Automation of Dosimetric Data Collection Using C# ESAPI for Intracavitary and Hybrid Intracavitary/Interstitial Brachytherapy Plans

Abstract

Dosimetric data extraction for complex brachytherapy plans can be a time-consuming and error-prone process. The Eclipse Scripting API (ESAPI) Microsoft .NET framework and its ability to access the ARIA treatment planning database makes it a powerful tool for researchers to incorporate automation into data collection. This automation allows for more efficient and error-free collection of dosimetric data, and the program can be iterated over multiple patients at one time. The main objectives of this study were (i) to develop a C# ESAPI script capable of automated data collection for multiple brachytherapy plans and (ii) to apply the script and examine dosimetric characteristics of a patient cohort with locally advanced cervical cancer treated using tandem and ovoid (T/O) applicators or hybrid tandem and ovoid applicators with needles (T/O + N). We developed a ESAPI script capable of receiving a list of patient IDs and creating a CSV dataset of dosimetric data of the requested patients. Eclipse commands were accessed to create a CSV dataset consisting of baseline demographic data, average V100, V125, D100, D90 as well as D0.1cc, D1cc, D2cc of organs at risk (bladder, rectum, sigmoid colon, small bowel) per fraction. We applied the ESAPI script to our single-institution retrospective cohort which included patients who underwent brachytherapy as part of definitive radiation treatment for cervical cancer between May 2017 and November 2022. Patients either received T/O or T/O + N brachytherapy boost. Statistical analysis was performed using two-sided t-tests to examine differences in descriptive dose metrics between the two techniques. The developed ESAPI script automatically generated a CSV file of baseline demographic data as well as dosimetric data for patients who underwent T/O or T/O + N brachytherapy boost. We demonstrated that the ESAPI script could acquire and analyze the data quickly (<5 s per test case) compared to time required to generate data via manual review (5 min per test case). This is a 98% reduction in time required to review and collect data per patient. The analyzed cohort included 72 patients with cervical cancer: 40 with T/O technique and 32 with T/O + N technique. Larger clinical target volumes (CTV) were seen in the group that received treatment with T/O + N technique versus T/O alone (37.4 cc vs 25.2 cc, p < 0.0007). Dose metrics including V100, V125, D100, D90 of the high-risk CTV as well as D0.1cc, D1cc, D2cc of the bladder, rectum, small bowel were not significantly different between techniques. However, the D0.1cc, D1cc, D2cc of the sigmoid colon were higher for patients who received T/O + N technique by 95, 74, and 64 cGy respectively (p < 0.005, p < 0.003 and p < 0.007). This study highlights the use of a newly developed ESAPI script specific to brachytherapy plans which allows for iteration over multiple patients at one time. We were able to quickly compare multiple dose metrics for two patient groups treated at our institution.