Optimizing image guidance frequency and implications on margins for gynecologic malignancies

BACKGROUND To analyze setup deviations using daily megavoltage computed tomography (MVCT) and to evaluate three MVCT frequency reducing protocols for gynecologic cancer patients treated with helical tomotherapy. METHODS We recorded the setup errors of 56 patients with gynecological cancer observed throughout their whole course by matching their daily MVCT with the planning CT. Systematic and random errors were calculated on a patient and population basis. We defined three different protocols corresponding to MVCTs from the first five fractions (FFF), the first ten fractions (FTF) or from the first and third weeks (505). We compared theoretical. setup errors calculated using these 5 or 10 early MVCT scans with the actual errors found with the remaining fractions to to analyze the residual deviations. RESULTS The total systematic (random) deviations had means of -2.0 (3.8)mm, 0.5 (3.4)mm, 0.5 (6.1)mm and -0.5° (0.9°) in vertical (V), longitudinal (LO), lateral (LA), and roll (R) directions, respectively. The proposed three MVCT protocols resulted in minor residual deviations. In all three protocols, 95% of all calculated residual deviations were less than or equal to 5 mm in all 3 directions. When examining the additional minimal CTV-PTV setup margins that were calculated based on these residual deviations, the 505 protocol would have allowed smaller margins than the FFF and FTF protocol, particularly in the V direction. CONCLUSIONS For patients with gynecologic cancer, the 505-protocol led to the lowest residual deviations and therefore might offer the best approach in reducing the frequency of pre-treatment MVCTs.