This is one of those questions which allows for three kinds of answer: a rather short, straight to the point one, which leaves almost the entire context out (and which I am giving you – feel free to ask if some points are obscure); a quite long one, which I am working on and may come together one day; and one consisting on a few volumes. It is also – to my knowledge – one that has not been addressed directly in any of the sources I have available (that is, a first check didn't produce anything related mainly to the Italian presence and military occupation in Anatolia). Therefore I have put this together relying for the most part on bits of different references, often just to be sure about a date and proper collocation of an event, which goes to explain the very long list for a very short answer.

 

Now, to answer your exact question, Italy didn't really return the southern portion of Anatolia to Turkey, since Italy had never properly acquired the southern portion of Anatolia, nor taken possession of it outside of a manifestly temporary occupation. Of course, the Italians were, at least in their Eastern Mediterranean policy, often over-reliant on the principle that “possession constitutes nine tenths of the law” - this had been, after all, the foundation of their persistent occupation of Dodecanese – so that one could legitimately suspect the series of small landings operated during May and June 1919 to entail a more permanent design. The first Italian landing, in Adalia, actually occurred on March 28^th – 300 men 3.00 pm (Comm. Elia to S. Sonnino, March 29^th 1919) – but the Italian status in Adalia was somewhat different, given previous obligations (of controversial recognition) between the Ottoman Empire and Italy; so that the most significant landings, both for their numerical entity and for their diplomatic impact were those of May.

These initiatives, taken at the peak of the diplomatic crisis between Italy and the Allies and Associate in Paris, carried off piecemeal before the execution of the Greek landing in Smyrna but after the Italians, at the time with their main delegation absent from the Peace Conference in protestation for Wilson's declaration of April 23^rd 1919, had received news of the authorization given to Venizelos by the Three Greats, were certainly intended to safeguard and substantiate with a small military presence the established Italian interests and claims in the region. Italy had after all three substantial reasons of conflict with Greece: the status of the Corfu Channel (of more significance for the Italians than the related matter of Southern Albania or Northern Epiros), the occupation of Dodecanese (the Rhodes Archipelago) and the Italian claim to Smyrna, which had been included in the agreements of St Jean de Maurienne (April 26^th 1917), but which the Allies regarded as dubious. The Italian landings occurred nonetheless still under armistice regime (between the Armistice of Mudros of October 30^th 1918 and the Treaty of Sevres of August 10^th 1920), and therefore were regarded as “not prejudicial” for any subsequent decision to be taken by the peace conference.

When the decision came – fifteen months later, with the most unfortunate of the 1919 peace treaties, the “stillborn” one signed in Sevres – most of the Italian troops had left the Turkish mainland already, while the Italian governments, increasingly concerned with internal unrest, forced to reduce the substantial deficit inherited from the war while looking for a very urgent resolution of the matter of Fiume and busy in the final stages of their accommodation with the Yugoslavs (the Treaty of Rapallo was signed on November 12^th 1920 and D'Annunzio driven out of Fiume five weeks later) no longer entertained any delusion of military occupation of Anatolia, and were satisfied with the maintenance of certain, more formal than substantial, perspective economical concessions. As for the upkeep of the extant Italian occupation zone, this was ensured in large part by the Italian policy of non-interference regarding the Turkish military operations (indeed the Italians, allegedly, begun to supply weapons to the Kemalist forces already before the ratification of Sevres entered discussion).

The Italian insistence on economical concessions – an ambiguous one, given the contemporary support for the proposal of lifting the capitulations which had been reintroduced with the armistice – was more of an attempt to salvage something of their diplomatic capital in that “liquidation of war assets” that the Ministry of F.S. Nitti was forced to pursue in light of the difficult internal and financial situation, than a true diplomatic strategy aimed at maintaining a presence in Turkey against the Turkish will. Indeed the size of the Italian expedition corp, which had reached 9,000 men at the end of June 1919 and had been maintained to the same level until March 3^rd 1920 (that is until the Conferences of London and San Remo had begun discussing the terms of the future Treaty of Sevres) had been reduced to 7,000 at June 3^rd 1920 (“under reduction to 4,500”). While this downsizing was in good measure a consequence of the general process of demobilization, neither the size of the Italian expedition corp (which appears to have been reduced to a mere appendix of the Dodecanese command between March and June) nor the internal situation allowed for the deployment of new troops – indeed the Italian forces were already strained by their efforts to maintain some form of occupation in Albania.

When both Turks and Greeks rejected the terms of the Treaty of Sevres, opening the main phase of the Greek-Turkish war, with the French troops pushed back from Cilicia, the Italians appear to have kept to their own for the most part, as they couldn't really do much else. In fairness, at least a part of the Italian establishment – for once – had seen the writing on the wall before hitting their noses on it. Italy's new Foreign Minister, the Count Carlo Sforza, who had replaced the more conservative Tommaso Tittoni and Vittorio Scialoja (after the definition of the Treaty of Sevres but before its signature), had been one of the first to support the opening of (informal) relations with the Kemalist fraction, already during his tenure as Italian Commissary to the “Sublime Porte” (which is to say the Ottoman Government in Constantinople) and to take note of the deep fragmentation of the Allied front on the issue of the Ottoman settlement. The signature of the Treaty of Sevres was therefore more of a formal necessity for the Italians (who indeed didn't even bother sending a high functionary, with the Ambassador to Paris, L. Bonin signing for His Majesty the King of Italy). The concurrent renunciation to the previous Tittoni-Venizelos agreement in July 1920 (and the signature of the annex convention Bonin-Venizelos) stood to represent the substantial disinterest of Italy for the supposedly inter-Allied (but in substance reduced to a Greek-British collaboration) initiative against the former Ottoman State and, de facto, Italy's reliance on a policy of cautious support of the Turks in order to gain at least some form of economical privilege over the region – or, as eventually happened with the Treaty of Lausanne of 1923, recognition for her possession of Dodecanese.