Abstract Book

S1082

ESTRO 37

Material and Methods Fifty head and neck cancer patients were treated with bilateral radiation therapy in a prospective trial. Salivary gland function was determined using two different methods: stimulated total salivary flow (F , averaged over 15 min) and dynamic salivary gland scintigraphy, where salivary ejection fraction (sEF) is determined after lemon juice stimulation at 15 min after TcO 4 - activity administration. Patients received image guided IMRT with radical or postoperative intent using 6 MV photons. The mean dose within the total volume of parotid glands is used as a dosimetric parameter. Both functional parameters are measured at baseline and 6 months after radiotherapy. Response is defined as relative proportions rF = F(6 mo)/F(baseline) or rEF = sEF(6 mo)/sEF(baseline) from baseline values. Additionally, rF results are compared with data reviewed in the QUANTEC salivary gland article (Deasy JO et al. IJROBP 76: Supplement, S58-S63, 2010). Results Averaged data points arranged in five groups are plotted as a function of mean parotid dose (stimulated flow: both parotids averaged, scintigraphic ejection fraction: per each gland) (Fig 1). When linear model for both total saliva flow and single gland ejection fraction are fitted a few interesting findings arise: - both rEF and rF lines intersect x axis at around 50 – 55 Gy indicating total loss of salivary function. - the rEF line intersects y axis at 1.28; this is explained either by linear model inadequacy at low doses (sigmoidal better?) or functional compensation of the remaining functional gland at 6 months after therapy. - the rF line intersects y axis at 0.8; this is explained by the omittance of other glands (submandibular and minor glands) from the flow model: the parotids are responsible of about 70 % of stimulated flow. - D 50 values for both curves are almost equal when calculated from the intersect point with y-axis: D 50 (rF) = 25.0 Gy, D 50 (rEF) = 26.7 Gy

Purpose or Objective For Head and Neck (HN) cancer patients the dose received by the skin is critical with respect to the risk of developing acute and severe skin toxicities. However, studies dealing with the quantitative relationship between the skin dose and the risk of developing acute or late toxicity are very rare. The aim of this work was to assess the correlation between individually recovered planning skin dose-volume/surface data and acute toxicities after Radiochemotherapy for HN cancer patients. Material and Methods The study involved 32 HN consecutive patients (oropharynx: 14, ipopharynx: 8, nasopharynx: 5, others: 5) previously treated with Helical Tomotherapy (HT) with radical intent (SIB technique: 54/66Gy to PTV1/PTV2 in 30fr). All patients received neo-adjuvant and/or concomitant chemotherapy (no patients received Cetuximab) and daily image-guidance. After recovery of CT planning data, a superficial body layer with a thickness of 2 mm (SL2) was automatically delineated with a cranial-caudal extension corresponding to the high-dose PTV2 (i.e.: receiving 66Gy). CTCAE v4.0 acute skin toxicities were reported for all patients. Absolute dose-volume histograms (DVH) of SL2 were calculated and average values for patients who developed severe/moderate (G3/G2) and mild/none (G1/G0) acute toxicities were assessed: DVH differences between the two groups (G3/G2 vs G1/G0) were analyzed by two-tails t-test to define the most discriminative regions of SL2 DVH. Results Average DVHs are shown in Figure 1. Fifty-eight % of patients experienced G2/G3 toxicity (rate of G3: 27%) against 42% for G1/G0. Differences in skin DVHs were significant in the range 53-63 Gy (Figure 2), suggesting that the fraction of SL2 receiving around 2 Gy/fr is highly correlated with the risk of skin acute toxicity. In the plateau of the DVH region with highly significant differences, p-values ranged between 0.005 and 0.01: of note, the mean values of V60 were 4 cc and 11 cc for the two groups, corresponding to a skin surface of 20 cm 2 and 55 cm 2 respectively. V60 was found to be highly discriminative (AUC:0.84, 95%CI: 0.64-0.95) with a best cut-off value equal to 3.3 cc: the rate of G2/G3 toxicity was equal to 82% and 11% for V60 ≥ 3.3 cc and < 3.3 cc respectively (p=0.003).

Conclusion A linear model as a function of mean dose within parotid glands fits quite well with the observations of total stimulated salivary flow, and a single parotid gland function described as scintigraphic ejectionn fraction. No data point at low doses <10 Gy exists and the linear model needs to be confirmed with the more comprehensive data set (e.g. unilateral neck treatments) to cover the whole dose range. EP-1990 Skin DVHs predict cutaneous toxicity in Head and Neck cancer patients treated with Tomotherapy M. Mori 1 , I. Dell'Oca 2 , S. Foti 2 , G.M. Cattaneo 1 , R. Calandrino 1 , N. Di Muzio 2 , C. Fiorino 1 1 San Raffaele Scientific Institute, Medical Physics, Milan, Italy 2 San Raffaele Scientific Institute, Radiotherapy, Milan, Italy

Conclusion Despite the limited number of patients, current results quantified for the first time the relationship between SL2