Comparison of the Predictability of Dentoalveolar Expansion with Different Aligner Systems in Orthodontics: A Longitudinal Clinical Study in Adult Patients

Abstract

Aligners represent a therapeutic option in orthodontics, offering advantages such as aesthetics, comfort, and individualized prescriptions for each patient. However, the predictability of maxillary expansion is subject to variability. The objective of this article is to evaluate the efficacy and predictability of aligners in maxillary expansion. One hundred adult patients were included in this study, divided into four groups, each assigned to a different brand of clear aligners: Angel Aligners, Invisalign, Spark, or Hey Smile. Digital models were obtained at three stages: initial STL (T1), prediction (T2), and final (T3) (before the first refinement). The models were measured to obtain linear distances between canines, first bicuspids, second bicuspids, and first molars. Statistical analysis was performed using SPSS 28.0. The best predictability was obtained in the lower arch (68.900%) for second bicuspids, while the worst accuracy was for canines (39.290% in the upper arch using Invisalign). Angel aligner showed the highest percentage of predictability (60.002%) among the evaluated brands, followed by Hey Smile (59.895%), Spark (59.275%), and Invisalign (57.153%). The results show that clear aligners are an effective treatment for transverse movements in both arches. However, further research is needed to improve predictability.

1. Introduction

The field of orthodontics has undergone a significant transformation since its beginning. The introduction of fixed mutibracket appliances during the 20th century was essential in standardizing orthodontic practice. The appearance of stainless-steel braces in the 1950s offered durability, flexibility, and a high level of customization. Moreover, they allowed for precise force application for optimal alignment [1]. In the early 2000s, clear aligners were introduced, providing an aesthetic and comfortable alternative to conventional fixed appliances. This system incorporates a wide range of mechanisms, methods, and materials, enabling the management of various types of malocclusions [2,3].

In 1997, Zia Chisti and Kelsey Wirth, two MBA students, founded Align Technology in Palo Alto, California, and launched Invisalign in the market in 1999. This system is based on 3D computerized imaging technology that is used to design treatment plans. Initially, clear aligners were used only for simple tooth movements, but as new materials, attachments, and more efficient staging protocols were developed, this system became applicable to all types of malocclusions. It is not used only for orthodontic cases but also for orthopedic and surgical applications [4,5,6,7,8].

Clear aligner brands include online platforms used to plan virtual treatment. These platforms provide tools that facilitate communication between the clinician and the CAD-CAM technician in order to coordinate force application, sequencing, and the expected dental positions for the design and manufacture of clear aligners [9]. Complex treatments are still difficult to resolve without auxiliary appliances, mainly due to the low predictability of some tooth movements. Clinical results often differ from those predicted in the initial virtual setup. This leads to a loss of time and the necessity of multiple refinements. Overall predictability ranges from 55% to 72% according to the literature. Particular less predictable movements include torque correction and rotation of rounded teeth [10,11,12,13].

Transverse deficiency is one of most common malocclusions in dentistry and consists of a reduced transverse width of the upper arch caused by skeletal factors, dental factors, or both [14]. The most characteristic signs are a posterior crossbite and crowding in the anterior region [15]. However, to choose an appropriate treatment, a rigorous diagnosis and evaluation of the etiology of the malocclusion—whether dentoalveolar or skeletal—are necessary. If the malocclusion is due to a skeletal factor, rapid maxillary expansion (RME) is required, which can be achieved with a skeletal expander in growing patients, with surgically assisted devices (MARPE or SARPE), or through surgical osseous distraction [14,16,17].

If the transverse deficiency is dentoalveolar or if the treatment plan involves skeletal camouflage in a mildly reduced transverse width in non-growing patients, expansion can be performed using orthodontics [14]. This can be accomplished using various orthodontic appliances, such as fixed multibracket systems, expansion plates, or clear aligners [18,19]. The expansion can occur through two types of tooth movement: bodily translation or buccal crown tipping. While the first method involves displacing both the crown and the root with the aid of rectangular or ellipsoid attachments, its predictability remains low. Studies have shown that most expansion results from crown inclination, which occurs due to the application of force at a point coronal to the center of resistance, generating a third-order movement (buccal tipping of the crown with limited root displacement) [20]. The accuracy of expansion with clear aligners ranges from 41% to 80%: 78.2% for the upper arch and 87.7% for the lower arch according to Houle at al. or 40.5% according to Kravitz et al. [21,22,23,24]. Ali S et al. do not recommend planning an expansion wider than 2–3 mm per side in order to minimize periodontal risk [25]. This is because excessive buccal tipping can have negative biological effects, such as bringing the roots too close to the cortical bone or causing gingival recession when certain limits are exceeded.

Digitalization has brought significant advances to dentistry, particularly in orthodontics. Many studies has confirmed that the accuracy of measurements on digital models is superior when compared to traditional techniques, which include the use of rulers or digital calipers [26,27]. Advantages of digital methods include the ease of storage (in addition to eliminating the physical space required for models, they also avoid problems such as fractures and material deterioration). Some authors favor the tools provided by clear aligner companies, as it has been shown that measurements taken between reference points on occlusal surfaces are more accurate using centroids or the gingival margin, as in software such as MeshLab or Meazure [28,29]. Galluccio et al. used Exocad (DentalCAD 2.4 Plovdiv) to measure STL models and analyze predictability in expansion movements [14]. In their study, Alvarado et al. concluded that Nemocast 3D facilitates diagnosis and the measurement of models [27]. Solano et al. also considered this software for measuring expansion predictability with aligners in their study [30].

The aim of this study was to compare the efficacy and predictability of maxillary and mandibular expansion with different aligner systems. The null hypothesis is that there are no differences between the results predicted by the virtual configuration in treatments involving dentoalveolar expansion with clear aligners using different systems.

2. Materials and Methods

2.1. Sample Selection

The initial sample consisted of 100 adult patients (60 female and 40 male) aged between 21 and 49 years (33.50 ± 1.43 years). Each patient was diagnosed with transverse deficiency requiring dentoalveolar expansion using aligners in both arches. The diagnosis was performed by an experienced orthodontist. The patients were evenly divided into four homogeneous groups of 25 patients, each treated with a different aligner brand: Angel Aligners (Angelalign Techonology Inc., Shanghai, China), Invisalign (Align Technology Inc., San Jose, CA, USA), Spark (Ormco Corporation, Orange, CA, USA), or HeySmile (Straumann Group, Basel, Switzerland). Thus, each group contributed to a quarter of the total sample. Group allocation was determined according to patient preference.

The study design was approved by the UCAM Bioethical Committee (CE052406) in accordance with ethical standards and the principles outlined in the Declaration of Helsinki. All patients were informed about the study procedures and provided written informed consent prior to participation.

The patient inclusion criteria were as follows: complete permanent dentition, adult patients, no use of auxiliary devices, requirement of transversal expansion in both arches, complete patient records, treatment requiring at least 15 aligners prior to any refinement, and therapy completed with adequate patient compliance (20–22 h per day).

The exclusion criteria included severe crowding (tooth size–arch length discrepancy greater than −5 mm), patients with periodontal disease, and cases requiring extractions.

There was baseline comparability among the four groups. Patients with skeletal Class I or mild skeletal Class II and III malocclusion (ANB between 0° and 4°) were selected.

2.2. Sample Size

For the sample size calculation, a specific formula for hypothesis testing related to effect size was used. A confidence level of 95% (α = 0.05), a statistical power of 80% (1 − β = 0.80), and an estimated effect size of d = 0.5 (moderate effect) were used. As a result, it was determined that a sample size of n = 22 per group is enough to detect statistically significant differences between paired measurements.

2.3. Study Design

For each patient, maxillary and mandibular models of the initial malocclusion (pretreatment model, T1) were acquired in STL format using an iTero Element 5D digital scanner (Align Technology, San José, CA, USA). Dental scanning was performed by a single trained operator following the manufacturer’s recommendations to ensure consistency. After sending the STL files to the selected brand for each patient, digital treatment planning was carried out using the appropriate software setup: iOrtho (5.2), Clincheck (6.0 Pro), Approver (R13.2), or HeySmile. After the planning reference model (T2) was acquired, it was used to calculate the prescribed moments by measuring the differences between the pretreatment model and the reference model.

Once the aligners were received, the necessary composite attachments were bonded using the templates provided by each company. The bonding protocol for the attachments was standardized across all cases to ensure reproducible and accurate placement, thereby optimizing aligner retention and the expression of the planned tooth movements. Both arches were bonded in the same appointment. Prophylaxis of the dental surfaces was performed, and the enamel was conditioned with 37% orthophosphoric acid etching gel (Ultra Etch IndiSpense, Ultradent Products Inc., South Jordan, UT, USA) for 15 s, followed by rinsing and drying. Cotton rolls were placed to control moisture, and Transbond Plus Self Etching Primer (3M Unitek, Monrovia, CA, USA) was applied. The attachment template was filled with Tetric EvoCeram Aligner nanohybrid composite resin (Ivoclar Vivant, Schaan, Liechtenstein) and was light-cured for 30 s per attachment.

Once the first set of aligners was finished, the attachments were removed, and a new dental scan of both the maxillary and mandibular arches was performed by the same operator who conducted the initial scans. A third STL file was obtained (pre-finishing model, T3), which was used to calculate the achieved movements by measuring the differences between the reference model and the pre-finishing model. All patients completed this first phase of aligner treatment.

2.4. Analysis of Digital Models

For each patient, maxillary and mandibular models of the initial malocclusion (pretreatment model (T1) at baseline) at the final step of digital planning (reference model, T2, approximately one month later) and at the end of the initially prescribed series of aligners (pre-finishing model (T3), approximately six months later) were acquired in STL format. Thus, a total of 600 models were analyzed. Each model was imported into and analyzed using Exocad software (GmbH, Darmstadt, Germany).

Transversal distances between anatomical points were defined on each arch:

• Inter-canine distance (IC): measured between the tips of the canine cusps;
• Inter-premolar distance (IP1): measured between the buccal cusps of the first premolars;
• Inter-premolar distance (IP2): measured between the buccal cusps of the second premolars;
• Inter-molar distance (IM): measured between the mesiobuccal cusps of the first molars.

Measurements were performed by drawing a straight line in the 3D space of each model. Thus, for each patient, 24 values were obtained: initial width, prescribed width, and achieved width of Ic, Ip1, Ip2, and Im on both upper and lower arches, as shown in Figure 1. The measurements were performed by a blinded, experienced orthodontist. Each STL file was a assigned a number, and the orthodontist who performed the measurements only received the STL files, along with the corresponding numbers. The measurements were recorded in an Excel sheet using these assigned numbers. The data were imported into Excel (Microsoft, Redmond, WA, USA) for further analysis. The measurement method is an established technique employed by Mario, P. et al. [10].

2.5. Analysis of Measurements

After obtaining the values, the following data were analyzed:

• Predicted expansion: The difference between the reference model and the pretreatment model.

  Predicted expansion = |T2 − T1|
• Achieved expansion: The difference between the prefinishing model and the pretreatment model.

  Achieved expansion = |T3 − T1|
• Accuracy: The difference between achieved and predicted results, expressed as an absolute value and percentage (%).

  $$\mathrm{Accuracy}= \left|{\displaystyle \frac{T3-T1}{T2-T1}}\right| \times 100$$

2.6. Statistical Analysis

Statistical analysis was performed using SPSS Statistics v28.0 software (IBM Corp., Armonk, NY, USA). Means and standard deviations were calculated for measurement of each arch, tooth type, and time point (T1, T2, and T3). Normality was assessed using the Shapiro–Wilk test. Since no value showed significant deviation from normality (p > 0.05), parametric statistical methods were applied.

Student’s t test was used to compare predicted measures and achieved measures within each group and to assess differences between groups. One-way ANOVA was used to simultaneously compare the four aligner brands for both the absolute value of the achieved expansion and the percentage of predictability. In all analyses, a significance level of s p = 0.05 (95% confidence interval) was applied. Additionally, clinical predictability was obtained as the percentage of expansion achieved relative to the predicted expansion for each type of tooth and aligner system. If it was statistically significant, a comparison between groups was conducted using the Bonferroni post hoc test.

3. Results

3.1. Characteristics of the Participants

The initial sample consisted of 100 adult patients (60 female and 40 male) aged between 21 and 49 years (mean age: 33.50 ± 1.43 years). Each patient was diagnosed with transverse deficiency requiring dentoalveolar expansion using aligners. The patients were evenly divided into four homogeneous groups, each treated with a different aligner brand (

Table 1. Descriptive statics: age and sex.

				Brand
		Invisalign	Spark	Angel Aligner	Hey Smile
		Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
Age		35.60 ± 7.02	32.52 ± 5.33	33.24 ± 6.12	32.64 ± 5.11
Sex (%)	Male	28.0%	44.0%	44.0%	44.0%
	Female	72.0%	56.0%	56.0%	56.0%

).

3.2. Analysis of the Interclass Correlation Coefficient (ICC)

The degree of agreement must be determined to assess the reliability of the measurements taken by the expert operator. The interclass correlation coefficient (ICC) between repeated measurements (T1 vs. T2) showed high reliability in most measurements (ICC > 0.90) in the premolars and in the upper and lower molars. A good level of agreement was observed in the canines, especially in the lower arch, where the lowest value (0.736) was found (

Table 2. Interclass correlation coefficients.

T1 vs. T2	Interclass Correlation
UPP-13–23	0.845
UPP-14–24	0.877
UPP-15–25	0.906
UPP-16–26	0.952
LOW-33–43	0.736
LOW-34–44	0.835
LOW-35–45	0.899
LOW-36–46	0.903

).

3.3. Analysis of Prescribed and Achieved Movements

Table 3. Predictability of the different aligner brands in relation to tooth movements by tooth group and arch. *: statistically significant differences with p < 0.05. **: with p < 0.01. Different superscript letters in the rows indicate in which groups the significant differences occurred according to Bonferroni post hoc tests. A = p < 0.05 vs. Invisalign. B = p < 0.05 vs. Spark. C = p < 0.05 vs. Angel Aligner. D = p < 0.05 vs. Hey Smile.

Type of Teeth		T2–T1		T3–T1		T3–T2			Clinical Accuracy
		Mean ± SD		Mean ± SD		Mean ± SD		p-Value	%
UPP 13–23	Invisalign	2.80	1.76	1.10 D	1.49	1.70 B	1.30	<0.01 *	39.29
	Spark	2.26	1.84	1.20	1.18	1.06 A	1.08	<0.01 *	53.10
	Angel Aligner	2.66	1.79	1.52	1.19	1.14	1.61	<0.01 *	57.14
	Hey Smile	3.39	1.62	1.84 A	1.69	1.54	0.61	<0.01 *	54.28
UPP 14–24	Invisalign	4.14	2.11	2.48 B	1.85	1.66 C	1.49	<0.01 *	59.90
	Spark	3.02	2.26	1.68 A	1.78	1.34	1.34	<0.01 *	55.63
	Angel Aligner	3.34	2.38	2.09	1.41	1.25 A	1.79	<0.01 *	62.57
	Hey Smile	3.71	1.47	2.24	1.37	1.47	0.55	<0.01 *	60.38
UPP 15–15	Invisalign	3.87	1.94	2.42	1.56	1.46 B	1.42	<0.01 *	62.53
	Spark	2.90	3.22	1.89	1.55	1.01 A	2.76	<0.01 *	65.17
	Angel Aligner	2.65	1.88	1.56 D	1.18	1.09	1.38	<0.01 *	58.87
	Hey Smile	3.96	1.83	2.57 C	1.81	1.39	0.72	<0.01 *	64.90
UPP 16–26	Invisalign	2.85	1.61	1.60	1.45	1.25	1.34	<0.01 **	56.14
	Spark	2.88	2.26	1.58	1.28	1.31 C	1.62	<0.01 **	54.86
	Angel Aligner	2.47	2.05	1.29 D	1.67	1.18 B,D	1.84	<0.01 **	52.23
	Hey Smile	2.94	1.72	1.62 C	1.09	1.32 C	1.31	<0.01 **	55.10
LOW 33–43	Invisalign	2.19	1.88	1.11 C	1.54	1.08	1.22	<0.01 **	50.68
	Spark	1.85	1.91	1.19	1.33	0.66	1.02	<0.01 **	64.32
	Angel Aligner	2.53	2.61	1.46 A	1.16	1.08	2.27	<0.01 **	57.71
	Hey Smile	2.60	2.04	1.43	1.24	1.17	1.07	<0.01 **	55.00
LOW 34–44	Invisalign	2.70	2.27	1.61 D	2.03	1.09	1.01	<0.01 *	59.63
	Spark	3.10	2.68	2.06	2.37	1.0 D	1.29	<0.01 *	66.45
	Angel Aligner	3.14	1.87	2.12	1.50	1.01	1.37	<0.01 *	67.52
	Hey Smile	3.70	2.21	2.41 A	1.85	1.2 B	0.80	<0.01 *	65.14
LOW 35–45	Invisalign	3.16	2.58	2.08	1.97	1.09	1.52	<0.01 *	65.82
	Spark	3.01	2.31	1.60 D	1.63	1.41	1.50	<0.01 *	53.16
	Angel Aligner	3.28	2.15	2.26	1.77	1.01 D	1.62	<0.01 *	68.90
	Hey Smile	4.00	1.63	2.64 B	1.10	1.3 C	0.86	<0.01 *	66.00
LOW 36–46	Invisalign	2.53	2.45	1.60	1.77	0.93 D	1.74	<0.01 *	63.24
	Spark	2.91	2.17	1.79 D	1.98	1.12	1.18	<0.01 *	61.51
	Angel Aligner	2.36	2.16	1.30	1.51	1.06	1.43	<0.01 *	55.08
	Hey Smile	3.29	3.12	1.92 B	2.54	1.37 A	0.99	<0.01 *	58.36

shows the means of the predicted (T2–T1) and achieved (T3–T1) expansion and the difference between them (T3–T2). Accuracy was calculated as a percentage.

When analyzing the accuracy among different aligner brands, it was found that the greatest absolute difference between T3 and T1 in the lower arch occurred at the second bicuspids with HeySmile, showing a mean change of 2.64 mm. In the upper arch, the largest change was also observed with HeySmile at the second bicuspids, with a mean of 2.57 mm. Conversely, the smallest changes were recorded with Invisalign, both in the upper and lower arches between canines, with mean changes of 1.10 mm and 1.11 mm, respectively. All the analyzed data showed p-values < 0.05, indicating that these changes were statistically significant. In absolute terms, the changes observed in the upper arch were greater than those in the lower arch.

To assess predictability, digital models of T3 and T2 were compared. All the analyzed data showed p-values < 0.05, indicating that the observed changes were statistically significant. In the upper arch, the highest predictability was found at the first bicuspids with Spark, showing a mean change of 65.17%, followed by HeySmile with mean a change of 64.90%. In lower arch, the greatest changes were between second and first bicuspids in Angel Aligners, with 68.9% and 67.52%, respectively. It can be seen that at time points T2–T1, there are no differences in the Bonferroni test for groups; however, the most relevant differences between pairs of groups are marked with letters (Table 3).

Figure 2 illustrates the predictability of the different aligner brands analyzed across each dental group.

Regarding the differences between dental groups (

Table 4. Predictability across different tooth groups. *: statistically significant differences with p < 0.05.

Type of Teeth	T2–T1	T3–T1	T3–T2		Clinical Accuracy
	Mean ± SD	Mean ± SD	Mean ± SD	p-Value	%
UPP-13–23	2.78 ± 0.29	1.42 ± 0.20	1.36 ± 0.09	<0.01 *	51.08
UPP-14–24	3.55 ± 0.65	2.12 ± 0.32	1.43 ± 0.33	<0.01 *	59.72
UPP-15–25	3.35 ± 0.37	2.11 ± 0.55	1.24 ± 0.18	<0.01 *	62.99
UPP-16–26	2.78 ± 0.39	1.52 ± 0.27	1.26 ± 0.12	<0.01 *	54.68
LOW-33–43	2.29 ± 0.16	1.29 ± 0.24	1.00 ± 0.40	<0.01 *	56.33
LOW-34–44	3.16 ± 0.84	2.05 ± 0.69	1.11 ± 0.15	<0.01 *	64.87
LOW-35–45	3.36 ± 0.69	2.14 ± 0.51	1.22 ± 0.18	<0.01 *	63.69
LOW-36–46	2.77 ± 1.24	1.65 ± 0.81	1.12 ± 0.43	<0.01 *	59.57

), regardless of the brand used, the highest accuracy was observed in the lower arch, with 61.11%, compared to 57.12% in the upper arch. The canines showed an accuracy of 53.5%; the first bicuspids, 62.29%; the second bicuspids, 63.34%; and, finally, the first molars presented an accuracy of 57.12%.

4. Discussion

The objective of this study was to evaluate the efficacy and predictability of expansion movements in treatment with aligners by comparing the digitally planned outcomes with the clinically achieved results. Four groups were analyzed, each corresponding to one of the four leading brands currently on the market: Angel Aligners, Invisalign, Spark, and HeySmile. The sample consisted of 100 patients, with each group comprising 25 individuals.

It was found that the lower arch was more predictable than the upper arch, with an accuracy of 61.16%, compared to 57.00%. Angel Aligner was the most accurate brand overall, while Spark demonstrated the highest consistency, as its values were consistently close to the maximum value and it never recorded the lowest result, with a mean accuracy of 59.28% The least predictable brand was Invisalign, with a mean accuracy of 57.15%. HeySmile showed a mean accuracy of 59.89%.

Kravitz et al. [22] conducted the first clinical study on the predictability of Invisalign in 2009. Their sample comprised 37 patients, and they reported an overall accuracy of 40.5% for anterior buccal expansion, with 36.0% specifically for canines. The same research group repeated the study in 2020 [23] to study if there were any improvements, finding a predictability of 57.6% in the maxillary arch (58.8% for canines, 66.3% for first bicuspids, 60.5% for second bicuspids, and 58,3% for first molars) and 57.6% predictability in the lower arch (67.9% for canines, 61.1% for first bicuspids, 69.7% for second bicuspids, and 53.6% for first molars). The results from their second study are comparable to those of the present investigation, in which we observed an accuracy of 57.12% in the upper arch and 61.11% in the lower arch. In both studies, bicuspids showed the highest predictability among the dental groups.

In 2017, Houle et al. [21] investigated expansion movements and concluded that the mean accuracy was greater at the cusp level than at the gingival level. Thus, expansion movement with clear aligners tends to produce a tipping movement rather than a bodily movement, a finding noted by several authors [14,20,31,32]. Macrí et al. [33] conducted a CBCT-based study and found that although expansion occurred both coronally and apically, the apical expansion was significantly less. In our study, expansion was assessed at the cusp level; however, it would be of interest to compare differences between apical and cusp levels in future investigations. Similarly to our study, there was better accuracy in the lower arch in the study of Houle et al. [21]. Nevertheless, a discrepancy in the dental groups exists because the accuracy in their study was higher in canines, as well as in Rosa Gay et al.’s [29] study. The results show higher values overall (82.98% in the upper arch and 98.5% in the lower arch) compared to the present study. In a study conducted in 2020, Ning Zhou et al. [34] also concluded that it is necessary to establish sufficient buccal torque in posterior teeth according to the level of expansion required.

In contrast to the high accuracy reported by Houle et al. [21], Loberto et al. [17] concluded that accuracy is variable and does not exceed of 50% during the first series of aligners. Therefore, to achieve an accuracy of at least a 70%, refinements are necessary. Ma S. et al. [35] conducted a systematic review and found that expansion movements with clear aligners are not entirely predictable, thus overcorrection should be planned. This has been further corroborated by authors such as Bouchant et al. and Putrino et al. [36,37]. In 2018, Charalampakis et al. [38] reported that there were not statistically significant differences between planned and achieved movements. They conducted a study and observed that the distance between the upper canines showed the highest discrepancy, attributing this to the longer roots and conical crown surfaces of these teeth. This finding is consistent with the results of our study, in which the accuracy for canines was 53.70%, the lowest among the different dental groups analyzed. They also noted that the planned expansion in bicuspids was highly accurate, but the magnitude was limited, with only 1.49 mm achieved for lower bicuspids and 1.76 mm for upper bicuspids. This contrasts with our data, as we achieved 2.09 mm for lower bicuspids and 2.12 mm for upper bicuspids.

In many studies, as in ours, greater accuracy was observed in bicuspids [13,14,33,35,39,40,41,42]. Santucci et al. [43] reported that accuracy decreases toward the posterior area of the arch and suggested that this is most likely due to root anatomy, increased cortical bone thickness, the greater resistance of the soft tissues, and a higher masticatory force. This finding is consistent with the conclusion of Rocha et al. [44] in their review, as well as with the results of the present study.

Premolars are more predictable, partly due to the anatomy of the root volume and bone density; it is easier to expand the premolar area than the molar area. It also has to do with the forces produced by a sequenced aligner; premolars are located in the middle area, making them easier to expand, as demonstrated in various finite element studies.

Vinzenzo D’Anto et al. [20] did not find statistically significant differences between the upper and lower arches; the mean expansion was 69% at the cusp level and 59.5% at the gingival level. Galuccio et al. [14] concluded that the mean expansion was of 70% in the upper arch and between 46% and 55% in the lower arch. In our study, we found that accuracy was higher in the lower arch (61.16%) than in the upper arch (57.01%).

In 2023, Castoflorio et al. [3] asserted that professional experience and attachment geometry could influence the accuracy of the movement. The frequency of aligner changes also affects the outcome: changing the aligner every 14 days decreases undercorrection by 12% compared to changing aligners every 7 days.

In this study, some limitations were found. First, patients were not randomly selected, and the sample size could have been larger. Additionally, the sample consisted of non-growing patients, which may affect the generalizability of the results to growing individuals. In future investigations, expansion could be analyzed in both the sagittal and axial planes to provide a more comprehensive assessment, and additional groups using different aligner brands could be included.

5. Conclusions

Based on the results of this study, it can be concluded that there are significant differences between planned movements and achieved movements. The changes observed in the lower arch were greater than those in the upper arch: the lower arch showed a clinical accuracy of 61.16%, compared to 57.01% for the upper arch. The most predictable brand in the lower arch was Angel Aligners, with an accuracy of 62.30%, while in the upper arch, it was Hey Smile, with 58.66%. The second most predictable brand was HeySmile, with 59.89%; followed by Spark, with 59.27%; and, finally, Invisalign, with 57.15%. The predictability of expansion at the clinical level needs to be improved to make it more effective in orthodontic expansion with aligners.