EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

A Combined Loss-driven Framework for Automated Parotid Segmentation in Head-and-Neck Computed Tomography

Aryan Tyagi, Anuj Kumar, Sandeep Singh, Mohini Manav, Soniya Pal

Journal of Medical Physics · 2025-07 · 50 (3) · 525-532

DOI: 10.4103/jmp.jmp_169_25 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Medical Physics, online erschienen am 29.09.2025. Verfasst von Aryan Tyagi und 4 weiteren Autor:innen. Beteiligte Einrichtung: University School of Automation and Robotics, Guru Gobind Singh Indraprastha University, New Delhi, India. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

This study presents a deep learning framework for automatic parotid segmentation using three-dimensional (3D) U-Net and attention-augmented 3D U-Net architectures trained with a novel combined loss function tailored for anatomical accuracy and class imbalance. Materials and Methods: A curated dataset of 379 noncontrast head-and-neck computed tomography scans with expert-verified contours was used. Two architectures a residual 3D U-Net and its attention-enhanced variant were implemented using TensorFlow. The networks were trained with both categorical cross-entropy and a proposed combined loss integrating modified Dice Score Coefficient (mDSC) and focal loss (FL) with weights 0.7 and 0.3. The models were evaluated using dice similarity coefficient (DSC), Intersection over Union (IoU), and categorical accuracy. A custom checkpointing strategy was designed to preserve model weights corresponding to both peak DSC and minimum validation loss. The code and pretrained models are hosted on a publicly available GitHub repository at: https://github.com/1aryantyagi/Segmentation-Paper.

Results

The 3D U-Net trained with the combined loss achieved a median Dice score of 0.8835 (left parotid) and 0.8709 (right), with mean IoU values of 0.7672 and 0.7358, indicating strong segmentation accuracy. The U-Net produced comparable results, supporting the combined loss’s consistency. Bland–Altman analysis confirmed reduced variability and improved agreement.

Conclusion

The integration of mDSC and FL within a 3D U-Net architecture significantly improves segmentation performance, robustness, and spatial precision. These findings support the clinical feasibility of the proposed framework for automated, reproducible parotid delineation in radiotherapy planning.

Schlagwörter

Stichwörter der Autor:innen

Computed tomography imagedeep learningloss functionmedical imageparotidradiotherapy planningsegmentation

Autor:innen und Institutionen

  • University School of Automation and Robotics, Guru Gobind Singh Indraprastha University, New Delhi, India.
  • Department of Radiotherapy, LLRM Medical College, Meerut, Uttar Pradesh, India.
  • Division of Medical Physics, Rajiv Gandhi Cancer Hospital and Research Centre, Rohini, New Delhi, India.
  • Department of Radiation Oncology, Andromeda Cancer Hospital, Kundli, Sonipat, Haryana, India.
  • Department of Radiation Oncology, Max Super Speciality Hospital, Shalimar Bagh, New Delhi, India.

Zitieren

APA

Tyagi, A., Kumar, A., Singh, S., Manav, M., & Pal, S. (2025). A Combined Loss-driven Framework for Automated Parotid Segmentation in Head-and-Neck Computed Tomography. Journal of Medical Physics, 50(3), 525-532. https://doi.org/10.4103/jmp.jmp_169_25

Vancouver

Tyagi A, Kumar A, Singh S, Manav M, Pal S. A Combined Loss-driven Framework for Automated Parotid Segmentation in Head-and-Neck Computed Tomography. Journal of Medical Physics. 2025;50(3):525-532. doi:10.4103/jmp.jmp_169_25

BibTeX
@article{Tyagi2025,
  author = {Tyagi, Aryan and Kumar, Anuj and Singh, Sandeep and Manav, Mohini and Pal, Soniya},
  title = {A Combined Loss-driven Framework for Automated Parotid Segmentation in Head-and-Neck Computed Tomography},
  journal = {Journal of Medical Physics},
  year = {2025},
  volume = {50},
  number = {3},
  pages = {525--532},
  doi = {10.4103/jmp.jmp_169_25}
}
Kostenlos starten →