of light towards PMTs); (B) a photo of the actual system. Please click here to view a larger version of this figure.
Figure 4: Computer-aided design software drawings showing the dimensions. (A) Drawing of the 90° elbow (B) Drawing of the one-piece acrylic light guide. (All the dimensions are in mm) Please click here to view a larger version of this figure.
Figure 5: Flow chart of the imaging procedure Please click here to view a larger version of this figure.
Figure 6: Intramedullary rod sensor in a reamed tibia of a rabbit cadaver. (A) Stainless steel rod coated with Gd2O2S:Eu incorporated epoxy and pH-sensitive Acrylamide-PEG gel; (B) coated stainless – steel rod inserted to the drilled hole in the rabbit tibia. This figure has been reproduced with permission from Uzair et al.14. Please click here to view a larger version of this figure.
Figure 7: Images of the rod sensor in the intramedullary canal. (A) Analyzed image for 620 nm wavelength; (B) analyzed image for 700 nm wavelength; (C) analyzed Ratio image (620/700); (D) plain radiograph of the rod sensor; (E) superimposed ratio image and plain radiography. This figure has been reproduced with permission from Uzair et al.14. Please click here to view a larger version of this figure.
Surface specific imaging | Equilibrium/non equilibrium | High resolution through tissue | Ionizing radiation | X-ray easily superimposed | Representative references | |
pH microelectrode | Yes, one point at a time | non-equilibrium, (Nernst Equation) fouling and time can cause drift | Yes, one point at a time | No | No | 23-25 |
Fluorescent pH indicators | Yes | Equilibrium | No | No | No | 18,26 |
MRI (CEST) | Yes | Equilibrium | 3D but slow | No | No | 27 |
XELCI | Yes | Equilibrium | Yes | Yes | Yes | 14,15,22 |
Table 1: Features of XELCI vs. other pH imaging techniques.
To be able to detect and study orthopedic implant-associated infections early to avoid complications from osteomyelitis and secondary surgical procedures, we have introduced XELCI as a novel, functional imaging technique. It is comparable with the currently available techniques for pH monitoring through tissue.
While positioning the sample for imaging, we use a laser cross head connected to polycapillary focusing optics with two intersecting line-shaped laser pointers at the 90° angle to align the elbow precisely underneath it. These lasers should be turned off prior to starting a scan to eliminate any unwanted light reaching the detector other than the light generated by the scintillators. In terms of distance, the user measures the distance from the tip of focusing optics to the top of the sample to be around 5-5.5 cm before initiating the experiment. This can be achieved either by raising the X-ray source manually or by moving the stage up and down. The same procedure is used for a live animal or measurement through tissue, except we put the animal under anesthesia and run an isoflurane gas tube to ensure it says asleep. We monitor its temperature and pulse periodically and add additional bedding and tape to properly position the animal anatomy, especially the angle, and we estimate 5-5.5 cm focal position to be the depth of the implant rather than just using the distance to the top surface of the skin. Often a plain X-ray image or coarse scan is taken to ensure the correct location and later superimpose with the chemical XELCI image taken at the same spot. For sensors not covered with tissue, the X-ray source current is generally decreased from the maximum of 600 µA at 50 kV to as low as 15 µA to make sure the PMTs do not saturate. The X-ray source and its condition is periodically monitored by Clemson university Radiation Safety. The key for X-ray controller is only used by XELCI users and it is always kept away to prevent any accidental powering on/off. Also, the push button interlock at the front door of the enclosure should be carefully secured before turning on the X-ray. If the interlock isn't working properly, the user will not be able to power on the X-ray, and it will generate errors. Throughout the experiment while the X-ray is on, an orange light is turned on to warn everyone that the X-ray is running.
Typically both a low-resolution scan and a high-resolution scan are run. The low-resolution scan is with a larger step size and the high-resolution scan is with a smaller step size. The time taken for a scan largely depends on three factors: the step size, speed of the scan, and the area of the scan. For example, it takes ~20 min to scan an area of 15 mm x 15 mm for a high-resolution image at a slow scan speed of 1 mm/s and at a step size of 200 µm. Scanning the same area at a lower resolution and faster scan speed reduces the time (e.g., 1 mm step size and 5 mm/s speed takes around 1 min). Additional time is needed prior to scanning to properly set up and position the animal or specimen and understating where the implant is located. For animal experiments, body temperature is monitored to ensure the temperature does not drop too much during anesthesia. Heating from X-ray radiation is negligible as the X-ray dose for the scan used in this study is small. Assuming a 1 Gy = 1 J/kg local X-ray dose, and a heat capacity of 3.45 kJ/kg K for muscle17, the maximum temperature increase from radiation absorption would be less than a mK
Equation 1
Q- heat energy
m- mass
c- specific heat capacity
ΔT- change in temperature
According to the above calculation, the increase in temperature is negligible. Additionally, even this small temperature increase would rapidly dissipate through blood circulation.
The PMTs have an active photo cathode diameter of 22 mm; this large area facilitates light capture from a large diffuse area under the skin. To reduce the dark current from thermally induced electron emission, the PMTs are allowed to cool with cooler underneath. We have a one-piece acrylic light guide that bifurcates into two streams leading to two different photomultiplier tubes (PMTs) for signal detection. This improvement for the system allowed us to enhance the light collection efficiency and thus detect signals through bone and tissue. Previously this imaging technique was used for monitoring infections on the surface of the orthopedic implants and we could successfully image pH changes14,15. We generated high signal/noise images of a modified intramedullary rod through about 2 cm of bone and tissue in the rabbit tibia. In general, light attenuation by scattering and absorption increases exponentially with tissue thickness. There is thus a tradeoff between tissue thickness imaged through, X-ray dose/scan time, and spatial resolution.
This imaging techniques described here involves both an implanted probe and a scanner. The probe contains X-ray scintillators that generate visible and NIR light when irradiated by X-rays. It has a chemically sensitive layer fabricated over the scintillator layer that affects the light intensity or spectrum. The chemically sensitive layer chosen for this application is Bromothymol Blue which has pH-dependent absorbance at 600 nm wavelength and almost constant (pH independent) absorbance at 700 nm light which overlaps with the scintillator emission (620 nm and 700 nm). Bromothymol Blue or Bromocresol green has a pKa value in the pH range we are interested in monitoring implant-associated infections. During analysis of pH microenvironment of bacterial biofilm with pH-sensitive fluorescent ratiometric dye, pH can vary in between 5.6 (within the biofilm) and pH 7 (in the surrounding bulk fluid)18. Implanted probe also has a scintillator layer without the pH indicating dye which acts as a spatially distinct reference region. Moreover, the scintillator particles we are using has prominent emissions at ~600 nm and ~700 nm wavelengths. The absorbance spectrum of the above dyes overlaps with the emission spectrum of the Gd2O2S:Eu particles. Both the exciting X-rays and red/NIR luminescence photons are utilized in in vivo biomedical imaging as they can propagate through tissue19,20,21.
Figure 1 shows three specific characteristics of the imaging system used in this study: X-ray resolution, chemical specificity, and implant surface specificity. Those characteristics are comparable to the currently available imaging techniques such as SPECT, PET, MRI, US, etc22. Although these techniques provide high-resolution structural/anatomical information in high resolution, we could not map/image chemical changes at an implant surface. This technique combines chemical sensitivity from optical indicator dyes with implant specificity from the coating. The spatial resolution from the X-ray provides a unique, low background and higher spatial resolution detection/mapping of chemical concentrations at implant surfaces through bone and tissue.
X-ray excited luminescence chemical imaging provides a unique way to study the local chemical environment at an implant surface with a high spatial resolution for studying infections. In the future, the approach can be generalized to monitor other analytes by selecting different indicator dyes. It can be applied potentially for other diseases and conditions using injected or implanted medical devices coated with scintillator particles and indicator dyes.
The authors have nothing to disclose.
The authors would like to thank Clemson University, COMSET and Clemson SC BioCRAFT. The XELCI setup was initially developed with funds from NSF CAREER CHE 12255535 and later by NIH NIAMS R01 AR070305-01.
90 degree elbow | Produced in Hilltop Technology Laboratory, 51 Parker, Irvine,CA | ||
Bromo Cresol Green | Sigma-Aldrich | 45ZW10 | |
Bromo Thymol Blue | Sigma | 76-59-5 | |
ElectraCOOL Advanced thermoelectric cool plate | Pollock industries, White River, VT, USA | TCP 50 | |
Ethanol | Beantown Chemical, 9 Sagamore Park Road Hudson, NH 03051 |
64-17-5 | |
Gadolinium Oxysulfide Europium doped (Gd2O2S:Eu) particles-~8.0 µm | Phosphor Technologies Inc., Stevenage, England | UKL63/N-R1 | |
LabVIEW | National Instruments, Austin, TX | ||
Motorized Linear Vertical Stage Model (for Z axis) | Motion Control, Smithtown, NY, USA | AT10-60 | |
National instruments c-DAQ 9171 | National Instruments, Austin, TX | NI cDAQ™-9171 | |
One piece acrylic light guide | Produced in Hilltop Technology Laboratory, 51 Parker, Irvine,CA | ||
pH 4 buffer | VWR BDH Chemicals | BDH5024 | |
pH 8 buffer | VWR BDH Chemicals | BDH5060 | |
Phosphate Buffer Solution | MP Biomedicals, Irvine, CA. USA | 2810305 | |
Photo multiplier tubes Model P25PC-16 | SensTech, Surrey, UK | Model P25PC-16 | |
Staphylococcus aureus subsp. aureus Rosenbach | American Type Culture Collection (ATCC), Manassas, VA | ATCC 25923 | |
Tryptic Soy Agar | Teknova, Hollister, CA, USA | T0520 | |
Tryptic Soy Broth | EMD Millipore, Burlington, MA, USA | 1005255000 | |
X-ray source-iMOXS | Institute for Scientific Instruments GmbH, Berlin, Germany | ||
X,Y motorized stage-30 cm x 15 cm x 6 cm travel | Thorlabs Inc., Newton, NJ, USA | LTS300 and LTS150 |