Influence of impression technique and screw access hole configuration on distortion in conventional impressions of prepared CAD-CAM abutment geometries

Abstract

PURPOSE

The aim of this study was to evaluate the influence of impression technique and screw access hole (SAH) configuration on distortion in conventional impressions of prepared CAD-CAM custom abutment geometries.

MATERIALS AND METHODS

A master cast with implants placed at the maxillary first premolar and first molar was fabricated. CAD-CAM abutment geometries were designed with SAH positions corresponding to 0°, 15°, and 30° buccal angulation designs. In the first experiment, six groups representing different polyvinyl siloxane impression techniques were tested. In the second experiment, groups were formed based on SAH position and filling (filled vs unfilled). Fabricated stone casts were scanned, and STL datasets were superimposed onto reference data. Total, horizontal, vertical, and angular deviations were analyzed using the Kruskal-Wallis test and two-way ANOVA.

RESULTS

The 2-step ML technique showed lower total deviation than the 1-step-L technique (24.35 µm [IQR, 6.28] vs 82.37 µm [IQR, 68.02]; P < .05). Unfilled SAHs showed reduced horizontal deviation but increased vertical and angular deviations. Angular deviation ranged from 3.60° to 4.37° in the filled SAH groups and from 5.05° to 5.13° in the unfilled SAH groups.

CONCLUSION

Impression technique and SAH configuration were associated with differences in impression distortion patterns. Within the limitations of this in vitro study, the 2-step triple phase technique showed reduced deviation, whereas increased implant angulation and unfilled SAHs were associated with greater vertical and angular deviations.

INTRODUCTION

Accurate implant impressions are essential for achieving a proper prosthetic fit, minimizing the need for occlusal adjustments, and reducing chairside time.¹ A well-adapted prosthesis contributes to long-term clinical success by preventing mechanical complications such as screw loosening and fracture, while also reducing the risk of biological complications including peri-implant mucositis and marginal bone loss.²,³

The accuracy of implant impressions can be influenced by several factors, such as the impression technique, interimplant angulation, and the splinting of impression copings. Among these, the type of impression material plays a particularly important role.⁴ For implant-level impressions, materials with high dimensional stability and tear resistance, such as polyether and polyvinyl siloxane (PVS), are widely used.⁵ Among elastomeric materials, addition silicone PVS is favored for its superior dimensional accuracy, elasticity, and long-term stability.⁶,⁷

While implant-level impressions remain standard, the increasing use of internal connection implants has increased interest in understanding how impression procedures may affect the reproduction of prepared abutment geometries. In internal implant-abutment connection designs, minor axial displacement under functional loading has been reported.⁸ This issue can be mitigated by applying functional loading through provisional prostheses prior to making impressions of prepared CAD-CAM custom abutments (impressions of prepared abutments). Such impressions may provide advantages in capturing geometric features of abutments under controlled conditions.⁹ In clinical situations involving positional discrepancies or modifications of an existing prepared abutment, impressions of prepared abutments may be required to reproduce the current abutment geometry. Such situations may include remaking fractured implant-supported restorations or restoring an existing abutment geometry without replacing the abutment.

Impressions of prepared abutments require greater precision due to the absence of impression copings or lab analogs and the need for accurate reproduction of the finish line. As with natural teeth, the wash technique using additional type PVS - combining putty with a low-viscosity wash material - is commonly used. Nevertheless, custom prepared abutments differ significantly from natural teeth in shape, often having a rounded rectangular form and a screw access hole (SAH), which may influence impression distortion.

Previous studies have reported that 2-step impression techniques generally result in higher accuracy than 1-step techniques.¹⁰,¹¹ This is attributed to the controlled thickness of the wash material and the hydraulic pressure generated during seating, which enhances the capture of fine details.¹² Depending on the number of viscosities involved, PVS impressions can be further classified into single phase or dual phase techniques.¹²,¹³ Most previous evaluations of 2-step impressions have involved dual phase protocols using putty and light body materials. However, the combination of materials with significantly different viscosities, such as putty and light body, may result in interfacial discrepancies and suboptimal adaptation, representing critical limitations of this technique.¹⁴,¹⁵,¹⁶

Volumetric distortion of PVS materials, including polymerization shrinkage and elastic rebound, has been reported to cause systematic deviations in both the vertical and horizontal dimensions of stone dies.¹⁷,¹⁸,¹⁹,²⁰ To overcome distortion, modified techniques such as the 2-step triple phase method - also known as the matrix impression technique - have been proposed. This approach involves creating a putty matrix that is internally relined with a medium viscosity material before the final wash is applied.²¹ Despite its potential advantages, few studies have investigated the effect of such modified techniques on the accuracy of impressions of prepared abutments.²²

Computer aided design (CAD) - computer aided manufacturing (CAM) customized abutments can be designed to compensate for angulation and provide optimal prosthetic contours and emergence profiles.²³ In customized abutments, the SAH introduces a discontinuity in the external surface geometry, potentially creating localized irregularities or undercut areas that may influence impression material behavior.

Although numerous studies have investigated the accuracy of implant-level impressions, relatively few have evaluated impressions of prepared abutments under different impression techniques and abutment configurations. This in vitro study evaluated the effect of impression technique and SAH configuration on the accuracy of conventional impressions of prepared CAD-CAM custom titanium abutments. The first null hypothesis was that impression technique would not affect impression accuracy. The second null hypothesis was that SAH configuration and filling status would not influence impression distortion under the selected technique.

MATERIALS AND METHODS

A maxillary partially edentulous model, missing the right premolars and first molar, was digitally designed using CAD software (DentalCAD, exocad GmbH, Darmstadt, Germany) and fabricated using a fused deposition modeling 3D printer (F370, Stratasys, Eden Prairie, MN, USA) and printing material (PC-ABS, Stratasys, Eden Prairie, MN, USA). Implants (TS II, Osstem Implant Co., Seoul, Korea) were placed at the first premolar (Ø4.5 × 10 mm) and first molar sites (Ø5.0 × 10 mm) using a surgical guide stent generated with implant planning software (Implant Studio, 3Shape, Copenhagen, Denmark) and printed using a 3D printer (Dental 3D Printer-P, Dentium, Seoul, Korea).

Simulated CAD-CAM custom abutment geometries with standardized external contours were designed. The SAH position was derived from the planned molar angulation design; therefore, the experiment evaluated angulation-derived abutment geometry and SAH position rather than clinical implant angulation or implant-abutment connection behavior. All abutments were designed with a 1-mm supragingival margin to eliminate soft tissue interference and were digitally aligned within a partial-arch master cast design incorporating anterior-posterior vertical stops. To eliminate any potential micromovement between the abutments and the resin cast during impression procedures, a one-piece model incorporating abutment geometries was fabricated using 3D printing material (Denprin P, B&L Biotech Inc., Seoul, Korea) (Fig. 1). The experimental model was designed to isolate geometric variables under controlled conditions rather than to replicate clinical material properties. Master casts were scanned using a desktop scanner (Freedom HD, Dentium, Seoul, Korea) to obtain reference STL datasets.

Fig. 1. Master cast design with integrated simulated abutment geometries and controlled SAH positions based on implant angulation.

The study was structured as two sequential experiments rather than a full factorial design in order to isolate the primary effect of impression technique before assessing geometry-related variables under controlled conditions. Experiment 1 evaluated six PVS impression techniques using total, horizontal, and vertical deviations. Experiment 2 was limited to the selected 2-step ML protocol and assessed the effects of SAH position and filling status using total, horizontal, vertical, and angular deviations. In the first experiment, the accuracy of impressions was evaluated according to different impression techniques by combining 1-step and 2-step methods with various PVS material viscosities (putty, heavy body, medium body and light body) (Spident Co., Incheon, Korea), resulting in 6 experimental groups (Fig. 2). Based on the technique demonstrating the most favorable performance, the second experiment was conducted to explore the influence of SAH configuration and filling status on impression distortion.

Fig. 2. Experimental grouping based on impression technique.

The 2-step ML technique was used in the second experiment. Molar abutment geometries were designed with SAH positions corresponding to 0°, 15°, or 30° buccal angulation relative to the premolar reference, and each SAH was either filled (“F”) or left unfilled (“nF”), yielding six experimental groups.Thus, Experiment 2 was limited to the selected 2-step ML protocol. To simulate clinical conditions, a 2-mm vertical depth was maintained for all SAHs. The lower part of the SAH was filled with polytetrafluoroethylene (PTFE) tape, and the upper part was sealed with light-polymerizing provisional resin (Quicks Flow, DentKist, Gunpo, Korea) and light polymerized. Impressions were taken using disposable plastic trays (PB, Seil Global Co., Busan, Korea), which were stabilized using a custom-designed positioning jig (DentalCAD). The jig was printed and fixed to the master cast using resin cement (G-CEM ONE, GC Corp., Tokyo, Japan).

All impressions were made, allowed to set, removed, and poured in the same air-conditioned laboratory at room temperature (23 ± 2℃). Following the manufacturer’s recommended setting time, all impressions were removed vertically and poured with type IV dental stone after the same elapsed time. The stone was mixed based on the powder and water ratio recommended by manufacturer’s instructions using a vacuum mixer. A total of 60 casts (10 per group) were fabricated in the first experiment. Another 60 casts were produced in the second experiment using the selected impression technique. An a priori sample size estimation was performed based on total deviation as the primary outcome of the first experiment. Using the observed effect size from the pilot dataset (Cohen’s f = 0.57) with α = .05 and a power of 0.80 for a one-way ANOVA with 6 groups, the required total sample size was approximately 46 specimens (≈8 per group). Therefore, 10 specimens per group were included in each group.

All stone casts were scanned (Freedom HD), and the STL datasets were imported into 3D inspection software (GOM Inspect 2018, GOM GmbH, Braunschweig, Germany). The STL datasets were superimposed onto the reference STL using the Prealignment function with additional best-fit computation. The region of interest (ROI) was standardized across all datasets and extended from the cusp tip of the first premolar to the mesiobuccal cusp tip of the second molar. Total deviation was defined as the software-reported “Deviation” value obtained after this alignment procedure and was used as a global discrepancy index within the predefined ROI, rather than as a direction-specific linear deviation (Fig. 3A). Total deviation represents the global discrepancy within the defined region of interest rather than abutment-specific deviation. Horizontal deviation was determined from the distance between the centroids of the premolar and molar abutment occlusal surfaces using local-fit alignment (Fig. 3B).²⁴,²⁵,²⁶ Vertical deviation was defined as the difference in the Z-axis position of the centroid of the molar abutment’s occlusal surface between the test and reference scans (Fig. 3C). In the second experiment, angular deviation was additionally measured to evaluate the effect of the SAH on the impression angulation. Angular deviation was calculated from the angle formed between a line perpendicular to the occlusal surface and the abutment axis (Fig. 3D).

Fig. 3. Deviation measurement. (A) Total deviation, (B) Horizontal deviation, (C) Vertical deviation, (D) Angular deviation.

Before STL alignment and deviation measurements, all STL files were anonymized and coded, and the examiner performing the measurements was blinded to group allocation. The ROI and alignment procedure were predefined and applied identically to all datasets. For horizontal and vertical measurements, both directional and absolute deviations were calculated. Directional deviation retained the algebraic sign to indicate expansion or contraction relative to the reference dataset, whereas absolute deviation represented the magnitude of discrepancy irrespective of direction. Absolute deviation values were primarily used for intergroup comparisons of overall accuracy. Because horizontal deviations did not demonstrate consistent directional trends and were uniformly positive, absolute values were used for statistical comparison of horizontal accuracy, whereas directional values were additionally examined to identify potential systematic distortion patterns.

Statistical analysis was performed using SPSS (v25.0, IBM Corp., Armonk, NY, USA). Data normality and homogeneity of variance were evaluated using the Shapiro-Wilk and Levene tests. In Experiment 1, deviation variables were analyzed using the Kruskal-Wallis test with Bonferroni-adjusted pairwise comparisons. In Experiment 2, horizontal deviation was analyzed using two-way ANOVA with Tukey HSD post hoc tests, whereas total, vertical, and angular deviations were analyzed using the Kruskal-Wallis test with Bonferroni-adjusted pairwise comparisons. Unless otherwise specified, statistical comparisons were performed using absolute deviation values. Multiple regression analysis was also conducted to assess effect sizes. Intra-examiner measurement reliability was assessed using repeated measurements of randomly selected STL datasets, and intraclass correlation coefficients (ICCs) were calculated for the deviation measurements. Statistical significance was set at α = .05. For non-parametric data, values are presented as median and interquartile range (IQR). For parametric data, values are presented as mean and standard deviation (SD).

RESULTS

Intra-examiner measurement reliability was high, with ICC values ranging from 0.933 to 0.957 across the deviation measurements. In the first experiment, total deviation, defined as the primary indicator of overall three-dimensional accuracy, differed significantly among impression technique groups (P < .05) (Table 1). The 1-step-M group showed greater deviation than the 2-step-M group, and the 1-step-L group demonstrated greater deviation than all 2-step groups (2-step-M, 2-step-L, and 2-step-ML). No significant differences were observed in horizontal deviation (absolute values). For vertical deviation, the 1-step-L group exhibited greater deviation than the 2-step-L and 2-step-ML groups (P < .05) (Fig. 4). Comparison of deviation magnitude and variability showed that the 1-step-L technique demonstrated the greatest deviation, whereas the 2-step-ML technique demonstrated the lowest deviation magnitude and variability.

Table 1. Median (interquartile range, IQR) of deviation (µm) by different impression techniques.

Group Deviation	1-step-M	1-step-L	2-step-M	2-step-L	2-step-HL	2-step-ML
Total	36.61 (43.90)ab	82.37 (68.02)a	21.71 (1.95)c	22.33 (5.04)bc	27.68 (22.96)abc	24.35 (6.28)bc
Horizontal (directional)	57.04 (36.87)	15.69 (78.43)	36.04 (153.55)	50.22 (19.85)	60.61 (15.06)	61.91 (17.64)
Horizontal (absolute)	57.04 (36.87)	57.45 (169.38)	61.11 (33.55)	50.22 (19.85)	60.61 (15.06)	61.91 (17.64)
Vertical (directional)	0.23 (74.18)	-51.37 (129.84)	-7.78 (23.23)	-1.90 (25.07)	-11.48 (18.35)	-4.43 (12.38)
Vertical (absolute)	30.35 (68.92)ab	59.61 (65.87)a	12.35 (17.43)ab	10.60 (14.74)b	12.92 (30.01)ab	5.96 (9.03)b

Different superscript letters within each deviation type indicate statistically significant differences according to Bonferroni correction (P < .05).

Fig. 4. Vertical absolute deviation according to impression technique. Box-and-whisker plots show the median, interquartile range, and minimum-to-maximum values (n = 10 per group). 1-step-M = 1-step medium-body; 1-step-L = 1-step light-body; 2-step-M = 2-step medium-body; 2-step-L = 2-step light-body; 2-step-HL = 2-step heavy/light-body; 2-step-ML = 2-step medium/light-body. P < .05.

In the second experiment, total deviation did not differ significantly according to SAH position or filling condition (Table 2). Two-way ANOVA revealed significant main effects of SAH position (implant angulation) and filling condition on horizontal deviation (P < .05), without a significant interaction (P = .446) (Table 2). Horizontal deviation was lower in unfilled groups and progressively decreased as the angulation-derived SAH position shifted buccally (P < .05, Table 3). Post hoc analysis showed that the 0F group had greater horizontal deviation than the 0nF, 15nF, 30F, and 30nF groups, and the 15F group had greater deviation than the 15nF, 30F, and 30nF groups (P < .05).

Table 2. Two-way ANOVA results for horizontal deviation according to SAH position and filling condition.

	Type III Sum of squares	df	Mean square	F	Sig
Corrected model	15850.878	5	3170.176	14.493	P < .001
SAH position	11702.105	2	5851.053	26.749	P < .001
SAH filling	3789.765	1	3789.765	17.325	P < .001
SAH position × filling	359.008	2	179.504	.821	.446
Error	11811.926	54	218.739
Corrected total	27662.804	59

Table 3. Deviation values by SAH position and filling condition.

Group Deviation	0F	0nF	15F	15nF	30F	30nF
Total	24.35 (6.28)	17.02 (9.13)	15.93 (11.06)	16.21 (14.31)	15.06 (20.80)	16.38 (16.77)
Horizontal (absolute)	63.34 ± 22.02a	42.21 ± 16.31bc	50.33 ± 11.06ab	33.14 ± 11.70cd	23.90 ± 11.17de	14.54 ± 13.32e
Vertical (directional)	-4.43 (12.38)	9.20 (12.65)	-6.31 (10.52)	19.07 (20.44)	4.73 (21.23)	22.30 (15.39)
Vertical (absolute)	5.96 (9.03)a	11.88 (11.87)abc	6.31 (7.79)ab	19.07 (20.44)cde	6.30 (19.90)ab	25.24 (21.05)e
Angular (°)	3.60 (0.30)a	5.05 (0.38)ce	4.37 (0.29)cd	5.13 (0.28)e	4.31 (0.29)abd	5.12 (0.71)e

Values are presented as median (interquartile range) for total, vertical, and angular deviations and as mean ± standard deviation for horizontal deviation. Different superscript letters within each deviation type indicate statistically significant differences according to Bonferroni correction and Tukey HSD test (P < .05).

Multiple linear regression analysis confirmed that angulation-derived SAH position and filling condition were associated with horizontal deviation. Using 15° and unfilled SAH as the reference categories, the 0° SAH position showed greater horizontal deviation (β = 0.242, P = .021), whereas the 30° SAH position showed lower horizontal deviation (β = –0.494, P < .001). Filled SAHs showed greater horizontal deviation than unfilled SAHs (β = 0.370, P < .001) (Table 4). For vertical deviation, the 30nF group exhibited greater deviation than the 0nF, 0F, and 15F groups (P < .05), and the 15nF group showed greater deviation than the 0F and 15F groups. For angular deviation, all unfilled groups (0nF, 15nF, and 30nF) showed higher values than their corresponding filled groups (0F, 15F, and 30F) (P < .05) (Fig. 5). Vertical and angular deviations increased with greater implant angulation in the unfilled condition.

Table 4. Multiple linear regression analysis of horizontal deviation according to SAH position and filling condition.

Variables	B	Std. Err.	β	t	P	Tolerance	VIF
(Constant)	33.788	3.806		8.877	< .001
SAH position dummy (0 degree)	11.042	4.662	.242	2.369	.021	.750	1.333
SAH position dummy (30 degrees)	-22.518	4.662	-.494	-4.830	< .001	.750	1.333
SAH filling dummy (Filling)	15.895	3.806	.370	4.176	< .001	1,000	1.000

R² = 0.560, adjusted R² = 0.536, F = 23.760 (P = .000), Durbin-Watson (1.977).

Reference group: 1) SAH position 15 degrees, 2) unfilled SAH.

Fig. 5. Angular deviation according to SAH position and filling condition. Box-and-whisker plots show the median, interquartile range, and minimum-to-maximum values (n = 10 per group). F = filled SAH; nF = unfilled SAH; 0, 15, and 30 indicate SAH positions corresponding to 0°, 15°, and 30° angulation-derived designs. Different lowercase letters indicate significant differences (P < .05).

DISCUSSION

This study evaluated the influence of impression technique and SAH configuration on the distortion of conventional impressions of prepared CAD-CAM custom titanium abutments. Impression technique significantly affected overall deviation, and SAH configuration was associated with directional distortion patterns, leading to rejection of the first null hypothesis. The second null hypothesis was partially rejected, as screw access hole (SAH) configuration and filling condition did not significantly affect total deviation, but were associated with differences in horizontal, vertical, and angular deviations.

In the first experiment, 1-step techniques demonstrated greater total and vertical deviations than 2-step techniques, indicating reduced trueness and precision. Among the tested protocols, the 1-step-L technique exhibited the greatest deviation, whereas the 2-step-ML technique showed the highest accuracy. These findings are consistent with previous reports suggesting that controlled wash space and staged material application improve dimensional accuracy.¹²,²¹ Insufficient wash thickness in the 1-step technique may have contributed to distortion, particularly along axial surfaces. A wash thickness of approximately 2 mm has been suggested as optimal for minimizing volumetric change, a condition more consistently achieved in the 2-step protocols.¹² In addition, differences in viscosity and interfacial adaptation between putty and light body materials may have contributed to reduced accuracy in the 1-step-L group, as previously described.¹⁶

The triple-phase 2-step technique demonstrated accuracy comparable to that reported in earlier studies evaluating matrix or modified putty-wash protocols.¹⁷,²¹ Although such techniques may require additional procedural steps, the protocol used in the present study was completed within a clinically feasible time, suggesting that improved accuracy can be achieved without substantial increases in chairside complexity.

Directional deviations showed negative vertical values and positive horizontal values, consistent with previous descriptions of volumetric distortion in polyvinyl siloxane impressions.¹⁷,¹⁸,¹⁹,²⁰ Elastic recovery and polymerization shrinkage toward the tray have been associated with reduced vertical dimensions and increased horizontal spacing in stone casts.¹⁷,²⁰ These mechanisms may explain the observed distortion patterns.

In the second experiment, total deviation did not differ according to SAH configuration; however, significant differences were observed in horizontal, vertical, and angular deviations. No significant interaction between implant angulation and SAH filling was detected. This discrepancy may be related to the alignment strategy. Total deviation was calculated using a best-fit algorithm that minimizes global discrepancies, whereas local-fit alignment may be more sensitive to regional deformation.²⁷,²⁸ Consequently, localized distortions associated with SAH configuration may not have been fully reflected in the global deviation value. Because both reference and test datasets were acquired using the same desktop scanner, the measured deviations reflect relative differences among experimental groups rather than absolute geometric accuracy, and therefore should be interpreted as reflecting overall distortion patterns rather than isolated abutment accuracy.

Although SAHs are typically filled in clinical practice, unfilled conditions were included to isolate the geometric effect of the hole on impression distortion under controlled experimental conditions. Horizontal deviation decreased with increasing implant angulation and in the absence of SAH filling. Therefore, the lower horizontal deviation observed in the unfilled SAH groups should not be interpreted as superior overall accuracy. Rather, this finding may reflect redistribution of distortion among different spatial axes instead of improved geometric accuracy. In contrast, vertical and angular deviations increased in unfilled groups, particularly at greater implant angulations. The presence of impression material within an unfilled SAH may increase mechanical interlocking during impression removal, potentially contributing to localized deformation. In addition, the buccal relocation of the SAH with increasing implant angulation may create geometric undercuts that alter the path of impression removal and concentrate strain within the elastomeric material. Although these mechanisms were not directly evaluated, previous studies have shown that SAH geometry influences mechanical retention in cement-retained restorations.²⁵,²⁶

Regression analysis further supported the influence of implant angulation on horizontal deviation, suggesting that angulation may have a greater effect on distortion patterns than filling status under the tested conditions. Greater angulation-derived buccal relocation of the SAH was associated with more pronounced directional changes, particularly when the SAH was located buccally. These findings are consistent with reports indicating that buccolingual implant angulation can increase impression error compared with parallel configurations.²⁹

Although several intergroup differences were statistically significant, their clinical relevance remains uncertain. In the present study, median total deviation values ranged from 15.06 to 82.37 µm, and the angular difference between corresponding filled and unfilled SAH groups ranged from approximately 0.76° to 1.45°. These values are lower than the commonly cited 120-µm marginal discrepancy reference for fixed dental restorations; however, that threshold refers to restoration fit rather than impression distortion.³⁰ Moreover, the clinically acceptable misfit threshold for implant-supported restorations remains undetermined.³ Therefore, the observed micrometer- and degree-level differences should be interpreted as experimental distortion patterns under controlled conditions, rather than as direct evidence of clinically meaningful differences in prosthetic fit, marginal adaptation, occlusal adjustment, or restoration remanufacturing.

This study has several limitations. As an in vitro investigation using a one-piece model incorporating simulated abutment geometries, this study did not reproduce clinical variables such as saliva, soft tissue compliance, patient movement, limited intraoral access, or tray-seating variability. In addition, implant-abutment connection behavior was not directly evaluated. Although standardized positioning jigs and vertical stops were used to enhance reproducibility, deviation measurements were dependent on the selected superimposition protocols, which may have influenced the distribution of directional discrepancies. Because the present study used a sequential design rather than a full factorial design, the effects of SAH configuration and filling status were evaluated only under the selected 2-step ML protocol and should not be generalized to other impression techniques.

Intraoral scanning is increasingly used in implant prosthodontics; however, optical data acquisition may be influenced by limited intraoral access and the reflective properties of metallic restorative components.³¹ Under complex clinical conditions or complete-arch situations, conventional elastomeric impressions may remain relevant. Within the limitations of this study, careful selection of impression technique and consideration of SAH configuration may help explain distortion patterns observed in impressions of prepared CAD-CAM custom abutment geometries under controlled experimental conditions. However, these findings should not be interpreted as evidence that staged impression techniques or SAH sealing reduce distortion in clinical situations, and further clinical validation is required.

CONCLUSION

This in vitro study demonstrated that impression technique, SAH position, and filling condition were significantly associated with differences in impression distortion under controlled experimental conditions. Among the evaluated protocols, the 2-step triple phase technique showed the lowest deviation, whereas the 1-step dual phase technique showed the greatest deviation. Greater angulation-derived buccal relocation of the SAH and unfilled SAHs were associated with greater vertical and angular deviations. Although horizontal deviation was reduced in unfilled SAH groups, this should not be interpreted as improved overall accuracy because vertical and angular deviations increased. These findings should be interpreted as laboratory-based distortion patterns rather than evidence of clinical effectiveness.