Methodology for Estimating Asset Life from Consecutive Condition Assessments
Table 1
The table below shows example (made up) PCI data for road segments reinspected after 5 years.

The last two points are meant to represent roads that have been reconstructed in between inspections are excluded from the analysis. The remaining points are graphed below.
A quick bit of high school math
Rate of Change of Condition = A * PCI Score + B
You could do a least mean squares fit if you want to, but for this example I have just picked a line by eye.
(It would be nice if real data follows an obvious pattern, but I suspect real data will be equally scattered).
The red line runs through the points (0,2) & (157,7.2)
Which gives the equations:
2 = A * 0 + B  (1)
7.2 = A * 157 + B  (2)
Therefore B = 2 and A = (7.22)/157 = 0.033121019
Which gives the equation: Rate of Change of Condition = 0.033121019 * PCI Score + 2
It follows that : PCI (year n+1) = PCI (year n)*1.03312019 + 2
This yields the following graph
And the above graph not only gives you a way of semiscientifically determining a useful life for road pavements, it also lets you estimate the expected remaining useful life of any road segment you have a condition score for.
Depending on your scoring system and construction standards an assets initial condition may not be zero.
If for example a typical road has a PCI of 10 immediately after construction, and you have adopted a service level that roads should be reconstructed when the PCI is greater than 200, a typical road would have a useful life of 40 years.
For a better result you would probably need to create graphs for a series of road categories and traffic volumes.