Search

Purpose: Morphologic imaging techniques perform poorly in assessing the response to preoperative radiotherapy (RT), mainly because of desmoplastic reactions. The aim of this study was to investigate the potential of sequential 18-fluoro-2-deoxy-d-glucose (18FDG-PET) in assessing the response of rectal cancer to neoadjuvant RT and to determine which parameters can be used as surrogate markers for histopathologic response. Methods and Materials: 18FDG-PET scans were acquired before and during the 5th week after the end of RT. Tracer uptake was assessed semiquantitatively using standardized uptake values (SUV). The percentage differences (%Δ) between pre- and post-RT scans in SUVmax, SUVmean, metabolic volume (MV), and total glycolytic volume (tGV) were calculated. Results: Forty-five consecutive patients with histologically confirmed rectal adenocarcinoma were enrolled. After neoadjuvant RT, 20 of the 45 patients were classified as histopathologic responders and 25 as non-responders. Intense 18F-FDG uptake was seen in all tumors before neoadjuvant RT (average SUVmax 12.9 ± 6.0). When patients were classified as histologic responders and nonresponders, significant differences in %ΔSUVmax (55.8% vs. 37.4%, p = 0.023) and %ΔSUVmean (40.1% vs. 21.0%, p = 0.001) were observed between the two groups. For %ΔMV and %ΔtGV, decreases were more prominent in responders but were not significantly different from those in nonresponders. As demonstrated by receiver operating characteristic analysis, %ΔSUVmean was a more powerful discriminator than was %ΔSUVmax. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value for optimal threshold of %ΔSUVmean (24.5%) were 80%, 72%, 76%, 70%, and 82% respectively. Conclusion: Sequential 18FDG-PET allows assessment of the response to preoperative RT. Both %ΔSUVmean and %ΔSUVmax correlate with histopathologic response and can be used to evaluate and compare the effectiveness of different neoadjuvant treatment strategies. The maximum accuracy figures and the positive predictive value figures for both Δ%SUVmean and Δ%SUVmax are, however, too low to justify modification of the standard treatment protocol of an individual patient. [Copyright &y& Elsevier]