ESTRO 2022

Session Item

Sunday

May 08

14:15 - 15:15

Mini-Oral Theatre 1

13: Implementation of new technology

Chair: , Italy;

Co-Chair: Stefanie Ehrbar, Switzerland

Session Code: 2430

Session Type: Mini-Oral

Track: Physics

Pre-study QA reduces delineation errors and minimize dose to OAR in a central lung tumour SBRT trial

Lone Hoffmann, Denmark

Presentation Number: MO-0550

Abstract

Abstract Title:

Pre-study QA reduces delineation errors and minimize dose to OAR in a central lung tumour SBRT trial

Authors:

Lone Hoffmann¹, Gitte F Persson², Lotte Nygård³, Tine B Nielsen⁴, Svetlane Borrisova⁵, Fatma Gaard-Petersen⁵, Mirjana Josipovic⁶, Azza A Khalil⁷, Rasmus Kjeldsen⁸, Marianne M Knap⁹, Charlotte Kristiansen¹⁰, Ditte S Møller⁹, Wiviann Ottosson⁵, Hella Sand¹¹, Rune Thing¹², Mette Pøhl¹³, Tine Schytte¹⁴

¹Aarhus University Hospital, Dept of medical Physics, Aarhus N, Denmark; ²Copenhagen University Hospital, Dept of Oncology, Herlev-Gentofte Hospital, , Copenhagen, Denmark; ³Copenhagen University Hospital, Dept of Oncology, Rigshospitalet, Copenhagen, Denmark; ⁴Odense University Hospital, Laboratory of Radiation Physics, Odense, Denmark; ⁵Copenhagen University Hospital, Dept of Oncology, Herlev-Gentofte Hospital, Copenhagen, Denmark; ⁶Copenhagen University Hospital, Dept of Oncology, Rigshospitalet, Copengaheg, Denmark; ⁷Aarhus university Hospital, Aarhus, Dept. of Oncology, Aarhus, Denmark; ⁸ Aalborg University Hospital, Dept. of Oncology, Aalborg, Denmark; ⁹Aarhus university Hospital, Dept. of Oncology, Aarhus N, Denmark; ¹⁰University Hospital of Southern Denmark, 8. Dept of Oncology, Lillebaelt Hospital, Vejle, Denmark; ¹¹Aalborg University Hospital, Dept of Medical Physics, Aalborg, Denmark; ¹²University Hospital of Southern Denmark, Dept of Oncology, Lillebaelt Hospital, Vejle, Denmark; ¹³ Copenhagen University Hospital, Dept of Oncology, Rigshospitalet, Copenhagen, Denmark; ¹⁴Odense University Hospital, Odense, Dept of Oncology, Odense, Denmark

Show Affiliations

Purpose or Objective

Stereotactic Body RT (SBRT) of central lung tumours are associated with a high risk of serious toxicity. Our multicentre STRICT LUNG trial evaluates safety and efficacy of SBRT for central tumours treated inhomogeneously to 56Gy in 8 fractions. We prioritized organs at risk (OAR) constraints over target coverage, making the dose plans very sensible to OAR delineation variations. Therefore, the Quality Assurance (QA) programme focused on adherence to protocol in OAR delineation, the dosimetric effects of interobserver variation and priorities in dose planning.

Material and Methods

The trial consensus and QA process consisted of two parts: Delineation of bronchial tree and mediastinal tissue; Dose planning.

For the delineation (D) part, CT scans with GTV delineated were distributed to seven physicians. Delineations were performed in three rounds, each with one new case.

D1: Delineation based on guidelines, followed by consensus meeting

D2: Delineation followed by consensus meeting leading to creation of an atlas

D3: Delineation

Delineation variability was measured as mean (H_mean) and maximum (H_max) Haussdorff distance. To evaluate the dosimetric effect of OAR delineation variability for each case, seven simulated treatment plans, repeatedly using each of the OAR structure sets from the seven physicians, were made and compared. All plans were made by one physicist using the same planning parameters.

For the treatment planning (TP) part, CT scans with GTV, PTV and OAR delineated were distributed to six dose planners. Two rounds were conducted.

TP1: Five cases. >95% of PTV should be covered by 95% of 56Gy. Dose to the GTV should be as high as possible. Constraints were set for all OAR and had higher priority than GTV dose and PTV coverage. Discussion at a consensus meeting. Maximum dose of 85Gy was set for the mean dose to the GTV.

TP2: Dose plans for two new cases.

Results

The median H_mean was ≤1mm. The median H_max was significantly shorter (p<0.05) for D3 (13mm), than for both D1 (21mm) and D2 (19mm). OAR delineation variation between centres resulted in high variabilities in OAR dose for simulated plans. In six plans for each of D1-D3, the variability in lobar bronchus delineation led to potential overdosage (constraint: D0.03cc<45Gy), with max D0.03cc of 58Gy (D1), 50Gy (D2) and 50Gy (D3) (Fig 1). For the mediastinal tissue, the constraint (D0.03cc<45Gy) was violated for the majority of the delineations in all three rounds, with max D0.03cc of 84Gy (D1), 85Gy (D2) and 72Gy (D3). 10 (D1), 14 (D2) and 4 (D3) plans led to mediastinal tissue dose >70Gy (Fig 1). For the dose planning study, the range of the standard deviation for GTV mean dose was 12.8-18.5Gy in TP1 and 2.8-3.5Gy in TP2 (Fig2).

Conclusion

Even small variations in OAR delineation led to high over-dosage of OAR. Extensive QA with consecutive delineation and dose planning rounds decreased variability in OAR delineation and dose planning and emphasises the need of thorough QA while conducting a multicentre trial.