行業指南

放射治療規劃中的人工智慧

AI in radiation therapy planning uses deep learning to outline tumors and nearby organs on CT or MRI scans (auto-contouring) and to predict or generate dose plans, which can cut hours of manual work to minutes.

聆聽此指南AI 語音敘述 · ~3 min

本頁閱讀時間3分鐘
  1. 概述
  2. 深入探討
  3. 現實世界的實施
  4. The Future of AI in Radiation Therapy Planning
  5. 付諸實踐
  6. 不斷探索
  7. 常見問題

概述

It matters because faster, more consistent planning can shorten the wait before treatment and makes adaptive radiotherapy practical. Radiation oncologists and medical physicists still review, edit and approve every contour and plan.

深入探討

Radiation therapy tries to deliver a high dose to cancer while sparing healthy tissue. Planning starts with contouring: outlining the visible tumor (gross tumor volume, GTV), areas of likely microscopic spread (clinical target volume, CTV), a margin for setup and motion (planning target volume, PTV), and organs at risk (OARs) such as the heart, lungs, spinal cord or rectum. Manual contouring can take hours for complex sites like head and neck, and different clinicians draw noticeably different outlines. Auto-contouring was first done with atlas-based methods, which warp outlines from reference patients onto a new scan. Deep learning models, often U-Net-style networks, now generally perform better and are built into commercial systems from vendors such as Varian, Elekta, RaySearch, Siemens Healthineers and Mirada. Microsoft's InnerEye research project, developed with Addenbrooke's Hospital in Cambridge, released open-source tools in this area. OAR contouring is the most mature use. Target contouring is harder because it depends on clinical judgment, pathology and imaging beyond the scan. Dose planning uses AI in two main ways. Knowledge-based planning, such as Varian's RapidPlan, learns from prior approved plans to predict achievable dose-volume histograms for a new patient. Deep learning dose prediction and automated planning go further by estimating the 3D dose distribution and driving the optimizer. Adaptive radiotherapy depends on this speed. Systems such as Varian Ethos and MR-linacs like Elekta Unity re-image the patient at treatment and adapt the plan. A common misconception is that AI plans treatment on its own. Physicians approve contours and prescriptions, and medical physicists check plans and run quality assurance before delivery.

現實世界的實施

A head and neck cancer clinic uses auto-contouring to draw more than 20 organs at risk, such as the parotid glands and spinal cord, and the dosimetrist edits a few slices instead of drawing them all by hand.

A prostate cancer center uses knowledge-based planning that predicts achievable dose limits for the bladder and rectum from past approved plans, giving planners a realistic starting target.

An adaptive therapy system re-contours organs on the day's imaging scan before each session, and the care team reviews and approves the updated plan while the patient waits on the table.

A physics team compares a new auto-contouring tool against local clinicians' contours on a test set of past patients before using it clinically, noting where it struggles, such as after surgery.

The Future of AI in Radiation Therapy Planning

Auto-contouring for organs at risk is already routine in many centers, and wider use of automated planning and online adaptive therapy is likely as evidence and software mature. Harder problems include reliable target contouring, handling unusual anatomy, and building automatic checks that flag when a contour or plan looks unlike anything the model was trained on. Professional bodies are developing guidance for commissioning and ongoing quality assurance of AI tools. The realistic direction is faster planning and more frequent plan adaptation, with physicians and physicists spending their time reviewing, catching errors and handling complex cases.

付諸實踐

為什麼這很重要

  • 背景與規則. 產業背景決定了人工智慧創意能否與現實接觸。
  • 品質管控. 領域約束會影響可接受的錯誤率和監督模型。
  • 配裝選擇. 成功的部署使技術能力與第一線工作流程保持一致。

留意

  • 監理要求可能會使原本強大的原型失效。
  • 歷史資料可能會編碼損害特定社區的偏見。
  • 遺留系統可能會造成整合瓶頸和隱性成本。

後續步驟

  1. 讓領域專家參與從問題框架到評估的整個過程。
  2. 在啟動前設計審計追蹤和文件。
  3. 儘早驗證合規性和安全義務。
  4. 分階段推出,並有明確的停止和回滾標準。

不斷探索

免費電子報

獲取每日 AI 簡報

每個工作日上午提供三則經過驗證的 AI 新聞,以淺顯易懂的英文撰寫。永久免費,無廣告。

每個工作日一封電子郵件。一鍵取消訂閱。我們絕不出售或分享您的地址。

測試一下自己

Take the AI in Radiation Therapy Planning quiz

每個答案皆有即時回饋,通過課程後可獲得具備可驗證 ID 的可分享證書。

開始測驗

支持免費的 AI 教育。AI Understanding 是一個 501(c)(3) 非營利組織——絕無廣告,絕無付費牆。 捐款

常見問題

What is AI in Radiation Therapy Planning?

AI in radiation therapy planning uses deep learning to outline tumors and nearby organs on CT or MRI scans (auto-contouring) and to predict or generate dose plans, which can cut hours of manual work to minutes. It matters because faster, more consistent planning can shorten the wait before treatment and makes adaptive radiotherapy practical. Radiation oncologists and medical physicists still review, edit and approve every contour and plan.

What does auto-contouring do in radiation therapy planning?

Auto-contouring draws outlines of targets and nearby organs, which saves hours of manual contouring.

Which contouring task is currently most mature for AI?

Organs at risk have consistent, visible anatomy. Targets depend more on clinical judgment and information beyond the scan.

What does the planning target volume (PTV) add to the clinical target volume?

The PTV adds a margin so the target still gets its dose despite small setup differences and movement.

How does knowledge-based planning such as RapidPlan work?

Knowledge-based planning uses previous good plans to set realistic dose goals for a new patient.

Why is a high Dice score not enough to prove a contour is clinically acceptable?

Dice measures overall overlap. Clinical acceptability and editing time depend on where errors are, which is why surface metrics and expert review matter.