ESTRO 2020 Abstract book

S356 ESTRO 2020

Dasu 1,5 , A.L. Grosu 3 1 Stockholm University, Medical Radiation Physics- Dept. Physics, Stockholm, Sweden ; 2 Skandion Clinic, Skandion Clinic, Uppsala, Sweden ; 3 Medical Center- Medical Faculty Freiburg- German Cancer Consortium DKTK Partner Site Freiburg, Dept. Radiation Oncology, Freiburg, Germany ; 4 University Medical Center, Dept. Nuclear Medicine, Freiburg, Germany ; 5 Karolinska Institutet, Dept. Oncology and Pathology, Stockholm, Sweden Purpose or Objective To investigate the trends of the evolution of the hypoxic compartment of head and neck (H&N) tumours imaged with FMISO PET before and during the course of radiotherapy and the potential correlations between the extent and the severity of tumour hypoxia and the outcome of the treatment. Several parameters describing the change in tumour hypoxia were introduced and analysed with respect to response predictions. Material and Methods Thirty patients with locally advanced H&N Squamous Cell Carcinoma treated with concomitant chemoradiotherapy (70 Gy in 35 fx and three cycles of 100 mg/m 2 cisplatin) were imaged three times with 18 FMISO PET/CT: before the start of the treatment and in weeks 2 and 5 during the treatment. The scans were co-registered to the planning CT in a research version of the treatment planning system RayStation. The uptake in the images was converted into partial oxygen pressure, pO 2 , by applying a previously developed conversion function at voxel level. The hypoxic compartment was determined by applying a thresholding method on the 10 mmHg level. The Receiver Operating Characteristic (ROC) method was used to seek correlations between the loco-regional recurrence (LRR) and several parameters describing the evolution of the hypoxic compartment: minimum oxygen level in paired pO 2 maps (min pO 2 (1;2), min pO 2 (1;3), min pO 2 (2;3)) and volume of the hypoxic compartment in paired pO 2 maps (HTV1;2, HTV1;3, HTV2;3) at the three different time points. Results The median of the differences between the paired observations for the minimum pO 2 values and the hypoxic tumour volumes at the different time points resulted as statistically significantly different from 0 in all the considered cases (min pO 2 (1;2) with p=0.005, min pO 2 (1;3) with p=0.0001, min pO 2 (2;3) with p=0.027; HTV1;2 with p=0.001, HTV1;3 with p=0.0001, HTV2;3 with p=0.016). The ROC analysis correlating the percent difference between HTV1 and HTV2 with LRR are shown in Figure 1(A) (AUC=0.75, p-value= 0.036). Corresponding absolute values of the difference between the hypoxic volumes in the first and second week also resulted in a statistically significant AUC value of 0.72 (p-value=0.019). Other combinations of percent or absolute differences between HTVs for the different time points did not result in statistically significant AUC values. The evolution of the hypoxic compartment with time was also studied for each individual patient. The slope of the linear regression curve fitting the datapoints was calculated. Correlation between the slope and LRR were sought and results of the ROC analysis are presented in Figure 1(B) (AUC= 0.74, p- value=0.008).

Conclusion The mean escalated dose changed less than 3% in the majority of the patients showing high dosimetrical stability. In 22% of the patients, OF 02 <70% resulted in a decrease of escalated dose of more than 3%. In these patients, large decrease in SUVpeak close to background level increased delineation uncertainty or necrosis/atelectasis resolved. OC-0583 Time evolution of the hypoxic compartment on sequential FMISO PET images M. Lazzeroni 1 , A. Ureba 2 , F. Schiavo 1 , N. Wiedenmann 3 , N.H. Nicolay 3 , M. Mix 4 , B. Thomann 3 , D. Baltas 3 , I. Toma-