Med Phys
-
Spectral CT using a dual layer detector offers the possibility of retrospectively introducing spectral information to conventional CT images. In theory, the dual-layer technology should not come with a dose or image quality penalty for conventional images. In this study, we evaluate the influence of a dual-layer detector (IQon Spectral CT, Philips Healthcare) on the image quality of conventional CT images, by comparing these images with those of a conventional but otherwise technically comparable single-layer CT scanner (Brilliance iCT, Philips Healthcare), by means of phantom experiments. ⋯ At equivalent dose levels, this study showed similar quality of conventional images acquired on iCT and IQon for medium-sized phantoms and slightly degraded image quality for (very) large phantoms at lower tube voltages on the IQon. Accordingly, it may be concluded that the introduction of a dual-layer detector neither compromises image quality of conventional images nor increases radiation dose for normal-sized patients, and slightly degrades dose efficiency for large patients at 120 kVp and lower tube voltages.
-
Comparative Study
Machine learning algorithms for outcome prediction in (chemo)radiotherapy: An empirical comparison of classifiers.
Machine learning classification algorithms (classifiers) for prediction of treatment response are becoming more popular in radiotherapy literature. General Machine learning literature provides evidence in favor of some classifier families (random forest, support vector machine, gradient boosting) in terms of classification performance. The purpose of this study is to compare such classifiers specifically for (chemo)radiotherapy datasets and to estimate their average discriminative performance for radiation treatment outcome prediction. ⋯ Random forest and elastic net logistic regression yield higher discriminative performance in (chemo)radiotherapy outcome and toxicity prediction than other studied classifiers. Thus, one of these two classifiers should be the first choice for investigators when building classification models or to benchmark one's own modeling results against. Our results also show that an informed preselection of classifiers based on existing datasets can improve discrimination over random selection.