ESTRO 2020 Abstract book

S994 ESTRO 2020

radiotherapy of head and neck cancers could be achieved by adding FAPI PET-CT information. Further and larger studies are warranted. PO-1708 Retrospective automatic detection of calcified plaque in heart on planning CT:viability and benefits M. Lizondo 1 , N. Jornet 2 , J. Fuentes-Raspall 3 , D. Viladés- Mendel 4 , R. Leta 4 , A. Latorre-Musoll 2 , P. Delgado-Tapia 2 , P. Carrasco 2 , J. Pérez-Alija 2 , P. Gallego 2 , P. Simón 2 , A. Ruiz-Martínez 2 , M. Adria 2 , I. Valverde-Pascual 2 , M. Barceló 2 , N. Garcia 2 , M. Ribas 2 1 Institut de Recerca Hospital de la Santa Creu i Sant Pau, Servei de Radiofísica i Radioprotecció, Barcelona, Spain ; 2 Hospital de la Santa Creu i Sant Pau, Servei de Radiofísica i Radioprotecció, Barcelona, Spain ; 3 Hospital de la Santa Creu i Sant Pau, Servei d'Oncologia Radioteràpica, Barcelona, Spain ; 4 Hospital de la Santa Creu i Sant Pau, Servei de Cardiologia, Barcelona, Spain Purpose or Objective We hypothesise that patients with calcified plaques in coronary arteries before radiotherapy treatment are more likely to suffer from cardiotoxicity after the treatment. The aim of this study is to explore the viability of an automatic calcium plaques detection and classification method using planning CTs. The hypothesis was validated for the group with higher Agaston Score. Material and Methods We analysed 760 breast cancer patients treated with radiotherapy between 2013 and 2018. CT planning images were acquired using a large‐bore multislice CT scanner (Brilliance™ CT Big Bore v3.6; Philips Healthcare) in supine position with voxel dimensions 2×2×5mm 3 . The technique used was: 3DCRT being still the standard technique, and IMRT (2013) used in patients not fulfilling dose objectives and constraints. DBHI (2015) was used in left breast patients. Patients were treated on Clinac2100CD equipped with RPM respiratory motion system (Varian). Mean [min, max] heart dose and V25 planned were 3.4 Gy [0.2 Gy,20.3 Gy] and 3.6 % [0.0%, 28.3%] respectively. Maximum density value inside the heart structure was extracted from Eclipse by scripting for each patient. Following the density factor used in Agatston Score (Coronary Artery Calcium Scoring), patients were classified in 4 risk groups depending on the maximum density value inside the heart structure: • 30-199 HU: 1 • 200-299 HU: 2 • 300-399 HU: 3 • 400+ HU: 4 Patients with a density factor >=3 were manually verified by a medical physicist trained by a cardiologist to test the automated tool. Our prospective breast data base (RedCap), where cardiac events are recorded, and medical histories of patients with a density factor equal to 4 were reviewed. Coincidences between cardiac events and planning CT Agaston score 4 were recorded. Results Fig.1a shows the patient’s distribution by risk group. 62% of the patient did not present any calcified plaque. The verification of the automatic tool resulted on a correct classification in Density Factor Group >=3 of 63% (62%) of the patients were well-classified in group 4 (3). Problems were due to a bad contouring of the heart adding bone areas (about 10%) and to an inclusion in the heart structure of the chemotherapy device (about 25%) (Fig.2). The corrected percentages are shown in Fig.1b. 5 out of the 30 patients in group 4 developed adverse cardiac events after radiotherapy. In all cases, symptoms preceded the diagnosis which included endocarditis, atrial fibrillation, syncope (due to a stenosis of the right internal

Heidelberg, Heidelberg Ion-Beam Therapy Center HIT, Heidelberg, Germany ; 4 University Hospital Heidelberg, Department of Nuclear Medicine, Heidelberg, Germany Purpose or Objective Tumor stroma including carcinoma-associated fibroblasts (CAFs) makes up to 90% of epithelial carcinomas. CAFs can be identified by overexpression of the Fibroblast activation protein (FAP). Hence, visualization of CAFs via FAP seems to be a promising opportunity for tumor detection and contouring for radiotherapy. Here, we are introducing PET-CT with 68 Ga-radiolabeled inhibitors of Fibroblast Activation Protein (FAPI) for the imagining of diffusely growing head and neck cancers. Material and Methods FAPI PET-CT was performed without complications prior to radiotherapy in addition to contrast enhanced CT (CE-CT) and MRI on fourteen patients with head and neck cancers. First, for tissue biodistribution analysis, volumes of interest were defined to quantify SUV mean and SUV max in tumor and healthy parenchyma. Second, using four thresholds of 3, 5, 7 and 10-fold increase of FAPI enhancement in the tumor as compared to normal tissue, four different gross tumor volumes (FAPI-GTV) were created automatically. These were compared to GTVs created with conventional CE-CT and MRI (CT-GTV). Results

The biodistribution analysis revealed that tumorous lesions showed high FAPI avidity (e.g. primary tumors, SUVmax 14.62 ± 4.44; SUVmean 7.41 ± 2.39). In contrast, low background uptake was measured in healthy tissues of the head and neck region, (e.g. salivary glands: SUVmax 1.76 ± 0.31; SUVmean 1.23 ± 0.28). Considering radiation planning, CT-GTV was of 27.3ml (range 9.1 – 266.5ml), whereas contouring with FAPI resulted in significantly different GTVs of 67.7ml (FAPI x3, p = 0.0134), 22.1ml (FAPI x5, p = 0.0419), 7.6ml (FAPI x7, p = 0.0001) and 2.3ml (FAPI x10, p = 0.0001). Taking these significant disparities between the two GTVs into consideration, we merged FAPI-GTVs with CT-GTVs. This resulted in median volumes, that were, as compared to CT-GTVs, significantly larger with FAPI x3 (54.7ml, +200.5% relative increase, p = 0.0005) and FAPI x5 (15.0ml, +54.9%, p = 0.0122). Furthermore FAPI-GTVs were often not covered by CE-CT based planning target volumes (CT- PTVs). Conclusion We present first evidence of diagnostic and therapeutic potential of FAPI-ligands in head and neck cancer radiation oncology. This novel, complication-free imaging modality provides crucial information regarding tumor spread and generates high contrast images with eminent tumor-to- background ratios. In contrast to FDG PET-CT, no tracer uptake is seen in surrounding healthy parenchyma or inflammatory tissue. Consequently, automated, precise and comprehensive target volume delineation for