ESTRO 37 Abstract book

S93

ESTRO 37

Conclusion The treatment-induced change in ADC during the first 3 weeks of nCRT for esophageal cancer seems most promising for prediction of histopathologic response. Our series is currently further extended to verify these results. OC-0179 Weekly, early assessment trends of FDG PET metrics and their relation to overall survival M. La Fontaine 1 , N. Bruin 1 , S.R. Van Kranen 1 , W. Vogel 2 , J. Knegjens 1 , J. Belderbos 1 , J. Van de Kamer 1 , J.J. Sonke 1 1 Netherlands Cancer Institute, Radiotherapy, Amsterdam, The Netherlands 2 Netherlands Cancer Institute, Nuclear Medicine, Amsterdam, The Netherlands Purpose or Objective Early response assessment of FDG PET may impact treatment options for patients with NSCLC. However, uncertainty exists in early response assessment in relation to the choice of imaging time point during treatment. In addition, the response of PET metrics from the beginning through the end of treatment is still unclear. This study aimed to investigate the response of PET metrics throughout treatment using weekly FDG PET scans. PET metrics were investigated for time points and trends associated with overall survival. Material and Methods The study included 22 patients with stage II-III NSCLC, who underwent concurrent chemoradiotherapy in 24 fractions of 2.75 Gy. During treatment, patients received five weekly FDG PET/CT scans from day 0 of treatment through day 28. Various PET metrics were measured at each time point, including SUVmax, total lesion glycolysis (TLG), SUVpeak, tumor volume, and SUVmean. Metrics were tested against one another for independence (r < 0.7). Linear regression was performed to determine the slope of each metric during treatment. For Kaplan-Meier survival analysis, the metrics were separated based on either the median slope or median metric value. Imaging time points were tested for PET metric association with overall survival with the Cox hazard ratio (HR) reported for significant differences. Results SUVmax and TLG were independent metrics for analysis. SUVmean (r=0.85) and SUVpeak (r=0.98) were highly correlated to SUVmax, while tumor volume was highly correlated to TLG (r = 0.83). Linear regression over the time points resulted in high correlation coefficients for both SUVmax (median: -0.90 [IQR:-0.95- -0.72]), and TLG (median: -0.95 [-0.97- -0.93]) (Figure 1). No associations to overall survival were found with SUVmax for any time point, slope, nor relative response. Overall survival was associated with a decrease as TLG increased. The absolute TLG slope was significantly associated with overall survival (p=0.001, HR =0.01), although absolute TLG values were not significant throughout treatment. In contrast, the relative TLG slope was marginally associated (p=0.09, HR=0.28) with overall survival, with relative values (Figure 2) being significant for the final two weeks of treatment for day 21 (p=0.01, HR=0.10), and day 28 (p=0.02, HR=0.12).

Conclusion The study found FDG PET metrics of tumors to have high correlation coefficients over four weeks of imaging with the median tumor slope of TLG being associated with overall survival. In addition, the study demonstrated that differences in overall survival may be found after two weeks of treatment. These results may be useful for future studies when selecting imaging time points and PET metrics for association with outcome in NSCLC patients. OC-0180 Parotid gland fat related Magnetic Resonance Image biomarkers improve prediction of late xerostomia L.V. Van Dijk 1 , M. Thor 2 , R.J.H.M. Steenbakkers 1 , A. Apte 2 , T. Zhai 1 , R. Borra 3 , N. Lee 4 , J.A. Langendijk 1 , J.O. Deasy 2 , N.M. Sijtsema 1 1 University Medical Center Groningen, Radiation oncology, Groningen, The Netherlands 2 Memorial Sloan Kettering Cancer Center, Department of Medical Physics, New York, USA 3 University Medical Center Groningen, Department of Radiology, Groningen, The Netherlands 4 Memorial Sloan Kettering Cancer Center, Department of Radiation Oncology, New York, USA Purpose or Objective Our previous studies showed that image biomarkers (IBMs) of the parotid gland from CT and PET images were associated with a higher risk of xerostomia 12 months after radiotherapy (Xer 12m ), and significantly improved the prediction of Xer 12m . The predictive IBMs indicated heterogeneity and metabolic functionality, subsequently