How about using split with a third argument limiting the resultant fields to two, splitting on the first comma and space that is followed by a digit. It would of course fail if the item itself contained <comma><space><digit>. Here, I printf'ed both fields so you can see where the demarcation lies.

use strict;
use warnings;

while( <DATA> )
{
    chomp;
    my( $item, $index ) = split m{,\s(?=\d)}, $_, 2;
    printf qq{%-42s%s\n}, $item, $index;
}

__END__
sclerosing 1954, 5-7, 54, 59f-60d, 90, 114
cribriform carcinoma, invasive, 89, 91-94, 112
comedo-type DCIS, 25-26
comedo-type necrosis, LCIS with, 55, 59, 65
complex sclerosing lesions (radial scar), 8-9, 54, 59, 90
[download]

Here is the output.

sclerosing 1954                           5-7, 54, 59f-60d, 90, 114
cribriform carcinoma, invasive            89, 91-94, 112
comedo-type DCIS                          25-26
comedo-type necrosis, LCIS with           55, 59, 65
complex sclerosing lesions (radial scar)  8-9, 54, 59, 90
[download]

I hope this is helpful.

Cheers,

JohnGG