I have just finished a refresher course in statistics, and am now moving onto more serious analytics tools. I'm looking forward to this practical application of data science that utilizes my domain knowledge and experience in medicine, as well as my work in producing medical media.
Here's the task:
 |
| 4-chamber Cine MRI |
"The 2015 Data Science Bowl challenges you to create an algorithm to automatically measure end-systolic and end-diastolic volumes in cardiac MRIs."
The purpose of these measurements is to determine cardiac output (volume of blood pumped per minute) by way of the ejection fraction (EF). "EF is the percentage of blood ejected from the left ventricle with each heartbeat." The end-systolic and end-diastolic volumes are used to derive the EF, which quantifies the heart's ability to pump blood.
Cardiac Output and Ejection Fraction
When the competition was started yesterday morning, what immediately came to mind was my experience as a surgery resident. Patients were scheduled for elective procedures and it was my job to ensure that they were prep'ed and cleared for surgery.
A patient's primary care physician might note that he or she had a cardiac history that needed to be addressed given the stresses of general anesthesia. EF was necessary to be determined.
This required a cardiac consult, which would involve a MUGA (multigated acquistion) scan (now MRI video is used), for the cardiologist to determine the EF. It was the number the anesthesiologist would look for on the chart before the patient was brought to the OR.
The Cleveland Clinic Web site provides an excellent discussion of the role of ejection fraction in determining cardiac health.
Frank-Starling Mechanism and a rubber band (a closer look why cardiac output and ejection fraction are important)
I remember reading my physiology text in medical school seeing this rubber band analogy to explain the Frank-Starling mechanism for heart function.
The heart muscle, specifically the left ventricle that has the major task of pumping throughout the entire body, stretches as venous blood is returned to the heart. Think of this as a rubber band stretching. Then the muscle contracts. This is the rubber band being allowed to apply a force as it contracts.
In this way, the heart can produce more forceful contractions if more blood is being returned from its trip through the body's circulatory system.
Systole is when the heart chamber (in this case we're looking at the left ventricle) is completely contracted, and diastole is when the the chamber is stretched and about to apply the force of contraction. As you can see in the photos, diastole 1 is moderate filling of the chamber, and diastole 2 is a greater filling. With diastole 2 in this example, you see how the rubber band (representing the heart muscle fibers) is stretched more and will apply a greater force when it contracts then with diastole 1. In the heart, the more the filling, the greater to force of contraction and therefore greater cardiac output.
"It's the heart's intrinsic way of synchronizing cardiac output with an increased venous return," says the author of this YouTube video.
With a diseased heart, this mechanism breaks down and the blood returning to the heart is not being adequately pumped to supply the oxygen needs of the body.
Good luck to all those teams participating. I'll be participating under the team name "Heartland Data."
