Radiopharmaceuticals labeled with short-lived positron-emitting or gamma-emitting isotopes are injected into patients just prior to performing a positron emission tomography (PET) or single photon emission computed tomography (SPECT) scan. These imaging modalities are widely used in clinical care, as well as in the development and evaluation of new therapies via clinical trials. Unlike ordinary pharmaceuticals, the short lifetime of radiopharmaceuticals requires that they be produced in relatively small batches close to the geographical location where the patient is scanned.

Since PET tracers are classified as drugs by regulatory agencies, they must pass stringent quality control (QC) tests after their production to ensure patient safety prior to injection. Performing and documenting these tests is cumbersome and time-consuming, and requires an array of expensive analytical chemistry equipment and significant dedicated lab space, and there is considerable interest in the development of automated and lower-cost approaches.

By replacing conventional techniques with lab-on-a-chip technologies, it may be possible to achieve further reductions in the size, cost, and complexity of automated QC testing platforms. While some advances have been made for some of the many QC tests, high-resolution miniaturized methods suitable for assessment of chemical or radiochemical identity and purity are notably missing. We have been exploring microchip capillary electrophoresis (MCE) as a potential means to fill this gap. We have shown the potential to perform chemical identity and purity analysis by successful separation of the PET tracer [18F]FLT from all of its side products, with comparable limit of detection as HPLC. More recently, we demonstrated first-in-field work to add radiation detection to the MCE system, using a high-resolution positron detector to perform the radiochemical identity analysis of PET tracers.

Microfluidic systems are also useful for the synthesis of radiopharmaceuticals such as radiolabeled peptides and antibody fragments, which provide a means to image disease-specific targets with extremely high specificity. To substantially reduce the high cost of radiolabeling, we have explored the feasibility of using a microdroplet reactor approach. As an example, a thiol-containing RGD peptide was labeled with fluorine-18 in a site-specific manner via the maleimide-based prosthetic group, [18F]FBEM on a microfluidic chip. We have also explored the site-specific radiofluorination of engineered antibody fragments (diabody) for ImmunoPET imaging via this microfluidic approach.

These studies, and other work on small-molecule synthesis, have shown that there are advantages in performing radiochemistry in microdroplets. Active means of manipulating microdroplets, such as electrowetting-on-dielectric (EWOD) provides more flexibility of reaction implementation than passive methods, but EWOD devices are well known to suffer from a charging effect and the dielectric and hydrophobic layers have been found to suffer defects during harsh chemical reactions. Therefore, a novel droplet actuation mechanism, electro-dewetting, has been developed to address these challenges. Electro-dewetting uses an electric field formed inside a droplet to manipulate the adsorption of ionic surfactant molecules on the solid surface to change the contact angle. The underlying mechanism of this pheromone has been elucidated.

We believe that all the work described here shows the potential for dramatic miniaturization of the complete PET tracer production process, and ultimately increasing the availability of diverse PET tracers for research use, further development, and clinical translation via lowering the cost and complexity of tracer production.

Since the industrial revolution, automation has proven to be a vital component in the growth of any major industry. For positron emission tomography (PET), automation provides efficient, reproducible, and safe methods for routinely producing PET imaging agents, or tracers, using short lived radioisotopes. Over the years, many automated radiosynthesizers have been developed to automate the required production steps up through preparation of the tracer for intravenous injection into the patient or subject, both for clinical and preclinical applications. However, many commercial radiosynthesizers have temperature, pressure, and reagent compatibility limitations, which have required synthetic radiochemists to tailor and optimize synthesis protocols to fit within the constraints of a particular automated radiosynthesizer. Furthermore, the fluidic pathways on these systems generally need to be reconfigured for each tracer one wishes to make. To overcome these limitations, our group has developed a novel, three-reactor, automated radiosynthesizer (ELIXYS) based on the principle of dynamic fluid paths, which is achieved by actuating the reactors to different positions and pneumatically sealing them against a chemically inert gasket. The mechanism enables compatibility with high pressures and obviates the need for large numbers of pressure-limiting valves and fittings. Coupled with a chemist-friendly user interface, it also enables users to quickly automate known radiosyntheses and develop new protocols without requiring specific hardware modifications for each tracer. In this dissertation, I present the design, characterization, validation, and implementation of the system and discuss the capabilities, limitations, and overall work I performed with the system in automating 18F-labeled PET tracer radiosyntheses. I also include the translation of several known tracers onto the system, automation of tracers not previously automated, as well as characterization data helpful for others to more efficiently utilize this and other automated radiosynthesizers. Overall, radiosyntheses of 18F-labeled PET tracers performed on the ELIXYS resulted in similar or improved radiochemical yields and synthesis times compared to both automated and manual radiosyntheses, without the need for any hardware reconfiguration to make different tracers.

Positron emission tomography (PET) is a highly sensitive non-invasive imaging modality that uses small masses of radiolabeled “tracer” compounds to probe specific biochemical pathways within the whole body. PET is used extensively in clinical diagnostics, basic research, drug development, treatment monitoring, and in the new field of theranostics. To minimize radiation exposure to the radiochemist, the preparation (synthesis, purification and formulation) of PET tracers is performed using automated instruments known as radiosynthesizers. Because current systems are quite expensive, bulky (requiring operation in a specialized radiation-shielded hot cell) and consume tens to hundreds of times as much precursors and reagents compared to the amount of tracer needed for imaging, there has been intense interest in microfluidics in this field. A particular promising avenue has been the use of droplet-based radiochemistry platforms that can produce similar quantities as conventional systems, yet reduce volume and reagents by 100x, reduce synthesis times, and are small enough that they could be self-shielded and operated as a benchtop instrument in a normal laboratory. In addition to advantages for tracer production, this approach enables parallel syntheses by performing multiple reactions together on a single chip, paving the way to rapid, high-throughput radiochemistry for basic investigations, synthesis optimization, and tracer development. In one project, I designed a droplet chemistry heating platform involving a fast ceramic heater for accurate temperature control for a singular droplet synthesis. This initial platform was used in the development of the droplet synthesis of [18F]AMBF3-TATE, realizing an Isolated radiochemical yield (RCY) of 16�1% (n=5) with via isotopic exchange radiofluorination. I also developed a robotic system to perform fully-automated droplet reactions (up to 64 at a time). A Cartesian 3-axis gantry was built that could quickly move a fluidics head (for reagent dispensing and crude product collection) across a 20cm x 40cm workspace containing a set of four multi-reaction chips (each with 16 sites) each controlled by an independent heater, a pipette tip rack, microwell plates for reagents and reaction products, and a set of thin-layer chromatography (TLC) plates for reaction analysis. The system was extensively tested and characterized, and a scripting language was developed to allow high-throughput studies to be programmed as a series of intuitive operation sequences applied to sets of reaction sites. Consistency was assessed by performing a series of identical reactions (synthesis of [18F]Fallypride) which exhibited crude RCY of 79�5% (n=16). This was lower than the manual optimal conditions (90�1%, n=4), and with slightly higher variation. In a second project, I explored microfluidic technology related to the analysis of radiotracers. After the synthesis of a PET tracer, the resulting product must be purified, formulated in saline and finally must pass a number of quality control (QC) tests to ensure safety of the tracer and identity of the product. These tests are typically performed using equipment such as high-performance liquid chromatography (HPLC) which is an expensive instrument that requires significant table space for the device. We designed a “hybrid” capillary and PDMS microfluidic chip for the performance of these QC tests using capillary electrophoresis (CE). I helped develop a volumetric injection chip to inject a 4nL volume of analyte into the separation capillary in a highly repeatable manner. Using this device, injected sample peak relative standard deviation (RSD) was found from the UV absorbance detector electropherogram to be <1.5%, allowing for baseline separation (i.e. resolution > 1.5) of FLT from known synthesis impurities, and enabling impurity species to be quantified independently. The major limitation towards this device performing some of the required QC tests was the lack of a method for in situ radiation detection. Therefore, I added an avalanche photodiode (APD) to the previous device for the purpose of detecting radioactive species. This detector provides highly localized detection due to the short positron range and thus narrow peak widths in the radiation detector electropherogram. The radiation detector enabled confirmation of radiochemical identity, and radiochemical purity, and a limit of quantitation of 114 MBq/mL (3.1 mCi/mL), suitable for testing batches of tracers at higher levels of radioactivity concentration.

The increasing diversity of, and demand for, PET radiotracers have stimulated development of emerging technologies such as microfluidics for more flexible and efficient radiotracer supply. Since the mid-2000s, various microfluidic radiosynthesizers have been reported, demonstrating the potential for smaller footprint and reduced need for radiation shielding. However, there are still a few challenges preventing microfluidic systems from being adopted into the mainstream to replace conventional macroscopic radiosynthesizers.

One of the significant challenges for batch microfluidic reactors has been the incompatibility with harsh radiosynthesis conditions (i.e., high pressures, high temperatures, and the use of organic solvents), limiting the diversity of PET tracers that can be synthesized. To overcome this challenge, a batch microreactor was developed that incorporates phase-change microvalves - a type of valve with the ability to withstand remarkably high pressures. A new phase-change material was used that is compatible with the organic solvents and anhydrous conditions required in radiochemistry. The radiofluorination step in the synthesis of 1-(2'-deoxy- 2'-[¹⁸F]fluoroarabinofuranosyl)cytosine ([¹⁸F]FAC), a radiotracer with a very challenging synthesis, was successfully demonstrated.

Another challenge with small-volume reaction platforms is the lack of online chemical analytical methods to evaluate the reaction progression and to analyze the products. Such online analytical tools are essential to accelerate the development and optimization of tracers on the microfluidic platform. Furthermore, such capabilities could potentially enable integration of the quality control testing required after synthesis to ensure the product is safe for injection, revolutionizing the current pipeline of PET tracer production steps. Detection of electrical properties of droplets in electrowetting-on-dielectric (EWOD) digital microfluidic devices was investigated as a means of achieving these goals. For the first time, conductivity measurement of sample droplets in EWOD platform with high sensitivity and wide dynamic range (3 orders of magnitude of conductivity) was demonstrated. As an example application, this technique was applied to perform on-chip conductometric measurements of a HCl-NaOH neutralization reaction. This reaction is similar to reactions and processes encountered in the production of many radiotracers (e.g., the hydrolysis/deprotection reaction, and the pH neutralization procedure at the end of synthesis).

By addressing two important limitations in microfluidic radiochemistry, the work in this dissertation expands the capabilities of microfluidic platforms for diverse radiotracer synthesis development and production.

Positon emission tomography (PET) is an imaging modality capable of visualizing biomolecules in vivo and can be used to aid in disease diagnosis, staging of disease severity, and monitoring of disease response to treatment. PET relies on the use of tracers (i.e. biomolecules labeled with radionuclides) for imaging. Due to the short half-life of the radionuclides used, PET tracer production is typically performed right before an imaging event. Production of a PET tracer can be broken down into three major parts: production of the radionuclide, radiochemical synthesis of the tracer, and, lastly, purification, formulation and quality control testing of the tracer.

Several groups, including our own, have looked into leveraging the benefits of microfluidics (reduced system size, finer control of reaction parameters, reduced reagent consumption) towards the production of PET tracers. These microfluidic versions of commonly-used PET tracer production equipment enable users to scale up or down the amount of tracer that is produced at a given time. Compared to current commercial systems which are designed to synthesize large batches of PET tracers for clinical applications, these microfluidic systems can offer substantial cost savings by enabling production of smaller batches based on user needs.

Microfluidic approaches have already been applied successfully towards radiochemical synthesis, purification, and quality control testing of tracers. There, however, still exists several steps in tracer production that could benefit from microfluidic technologies. Evaporation of solvent during the concentration and formulation steps following tracer purification is currently performed on slow and bulky rotary evaporators. Microfluidic technology could aid in size reduction while also leveraging microfluidic advantages such as faster heat transfer to increase evaporation rates. Part of this dissertation is focused on applying microfluidic technologies towards the concentration and formulation of PET tracers following purification.

Despite successful design and operation of microfluidic radiosynthesizers, one main limitation that microfluidic radiosynthesizers face is differences between radionuclide volumes (mL) and microfluidic reactor volumes (μL). This disconnect limits the amount of starting radioactivity that can loaded into microscale reactors for radiosyntheses. To address this need, work presented in this dissertation also focuses on the design, fabrication, and testing of an automated microfluidic radionuclide concentrator based on miniaturized anion and cation exchange cartridges enabling the concentration of various types of radionuclides (e.g. fluorine-18 and gallium-68). Concentrated radionuclides in microliter volumes enables microfluidic synthesis of a diverse range of PET tracers in large quantities.

We have also pursued further development of microreactors leading to new advancements to improve the radiosynthesis step during tracer production. Previously our group has demonstrated synthesis of a diverse range of PET tracers using a microfluidic radiosynthesizer based on electrowetting on dielectric (EWOD). Despite advantages of compact size and reduced reagent consumption, a limitation was the cost and complexity of the single-use EWOD chips used during production. In this dissertation, we combat these limitations through the design and fabrication of an automated microfluidic radiosynthesizer based on patterned wettability. This new platform uses reaction chips that are easier to fabricate (compared to EWOD), are a fraction of the cost, and are significantly easier to operate.

Lastly, in this dissertation we demonstrate successful integration of our radionuclide concentrator with our newly design radiosynthesizer. We perform synthesis of [18F]fallypride using high starting activities and demonstrate the ability to produce tracers in high quantities. This integrated platform thus enables both production of low and high quantities of tracer depending on user needs.

The developments presented in this dissertation represent tools for performing portions of the whole PET tracer production process in a more cost effective and efficient manner. Future work will be focused on the successful integration of all components (both microfluidic and conventional) necessary for PET tracer production to enable automated, reliable, high yielding radiosyntheses of clinical grade PET tracers.