Integrated 18F-FDG PET/perfusion CT for the monitoring of neoadjuvant chemoradiotherapy in rectal carcinoma: correlation with histopathology

Fischer, Michael ; Vrugt, Bart ; Alkadhi, Hatem ; Hahnloser, Dieter ; Hany, Thomas ; Veit-Haibach, Patrick

In: European Journal of Nuclear Medicine and Molecular Imaging, 2014, vol. 41, no. 8, p. 1563-1573

Add to personal list
    Purpose: The aim of this study was to prospectively monitor changes in the flow-metabolic phenotype (ΔFMP) of rectal carcinoma (RC) after neoadjuvant chemoradiotherapy (CRT) and to evaluate whether ΔFMP of RC correlate with histopathological prognostic factors including response to CRT. Methods: Sixteen patients with RC (12 men, mean age 60.7 ± 12.8years) underwent integrated 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET)/perfusion CT (PET/PCT), followed by neoadjuvant CRT and surgery. In 13 patients, PET/PCT was repeated after CRT. Perfusion [blood flow (BF), blood volume (BV), mean transit time (MTT)] and metabolic [maximum and mean standardized uptake values (SUVmax, SUVmean)] parameters as well as the FMP (BF × SUVmax) were determined before and after CRT by two independent readers and correlated to histopathological prognostic factors of RC (microvessel density, necrosis index, regression index, vascular invasion) derived from resected specimens. The diagnostic performance of ΔFMP for prediction of treatment response was determined. Results: FMP significantly decreased after CRT (p < 0.001), exploiting higher changes after CRT as compared to changes of perfusion and metabolic parameters alone. Before CRT, no significant correlations were found between integrated PET/PCT and any of the histopathological parameters (all p > 0.05). After CRT, BV and SUVmax correlated positively with the necrosis index (r = 0.67/0.70), SUVmax with the invasion of blood vessels (r = 0.62) and ΔFMP with the regression index (r = 0.88; all p < 0.05). ΔFMP showed high accuracy for prediction of histopathological response to CRT (AUC 0.955, 95% confidence interval 0.833-1.000, p < 0.01) using a cut-off value of −75%. Conclusion: In RC, ΔFMP derived from integrated 18F-FDG PET/PCT is useful for monitoring the effects of neoadjuvant CRT and allows prediction of histopathological response to CRT.