QCD factorization for the DIS structure functions at small x

We suggest a new form of QCD factorization for inclusive processes at high energies and then reduce it to K T -factorization. In turn, K T -factorization can be reduced to collinear factorization when the unintegrated parton distributions have at least one maximum in k ⊥. This property of the parton distributions can be checked with studying data of experiment. The sharper the maximum is, the more accurate the transition to collinear factorization can be done. We apply our results to deduce theoretical restrictions requirements on fits for parton distributions in both K T - and collinear factorizations contrary to the present situation where the fits are introduced from purely phenomenological consideration so that any formula for them is acceptable if it matches experimental data.