Cancer is a worldwide problem and one of the leading causes of death. Increasing prevalence of cancer, particularly in developing countries, demands better understandings of the effectiveness and adverse consequences of different cancer treatment regimes in real patient populations. Current understandings of cancer treatment toxicities are often derived from either “clean” patient cohorts or coarse population statistics. Thus, it is difficult to get up-to-date and local assessments of treatment toxicities for specific cancer centers. To address these problems, we propose a novel and efficient method for discovering toxicity progression patterns in the form of temporal association rules (TARs). A temporal association rule is defined as a rule where the diagnosis codes in the right hand side (e.g., a combination of toxicities/complications) are temporally occurred after the diagnosis codes in the left hand side (e.g., a particular type of cancer treatment). Our method develops a lattice structure to efficiently discover TARs. More specifically, the lattice structure is first constructed to store all frequent diagnosis codes in the dataset. It is then traversed using the paternity relations among nodes to generate TARs. Our extensive experiments show the effectiveness of the proposed method in discovering major toxicity patterns in comparison with the temporal comorbidity analysis. In addition, our method significantly outperforms existing methods for mining TARs in terms of runtime.
- Cancer treatment
- Data mining
- Pairwise association analysis
- Temporal association rules