Please click here to view the model.
Both concurrent chemoradiotherapy (CCRT) and induction chemotherapy (ICT) followed by CCRT are recommended for advanced nasopharyngeal carcinoma (NPC). We aimed to develop a radiogenomics-based clinical decision support system (CDSS) to inform treatment decisions. First, a radiomics-based fusion nomogram (RFNCCRT) for prognostic prediction of patients receiving CCRT was constructed using machine learning based on magnetic resonance imaging and clinical features associated with disease-free survival. Second, a reported six-gene signature (GSICT) was used to predict ICT benefits in low- and high-risk patients stratified by the RFNCCRT. Ultimately, a radiogenomics-based CDSS was established by categorizing patients into treatment-specific groups according to their predicted outcomes with CCRT and ICT benefits. The manuscript of this study has been submitted to the Lancet Digital Health, titled“Development and validation of a radiogenomics-based clinical decision support system for advanced nasopharyngeal carcinoma: a multicenter retrospective study”.
The browser-based tool for a radiogenomics-based CDSS for advanced NPC
Figure 1. The radiomics-based fusion nomogram (RFNCCRT). The RFNCCRT predicts the probability that a patient will not experience disease progression within 3 and 5 years after receiving CCRT. For example, for a patient who is newly diagnosed as stage III (T3N2M0) NPC, with the retropharyngeal lymph node (RLN) volume < 3mL and a radiomic siganture score of -0·5, the total point is about 32·5 (0+12·5+0+20) and the corresponding 3-year and 5-year DFS rate are about 89%, 85%, respectively, when receiving CCRT. NPC, nasopharyngeal carcinoma.
Figure 2. The schematic diagram of the radiogenomics-based CDSS. NPC, nasopharyngeal carcinoma; RFNCCRT, radiomics-based fusion nomogram for predicting prognosis with CCRT; GSICT, 6-gene signature for predicting ICT benefit; CCRT, concurrent chemoradiotherapy; ICT, induction chemotherapy.