Exploiting Machine Learning for Predicting Nodal Status in Prostate Cancer Patients

Prostate cancer is the second cause of cancer in males. The
prophylactic pelvic irradiation is usually needed for treating prostate
cancer patients with Subclinical Nodal Metestases. Currently, the physi-
cians decide when to deliver pelvic irradiation in nodal negative patients
mainly by using the Roach formula, which gives an approximate estima-
tion of the risk of Subclinical Nodal Metestases.
In this paper we study the exploitation of Machine Learning techniques
for training models, based on several pre-treatment parameters, that
can be used for predicting the nodal status of prostate cancer patients.
An experimental retrospective analysis, conducted on the largest Italian
database of prostate cancer patients treated with radical External Beam
Radiation Therapy, shows that the proposed approaches can effectively
predict the nodal status of patients.