SQL & PL/SQL :: Using Correlated Values In Nested Subquery

select rtrim(xmlagg(xmlelement(e, table_name||',')).extract('//text()'),',')
from (
select distinct table_name from tr_products tr join all_instruments v
on tr.PROD_COA_ID = v.PROD_COA_ID
where tr.LEVEL_06_VALUE like t1.COLUMN_VALUE||'%'
or tr.LEVEL_05_VALUE like t1.COLUMN_VALUE||'%'
or tr.LEVEL_04_VALUE like t1.COLUMN_VALUE||'%'
or tr.LEVEL_03_VALUE like t1.COLUMN_VALUE||'%'
or tr.LEVEL_02_VALUE like t1.COLUMN_VALUE||'%'
or tr.LEVEL_01_VALUE like t1.COLUMN_VALUE||'%')

This is an excerpt of a code of a huge query, I didn't want to write it all here because there is no point in doing that since this is the part of the query that I'm actually having problem with, and I think you will be able to understand my question just by looking at this piece of code.

Anyway, this here is actually a subquery in a select-clause of a main query, and apparently it has a (nested) subquery of its own (as seen in the code). I want to reach some values ("COLUMN_VALUE") that are correlated to the parent query (t1 is alias of a table in the from-clause of the main query), and as you can see, I'm trying to reach it from this level 2 subquery. As I have recently found out, Oracle does not allow this, it only tollerates correlation down to level 1.

Here is a scheme of my query:

select something1,
(select something2 from
(select something3 from some_table2 t2
where t1.value = t2.value))
from some_table t1

can this be overridden somehow, is there a workaround for this limitation?