PURPOSE: Inconsistent and institution-specific naming of regions of interest (ROIs) in cancer radiotherapy planning limits interoperability, hinders automation, and introduces ambiguity in clinical workflows. This study aims to develop and evaluate a multistage, foundation model–based workflow for automated ROI classification and standardization in accordance with AAPM Task Group 263 (TG-263) nomenclature, while maintaining flexibility for customization to institution-specific naming conventions. MATERIALS AND METHODS: A four-stage modular pipeline integrating large language models (LLMs) and a Contrastive Language-Image Pretraining (CLIP) model was developed. The system sequentially performs (1) anatomic site classification, (2) ROI type classification into four categories, (3) standardization based on customizable protocols, and (4) laterality verification via CLIP. The pipeline operates on standard DICOM RTSTRUCT and CT data and requires no model retraining. We evaluated the workflow on 600 patients from three institutions across five disease sites (head and neck, breast, lung, cervix, rectum). Classification and renaming accuracy were assessed at both global and institution-site levels, with error taxonomy used to characterize failure modes. RESULTS: The workflow achieved an overall ROI type classification accuracy of 99.12% and a renaming accuracy of 97.92%. Classification performance was consistently high across all institutions and sites; renaming accuracy showed minor variation driven by protocol omissions, compound suffix ambiguity, and institution-specific naming conventions. CLIP-based laterality verification corrected all errors in a targeted stress test with deliberate left-right label inversions. CONCLUSION: This study demonstrates that foundation language models, augmented by a vision-language module for laterality verification, can robustly automate ROI classification and naming across institutions, offering a customizable and retraining-free solution for standardizing cancer radiotherapy workflows.
Background/Objectives: Multi-scenario calculational methods have been used to evaluate proton teletherapy plan robustness but few studies have been performed to determine the accuracy of these calculational methods. This study evaluates a multi-scenario method by comparing calculations to measurements made in phantoms that simulate the effects of possible uncertainties. Methods: Plans were made using four phantoms in which the delivered dose was highly sensitive to positional and penetration uncertainties. The effects of alignment and penetration uncertainties on the dose distributions of each of those phantoms were simulated by performing calculations using nine different uncertainty scenarios and comparing the calculations to measurements with induced physical alignment displacements. Measured dose distributions were obtained by exposing films placed inside the phantoms and extracting multiple linear profiles. The maximum and minimum doses obtained for each of the calculational scenarios were compared with the measured dose profiles. In addition, comparisons of DVHs for nominal and uncertainty scenarios were performed. Results: The results showed that, under the influence of uncertainties, the minimum dose for the four phantoms decreased by more than 20 Gy, the V95% coverage fluctuated by more than 10%, but the maximum dose parameter changed by less than 5 Gy. This was expected, as no margins for uncertainties were applied around the targets. The envelope bounded by the maximum and minimum possible calculated doses contained most of the measurements, although the shapes of the dose profiles displayed some mismatches for wedge and head phantoms. There were a few points where the measured maximum dose for bone and lung slab phantom cases was slightly higher than the maximum dose calculated from the nine scenarios. Conclusions: This study demonstrates that a nine-scenario method can adequately evaluate the robustness of simple mono-directional plans containing heterogeneities.
About Me
Comparison of a multi-scenario robustness evaluation method with measurement for proton teletherapy
Comparison of a multi-scenario robustness evaluation method with measurement for proton teletherapy