Blood glucose level normalization and accurate timing improves the accuracy of PET-based treatment response predictions in rectal cancer

• Published in: Radiotherapy and oncology Vol. 95; no. 2; pp. 203 - 208
• Main Authors: Janssen, Marco H.M, Öllers, Michel C, Stiphout, Ruud G.P.M. van, Riedl, Robert G, Bogaard, Jørgen van den, Buijsen, Jeroen, Lambin, Philippe, Lammering, Guido
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier Ireland Ltd 01.05.2010
• Subjects: Adult; Aged; Blood Glucose; Blood glucose level normalization; Female; Hematology, Oncology and Palliative Medicine; Humans; Male; Middle Aged; Neoplasm Invasiveness; Positron-Emission Tomography; Rectal cancer; Rectal Neoplasms - blood; Rectal Neoplasms - diagnosis; Response prediction; Sequential FDG–PET–CT imaging; Time Factors; Blood glucose level normalization; Sequential FDG–PET–CT imaging; Rectal cancer; Response prediction
• ISSN: 0167-8140; 1879-0887
• DOI: 10.1016/j.radonc.2010.01.021

Abstract Purpose To quantify the influence of fluctuating blood glucose level (BGLs) and the timing of PET acquisition on PET-based predictions of the pathological treatment response in rectal cancer. Material and methods Thirty patients, diagnosed with locally advanced-rectal-cancer (LARC), were included in this prospective study. Sequential FDG–PET–CT investigations were performed at four time points during and after pre-operative radiochemotherapy (RCT). All PET-data were normalized for the BGL measured shortly before FDG injection. The metabolic treatment response of the tumor was correlated with the pathological treatment response. Results During RCT, strong intra-patient BGL-fluctuations were observed, ranging from −38.7 to 95.6%. BGL-normalization of the SUVs revealed differences ranging from −54.7 to 34.7% ( p < 0.001). Also, a SUVmax time-dependency of 1.30 ± 0.66 every 10 min (range: 0.39–2.58) was found during the first 60 min of acquisition. When correlating the percent reduction of SUVmax after 2 weeks of RCT with the pathological treatment response, a significant increase ( p = 0.027) in the area under the curve of ROC-curve analysis was found when normalizing the PET-data for the measured BGLs, indicating an increase of the predictive strength. Conclusions This study strongly underlines the necessity of BGL-normalization of PET-data and a precise time-management between FDG injection and the start of PET acquisition when using sequential FDG–PET–CT imaging for the prediction of pathological treatment response.