Learning to Optimize Radiotherapy Plans via Fluence Maps Diffusion Model Generation and LSTM-based Optimization

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Poles, Isabella, Arberet, Simon, Gao, Riqiang, Kraus, Martin, Santambrogio, Marco D., Ghesu, Florin C., Kamen, Ali, Comaniciu, Dorin
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910217126543360
author Poles, Isabella
Arberet, Simon
Gao, Riqiang
Kraus, Martin
Santambrogio, Marco D.
Ghesu, Florin C.
Kamen, Ali
Comaniciu, Dorin
author_facet Poles, Isabella
Arberet, Simon
Gao, Riqiang
Kraus, Martin
Santambrogio, Marco D.
Ghesu, Florin C.
Kamen, Ali
Comaniciu, Dorin
contents Volumetric Modulated Arc Therapy (VMAT) is a cornerstone of modern radiation therapy, enabling highly conformal tumor irradiation and healthy-tissue sparing. Yet, its planning solves inverse and nested optimization for multi-leaf collimators, monitor units and dose parameters, while enforcing their consistency to ensure mechanical deliverability. Nevertheless, this process often requires repeated re-optimization when treatment configurations change, resulting in substantial planning time per patient. To address these problems, we present a diffusion-driven Learning-to-Optimize (L2O) method for end-to-end VMAT planning. A distribution-matching distilled diffusion model learns a clinically feasible manifold of fluence maps, enabling their one-shot generation. On top of this, an LSTM-based L2O module learns gradient update dynamics to swiftly refine fluence maps toward prescribed dose objectives during inference. Experimental results on clinical and public prostate cancer cohorts demonstrate improved planning efficiency, flexibility, and machine deliverability over currently available end-to-end VMAT planners.
format Preprint
id arxiv_https___arxiv_org_abs_2605_13713
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Learning to Optimize Radiotherapy Plans via Fluence Maps Diffusion Model Generation and LSTM-based Optimization
Poles, Isabella
Arberet, Simon
Gao, Riqiang
Kraus, Martin
Santambrogio, Marco D.
Ghesu, Florin C.
Kamen, Ali
Comaniciu, Dorin
Computer Vision and Pattern Recognition
Image and Video Processing
Volumetric Modulated Arc Therapy (VMAT) is a cornerstone of modern radiation therapy, enabling highly conformal tumor irradiation and healthy-tissue sparing. Yet, its planning solves inverse and nested optimization for multi-leaf collimators, monitor units and dose parameters, while enforcing their consistency to ensure mechanical deliverability. Nevertheless, this process often requires repeated re-optimization when treatment configurations change, resulting in substantial planning time per patient. To address these problems, we present a diffusion-driven Learning-to-Optimize (L2O) method for end-to-end VMAT planning. A distribution-matching distilled diffusion model learns a clinically feasible manifold of fluence maps, enabling their one-shot generation. On top of this, an LSTM-based L2O module learns gradient update dynamics to swiftly refine fluence maps toward prescribed dose objectives during inference. Experimental results on clinical and public prostate cancer cohorts demonstrate improved planning efficiency, flexibility, and machine deliverability over currently available end-to-end VMAT planners.
title Learning to Optimize Radiotherapy Plans via Fluence Maps Diffusion Model Generation and LSTM-based Optimization
topic Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2605.13713