Select text between two patterns with sed, when the patterns are repeated

I would like to select the text between two patterns, when some of the patterns can be repeated as follows.

Here is the input:

Blalala
PAT1
'Omfoem From 
balanf PAT1 This is the
text that I want
to get PAT2: apples
Whatever: oranges

Here is the output I would like:

This is the
text that I want
to get

I have tried using this script (I am using OSX):

gsed -e 's/PAT1/\nPAT1\n/' -e 's/PAT2/\nPAT2\n/' file1.txt | sed -n '/PAT1/,/PAT2/{//!p;}'

But it outputs the following:

'Omfoem From 
balanf 
PAT1
This is the
text that I want
to get 

In this particular case, I could apply sed -n '/PAT1/,/PAT2/{//!p;}' a second time and I would get the right output.

However, I am looking for a script that works in different situations: if any of the patterns is repeated, independently of the number of times, and if the pattern are not repeated at all.

In any of these cases, I would like to be able to extract the text between pattern PAT1 and PAT2 when PAT1 and PAT2 are the closest to each other.