Aim: to investigate changes in the upper and lower dental arches caused by orthodontic treatment and changes that occured in long-term postretention period. Another aim was to compare the differences in different time periods between dental arches treated with and without extraction as well as to establish predictors of long-term stability after orthodontic treatment with and without extraction. Materials and methods: This retrospective study included 103 patients with Class I and Class II malocclusions treated with fixed appliances in both arches with and without extraction. Inclusion criteria were treated patients with dental and skeletal class I and class II, at least 5 years out of retention with full documentation and study models of satisfiying quality. Exclusion criteria were patients with dental and skeletal class III, orthognathic patients, patients with congenitally missing teeth (except third molars), patients with circumferential supracrestal fiberotomy performed and patients who had prosthetic work done in T3 time period. The sample was collected from a database of a private orthodontic office and the patients were treated by an experienced orthodontist during the 1970s and 1980s with comprehensive orthodontic treatment with fixed appliances in both arches. After orthodontic treatment, the dental arches were retained with a lower bonded retainer from canine to canine and a Wraparound retainer in the upper arch. The retention period lasted for three years. After three years, all retention appliances were removed and interproximal reduction was performed on the lower incisors (the intercanine segment) where this had not been done during treatment. The dental casts were obtained at pretreatment (T1), posttreatment (T2), and long- term postretention period of at least 5 years (T3). The study models were scanned with Ortho Insight 3D scanner (Motion View LLc, Hixon, USA), digitalized and converted to STL files. Linear and angular measurements were evaluated in Mathlab enviroment, a programme specifically constructed for this purpose. The measurments were determined for the upper and lower arch in three time periods as follows: the intercanine width (IK), intermolar width (IM), Little's Irregularity Index (LII), the interdental angle (ID), contact angle (CA) and rotational angle (RA). The measurements were performed by an experienced and calibrated examiner . Thirty days after the initial measurements, 40 dental casts were randomly selected to check for intraexaminer reliability. V Statistical analyses were performed in STATISTICA 64, version 10 for Windows. Results: The study sample comprised 70.9% female and 29.1% male patients. Class I in 71.8% of cases and class II in 28.2%. The average postretention time was 17.2 years (±6.5) with an average retention time of 3.4 (±1.17) years. Extraction was performed in 55 patients while 48 (46.6%) received nonextraction treatment. The Bonferroni Post Hoc test showed that LII in the upper and lower arch in T1 was statistically significantly higher in the extraction group (p<0,001). In T3 in the upper and lower arch LII was slightly higher in extraction cases but remained under 2.05mm. In the upper arch 92.2% of patients and 82.5% in the lower arch showed satisfactory alignment under 3mm. Average LII at T3 in upper arch was 1.69 ± 1.27 mm in extraction and 1.10 ± 0.79 mm in nonextraction cases. Average LII at T3 in the lower arch was 2.05 ± 1.64 mm in extraction and 1.52 ± 1.27 mm in nonextraction cases. Average LII in postretention period (T3) in the upper and lower arch showed a negative correlation with IM T3 in the upper arch (Pearson,N=103, p=0.047), while IK in the upper and lower arch in T3 correlated with IM T3 in the upper and lower arches (N=103, p<0,001). Intercanine width in both arches showed similar changes, it increased during orthodontic treatment period (T1-T2) , and returns to initial values in the posttreatment and postretention period (T2-T3). Females and extraction cases show more posttreatment changes. Intermolar width has been shown as the most stable variable. The angular parameters showed greater values in nonextraction cases, even though this difference hasn't proved statistically significant. The extraction cases showed narrower arches before orthodontic treatment. The postretention changes of interdental and contact angles didn't show any significant differences. In extraction cases, the initial ID of lower canines showed a negative correlation with LII posttreatment change in the lower arch (Pearson, N=48, p=0.02). In nonextraction cases, the initial ID of upper canines showed a negative correlation with LII in the upper arch in T3 (Pearson, N=47, p=0.045). In the extraction cases treatment change of the interdental angle of lower canines showed positive correlation with lower LII (Pearson, N=48, p=0.01). In the lower arch rotational angles in the postretention period showed a tendency to return to initial values. This was not registered in the upper arch. The long-term rotational changes of upper and lower anterior teeth were under 3 degrees and clinically insignificant. Conclusion: Clinically relevant long-term stability was found in extraction and nonextraction cases in both arches twenty years after orthodontic treatment with Little's Irregularity Index value under 3mm. The lower arch, both in extraction and nonextraction cases, showed greater instability. VI In the postretention period, intercanine width in extraction cases showed a tendency to return to initial values, while in nonextraction cases intercanine width showed better stability. In females treated with extractions, intercanine width tended to return to initial values, while males showed stable results. Intermolar width changes during orthodontic treatment were bigger that postretention changes of intermolar changes. Intermolar width in nonextraction cases showed to be the most stable linear variable. After orthodontic treatment, the smallest contact angle in the lower arch was measured between canines and lateral incisors and, in the upper arch, smallest contact angle was measured between central incisors. The initial arch form in extraction cases was tapered, while in nonextraction cases it was more ovoid. Rotational angles in the lower arch showed a tendency to return towards initial values, while this was not shown in the upper arch. Postretention changes of rotational angles were less than three degrees and can be considered clinically insignificant. Long- term postretentional changes of dental arches were inevitable, but in a clinical sense those changes are insignificant. Intermolar width and its change during orthodontic treatment can be a good predictor of stability in extraction cases.