RESPONSES IN CAPS

Sorry, this is not my field but I'm just trying to get a general idea of how Kaplan Meier works. I'm reading a paper that reports 5-year OS, DSS, etc for a retrospective study, but it states that the median follow up period was 16 months. Can someone explain in simple terms how:

1. you can know the 5 year OS if the median follow up is way less than 5 years? K-M PROVIDES THE SURVIVAL CURVE WHICH AT EACH POINT T ALONG THE X-AXIS REPRESENTS THE PROBABILITY OF SURVIVING TO TIME T. THE MEDIAN FOLLOWUP APPEARS TO BE A COMBINATION OF TIMES TO THE EVENTS (I.E. WE KNOW THE TIME THE PATIENT DIES RELATIVE TO STARTING THE TRIAL) AND THE TRUNCATED OR CENSORED TIMES. TYPICALLY ONE USES THE MEDIAN SURVIVAL TIME TO DESCRIBE THE DATA AND NOT THE MEDIAN FOLLOWUP. CERTAINLY IF AT LEAST ONE SUBJECT IS FOLLOWED FOR AT LEAST 5 YEARS, YOU CAN GET AN ESTIMATE OF SURVIVAL AT 5 YEARS.

2. If, and how, a short median follow up time can bias the results? YOU WOULD NEED TO KNOW THE DISTRIBUTION OF CENSORED AND UNCENSORED (REACHED EVENT) DATA. HOWEVER, WHAT IS REALLY IMPORTANT IS THE NUMBER OF SUBJECTS WHO DO REACH THE 5 YEAR POINT RELATIVE TO THE NUMBER YOU STARTED WITH. WITH SMALL NUMBERS WHO REACH 5 YEARS, THE CONFIDENCE INTERVAL AT THAT POINT WOULD BE REALLY LARGE, AND RESULTS MAY NOT BE RELIABLE. 

thanks.