Encrypted login | home

Program Information

Automated and Robust Dose Restoration in IMPT: Reaching Dose Stability Under Anatomical Changes in Head and Neck Cancer Patients

no image available
K Bernatowicz

K Bernatowicz1*, X Geets1 , E Sterpin1,2 , ,(1) Universite catholique de Louvain, Imagerie Molucalaire et Radiotherapie Experimental, Brussels, ,(2) KU Leuven, Leuven, Vlaams-Brabant


SU-I-GPD-T-167 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To automatically restore an IMPT robust plan dose, containing clinically approved dose-volume trade-offs, on a daily CT image.

Methods: The dose restoration (DR) approach was tested on three clinical head and neck cases with three CTs (week 1-3). Pencil-beam proton plans with four beams were optimized in Raystation.v.5 using simultaneous integrated boost, with median doses of 70Gy and 56Gy prescribed to the therapeutic and prophylactic CTVs (nodal and primary). Plans were optimized using: (P1) PTV (=CTV+4mm), (P2) same PTV and robust optimization (3% HU changes), and (P3) CTV and robust optimization (3% HU and 4mm setup errors). DR was guided using initially approved structures, rigidly projected on new anatomy, and a set of iso-dose volumes generated from an initial plan. Iso-dose volumes were generated every 2% dose increments within 95-107% of 70Gy. A 5% increment was used otherwise. DR plan was obtained automatically in <5min (PTV-based plans), without manual adjustments of optimization settings.

Results: P1-P2 plans were evaluated by calculation on CT1-3. Robustness of P3 plans was judged on worse-case scenario. For PTV 70Gy, V95≥98% and V107≤2% was not satisfied in 10/27 of initial (non-adapted) plans (V95%=97.03±1.4%, V107%=4.69±0.3%), which was regained with 27/27 DR plans. Dose error (DE) histograms in PTV 70Gy and PTV 56Gy shown reduced median DE5% in DR plans (2.4% and 3.8% compared to 3.7% and 7.4% for non-adapted initial plans). Dose was also stabilized in organs at risk, with median DE5% below 5%, as specified by requested iso-dose volume levels.

Conclusion: DR is a step towards an on-line adaptive strategy. On contrary to re-optimization on daily image using objectives of initial plan, DR reproduces dose gradients of initial plan in whole patient and hence has a potential to by-pass the need of physician’s approval of a ‘daily-plan’, as long as dose stability can be assumed.

Funding Support, Disclosures, and Conflict of Interest: Funding disclosure. Kinga Bernatowicz works as a Post-doctoral researcher at UCL (Universite catholique de Louvain, Belgium) and is supported 50% by the Baillet-Latour Fund and 50% by IBA in terms of a joint collaboration project between UCL and IBA (ImageX-research).

Contact Email: