What we can do is to ignore the last t digits of the numbers, then to apply the pseudo-polynomial time algorithm, and then to return the corresponding solution in the original instance.
By simple computation you can prove two things.
The first thing you can prove is that if you first cut t digits and then you return the corresponding solution in the original instance, then you do not loose so much. Actually, you can prove that for any performance ratio you want to reach you can compute the number of digits you have to cut so that the return solution has the decided performance ratio.
The second think you can prove is that the time complexity of this algorithm is polynomial in the and in the performance ratio you want to obtain.
