Portal de Periódicos da CAPES

OC-0275: Critical evaluation and comparison of CT vs MR based HRCTV and OAR contouring during IGBT for cervical cancer

2015; Elsevier BV; Volume: 115; Linguagem: Inglês

10.1016/s0167-8140(15)40273-7

1879-0887

ChiraRanjan Khadanga, Umesh Mahantshetty, S. Engineer, Shailendra Chopra,

MRI in cancer diagnosis

Materials and Methods: 20 cervical cancer patients who underwent treatment at Royal Marsden Hospital were identified retrospectively.Both EBRT and BT plans were retrieved from radiotherapy archive.The plans were anonymised and transferred to VODCA-RT (MSS GmbH,Hagendorn,CH) which is a dosimetric analysis software.The rectum was recontoured using standardised anatomical definition of 2.5cm from anal canal to the rectosigmoid junction.Posteriorly cut DSM were generated using VODCA-RT for both EBRT and BT.The raw data was then organised using in-house R script before being normalised to 21x21 pixels using MATLAB (Mathworks, Natick, MA).Assuming an α/β ratio of the linear quadratic model of 3Gy for the rectum, equivalent BED dose in 2Gy conversion was made using the equation shown below.Dose values of corresponding pixels were then summated to derive the respective DSM.Maximum rectal toxicity such as bleeding and proctitis within 6 months of treatment were also recorded using CTCAE v4.0.Statistical analysis was performed using MedCalc software v14.10.2.Results: Combined rectal DSM were successfully generated for all 20 patients using our methodology.An example of a combined DSM is illustrated in figure 1.As expected, there was significant variation between the mean rectal volume during EBRT and BT (67.71cc and 53.99cc respectively, p=0.0189) due to presence of brachytherapy applicator and rectal retractor.There was also a mean difference of 0.59cm between EBRT and BT rectal cranio-caudal length (p=0.049).This is most likely to be due to true anatomical variation between both time points e.g.angulation of upper rectum.Normalisation to 21x21 pixels was therefore felt to be a good compromise to account for these anatomical differences whilst maintaining relatively accurate spatial dose distribution following dose summation.Consistent with previously published study by Buettner et al, ROC analysis on our data also demonstrated that lateral extent (rectal circumference) receiving >60Gy is a good predictor of risk of G1-3 proctitis (AUC=0.879,p<0.0001).Conclusions: Rectal dose summation using DSM for cervical EBRT and BT is feasible.This method retains valuable spatial information which is closely correlated to toxicity.We propose a follow-up study with a bigger cohort incorporating organ motion for further validation of DSM as alternative eassessment of rectal constraints in cervical cancer patients.