H2R, there is good reason why clinical trial survival analysis is done this way, and it is not biased as you think. That is why LL reports 13% alive at 5 years, while the actual number known alive at 5 years was only 10.8%.
Trials will almost always have incomplete data. Not just LTFU, but patients who were still alive at the end of the trial but had not been in the trial long enough to have complete data.
You do now want to assume they died or taht they lived the full 60 months. Nor do you want to throw them away as that creates a time bias where the median (and tail) will increase if you wait.
The K/M curve is nearly universally used to handle this. It provides the best estimate of what the real numbers would be if you had the full data.
The curve is very simple and does not rely on any magic to create. Every month the curve drops by a percent that is the same percent as died out of the then at risk population,.
The assumption you are making in graphing at risk over time is that every patient who was not measured alive at 5 years was dead. Yet not all the trials in question even reported 5 year data.
Market Data 
Markets 
AI
