PostgreSQL POSIX regular expression problem

* Trying to build a regular expression syntax that doesn't drop the last digit (5) at the very end of the phone number.

testdb=# select substring('OfficeName: abc office supply PhoneNumber: [commercial] 865-999-5695 ' from 'PhoneNumber: \[commercial](.*?)(,|(\d\s+))');
substring
--------------
865-999-569
(1 row)

* Removing the space at the end returns empty, but I need the complete phone number: 865-999-5695

testdb=# select substring('OfficeName: abc office supply PhoneNumber: [commercial] 865-999-5695' from 'PhoneNumber: \[commercial](.*?)(,|(\d\s+))');
substring
-----------
(1 row)

* These three examples work fine using a slight variation of the regex clause.

testdb=# select substring('OfficeName: abc office supply PhoneNumber: [commercial] 865-999-5695 abc' from 'PhoneNumber: \[commercial](.*?)(,|(\s(\w+)$))');
substring
---------------
865-999-5695
(1 row)

testdb=# select substring('OfficeName: abc office supply PhoneNumber: [commercial] 865-999-5695, abc' from 'PhoneNumber: \[commercial](.*?)(,|(\s(\w+)$))');
substring
---------------
865-999-5695
(1 row)

testdb=# select substring('OfficeName: abc office supply PhoneNumber: [commercial] 865-999-5695,' from 'PhoneNumber: \[commercial](.*?)(,|(\s(\w+)$))');
substring
---------------
865-999-5695
(1 row)

* but if I remove the last word and the comma. again I get null, when I need the phone number. The phone number may be the last value at the end of the string or it may not be.

testdb=# select substring('OfficeName: abc office supply PhoneNumber: [commercial] 865-999-5695' from 'PhoneNumber: \[commercial](.*?)(,|(\s(\w+)$))');
substring
-----------

(1 row)

thanks

Fred