Feasibility of position averaged planning-CT for head-neck tumours

Curative intent fractionated (chemo)radiotherapy for head-neck tumours is
delivered with very high spatial accuracy, with patients positioned in a
personalized rigid mask during all imaging and treatment. However, despite this
mask changes in the location of tissues can occur. This includes setup
variations between separate positionings, and incidental motion such as flexing
of the neck, swallowing, and coughing in treatment position. Because the
planning CT (pCT) that is used for target delineation is prone to these issues,
it may not optimally represent the average location of tissues during
treatment. Furthermore, discrepancies between the location of tissues on pCT
and other imaging (such as PET) may cause confusion in multimodal target
definition. These geometrical uncertainties can lead to suboptimal target
definition and dose delivery, and are currently compensated by a uniform PTV
expansion of 3 mm in all directions.

Van Kranen et al. recently demonstrated the concept of cone beam CT (CBCT)
based position averaged planning CT using non-rigid deformation maps from
consecutive CBCT scans that were acquired during the first fractions of
radiotherapy (pa-pCTCBCT). The pa-pCTCBCT was able to represent the average
anatomical position of tissues during treatment, it revealed systematic
deviations of standard pCT relative to the averaged anatomy of 2.5mm for bony
structures and 3.4 mm for soft tissues, and re-planning of treatments based on
pa-pCTCBCT resulted in lower systematic deviations. However, this approach is
only applicable to re-planning during an ongoing treatment and it cannot
contribute to the initial phase of (multimodal) target definition. We propose
to apply lowdose CT (LDCT) based position averaging to the planning CT
(pa-pCTLDCT), to achieve better representative target definition with less risk
on systematic errors in treatment delivery, to provide easier correlation with
PET, and to test the potential for patient- and location specific PTV margins.

To determine the usefulness of pa-pCTLDCT to depict the average anatomy in the
head/neck area during the first week of treatment, compared to standard pCT.

Secondary objectives are to determine if pa-pCTLDCT has better correlation than
standard pCT with the location of metabolic lesions on FDG PET, and to
determine the technical feasibility of deforming PET+LDCT to pa-pCTLDCT and of
deriving patient-specific and location-specific anatomical variations from
pa-pCTLDCT.

Monocenter prospective observational pilot study

No significant side effects are expected. Five additional LDCTs of the
head-neck area will give an additional radiation burden of 5 x 2 mSv = 10 mSv,
this is in the range of standard diagnostic procedures and is not considered a
relevant risk for these patients with cancer who will receive high dose
external beam radiotherapy. Patients will spend an additional 30 minutes in the
hospital on the day of pCT imaging to receive the additional LDCTs (performed
in the same session). Other diagnostic procedures and treatment will not be
changed or delayed by participation in this study.