“…Then there exist α ∈ (0, 1] and constants r * ∈ (0, 1], C > 0, such that the following holds. If u ∈ L p (Q − 1 ) satisfies (1.1) with the associatedω defined in (1.2), then we have: i) Pointwise BMO estimate iii) Pointwise control on the solution Ifω is Dini, thenÑ(u, ·) is Dini, and there exists a caloric polynomial P 0 (i.e., a solution of (P 0 ) t = ∆P 0 ) of degree less than or equal to two in space and of degree less than or equal to one in time, such that for every r ∈ (0, r * ] there holds Remark 1.5 Notice that our definition ofω(r) differs from the analogue given in [13], not only because we consider here the parabolic problem instead of the elliptic one, but also because there is no supremum in this new definition. From that point of view, estimate (1.5) is finer than the one given in [13], and than the ones that can be found in the classical literature.…”