Quantification of changes to dose distribution caused by positional corrections from verification image matching to rectal and anal cancer patients during IMRT treatment

Background and Objectives: Intensity modulated radiotherapy (IMRT) for the treatment of rectal and anal cancers is considered in this paper. In order to replicate the patient’s position at the planning CT stage, cone beam CT (CBCT) scans are taken at certain fractions and the couch moved according to verification image matching of the patient’s bony anatomy. Compared to the original treatment plan delivery, there are discrepancies in dose due to this movement. By using software to replicate the movements made on set, the dose distributions between the original and treatment plans can be compared. This study quantifies these differences and determines if they introduce a clinically or statistically significant dose difference. The results will be used to inform discussions about whether different technologies (such as adaptive radiotherapy) might offer superior treatment for these patients.

Materials and Methods: To determine quantitatively if plans are clinically acceptable, they are analysed using dose volume histograms (DVHs) and comparing the optimal dose constraints given in departmental protocols to the actual values obtained for each plan. The biological effective dose (BED) and equivalent dose from two Gray fractions (EQD(2)) from the clinical plan and new plan with movements are calculated and compared in some cases. Further analysis involves the equivalent uniform dose (EUD), normal tissue complication probability (NTCP) and tumour control probability (TCP). Wilcoxon matched pairs signed ranks tests is used (significance threshold p = 0.05) for statistical analysis and to retain or reject the null hypothesis that “there is no difference between the median dose for a specified condition between the original plan and that delivered after image verification matching”.

Results: Three of the ten patients analysed show clinically significant differences where dose constraints are breached. Further statistically significant differences were seen on some volumes but in many cases this did not translate into clinical significance. Twenty of the 84 fractions analysed (23.8%) showed an organ at risk (OAR) that breached a tolerance in the manual alignment case that wasn’t present in the isocentric. DVH and NTCP analysis do not always corroborate in the results which is explained by the use of the whole curve with the NTCP model but only certain dose levels when considering the DVH method.

Conclusion: Since it cannot currently be predicted which patients or organs will be adversely affected by couch movements during patient set up, it is concluded that the possibility of poorer dosimetry in the form of a higher NTCP or lower TCP must be considered in all cases. This is unlikely to cause large differences but may be clinically signi icant. The null hypothesis is therefore rejected and new methods recommended for consideration for future clinical practice.