CORRELATION BETWEEN AREAS OF HIGH FDG UPTAKE ON PRE-TREATMENT PET/CT AND PREFERENTIAL SITES OF LOCAL RELAPSE AFTER CHEMO-RADIOTHERAPY FOR HEAD AND NECK SQUAMOUS CELL CARCINOMA. A. Chaput*1, J. Calais*2,3, P. Robin1-4, S.Thureau3, D. Bourhis1, R. Modzelewski3, U. Schick5, P. Vera3, P-Y. Salaün1-4, R. Abgral1-4. These two authors contributed equally to this work 1- Department of Nuclear Medicine, Brest University Hospital, Brest, France 2- Department of Nuclear Medicine, Bichat University Hospital, Inserm 1148, DHU FIRE, Assistance Publique - Hôpitaux de Paris, Paris, France 3- Department of Nuclear Medicine and Radiology, Henri Becquerel Center, QuantIF (LITIS EA 4108 – FR CNRS 3638), Rouen University Hospital, Rouen, France. 4- European University of Brittany, EA3878 GETBO, IFR 148, Brest, France 5- Department of Radiotherapy, Brest University Hospital, Brest, France. . PURPOSE The potential benefits of 18F-Fluoro-2-deoxy-D-glucose positron emission tomography (FDG PET/CT) imaging for radiotherapy (RT) planning treatment of head and neck squamous cell carcinoma cancer (HNSCC) are increasingly being recognized. It has been suggested that intra-tumor sub-volumes with high FDG avidity (“hotspots”) are potential target for selected dose escalation [1-2]. Our aims were to demonstrate that pre-RT FDG PET/CT can identify intra-tumor sites at increased risk of local relapse (LR) after RT and to determine an optimal maximum standardized uptake value (SUVmax) threshold to delineate smaller RT target volumes that would facilitate RT dose escalation without impaired tolerance. METHODS The study included 72 consecutive patients with locally advanced HNC treated by radiotherapy ± concurrent chemotherapy. All patients underwent FDG PET/CT at initial staging (PETA) and during systematic follow-up (PETR) in a single institution. FDG PET/CT acquisitions were co-registered on the initial CT scan with a rigid method. Various sub-volumes (AX; X=30, 40, 50, 60, 70, 80 and 90% SUVmax thresholds) within the initial tumor and in the subsequent LR (RX; X=40 and 70% SUVmax thresholds) were pasted on the initial PET/CT and compared together [Dice, Jaccard, overlap fraction (OF), common volume/baseline volume, common volume/recurrent volume] (Fig.1) Fig. 2 : Patterns of relapses Fig. 1 Typical A70 R40 sub-volumes overlapping estimation after co-registration and reports. Numbers indicate the SUVmax thresholds in percentage. RESULTS Nineteen patients (26%) had LR (Fig.2). Using a 40% of SUVmax threshold, initial metabolic tumor volume (MTV) was significantly higher in all relapse patients (local and distant relapse) than in controlled patients (mean: 11.3±9.8 vs. 5.1±4.9 cc, p=0.001) as well as total lesion glycolysis (TLG) (mean: 134.6±116 vs. 60.6±80.4, p=0.002). For both using methods, the overlap index between A30, A40, and A50 sub-volumes on PETA and the whole metabolic volume of recurrence R40 and R70, on PETR showed a moderate agreement (between 0.40 and 0.60) (Fig.3). Example 1: 45-year-old woman with an oropharynx squamous cell carcinoma T4N2bM0. The A50 sub-volume was reported on PETR and the R40 and R70 sub-volumes on PETA. The A50∩R40 OF index was calculated at 0.53, meaning that half of R40 was included in A50. The A50∩R70 OF index was calculated at 0.77 meaning that three quarters of R70 was included in A50. Example 2: 49-year-old woman with a T4N2cM0 carcinoma. The overlap index were calculated all inferior to 0.4. But we can visually see the obvious mis-registration due to tumor decreasing and glottic position. High 18F-FDG uptake sub-volume of initial tumor (i.e. A70) is on the same side that the core of the residual tumor (i.e. R70) and close from it. We can suppose that with an accurate co-registration, overlap index would have been much higher. Fig. 3: Histogram of the mean values of overlap indices for various SUVmax thresholds to delineate the volumes on PETA (baseline, AX) and PETR at relapse, (R40 and R90). See the text for a description of the indices. The data are the same as those in Tables 4 and 5. CONCLUSIONS We confirm that TLG and MTV are independently correlated with recurrence-free survival in patients with HNSCC [3-4]. . Due to sub-optimal co-registration, we could not reach high overlap index values between initial tumor and recurrence sub-volumes. Further larger prospective studies with FDG-PET/CT performed in the same RT position and with a validated elastic registration method are needed. 1. Calais et al. J Nucl Med 2015;56:196-203. 2. Calais et al. Eur J Nucl Med Mol Imaging 2015;42:858-67. 3. Abgral et al. Eur J Nucl Med Mol Imaging 2014;41:659-67 4. Abgral et al. Head Neck 2016;38:600-8 P2707