Validation of different automated segmentation models for target volume contouring in postoperative radiotherapy for breast cancer and regional nodal irradiation.

Introduction: Target volume delineation is routinely performed in postoperative radiotherapy (RT) for breast cancer patients, but it is a time-consuming process. The aim of the present study was to validate the quality, clinical usability and institutional-specific implementation of different auto-segmentation tools into clinical routine.
Methods: Three different commercially available, artificial intelligence-, ESTRO-guideline-based segmentation models (M1-3) were applied to fifty consecutive reference patients who received postoperative local RT including regional nodal irradiation for breast cancer for the delineation of clinical target volumes: the residual breast, implant or chestwall, axilla levels 1 and 2, the infra- and supraclavicular regions, the interpectoral and internal mammary nodes. Objective evaluation metrics of the created structures were conducted with the Dice similarity index (DICE) and the Hausdorff distance, and a manual evaluation of usability.
Results: The resulting geometries of the segmentation models were compared to the reference volumes for each patient and required no or only minor corrections in 72 % (M1), 64 % (M2) and 78 % (M3) of the cases. The median DICE and Hausdorff values for the resulting planning target volumes were 0.87-0.88 and 2.96-3.55, respectively. Clinical usability was significantly correlated with the DICE index, with calculated cut-off values used to define no or minor adjustments of 0.82-0.86. Right or left sided target and breathing method (deep inspiration breath hold vs. free breathing) did not impact the quality of the resulting structures.
Conclusion: Artificial intelligence-based auto-segmentation programs showed high-quality accuracy and provided standardization and efficient support for guideline-based target volume contouring as a precondition for fully automated workflows in radiotherapy treatment planning.
Competing Interests: The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Our radiotherapy department has on-going research collaboration with RaySearch Laboratories AB. All authors affirm that they have no financial, non-financial or personal interest or belief in the subject matter or materials discussed in this manuscript, that might jeopardize their objectivity. EM received speaker fees from Elekta outside the submitted work. SK received speaker fees from Siemens Healthineers. JL received travelling support from Micropos Medical and from RaySearch Laboratories outside the submitted work. FW received speaker fees from AstraZeneca, Varian Medical Systems and Merck Sharp & Dohme and travel support for attending meetings from Varian Medical Systems and Micropos Medical outside the submitted work. PHS received support from Physician Scientist Program of the Medical Faculty (University of Heidelberg), grants from Dietmar-Hopp-Foundation and fees from NovoCure GmbH outside the submitted work. JD reports grants from CRI The Clinical Research Institute, grants from View Ray Incl., grants from Accuray International, grants from Accuray Incorporated, grants from RaySearch Laboratories AB, grants from Vision RT limited, grants from Merck Serono GmbH, grants from Astellas Pharma GmbH, grants from Astra Zeneca GmbH, grants from Siemens Healthcare GmbH, grants from Solution Akademie GmbH, grants from Eromed PLC Surrey Research Park, grants from Quintiles GmbH, grants from Pharmaceutical Research Associates GmbH, grants from Boehringer Ingelheim Pharma GmbH Co, grants from PTW-Frieburg Dr. Pychlau GmbH, grants from Nanobiotix A.a., grants from IntraOP Medical, outside the submitted work. LK reports, personal fees from Accuray Inc., and Novocure GmbH outside the submitted work. JHR reports honoraria and travel reimbursement by Viewray Inc., Pfizer Inc and IntraOP Medical as well as grants from IntraOP Medical and Varian Medical Systems outside the submitted work. BG, FRW, DN, LH, LAD, SH, MV have nothing to declare.
(© 2024 The Author(s).)