Something like this?

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11116556
use warnings;
use Path::Tiny;

my $dictionary = path( '/usr/share/dict/words' );
my @tiles  = qw( a a d r t s c );
my $string = 'f   e  ';
my $pat = join '', map "$_?", sort @tiles, $string =~ /\w/g;
my $tilepat = join '', map "$_?", sort @tiles;
my $letters = join '', @tiles, $string =~ /\w/g;
print "$pat\n$tilepat\n";

my @matches =
  grep { ( join '', sort split // ) =~ /^$pat$/ }
  $dictionary->slurp =~ /^[$letters]{2,}$/gm;

my @places;
($string =~ tr/ /./r) =~
  /(?<!\w).{2,}(?!\w)(?{ push @places, $& })(*FAIL)/;
@places = grep /\w/, @places;
use Data::Dump 'dd'; dd \@places;

my @matches1 = extension(1);
print "@matches1\n\n";

my @matches2 = extension(2);
print "@matches2\n\n";

sub extension
  {
  my $id = shift;
  my @found;
  for my $placepat ( $id == 1 ? @places : map expand($_), @places )
    {
    for my $match ( @matches )
      {
      $match =~ /^$placepat$/ or next;
      my $newtiles = $match & ($placepat =~ tr/.a-z/\xff\0/r);
      (join '', sort $newtiles =~ /\w/g) =~ /^$tilepat$/ and
        push @found, $match;
      }
    }
  return @found;
  }

sub expand
  {
  grep /\w/, glob join '', map { /\w/ ? "{$_,.}" : $_ } split //, shif
+t;
  }
[download]