Online

ESTRO 2020

Session Item

Physics track: Dose measurement and dose calculation
9319
Poster
Physics
00:00 - 00:00
EPID 2D transit In Vivo Dosimetry: Can relevant anatomy and positioning differences be detected?
Núria Jornet, Spain
PO-1365

Abstract

EPID 2D transit In Vivo Dosimetry: Can relevant anatomy and positioning differences be detected?
Authors: Mar Adria.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Pablo Carrasco de Fez.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Paula Delgado.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Nuria Espinosa.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Pedro Gallego.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Núria Jornet.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Artur Latorre-Mussoll.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Jaime Perez-Alija.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Montserrat Ribas Morales.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Agusti Ruiz.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Pablo Simon.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain), Itziar Valverde.(Hospital de la Santa Creu i Sant Pau, Medical Physics, Barcelona, Spain)
Show Affiliations
Purpose or Objective

To study the sensitivity of a 2D EPID transit IVD system to detect imposed anatomy and positioning differences under controlled experimental conditions.

Material and Methods

A sensitivity test using QUASAR phantom (Modus) was designed mimicking a lung case. One configuration (Fig1a) was CT-scanned for treatment planning.  Three plans were designed to treat the PTV (2.5 cm diameter sphere + 1.5 cm margin): 2-field 3DCRT, 6-field IMRT and 1-partial arc VMAT. Eclipse V13.5 with AAA dose calculation algorithm was used. These plans were then delivered on the original configuration and with eleven modified phantom settings (Fig 1b-l). For each fraction, CBCT, logfiles and EPID integrated images were recorded and exported to PerFRACTION™ (Sun Nuclear Corporation). 2D transit dose measurements were compared to the predicted 2D dose for the first fraction (1a). For the remaining fractions, the reference for comparison was set to fraction 1.  Two metrics were used: global gamma 3%-3mm, global dose differences 3%; threshold 10%. The sensitivity of the 2D transit IVD was assessed. At the same time, by using the CBCT and logfiles, the impact on the imposed anatomy and position differences on the DVHs for the CTV, PTV, and two structures were evaluated. The 2D IVD pass rates were compared to the 3D dose calculations on the CBCT to assess if the in vivo technique would be capable of detecting dose differences larger than 2% in delineated structures.




Results

A 3% dose systematic difference was found between the predicted and measured 2D EPID dose for each field for configuration (1a) for all treatment plans. This difference was attributed partially to a reference dose offset of 1.2% on the EPID calibration day and partially to the EPID dose prediction model.

The sensitivity of the 2D in vivo method to anatomy changes, per field, using gamma 3%-3mm pass rate 95%, was 81%, 97% and 80% for 3DCRT, IMRT and VMAT while it was 94%, 97% and 100% if the comparison was made in terms of dose differences.  If the mean for all fields was used, the sensitivity increased to 100% for 3DCRT and IMRT plans.

The dose differences of the imposed changes, together with the 2D IVD pass rates, are shown in fig 2 for one metric (dose difference> 3%)
Conclusion


The PerFRACTION module for 2D IVD performs well regardless of the treatment technique, being capable of identifying all the changes imposed on the phantom. However, the systematic shift from predicted to measured 2D dose must be minimized to use predicted 2D dose as the reference. 

The sensitivity of the system to detect clinically relevant changes depends on the dose distribution with respect to the volumes of interest and changes. Increasing the distribution of beam portals increases this sensitivity. From this phantom study it can be concluded that even if large differences in 2D passing rates (pass rate<50%) correctly identify dose differences larger than 4%, to track smaller changes, moving to other approaches other than the current dose or DTA metrics is needed