Cherenkov imaging is a novel technique that captures light emissions during radiation therapy, allowing for visualization of radiation treatments on patients, in real-time. We hypothesized that simply viewing the Cherenkov video images, both in real-time and with post-treatment review by radiation therapists, would identify events not previously reported in the existing QA program.In December 2020, Cherenkov imaging cameras were introduced into an academic medical center with an existing QA program consisting of a hospital wide incident reporting system that is anonymous, voluntary, and non-punitive. Events are reviewed monthly by a multidisciplinary group including representatives from radiation therapy, dosimetry, nursing, physics and radiation oncologists. The Cherenkov cameras were installed in each treatment bunker, positioned laterally on each side of the couch. The cameras provided continuous, real-time video images of the patients and visualization of the irradiated tissue. Live viewing of the treatments was provided via a dedicated monitor in the console room. All treatments were imaged with the exception of treatments where optical surface imaging lights were on for SGRT.During this 3-month period, 12 events were reported in the hospital-based incident reporting system. Events were reviewed and categorized as 3 operational/process improvement (e.g., scheduling errors), 3 other safety events (e.g., patient falls), 3 treatment planning errors (e.g., wrong shifts calculated for setup), 1 prescription transcription error, 1 treatment delivery error (a missed treatment field), and 1 simulation error (suboptimal immobilization equipment used). Aside from a patient fall, all events were deemed to have no detectable harm to the patient. In this same time period, review of Cherenkov images identified 3 treatment delivery events, which were not identified by other means. The first was an AP-PA thoracic spine treatment and on one day, the treating therapists noted the patient's chin in the treatment field. Treatment was stopped, the patient was re-positioned for the remainder of the treatment. Second was an AP-PA lumbar spine field that on post-treatment review was noted that the patient's hands moved into the field. The third was a 3-field sacrum plan that on post-treatment review was noted that for 7 of 10 fractions, the patient's left arm was positioned over the exit RPO beam. Physics review estimated that the uninvolved arm received approximately 3 Gy.Viewing of Cherenkov emission imaging by the treatment team identified delivery incidents due to non-ideal patient positioning during treatment and these events were not identified in the existing QA program. Future work will focus on determining incident rates detected by Cherenkov imaging and if this imaging can identify and/or avoid treatment delivery errors from reaching the patient. Automated detection and near real time notification of such events is a work in progress.
Read full abstract
Dec 1, 2021
Kenneth Groves 
Open Access