About & Announcements

Announcements

[October 2026] Best Paper Award goes to Blanca Rodriguez-Gonzalez. Congratulations to Blanca Rodriguez-Gonzalez and co-authors Ivan Escobar Corominas, Angel Torrado-Carvajal, Borja Rodriguez-Vila, Alejandro Mazal, and Norberto Malpica, whose paper Can We Trust Synthetic CT Algorithms? Uncertainty-Aware Evaluation of CBCT to CT Synthesis for Adaptive Proton Therapy received the MIART 2026 Best Paper Award. The award was presented at the closing session of the workshop. The work shows that uncertainty estimates for CBCT-to-CT synthesis can localise regions of increased HU and dose error, which adds useful reliability information to quality assurance for adaptive proton therapy. The slides are available, and the paper is listed on the papers page.
[October 2026] MIART 2026 has concluded, and the slides are now online. MIART 2026 took place on October 1, 2026 at MICCAI 2026 in Strasbourg, with two invited keynotes, nine oral presentations, and 27 posters. Many thanks to our keynote speakers, Stine Korreman and Matteo Maspero, to all presenting authors, to the reviewers, and to everyone who attended. Slides from both keynotes and all nine oral presentations are linked from the programme. We look forward to welcoming you again next year. Follow us on LinkedIn for news about the next edition.
[September 2026] Programme now online. The nine oral presentations are organised into three thematic sessions, alongside two invited keynotes. MIART 2026 takes place in Room Londrezs 1 (ground floor), 08:00 – 12:30 on October 1, 2026; see the full programme. The complete list of accepted papers is available on the papers page.
[September 2026] Twenty-seven papers accepted to MIART 2026. Following double-blind peer review, the organising committee has accepted 27 manuscripts, all of which are presented as posters, with 9 additionally selected for oral presentation. The full listing is available on the papers page. The accepted work originates from more than 40 universities, hospitals, and research institutes across 14 countries on three continents. Thematically, the accepted contributions span image synthesis and cross-modality translation (7 papers); segmentation and contour quality assurance (7); dose calculation, dose prediction, and plan optimisation (5); outcome, toxicity, and treatment-response modelling (4); motion management, longitudinal adaptation, and deformable registration (3); and agentic large language models for radiological image analysis (1). All accepted papers will be published in the MICCAI 2026 Springer proceedings. The committee thanks all authors for their submissions and the reviewers for their careful and timely assessments.
[August 2026] Decisions released. Author notifications were sent on August 21, 2026, and camera-ready papers were due August 27, 2026. Accepted papers will appear in the MICCAI 2026 Springer proceedings.
[August 2026] MIART 2026 is endorsed by ESTRO (European Society for Radiotherapy and Oncology). This highlights the growing importance of clinically impactful AI research at the intersection of medical image computing and radiation oncology.
[July 2026] Submission deadline extended! The paper submission deadline has been extended to July 17, 2026 (11:59 PM CEST), with submissions closing at 12:00 AM CEST on July 18, 2026. Submit your paper via OpenReview.
[June 2026] Submissions are now open! Submit your paper via OpenReview.
[May 2026] MIART 2026 has been accepted as an official MICCAI 2026 satellite event. Follow us on LinkedIn for updates.

[March 2026] The MIART 2026 workshop proposal has been submitted. Further details will be announced in due course.

About the Workshop

Radiotherapy (RT) is a cornerstone of modern oncology, accounting for approximately 40% of curative cancer treatments. RT uses targeted radiation to destroy cancer cells; however, healthy tissues in the radiation path may also be affected. The fundamental challenge is therefore to precisely deliver dose to the tumour while protecting nearby organs. Unlike diagnostic imaging, where the primary goal is detection, RT operates as a complex, multi-stage therapeutic ecosystem requiring the seamless integration of longitudinal imaging, precise anatomical definition, radiation physics, and biological response modelling.

The opacity of deep learning models is a particular concern in RT, where algorithmic decisions directly govern the physical delivery of high-dose radiation. A geometric error in AI-based contouring or a hallucination in image synthesis does not merely result in a misdiagnosis; it can lead to catastrophic geographic misses, reducing the probability of cure and increasing the risk of severe toxicity.

The MIART Workshop aims to establish a dedicated forum within the MICCAI community for researchers applying AI to radiation oncology. The workshop advances a vision of therapeutic AI that is physics-aware and biologically grounded, bridging the gap between data-driven discovery and clinical intervention to support personalised radiotherapy.

Workshop Objectives

Community Unification – Establish a centralised home within MICCAI for the dispersed radiotherapy AI research community.
Holistic Optimisation – Advance end-to-end optimisation across the full therapeutic pipeline, moving beyond isolated sub-task solutions.
Domain Integration – Promote the principled incorporation of physics, biology, and clinical constraints into deep learning models.
Comprehensive Modelling – Expand predictive AI beyond binary outcomes towards detailed safety, toxicity, and outcome modelling.

← Back to Home Papers →