Impact of rigid and nonrigid registration on the determination of 18F-FDG PET-based tumour volume and standardized uptake value in patients with lung cancer

Assessment of the metabolically active tumour tissue by FDG PET is evolving for use in the diagnosis of non-small-cell lung cancer (NSCLC), in the planning of radiotherapy, and in follow-up and response evaluation. For exact evaluation accurate registration of PET and CT data is required. The registration process is usually based on rigid algorithms; however, nonrigid algorithms are increasingly being used. The influence of the registration method on FDG PET-based standardized uptake value (SUVmax) and metabolic tumour volume (MTV) definition has not yet been evaluated. We compared intra- and interindividual differences in SUV and MTV between rigid- and nonrigid-registered PET and CT acquired during different breathing manoeuvres. The study group comprised 28 radiotherapy candidates with histologically proven NSCLC who underwent FDG PET acquisition and three CT acquisitions (expiration - EXP, inspiration - INS, mid-breath-hold - MID). All scans were registered with both a rigid (R) and a nonrigid (NR) procedure resulting in six fused datasets: R-INS, R-EXP, R-MID, NR-INS, NR-EXP and NR-MID. For the delineation of MTVs a contrast-oriented contouring algorithm developed in-house was used. To accelerate the delineation a semiautomatic software prototype was utilized. Tumour mean SUVmax did not differ for R and NR registration (R 17.5 +/- 7, NR 17.4 +/- 7; p=0.2). The mean MTV was higher by 3 +/- 12 ml (p=0.02) in the NR group than in the R group, as was the mean tumour diameter (by 0.1 +/- 0.2 cm; p < 0.01). With respect to the three different breathing manoeuvres, there were no differences in MTV in the R group (p > 0.7). In intraindividual comparison there were no significant differences in MTVs concerning the registration pairs R-EXP (68 +/- 88 ml) vs. NR-EXP (69 +/- 85 ml) und R-MID (68 +/- 86 ml) vs. NR-MID (69 +/- 83 ml) (both p > 0.4). However, the MTVs were larger after NR registration during inspiration (R-INS 68 +/- 82 vs. NR-INS 78 +/- 93 ml; p=0.02). The use of nonrigid algorithms may lead to a change in MTV, whose extent is influenced by the breathing manoeuvre on CT. Nonrigid registration methods cannot be recommended for the definition of MTV if the CT scan is performed during inspiration. The choice of registration algorithm has no significant impact on SUVmax.