Photon iso-effective dose for cancer treatment with mixed field radiation based on dose-response assessment from human and an animal model: clinical application to boron neutron capture therapy for head and neck cancer.

Boron neutron capture therapy (BNCT) is a treatment modality that combines different radiation qualities. Since the severity of biological damage following irradiation depends on the radiation type, a quantity different from absorbed dose is required to explain the effects observed in the clinical BNCT in terms of outcome compared with conventional photon radiation therapy. A new approach for calculating photon iso-effective doses in BNCT was introduced previously. The present work extends this model to include information from dose-response assessments in animal models and humans. Parameters of the model were determined for tumour and precancerous tissue using dose-response curves obtained from BNCT and photon studies performed in the hamster cheek pouch in vivo models of oral cancer and/or pre-cancer, and from head and neck cancer radiotherapy data with photons. To this end, suitable expressions of the dose-limiting Normal Tissue Complication and Tumour Control Probabilities for the reference radiation and for the mixed field BNCT radiation were developed. Pearson's correlation coefficients and p-values showed that TCP and NTCP models agreed with experimental data (with r > 0.87 and p-values >0.57). The photon iso-effective dose model was applied retrospectively to evaluate the dosimetry in tumours and mucosa for head and neck cancer patients treated with BNCT in Finland. Photon iso-effective doses in tumour were lower than those obtained with the standard RBE-weighted model (between 10% to 45%). The results also suggested that the probabilities of tumour control derived from photon iso-effective doses are more adequate to explain the clinical responses than those obtained with the RBE-weighted values. The dosimetry in the mucosa revealed that the photon iso-effective doses were about 30% to 50% higher than the corresponding RBE-weighted values. While the RBE-weighted doses are unable to predict mucosa toxicity, predictions based on the proposed model are compatible with the observed clinical outcome. The extension of the photon iso-effective dose model has allowed, for the first time, the determination of the photon iso-effective dose for unacceptable complications in the dose-limiting normal tissue. Finally, the formalism developed in this work to compute photon-equivalent doses can be applied to other therapies that combine mixed radiation fields, such as hadron therapy.

Tumor control and normal tissue complications in BNCT treatment of nodular melanoma: a search for predictive quantities.

A previous work concerning tumor control and skin damage in cutaneous melanoma treatments with BNCT has been extended to include doses, volumes and responses of 104 subcutaneous lesions from all patients treated in Argentina. Acute skin reactions were also scored for these patients, and cumulative dose-area histograms and dose-based figures of merit for skin were calculated. Broadening the tumor response analysis with the latest data showed that the (minimum or mean) tumor dose is not a good predictor of the observed clinical outcome by itself. However, when the tumor volume was included in the model as second explicative variable, the dose increases its significance and becomes a critical variable jointly with the volume (p-values<0.05). A preliminary analysis to estimate control doses for two groups of tumor sizes revealed that for small tumor volumes (< 0.1cm(3)) doses greater than 20 Gy-Eq produce a high tumor control (> 80%). However, when tumor volumes are larger than 0.1cm(3), control is moderate (< 40%) even for minimum doses up to 40 Gy-Eq. Some quantities based on skin doses, areas and complication probabilities were proposed as candidates for predicting the severity of the early skin reactions. With the current data, all the evaluated figures of merit derived similar results: ulceration is present among the cases for which these quantities take the highest values.

A new approach to determine tumor-to-blood 10B concentration ratios from the clinical outcome of a BNCT treatment.

A new approach to determine the tumor-to-blood (10)B concentration ratio in boron neutron capture therapy (BNCT) is introduced. It is a statistical method, which uses maximum likelihood estimation on the clinical outcome of a BNCT treatment. Its performance is shown in a clinical case of cutaneous multiple nodular melanomas. The calculations involve a detailed dosimetry analysis, the determination of tumor control probabilities for the different nodules, the maximum likelihood estimation itself, and a parametric bootstrap to obtain confidence intervals for the tumor-to-blood ratio. The obtained ratio is 3.05 +/- 0.46 with a 95%-confidence interval. These results are consistent with those found in literature. Moreover, a single patient with multiple nodules proves enough to get statistically relevant results. The proposed method does not involve surgery and can be performed after a BNCT treatment without being invasive for the patient.