Session Item

Re-irradiation of locally recurrent nasopharyngeal carcinoma (rNPC) is usually performed with Intensity Modulated Radiation Therapy (IMRT). However, it is often associated with sub-optimal target coverage due to tight constraints of the previously heavily irradiated organ-at-risks (OARs). Although manual use of non-coplanar beams might improve the situation, the process is often un-intuitive and in-efficient. In recent years, there are growing interests in expanding the use of an automated non-coplanar volumetric modulated arc therapy technique ‘Hyperarc’ (HA), which was originally introduced for stereotactic radiosurgery, onto extra-cranial cases. This study aims to investigate the potential dosimetric advantages of HA in terms of target volume coverage, plan quality metrics and sparing of OARs in re-RT of rNPC.

Twenty patients with locally rNPC who have previously received a 70Gy-irradiation for primary disease were retrospectively included in this study. For each patient, treatment plans for the recurrent disease were computed using IMRT and HA. The first-priority cumulative BED limits for OARs were defined as 130% of the primary course BED limit. In case the first-priority constraints could not be fulfilled, the secondary-priority cumulative BED limits, which were defined as 150% of the primary course BED limit for all OARs, except 160% for temporal lobes, were followed. The planning goal was to give maximal permissible dose to PTV within the OAR dose limits. Dose delivered to target volumes, plan quality metrics (i.e. PTV V100%, conformity index CI, homogeneity index HI, high-dose spillage and gradient index GI), and dose delivered to the planning-risk volumes (i.e. brainstem, spinal cord, optic chiasm, lens, temporal lobes and optic nerves), internal carotid arteries (ICA) and the nasopharyngeal (NP) mucosa were evaluated.

HA plans significantly improved target volume coverage (GTV D95%: +9.9%, CTV D95%: +11.5%, PTV D95%: +12.1%) while achieving dose escalation by delivering higher average mean dose to target volumes (PTV: +6.8%, CTV: +6.8%, PTV: +5.5%). The HA plans were also associated with better quality in terms of PTV V100%, CI, high-dose spillage and GI without a significant difference in HI. Moreover, HA plans delivered significantly lower average maximum dose (Dmax) to spinal cord (-1.94Gy), brainstem (-3.34Gy), optic chiasm (-1.05Gy), left temporal lobe (-4.73Gy), encapsulated ICA (-0.88Gy) and non-encapsulated ICA (-12.8Gy). However, HA was also associated with a significantly higher Dmax to bilateral lens (+1.47Gy) and NP mucosa (+5.4Gy).

HA is dosimetrically feasible in re-irradiating rNPC patients with less under-dosage of target volumes when compared with IMRT. However, it must be reiterated that given the high-dose hotspots within the target volumes, high dose delivered to the NP mucosa and high fatality rate of carotid blowout, HA must be used with stringent patient selection and detailed dosimetric evaluation.