Change search
ReferencesLink to record
Permanent link

Direct link
Dose-volume histogram prediction using density estimation
Elekta Instrument AB, Sweden.
Linköping University, Department of Biomedical Engineering, Medical Informatics. Linköping University, Faculty of Science & Engineering. Elekta Instrument AB, Sweden.
2015 (English)In: Physics in Medicine and Biology, ISSN 0031-9155, E-ISSN 1361-6560, Vol. 60, no 17, 6923-6936 p.Article in journal (Refereed) Published
Abstract [en]

Knowledge of what dose-volume histograms can be expected for a previously unseen patient could increase consistency and quality in radiotherapy treatment planning. We propose a machine learning method that uses previous treatment plans to predict such dose-volume histograms. The key to the approach is the framing of dose-volume histograms in a probabilistic setting. The training consists of estimating, from the patients in the training set, the joint probability distribution of some predictive features and the dose. The joint distribution immediately provides an estimate of the conditional probability of the dose given the values of the predictive features. The prediction consists of estimating, from the new patient, the distribution of the predictive features and marginalizing the conditional probability from the training over this. Integrating the resulting probability distribution for the dose yields an estimate of the dose-volume histogram. To illustrate how the proposed method relates to previously proposed methods, we use the signed distance to the target boundary as a single predictive feature. As a proof-of-concept, we predicted dose-volume histograms for the brainstems of 22 acoustic schwannoma patients treated with stereotactic radiosurgery, and for the lungs of 9 lung cancer patients treated with stereotactic body radiation therapy. Comparing with two previous attempts at dose-volume histogram prediction we find that, given the same input data, the predictions are similar. In summary, we propose a method for dose-volume histogram prediction that exploits the intrinsic probabilistic properties of dose-volume histograms. We argue that the proposed method makes up for some deficiencies in previously proposed methods, thereby potentially increasing ease of use, flexibility and ability to perform well with small amounts of training data.

Place, publisher, year, edition, pages
IOP PUBLISHING LTD , 2015. Vol. 60, no 17, 6923-6936 p.
Keyword [en]
DVH prediction; machine learning; treatment planning; kernel density estimation
National Category
Medical Biotechnology
URN: urn:nbn:se:liu:diva-121747DOI: 10.1088/0031-9155/60/17/6923ISI: 000361123500020PubMedID: 26305670OAI: diva2:859351

Funding Agencies|Swedish Research Council [2012-4281]; Linnaeus Center CADICS

Available from: 2015-10-06 Created: 2015-10-05 Last updated: 2015-10-28

Open Access in DiVA

fulltext(766 kB)3 downloads
File information
File name FULLTEXT01.pdfFile size 766 kBChecksum SHA-512
Type fulltextMimetype application/pdf

Other links

Publisher's full textPubMed

Search in DiVA

By author/editor
Sjölund, Jens
By organisation
Medical InformaticsFaculty of Science & Engineering
In the same journal
Physics in Medicine and Biology
Medical Biotechnology

Search outside of DiVA

GoogleGoogle Scholar
Total: 3 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Altmetric score

Total: 179 hits
ReferencesLink to record
Permanent link

Direct link