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

Abstract

Purpose: 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. Approach: We train machine learning (ML) models to map (2-D) patient features to 3-D organat-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. Results: Different ML algorithms result in similar test mean absolute errors: ∼8 mm for liver LR, IS, and spleen AP, IS; ∼5 mm for liver AP and spleen LR; ∼80% for abdomen sDSC; and ∼60% 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 (þ5-mm error for spleen IS, −10% sDSC for liver). The automatic step to improve realism generally results in limited metric accuracy loss, but fails in one case (out of 60). Conclusion: Our ML-based pipeline leads to phantoms that are significantly and substantially more individualized than currently used human-designed criteria.

A computational framework for physics-informed symbolic regression with straightforward integration of domain knowledge

“…This is done by performing variation based on the linkage model, to capture genotypic interdependencies. Because a short yet accurate equation is needed in many physical use cases, GP-GOMEA has been adapted in several physical regression efforts 75 , 76 . This is despite the fact that it does not consider any domain knowledge or physical requirements aside from interoperability.…”

Parameterless Gene-pool Optimal Mixing Evolutionary Algorithms

“…Another example is adaptation of GOMEA for Genetic Programming (GP-GOMEA) [42]. Beside showing better performance than alternative GP algorithms on classical machine learning benchmarks, GP-GOMEA has been also successfully applied to a real-world medical problem, namely, a radiotherapy dose reconstruction [41], [44]. This application was noted with a Silver Humies award in 2021.…”

Symbolic regression as a feature engineering method for machine and deep learning regression tasks

“…Machine and deep learning (DL), achieving optimal performance and interpretability in regression tasks stands as a fundamental computational challenge, critical for applications spanning various fields of science and engineering [1][2][3][4][5][6][7]. The efficacy of the machine learning (ML) pipeline hinges on numerous components that govern its performance [8][9][10], explainability [11,12], and development efficiency [13,14].…”