Even if your disc file is exactly what you expect, you must fix your regex to match and remove the newlines from the unwanted lines.

use strict;
use warnings;
my $file = \do{my $disk_file =
     "Inventorier\n"
    ."Trade receivables\n" 
    ."Assets\n"
    ."This item includes impairment losses.\n"
    ;
};


open (IN,'<:encoding(UTF-8)', $file) or die "Could not open '$file'$!"
+;
*OUT = *stdout;


    while (my $text = <IN>) {

    #$text =~ s/^[a-z].*[a-z]$//gmi;
    $text =~ s/^[a-z].*[a-z]\n$//i;

      print OUT $text;

    }

  close (IN) or die "Could not close input file: '$file' $!";
[download]

Note that I have not made any other changes to your posted code (except to remove the looping and the output files.) Even though this should 'work', I prefer LanX's first suggestion.