ESTRO 2022 - Abstract Book

S556

Abstract book

ESTRO 2022

number of clusters. The package used for these analyses were SuperML, factoextra and ClusterR. Cox model and Logrank test were used for PFS analysis. All analyses were conducted with R (4.0.2) and Stata software (16). Results Five of clusters were identified as sharing similarities of metabolic information and are predictive of PFS. Clusters 2 and 4 revealed metabolic abnormalities (increase in Cho, stability or decrease in Cr, decrease in NAA and increase in Lac). The PFS was lower when cluster 4 or 2 are the dominant clusters in patient’s MRSI data. Table 1 shows that PFS were significantly different between clusters. Patients having foremost clusters 2 or 4 in their MRSI data have lower PFS.

Conclusion The preliminary results showed that MRSI can be used to reveal heterogeneity of the GBM in pre-radiotherapy examination. Groups of spectra sharing similarities in metabolite information would reflect the different components of tissue in the MRSI acquired volume. Clusters with metabolic abnormalities representative of tumor proliferation and hypoxia, have been identified as predictor of progression free survival. Comparison of our clustering results with neuroradiologist segmentation of different tumor compartments on multimodal MRI data is ongoing.

Proffered Papers: Application of functional & quantitative imaging

OC-0627 TCP predictions of an automated dose painting strategy based on FDG and FMISO PET imaging

M. Lazzeroni 1,2 , A. Ureba 3,2 , N.H. Nicolay 4 , A. Ruehle 4 , B. Thomann 4 , D. Baltas 4 , M. Mix 5 , I. Toma-Dasu 6,2 , A.L. Grosu 4

1 Stockhoolm University, Department of Physics, Stockholm, Sweden; 2 Karolinska Institutet, Department of Oncology and Pathology, Stockholm, Sweden; 3 Stockholm University , Department of Physics, Stockholm , Sweden; 4 Medical Center, Medical Faculty Freiburg, German Cancer Consortium (DKTK) Partner Site Freiburg, Department of Radiation Oncology, Freiburg, Germany; 5 University Medical Center, Department of Nuclear Medicine, Freiburg, Germany; 6 Stockholm University, Department of Physics, Stockholm, Sweden Purpose or Objective Aim of this study was to assess the Tumour Control Probability (TCP) of a personalized Dose Painting By Contour (DPBC) strategy based on the synergistic exploitation of the tumour clonogenic cell information derived from FDG PET images and the tumour oxygen distribution derived from FMISO PET images. Materials and Methods Thirty H&N cancer patients imaged with FDG and FMISO PET/CT before radiochemotherapy were analysed. FMISO PET images were converted into maps of oxygen partial pressure (pO 2 ) using a non-linear conversion function of radiotracer uptake previously derived. A dose distribution at voxel level considering the heterogeneity in radiosensitivity related to the oxygenation and an heterogenous distribution of clonogenic cells in the CTV was determined. The tumour clonogens information was retrieved from FDG PET images by applying a linear conversion of the radiotracer uptake, having as an anchor point the tumour cell density carrying capacity set to correspond to the maximum uptake level retrieved from the patient dataset. The CTV was segmented into hypoxic target volume (HTV), GTV-HTV and CTV-GTV. A dose escalation strategy consisting of uniform doses applied to the segmented targets (dose painting by contour, DPBC) aiming at 95% of TCP in the CTV was employed. Automated photon treatment plans were made in RayStation (v10, RaySearchLaboratories) with ±3mm setup errors (7 scenarios) for minimax robust optimization. The unsupervised planning strategy used the same objective functions for the whole patient population and a total dose delivered in 35 fractions with an integrated boost. A dosimetric evaluation of the treatment plans, accounting for target coverage and constraints for the organs at risk, was followed by an assessment of the TCP in the targets accounting for the dose distribution in the nominal plan and the radiosensitivity maps derived from combined FMISO and FDG PET information (Figure 1).