We are designing a decision system which will act as an assistant to doctors in the treatment of lung cancer patients. This Bayesian Belief Network, or BBN, will behave similarly to the Lung Cancer Assistant (LCA) created in the UK by compiling information from medical datasets and then graphically demonstrating causal relationships between the various variables. However, what is common to almost all medical datasets are the occurrence of missing data which will disturb the system. There are many tools available for handling missing data including multiple imputations (MI) and missingness indicators. The issue with these methods is that they are only viable for certain types of missing data and then inefficient for others. These types of missing data are related to the reason the data is absent: Missing at Random (MAR), Not Missing at Random (NMAR), and Missing Completely at Random (MCAR). When examining medical data, most of the missing values fall into MAR, NMAR, or on a spectrum of the two. Because of this variance in missing data within the medical datasets and the fact that sometimes data is missing from a dataset for unknown reasons, the focus of this research is the use of graphical models for processing the missing data prevalent in the lung cancer datasets so that the BBN can then be applied. The missingness graphs typically employ Directed Acyclic Graphs (DAGs) to depict the causal relationships and conditional dependencies. The models will also indicate the category the missing data falls into based on the graphical structure while being applicable to all types. By solving the missing data problem, the BBN can be fully utilized and will have improved accuracy in the treatment of lung cancer patients. This can then reduce time in diagnosis, reduce mortality, improve early detection of cancer, and especially benefit communities of limited health care resources.