Advertisement

Proof of Principle Subjective Analysis of a New Deep-Learning Model for MRI Autocontouring in Clinical Practice

      Delineation of organs at risk (OARs) is critical for minimizing toxicity to vulnerable normal tissues while delivering radiation therapy. This is especially true when the radiosensitive OARs are in close anatomical proximity to disease sites receiving definitive radiation doses, a common occurrence in head and neck (HN) patients. Unfortunately, accurate manual delineation of OARs is operator-dependent and time consuming. While atlas-based segmentation (ABAS) has been used for several years to expedite the process of OAR delineation, it is not always robust to patient variability and resulting contours can be insufficiently accurate or require time-consuming corrections. As such, deep-learning contouring (DLC) has emerged as a possible framework for the next generation of autocontouring tools which can produce contours that are more robust to unique patient anatomy and more closely approximate expert level manual contouring. While quantitative analysis provides some information regarding the quality of DLC versus ABAS or manual delineation, it may not be the best predictor of clinical acceptability or expert approval. Here we present data from a preliminary qualitative analysis of the performance of a new MRI-based DLC of a critical OAR in the HN: the parotid gland.
      To read this article in full you will need to make a payment
      ASTRO Member Login
      ASTRO Members, full access to the journal is a member benefit. Use your society credentials to access all journal content and features.
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect

      Comments

      Commenting Guidelines

      To submit a comment for a journal article, please use the space above and note the following:

      • We will review submitted comments as soon as possible, striving for within two business days.
      • This forum is intended for constructive dialogue. Comments that are commercial or promotional in nature, pertain to specific medical cases, are not relevant to the article for which they have been submitted, or are otherwise inappropriate will not be posted.
      • We require that commenters identify themselves with names and affiliations.
      • Comments must be in compliance with our Terms & Conditions.
      • Comments are not peer-reviewed.