ESTRO 2024 - Abstract Book

S230

Brachytherapy - Gynaecology

ESTRO 2024

Material/Methods:

The HDR plans of a consecutive cohort of cervical cancer patients were reviewed, and based on the dosimetric targets combining EBRT and HDR doses (d90 ≥ 85Gy, d2cc bladder < 80Gy, d2cc rectum/sigmoid/small bowel < 65Gy) splint into 3 groups. Group A d90 ≥ 85Gy and all OARs within constraints; group B d90 ≥ 85Gy and ≥ 1 OARs out with constraints, group C d90 < 85Gy regardless of OARs. The vector has n components () which are d2cc bladder/rectum/sigmoid/small bowel and 135 – d90 (higher d90 doses correspond to a lower value) – all equally weighted, and the lower a component value the more optimal the corresponding dose registered. We then calculated the magnitude of the vector: z=√ (x 1 2 +.... + x n 2 ). We hypothesised that the better plans should have a smaller z. For each patient the n-dimensional vector value z was calculated using these 5 variables. Statistical testing for differences amongst the 3 groups was carried out using the Kruskal-Wallis test with a significance level taken as p < 0.05.

Results:

The records for a total of 112 consecutive patients were reviewed and put into one of the 3 groups and the z value calculated for each. The n-dimensional vector z value was found to be significantly different among the 3 groups: Kruskal-Wallis H statistic = 59.873, p<0.0001 (df =2).

Group A

Group B

Group C

n

27

53

32

Mean z

128.22

134.85

141.78

126.9 (117.45 - 139.57)

134.72 (125.54 - 146.37)

142.05 (133.17 - 155.53)

Median z (range)

Standard deviation

5.24

4.19

5.39

Conclusion:

The z value can help to quantify the quality of brachytherapy plans and could help inform the choice between plans with variations in doses to the different OARs. Further work using larger cohort can help establish threshold measurements for the z value that are linked with clinical outcomes.

Keywords: n-dimensional vector, cervix

529

Digital Poster

Use of a radar plot to assess cervical HDR brachytherapy plans and correlate with clinical outcomes

Mark Anthony Zahra, Will Keough

Edinburgh Cancer Centre, Oncology, Edinburgh, United Kingdom