GRAD 2008 Poster

Intro ] [ GRAD2008 Pray Ruiz Poster.ppt ] [ GRAD 2008 Abstract.doc ] [ GRAD 2008 Abstract ]

Up: Presentations ]

In hopes of improving upon the recent work at WPI in predicting
survival and quality of life of pancreatic cancer patients we are
conducting a study of various state of the art machine learning
algorithms, meta-learning methods and dimensionality reduction
techniques. Through a thorough understanding of how machine learning,
meta-learners and dimensionality reduction function and interact we
aim to expertly apply and possibly improve them to produce superior
models capable of prediction accuracy unattained to date. 

Surgical resection of tumors is an important decision to be
made in the treatment of pancreatic cancer. While this
procedure can increase the five year life expectancy to 40%, up from
4%, it is debilitating and involves a long convalescence. A patient's
quality of life can be severely decreased during what may be the last
months of their lives. Being able to predict survival and quality of
life with a high degree of accuracy will provide much needed
information on which to base the decision and maximize survival and
quality of life for these patients. 

--- OLD ---

Found estimated for 2008:
Deaths:

Lung - 161,840
Colon - 49,960
Breast - 40,930
Pancreas - 34,290

Pancreatic cancer, the fourth leading cause of cancer deaths after
[LIST THE THREE TOP KILLING CANCERS HERE], has a one year survival rate of
19% and a five year survival rate of 4%. With surgical resection of
the tumor the five year survival rate increases to 40% but the 
procedure is often debilitating and involves a long convalescence,
often negating the benefits. Being able to predict survival and quality
of life with a high degree of accuracy will provide much needed
information on which to base treatment decisions to maximize survival
and quality of life for these patients.
 

by: Keith A. Pray
Last Modified: February 24, 2008 4:02 PM
© 2008 - 2007 Keith A. Pray.
All rights reserved.