Purpose: Cherenkov signals generated in patients treated with TSET can be recorded in-real time to monitor the dose distribution on patient body surface during the treatment. However, the front facing camera alone cannot predict dose accurately near periphery areas (e.g. Vertex) of the patient due to large angle emission of Cherenkov light.
Methods: The TSE patients were treated in Stanford technique at 500 cm source to patient midplane distance using dual-field 6-MeV electron beams at gantry angles 74° and 106° at 6 postures with HDTSE mode delivered by Varian Truebeam. Three ceiling mount time-gated intensified cameras were used to capture the Cherenkov signal generated from patient’s skin surface. Scintillator luminescent detectors (SLD) and optically stimulated luminescent dosimeters (OSLDs) were taped to patient’s chest, umbilicus, and side to record and verify the dose delivered to those locations. Prospective (geometrical) calibration was performed in 2-dimensions to correct the camera vignetting effect, lens correction and different pixel response. Dose was reconstructed from the Cherenkov imaging by normalizing intensity to the OSLD measurements at the same location.
Results: Perspective calibration was studied in 2D for Cherenkov imaging for all cameras and applied to acquired patient data. The corrected Cherenkov intensity can be converted to dose by normalizing to OSLD or SLD measurements at one location. Using additional Cherenkov cameras improved agreement in vertex from 40-50% low to within 5% for most of the patients compared to the in vivo dose measurements.
Conclusion: Cherenkov imaging provides valuable information about the dose distribution for TSET. Geometrical correction was established in 2D for Cherenkov imaging and dose distribution can be reconstructed from the Cherenkov imaging.