Virginia Gambetta, PhD student in the High Precision Proton Therapy research group at OncoRay, received for her work on Partial adaptation for online-adaptive proton therapy triggered by prompt gamma imaging the Young Researcher Award at the annual ESTRO conference 2024 in Glasgow. The jury awarded the best contribution by a junior scientist in the physics track of the conference based on both the outstanding scientific value of the abstract and the clarity, visual quality and layout of the oral presentation at the conference. The prize was sponsored by Elsevier and presented under the auspices of the ESTRO’s open access journal Physics & Imaging in Radiation Oncology (phiRO).
PGI-triggered instead of upfront evaluation-based online-adaptive proton therapy
The delivery of suboptimal dose distributions in proton therapy, e.g. due to random daily changes in organ filling or their position shifts, is today mitigated during the treatment plan generation by sufficiently large safety margins. These could be reduced in the future for the sake of healthy tissue sparing when adapting the treatment plan to the current imaged anatomy of the day while the patient is already on the treatment couch. One vision is to initiate such an online adaptation workflow only within those fractions where it is really necessary and to completely omit it otherwise by verifying the treatment beam in the patient instead. Such an in vivo treatment verification could be done by prompt gamma imaging (PGI), under clinical investigation in our institute, which measures the stopping point of the proton beam in the patient while delivering the first non-adapted treatment field. So instead of daily upfront imaging and subsequent evaluation of the adaptation need one would directly start treatment every day. This would not only save time, and therefore increase patient comfort while decreasing costs and humanpower, it would also reduce the imaging dose over the whole treatment course and thus reduce the associated secondary cancer risk for the patient.
Partial adaptation as good as adaptation of the complete treatment plan for most challenging setting
Such an in-treatment triggered workflow would only be persuasive if there are no relevant compromises in the total delivered dose distribution to the patient when adapting only the remaining part of the treatment plan after the PGI-monitored first field delivery. Virginia Gambetta had initially proposed this method of partial adaptation for speeding up the standard upfront online adaptation process and presented it in a first proof-of-concept simulation study for head and neck cancer patients. Now she tested, together with former OncoRay master student Victoria Pieta, the method of partial adaptation in the context of a PGI-triggered online adaptation in prostate cancer patients. Since typical proton therapy treatment plans for these patients consist of only two opposing treatment fields, it is one of the most challenging situation for partial adaptation when there is only one treatment field left for compensating an obviously suboptimal dose delivered by the first field. The awarded study was based on selected in-room CT imaging data of 10 treatment fractions that confirmed different rectal filling, femoral head position and/or weight loss causing treatment deviations. These deviations were actually detected in retrospectively evaluated PGI data from an ongoing clinical study at OncoRay and would have initiated an alert for an online adaptation demand if patients would have been treated with smaller safety margins. Indeed, it turned out for all fractions that the new concept of PGI-triggered partial adaptation compensates for the dosimetric influence of anatomical changes and especially ensures to deliver the prescribed dose to all parts of the tumor while well sparing the healthy organs. As the results were even as good as for an upfront adaptation of the complete plan, which was a priori not expectable, there would, in principle, be no need to wait for starting the daily treatment since effective adaptive workflows could be triggered later on if necessary. The work also highlighted the value of in vivo treatment verification to ensure optimal proton therapy and together with the concept of partial adaptation the team achieved an important step towards a so-called closed online-adaptive feedback loop.
