Hi soupman,
I don't consider myself a scholar, but I'd like to offer a response anyway, if that's OK.

In the well-regarded Paderewski Edition, that note is indicated as a D# being governed by the key signature still in effect and the absence of any accidental in this particular edition. (I play the D#.) The editorial board (Paderewski, Bronarski and Turczynski) primarily studied the composer's manuscripts, the original French edition (H. Lemoine) and the original German edition (F. Kistner). They also consulted the editions by Mikuli (Kistner), Klindworth (Bote & Bock), Scholtz (Edition Peters), Bulow (Universal Edition), Riemann (Steingraber), Pugno (Universal Edition), Brugnoli (Ricordi) and the venerable Urtextausgabe (Breitkopf & Hartel). So they looked at many "votes" and evidently the D# prevailed. They definitely saw the D natural option in the original French edition (more on that in a moment) and very likely must have seen as well the D flat option in Mikuli, and if so, apparently were not swayed by it, despite Mikuli's having been a student of Chopin (as was Klindworth).
Here is their commentary on the note based primarily on the manuscript, the German edition and structural analysis of the passage in the score in order to justify the probable correct note being a D#:
"Before the last semiquaver of this bar the French edition puts a natural, which is not in the manuscript or German edition. Although this version produces a smoother link with the figure in the next bar, it conflicts with the structural principle of this figure, in which the penultimate note (which is the same as the first) and the last note in each group of six semiquavers are harmonic notes."
So in other words, changing the note to a D natural (or the D flat you mention) would have made the note melodic, thus placing it out of context with the same last note--which serves a harmonic function--in the previous group of 16th notes, and, as well, out of context with all the other similar measures for the very same reason. While the voice leading value of a D flat is intriguing, there are also the same precedents in the score to consider in reaching for a justification for change.
One of the problems with these variances, which run rampant in Chopin editions, is that unlike Liszt, for example, who was very meticulous in examining his publisher's proofs, Chopin, obversely, was most often casual, cursory and careless about it. Thus there are many errors and unauthorized changes. Look at the debate for generations over the E or E flat in the third measure of Chopin's Prelude Op. 28, No. 20 in Cm! The second issue is the value of penciled changes of notes by Chopin in Mikuli's (or Klindworth's) scores entered during lessons. The unanswerable questions are: 1) where these legitimate corrections of misprinted notes in the score (due to Chopin's inattentiveness in reviewing, correcting, and approving publishers' proofs)?, or 2) was Chopin actually "re-composing" in the moment and on the spot, as he simply discovered a sound he liked even better, yet did not formally notify his publisher of the revision(s) for future reprintings? I imagine it's a difficult conundrum to unravel, even for musicologists.
I hope this is helpful.