Proper administration of a diagnostic radiopharmaceutical is essential for nuclear medicine image quality and quantification (1–8). Extravasation, an incomplete delivery of a radiopharmaceutical into the venous system, can compromise imaging, the patient's ensuing care, and sometimes lead to unnecessary additional radiation dose for repeat imaging studies. In addition, large extravasations can pose radiation harm to patients (9).

Studies suggest that an average of 15% (2 to 23%) of radiopharmaceutical administrations may result in an extravasation (5, 6, 10–14), but currently there is no routine quality control to ensure complete delivery of the radiopharmaceutical into the patient's circulation. However, with adequate training and ongoing injection quality monitoring, technologists are capable of improving injection quality significantly (15) and sustaining the results by implementing ongoing monitoring (16).

In 2016, our center participated in a multicenter ¹⁸F-FDG radiopharmaceutical administration QI Project, using Define, Measure, Analyze, Improve, Control methodology (15). The goal of the Project was to monitor injection quality, use analysis of factors that contribute to extravasations to guide improvements, remeasure rates in similar number of patients, and evaluate sustainability of interventions.

A quality improvement plan (QIP) was created on the basis of the analysis of contributing factors and in discussion with each technologist participating in the study. The QIP included three components: addition of an auto-injector to provide consistent infusion and flush parameters across injections, adjustment of uptake room setup to allow for improved access to left and right arms of the patient, and peer-led refresher training for venous access and injection technique. Following the QIP the technologists reduced the extravasation rates from 13.3% (Measure Phase) to 2.9% (Control Phase) (16). The 2.9% was a combination of auto injections extravasation rate (0.6%) and manual injection extravasation rate (7.1%). Ongoing monitoring ensured continued sustainability of results and further reduced manual injection extravasation rate to 3.4%.

At our clinic, general nuclear medicine procedures were performed in a different facility than PET/CT procedures, but the technologists who performed ^99mTc-MDP administrations were also the same ones who performed ¹⁸F-FDG administrations. These technologists were aware of the steps required to reduce ¹⁸F-FDG administration extravasations. As a result, we hypothesized that they would apply ¹⁸F-FDG quality improvement lessons-learned to ^99mTc-MDP administrations, and the resulting ^99mTc-MDP administration manual injection extravasation rate would not be statistically different from the ongoing manual injection extravasation rate (3.4%) in ¹⁸F-FDG administrations.

We tested our hypothesis by adding monitoring equipment 30 months after completing the PET/CT administration quality plan, and following quality improvement methodology, measured a baseline extravasation rate for ^99mTc-MDP administrations. Our findings suggest a new hypothesis: active monitoring of radiopharmaceutical administrations is critical to ensuring radiopharmaceutical administration quality.

The Institutional Review Board of Carilion has determined that quality improvement efforts for ¹⁸F-FDG do not meet the definition of research as defined by the federal government in 45 CFR 46.102(d) and therefore, no patient consent was required. No protected health information was collected or transmitted. Every patient undergoing routine ^99mTc-MDP bone scans was eligible to be included in the QI methodology. Patients followed standard clinical preparations for SPECT/CT imaging and were not exposed to additional radiation during this Project.

For 30 months after completion of the ¹⁸F-FDG quality improvement project we did not actively monitor the quality of ^99mTc-MDP administrations. We then introduced a Lara® System (Lucerno Dynamics, Cary, NC) designed for capturing ^99mTC emissions to help clinicians determine the quality of MDP administrations. The Lara® System is comprised of gamma scintillating sensors, adhesive pads, a reader, and software (Figure 1). After technologists gain venous access, they place the sensors on the injection arm and in a mirrored location on the other arm. The sensors identify the presence of excess radiopharmaceutical near the injection site and are connected to a reader that provides real-time display of activity in each arm. The reader also stores data and later transmits the recorded scan to a server for processing of time-activity curves (TACs).

Bone SPECT images were acquired with a Siemens Intevo Bold (Siemens Medical Solutions USA, Inc.) after standard injected activity of 25mCi ^99mTc-MDP using a LEHR collimator and matrix size of 256 × 256 with a zoom of 1. SPECT data were acquired using 30 views in a noncircular (body contour) orbit with an acquisition time of approximately 20 s per view. For CT Acquisition, data were acquired using 130 kVp with CareDose 4D and pitch of 1.5 and detector settings of 16 × 1.2 mm with 3 mm slice thickness.

Data were collected on 816 administrations during 16-month period and the extravasation rate was found to be 7.84%. A closer examination of the 16-month (four 4-month periods: A, B, C, D) and potential contributing factors revealed that a significant reduction in extravasation rate had occurred over these periods (Figure 2).

Period A (first four months of active monitoring) extravasation rate was not statistically different from the Measure Phase extravasation rate of the previously completed PET/CT QI Project: 12.75% compared to 13.3% (p-0.7925). Periods C (months 9–12) and D (months 13–16) extravasation rates were statistically lower than period A extravasation rate: 2.94% and 3.43% (p < 0.0001) compared with 12.75%. After analyzing contributing factors and implementing improvements, technologists achieved extravasation rates during Period C and D comparable to the longstanding manual ¹⁸F-FDG injection extravasation rate (3.4%).

Our contributing factors analysis suggests that during Period A, injection site location, venous access technique, and needle size, were associated with higher extravasation rates (Table 1). Further analyses suggest that when technologists stopped using the venous access technique of straight sticks (Table 2) and 23-gauge needles (Table 3), those factors were no longer associated with extravasation rates in Period D.

Our hypothesis that problem awareness and solution availability from the previously completed quality improvement project would result in radiopharmaceutical administration process improvement was not accurate. While those factors are important, they are not sufficient. At our center, active monitoring and the associated display of results have proven to be critical to quality improvement efforts to reduce and sustain radiopharmaceutical extravasation rates. Active monitoring provides information necessary to determine contributing factors and when combined with active management review of displayed results, leads to improved administration quality. Our findings related to active monitoring are also supported by another center's quality improvement results¹.

While ongoing active monitoring may be necessary for sustained quality administrations, it has many other implications.

Our initial hypothesis that awareness of a problem and the steps needed to correct would result in process improvement was incorrect. In our center, technologists trained in QI methodology had not applied lessons-learned from their ¹⁸F-FDG PET/CT QI Project to ^99mTc-MDP administrations. However, by requiring active monitoring and displaying results, the trained technologists determined contributing factors and reduced extravasation rates significantly. These findings are consistent with our previous experience in PET/CT QI findings from another center, and the well-known observer effect phenomenon. These findings also suggest that centers that want to improve their radiopharmaceutical extravasation rates can benefit by implementing active monitoring.