I am a scientific researcher working on lung cancer. I am trying to develop a machine learning method able to process just few early detection features and then predict if a patient with lung cancer will develop metastasis (or not).
In my current statistical model, I am using the values of the following features as input:
T (tumor size) feature of the TNM staging system
N (presence of cancer in the lymph nodes) feature of the TNM staging system
From the values of these features related to a patient, my machine learning method should predict if he/she will have metastasis or not.
Before proceeding with this project, I need to know if my hypothesis is correct.
1) Can the T (tumor size) feature and the N (presence of cancer in the lymph node) feature be detected by doctors and their biomedical machines on the first visit?
2) My method would be useful in case the tumor size, presence of cancer in the lymph nodes, and age of a patient were available, when metastasis has not been detected (yet). Is this scenario realistic?