Machine learning for the prediction of pseudorealistic pediatric abdominal phantoms for radiation dose reconstruction

Marco Virgolin, Ziyuan Wang, Tanja Alderliesten, Peter A. N. BosmanJournal of Medical Imaging 2020 · DOI

Abstract

<b>Purpose:</b> Current phantoms used for the dose reconstruction of long-term childhood cancer survivors lack individualization. We design a method to predict highly individualized abdominal three-dimensional (3-D) phantoms automatically. <b>Approach:</b> We train machine learning (ML) models to map (2-D) patient features to 3-D organ-at-risk (OAR) metrics upon a database of 60 pediatric abdominal computed tomographies with liver and spleen segmentations. Next, we use the models in an automatic pipeline that outputs a personalized phantom given the patient's features, by assembling 3-D imaging from the database. A step to improve phantom realism (i.e., avoid OAR overlap) is included. We compare five ML algorithms, in terms of predicting OAR left-right (LR), anterior-posterior (AP), inferior-superior (IS) positions, and surface Dice-Sørensen coefficient (sDSC). Furthermore, two existing human-designed phantom construction criteria and two additional control methods are investigated for comparison. <b>Results:</b> Different ML algorithms result in similar test mean absolute errors: <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mo>∼</mml:mo> <mml:mn>8</mml:mn> <mml:mtext> </mml:mtext> <mml:mi>mm</mml:mi></mml:mrow> </mml:math> for liver LR, IS, and spleen AP, IS; <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mo>∼</mml:mo> <mml:mn>5</mml:mn> <mml:mtext> </mml:mtext> <mml:mi>mm</mml:mi></mml:mrow> </mml:math> for liver AP and spleen LR; <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mo>∼</mml:mo> <mml:mn>80</mml:mn> <mml:mo>%</mml:mo></mml:mrow> </mml:math> for abdomen sDSC; and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mo>∼</mml:mo> <mml:mn>60</mml:mn> <mml:mo>%</mml:mo></mml:mrow> </mml:math> to 65% for liver and spleen sDSC. One ML algorithm (GP-GOMEA) significantly performs the best for 6/9 metrics. The control methods and the human-designed criteria in particular perform generally worse, sometimes substantially ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mo>+</mml:mo> <mml:mn>5</mml:mn> <mml:mtext>-</mml:mtext> <mml:mi>mm</mml:mi></mml:mrow> </mml:math> error for spleen IS, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mo>-</mml:mo> <mml:mn>10</mml:mn> <mml:mo>%</mml:mo></mml:mrow> </mml:math> sDSC for liver). The automatic step to improve realism generally results in limited metric accuracy loss, but fails in one case (out of 60). <b>Conclusion:</b> Our ML-based pipeline leads to phantoms that are significantly and substantially more individualized than currently used human-designed criteria.
BibTeX
@article{virgolin2020machine,
  author = {Marco Virgolin and Ziyuan Wang and Tanja Alderliesten and Peter A. N. Bosman},
  title = {{Machine learning for the prediction of pseudorealistic pediatric abdominal phantoms for radiation dose reconstruction}},
  journal = {Journal of Medical Imaging},
  year = {2020},
  doi = {10.1117/1.jmi.7.4.046501},
}
Figure