Lock versus merge: How software algorithms modulate the trueness of trim-rescan versus continuous digital scans in complex treatments

Abstract

Statement of problem

The trim-rescan function has been advocated for capturing multiple finish lines, yet its accuracy appears inconsistent across studies. This discrepancy may stem from fundamental differences in how the intraoral scanner (IOS) software program handles scan alignment and merging, a factor poorly understood in the context of multiple preparations.

The trim-rescan function has been advocated for capturing multiple finish lines, yet its accuracy appears inconsistent across studies. This discrepancy may stem from fundamental differences in how the intraoral scanner (IOS) software program handles scan alignment and merging, a factor poorly understood in the context of multiple preparations.

Purpose

This in vitro study aimed to investigate the influence of scanner-specific software algorithms by comparing the trueness of trim-rescan versus continuous scanning for a multi-abutment arch using 3 IOSs: the CS 3700 and Medit i700 (with "lock" feature) versus the CEREC Omnicam (without).

This in vitro study aimed to investigate the influence of scanner-specific software algorithms by comparing the trueness of trim-rescan versus continuous scanning for a multi-abutment arch using 3 IOSs: the CS 3700 and Medit i700 (with "lock" feature) versus the CEREC Omnicam (without).

Material and methods

A maxillary definitive cast with 8 prepared abutments was fabricated via 3-dimensional (3D) printing. A laboratory scanner (Medit T710) provided the reference cast. The cast was scanned (n=15 per group) using 3 IOSs (CS 3700, Medit i700, CEREC Omnicam) with both continuous and trim-rescan protocols. For trim-rescan, the protocol adhered to each scanner's unique software workflow. Trueness was assessed by computing the root mean square (RMS) error between test and reference scans after best-fit alignment. Data were analyzed using 2-way ANOVA, with complementary robust methods including Welch ANOVA and Games-Howell post hoc tests (α=.05). Effect sizes with 95% confidence intervals were calculated.

A maxillary definitive cast with 8 prepared abutments was fabricated via 3-dimensional (3D) printing. A laboratory scanner (Medit T710) provided the reference cast. The cast was scanned (n=15 per group) using 3 IOSs (CS 3700, Medit i700, CEREC Omnicam) with both continuous and trim-rescan protocols. For trim-rescan, the protocol adhered to each scanner's unique software workflow. Trueness was assessed by computing the root mean square (RMS) error between test and reference scans after best-fit alignment. Data were analyzed using 2-way ANOVA, with complementary robust methods including Welch ANOVA and Games-Howell post hoc tests (α=.05). Effect sizes with 95% confidence intervals were calculated.

Results

Scanner type significantly affected trueness (P<.001, partial η²=0.552, 95% CI [0.412 to 0.648]), with Medit i700 demonstrating the highest trueness overall. A significant main effect was also found for Technique (P=.015, partial η²=0.068, 95% CI [0.008 to 0.172]). The interaction between scanner type and scanning method was not significant (P=.852, partial η²=0.004, 95% CI [0.000 to 0.028]), indicating that the effect of scanning method was consistent across scanners. Post hoc analysis using the Welch t test to account for unequal variances revealed the trim-rescan method provided significantly better trueness only for the CEREC Omnicam (44.40 ±6.17 µm) compared to its continuous scan (53.87 ±15.77 µm); mean difference 9.47 µm, 95% CI [0.287 to 18.646], P=.044, Cohen d=0.791, 95% CI [0.039 to 1.529]. For CS 3700 and Medit i700, trueness between methods was not statistically different with P=.205 for Medit and P=.288 for CS 3700.

Scanner type significantly affected trueness (P<.001, partial η²=0.552, 95% CI [0.412 to 0.648]), with Medit i700 demonstrating the highest trueness overall. A significant main effect was also found for Technique (P=.015, partial η²=0.068, 95% CI [0.008 to 0.172]). The interaction between scanner type and scanning method was not significant (P=.852, partial η²=0.004, 95% CI [0.000 to 0.028]), indicating that the effect of scanning method was consistent across scanners. Post hoc analysis using the Welch t test to account for unequal variances revealed the trim-rescan method provided significantly better trueness only for the CEREC Omnicam (44.40 ±6.17 µm) compared to its continuous scan (53.87 ±15.77 µm); mean difference 9.47 µm, 95% CI [0.287 to 18.646], P=.044, Cohen d=0.791, 95% CI [0.039 to 1.529]. For CS 3700 and Medit i700, trueness between methods was not statistically different with P=.205 for Medit and P=.288 for CS 3700.

Conclusions

The effect of the trim-rescan function on scanning trueness was not universal but was critically dependent on the IOS and its underlying software algorithm. Scanners lacking a "lock" feature, such as CEREC Omnicam, may derive significant accuracy benefits from the trim-rescan technique for multiple abutments, whereas scanners with advanced registration locking may achieve equivalent accuracy with either method. Clinicians and researchers must consider software workflow design when selecting a scanning protocol.

The effect of the trim-rescan function on scanning trueness was not universal but was critically dependent on the IOS and its underlying software algorithm. Scanners lacking a "lock" feature, such as CEREC Omnicam, may derive significant accuracy benefits from the trim-rescan technique for multiple abutments, whereas scanners with advanced registration locking may achieve equivalent accuracy with either method. Clinicians and researchers must consider software workflow design when selecting a scanning protocol.