Brachytherapy is an effective local treatment for early-stage head and neck cancers. Mold irradiation is a method in which the source is placed in the oral cavity in sites where the soft tissue is thin and an irradiation source cannot be implanted. However, dose calculations based on TG-43 may be subject to uncertainty due to the heterogeneity of tissues and materials used for the irradiation of head and neck cancers.

In this study, we investigated the basic physical properties of different materials and densities in the molds, retrospectively analysed patient plans and verified the doses of intraoral mold irradiation using a dose verification system with MC simulations specifically designed for brachytherapy, which was constructed independently.

Dose–volume histograms were obtained with a treatment planning system (TG-43) and MC simulation and revealed a non-negligible difference in coverage of high-risk clinical target volume (HR-CTV) and organ at risk (OAR) between calculations using computed tomography values and those with density changes. The underdose was 10·6%, 3·7% and 5·6% for HR-CTV, gross tumour volume and OAR, respectively, relative to the treatment plan. The calculations based on the differences in the elemental composition and density changes in TG-43, a water-based calculation algorithm, resulted in clinically significant dose differences. The validation method was used only for the cases of complex small source therapy.

The findings of this study can be applied to more complex cases with steeper density gradients, such as mold irradiation.

The control of patient position, posture and respiratory movements during radiotherapy is important for effective and specific treatment of malignancy. We have developed an infrared (IR) interactive patient position guidance and acquisition control system for clinical use, comprising IR cameras, IR markers and dedicated software.

We evaluated the system with ten healthy volunteers and ten experienced operators. IR markers were placed on the body surface. Their positions were calculated using vectors of three translational and three rotational parameters, and the intrafractional error for each marker was acquired with and without respiratory motion. The inclusion of multiple positioning markers allowed for real-time visualisation of the patient posture, with feedback on misalignment and required postural adjustments.

The positioning time was 73 seconds (with a minimum period of 39 seconds), which was significantly shorter than for conventional line alignment. A comparison of positioning reproducibility between conventional line alignment and this system was <3·5 mm and was not patient dependent or operator dependent. An intrafractional error of displacement of up to 10·0 mm was found in the right iliac crest.

This IR interactive system was shown to be high utility and suitable for monitoring patient position, posture and respiratory movements during radiotherapy.