2383 - Clinical Observation and Timing Log-File Dose Reconstruction of Phantosmia in Pediatric Patients Undergoing Proton Craniospinal Irradiation
Presenter(s)
H. Stebbeds1, J. Zhang1, J. Becksfort1, A. M. Faught2, J. Gattuso1, S. Day3, T. E. Merchant1, and C. C. Chen1; 1St. Jude Children's Research Hospital, Memphis, TN, 2The Ohio State University Wexner Medical Center, Columbus, OH, 3University of Tennessee Health Science Center, College of Nursing, Memphis, TN
Purpose/Objective(s): Radiation-induced phantosmia, defined as an olfactory hallucination where a smell is perceived but not present in the environment, has been observed during radiotherapy, including proton craniospinal irradiation (CSI). It may provoke nausea, vomiting, heightened anxiety, and a decline in quality of life. Clinical observations of phantosmia and dose reconstruction based on treatment log file segments aligned with patient-reported phantosmia timing are presented.
Materials/Methods: A survey was conducted among non-anesthetized patients who received proton CSI between June 2024 and December 2025. A single patient who experienced phantosmia was recruited to use a clicker to record the time of onset and end of smell perception during treatment. Delivered doses corresponding to time intervals before, during, and after the reported phantosmia were reconstructed in the treatment planning system (TPS) using treatment log files. The workflow was experimentally validated using a QA field.
Results: Of 15 participants under CSI, 14 reported phantosmia, described as the scent of bleach or pool chemicals, during irradiation of the brain. Other sensory phenomena were also noted, such as blue lights. Dose reconstruction using treatment log files was validated in the TPS for a 100 MeV field by comparing measurements from a 2D ion chamber array collected over a 10-second interval. Gamma analysis showed a 99.2% passing rate for a ±3%/2mm criterion. Fractional dose (Rx:180 cGy(RBE) for the patient treatment was reconstructed using treatment log file data corresponding to the periods when the participant reported phantosmia via the clicker. The results revealed a high-dose region of 70–80 cGy(RBE) from a single field near the cribriform plate coinciding with the phantosmia event.
Conclusion: Dose reconstruction suggests an association between phantosmia and proton spots reaching regions near the olfactory nerve. During a brief time window, the modulated proton fluence contributed nearly 90% of the field dose at the time phantosmia was identified. The dose threshold associated with this sensory effect can be further estimated, and strategies to reduce or prevent phantosmia, should be explored and evaluated in future studies.