Dose Estimation Using 3D Transformer Models in High-Dose-Rate Brachytherapy for Cervical Cancer
-
By
-
Weiwei Guo
-
Wanwei Jian
-
Lin Zhu
-
Bailin Zhang
-
Qiang He
-
Geng Yang
-
Xuetao Wang
-
January 20, 2026
-
Clinical Scorecard: Dose Estimation Using 3D Transformer Models in High-Dose-Rate Brachytherapy for Cervical Cancer
At a Glance
| Category | Detail |
|---|
| Condition | Locally advanced cervical cancer |
| Key Mechanisms | High-dose-rate brachytherapy (HDRBT) combined with external beam radiation therapy (EBRT) |
| Target Population | Patients with locally advanced cervical cancer undergoing HDR brachytherapy |
| Care Setting | Oncology clinics specializing in radiation therapy |
Key Highlights
- Proposed a 3D transformer-based deep learning model for dose prediction in HDR brachytherapy.
- Study analyzed 96 CT-based treatment plans from 24 patients.
- Hybrid architectures combining CNNs and self-attention mechanisms enhance dose prediction accuracy.
Guideline-Based Recommendations
Diagnosis
- Use CT imaging for contour delineation of clinical targets and organs at risk (OARs).
Management
- Implement HDR brachytherapy with freehand interstitial needle insertion for customized dose distribution.
Monitoring & Follow-up
- Conduct dosimetric evaluation and analysis using DVH metrics and 3D gamma analysis.
Risks
- Consider procedural complexity and variability in needle insertion quality among oncologists.
Patient & Prescribing Data
96 CT-based treatment plans from 24 patients with cervical cancer.
Patients received 45.0–50.4 Gy EBRT followed by HDR brachytherapy with 6 Gy/fraction.
Clinical Best Practices
- Utilize knowledge-based planning (KBP) models to improve plan quality.
- Ensure treatment planning is optimized using hybrid inverse treatment planning and optimization algorithms.
References
