Dosiomics-based comparison of dose distributions in nasopharyngeal cancer patients: 3D-CRT versus Tomotherapy

Abstract

Background: This research was conducted to compare dosiomics features extracted from planning target volume (PTV) between the two three-dimensional conformal radiation therapy (3D-CRT) and helical tomotherapy (HT) techniques in nasopharyngeal cancer. Materials and Methods: 3D-CRT plans were designed for ten nasopharyngeal patients previously treated with HT. For both treatment techniques, the total prescription dose was 70 Gray in 33 fractions. At first, the dosimetric parameters, including mean dose, conformity index (CI), and homogeneity index (HI), were calculated for two techniques. Then, using 3D-Slicer software, dosiomics features were extracted from the dose matrix of PTV. Results: In comparing the plans regarding dosimetric parameters, HT plans had a lower mean dose and higher CI (p-value <0.01 and <0.001, respectively). HI had no statistically difference between the two groups (p>0.9). Among 93 features extracted from PTV70, only 40 features including eight features from the first-order group, ten features from the gray-level co-occurrence matrix group (GLCM), seven features from the gray-level dependence matrix group (GLDM), ten features from the gray-level run-length matrix group (GLRLM), four features from the gray-level size-zone matrix group (GLSZM), and one feature from the neighboring gray-tone difference matrix group (NGTDM) were found to be significantly different between the two groups. Conclusion: According to the results, the dosiomics features can distinguish the differences between dose distributions in different studied plans. However, more studies should be done in selecting the most suitable features for use in evaluating the quality of the treatment plans.