Author: Rex A. Cardan, Richard A. Popple 👨🔬
Affiliation: University of Alabama at Birmingham 🌍
Purpose:
Compliance with TG 263 naming conventions for target structures in radiation oncology remains a challenging task due to the complexity and variability of the protocol. Traditional validation methods using regular expressions often fail to capture all possible variations or exhibit performance limitations. This work presents the TG 263 Toolkit, developed in C#, to codify and validate target naming conventions, enabling robust compliance, accurate name generation, and parsing.
Methods:
The TG 263 Toolkit encapsulates all naming components specified by the protocol into a comprehensive class structure, including fields such as TargetBase, TargetIndicator, TargetQualifier, TargetLevel, and TargetDose. A generator function was implemented to produce 10,000 random but protocol-compliant target names, representing the broad spectrum of TG 263 naming conventions. A parser combinator framework validated the generated names against protocol rules. Additionally, 25 deliberately invalid target names were tested to evaluate the framework's rejection capabilities. Parsing of validated target names back into structured class instances was performed to verify consistency and correctness.
Results:
The framework successfully validated all 10,000 compliant target names, while correctly rejecting all 25 invalid names. Parsing tests revealed 562 instances with at least one incorrect property, highlighting opportunities for further refinement. The robust implementation uncovered an error in the TG 263 Excel file hosted by the AAPM. The toolkit's performance and accuracy demonstrate its utility for both generating and validating compliant names and parsing them into structured formats.
Conclusion:
The TG 263 Toolkit provides a validated, open-source solution for generating and validating target names compliant with TG 263, while supporting bidirectional parsing. This framework enhances protocol adherence, simplifies implementation, and sets the stage for future work extending to organs and risk structure naming conventions.