Your Cart Is Empty
✖No products in the cart.
This article was developed by the ICam Clinical Education Team in collaboration with full-arch clinicians, prosthodontists, digital specialists, and dental laboratory professionals. Our goal is to advance accuracy standards in digital implant workflows through verified, evidence-based practices.
Accurate complete-arch implant scans are essential for ensuring passively fitting implant-supported restorations. Misfits in implant-supported prostheses can cause static strain on prosthetic components, screw loosening or fracture, restoration chipping, and bacterial accumulation at the prosthetic interface. While some tolerance exists, clinicians should aim to minimize misfits and ensure high reliability of impressions or digital scans to avoid such complications.
Recently, the impact of digital intraoral scans (IOS) and dental photogrammetry (PG) on the accuracy of complete-arch digital implant scans has gained significant attention. A meta-analysis suggested IOS offers comparable or superior accuracy to conventional impression techniques, although most studies have been conducted in vitro. Similarly, limited clinical research exists on PG digital scans.
Although some studies indicate IOS is feasible for complete-arch implant scans, stitching errors—especially when capturing large soft tissue areas—can compromise accuracy. Approaches such as using horizontal scanning bridges or horizontal scan bodies enhanced by artificial intelligence have been explored to address these limitations, though results vary.
Extra-oral photogrammetry has been proposed to overcome stitching errors, as it uses a stereo camera system to measure spatial relationships between scan body targets. However, PG cannot capture soft tissue contours or occlusal relationships, requiring additional IOS scans or conventional impressions. Combining IOS with intraoral photogrammetry has been suggested to streamline workflows, but similar limitations regarding field of view and data merging remain.
Previous studies have shown PG yields superior trueness compared to IOS, but did not evaluate precision—a vital component of accuracy. Many studies rely on stone verification casts as a reference, yet these can introduce mean 3D deviations of 34–37 µm, potentially exceeding the error of the scans themselves. Alternatives, such as extra-oral industrial scanners, face limitations due to intraoral access and motion artifacts.
According to ISO 5725–1:1994, accuracy includes trueness (agreement with a reference value) and precision (consistency between repeated measures). Where a reference for trueness is unavailable, evaluating precision provides valuable insight. This study therefore investigated the precision of IOS versus PG for complete-arch implant scans, and whether jaw type (maxilla vs. mandible) influenced precision.
Null hypotheses tested:
There is no difference in precision between IOS and PG.
Jaw type does not affect the precision of IOS and PG in complete-arch scans.
This study was approved by the Harvard Faculty of Medicine, Committee on Human Studies (#IRB18–0660) and designed as a within-patient comparison of IOS and PG precision.
From August 2019 to December 2021, nineteen patients (14 maxillary arches, 5 mandibular arches) meeting inclusion criteria were recruited. All had completely edentulous arches with at least four bone-level implants planned for screw-retained abutment-level fixed implant-supported prostheses. Patients signed informed consent, and none dropped out.
All patients had multi-unit abutments (MUA) for screw-retained restorations (NC/RC ∅ 4.6 mm; Straumann, Basel, Switzerland). Protective caps or prostheses were removed before scanning. Lighting was controlled, and all scans were performed by one experienced operator in a single visit.
Calibration of the Trios 3 scanner (3Shape, Copenhagen, Denmark) was done before each patient. Scanning followed a defined strategy, capturing the ridge crest and soft tissue, with special attention to implant areas. Five IOS scans per arch were taken. Scanbodies (PEEK; CARES Mono Scanbody; Straumann) were hand-tightened and scanned. Acceptance criteria included no visible distortion, no voids, and accurate soft tissue reproduction. STL files were exported and implant positions extracted via blinded CAD analysis.
IOS scanbodies were replaced with photogrammetry scanbodies (ICamBodies; Imetric4D, Switzerland). Five PG scans per arch were obtained using an extra-oral photogrammetry device (ICam4D). The scan path ensured visualization of all scanbody surfaces. Acceptance criteria included all scanbodies turning green and capturing at least forty images per arch. Implant coordinates were automatically calculated and extracted.
Helmert transformations aligned repeated scans to minimize deviations and obtain precision metrics. Analyses included:
Spatial Fit
Deviation distances (x, y, z) between corresponding implants were calculated across repeated scans. The 3D Euclidean distance and horizontal/vertical distances were measured.
Virtual Sheffield Test
Adapted from Sheffield’s test, this measured deviations at the contralateral distal implant after aligning scans to the most distal implants on one side.
Cross-Arch Distance
Measured scanner precision in maintaining arch width between the most distal implants.
A power calculation indicated a sample size of 1076 observations per scanner achieves >99% power to detect a 40 μm difference. Descriptive statistics and generalized estimating equations (GEE) were used for comparisons. Significance was set at α = 0.05.
| Test | Observations | IOS Mean ± SD (µm) | PG Mean ± SD (µm) | p-value |
|---|---|---|---|---|
| Spatial Fit — Average 3D distance | 1076 | 61 ± 54 | 7 ± 5 | <0.0001 |
| Average vertical distance | 1076 | 21 ± 24 | 2 ± 2 | <0.0001 |
| Average horizontal distance | 1076 | 54 ± 51 | 6 ± 5 | <0.0001 |
| Average Maximum 3D distance | 19 | 152 ± 95 | 18 ± 10 | <0.0001 |
| Average Maximum vertical distance | 19 | 66 ± 55 | 6 ± 5 | <0.0001 |
| Average Maximum horizontal distance | 19 | 145 ± 90 | 17 ± 9 | <0.0001 |
| Cross-Arch Distance — average deviation | 95 | 89 ± 6 | 10 ± 1 | <0.0001 |
| Cross-Arch Distance — average maximum deviation | 95 | 181 ± 113 | 20 ± 15 | <0.0001 |
| Virtual Sheffield — average 3D deviation | 190 | 447 ± 420 | 38 ± 31 | <0.0001 |
| Virtual Sheffield — average maximum deviation | 19 | 807 ± 654 | 70 ± 41 | <0.0001 |
| Test | Jaw Type | IOS Mean ± SD (µm) | PG Mean ± SD (µm) | p-value |
|---|---|---|---|---|
| Spatial Fit | Mandible | 95 ± 69 | 8 ± 6 | 0.040 |
| Spatial Fit | Maxilla | 51 ± 44 | 6 ± 5 | 0.167 |
| Cross-Arch Distance | Mandible | 122 ± 87 | 14 ± 15 | 0.026 |
| Cross-Arch Distance | Maxilla | 77 ± 76 | 8 ± 7 | 0.138 |
| Virtual Sheffield | Mandible | 722 ± 624 | 42 ± 30 | 0.019 |
| Virtual Sheffield | Maxilla | 349 ± 256 | 36 ± 31 | 0.617 |
This study found PG to be significantly more precise than IOS across all precision tests. IOS showed greater variability, raising concerns about its reliability for fixed complete-arch implant restorations. PG offered greater repeatability with significantly lower deviations.
IOS accuracy can be impacted by stitching errors, particularly in edentulous arches lacking anatomical landmarks. PG avoids this by capturing scan body positions simultaneously. Jaw type also influenced IOS precision, with mandibular scans showing lower precision, likely due to saliva, movable mucosa, and fewer landmarks. PG was unaffected by jaw type.
The Virtual Sheffield test revealed IOS deviations exceeding clinically acceptable thresholds, underscoring the need for caution in clinical workflows. Limitations of this study include the use of a single scanbody type and IOS protocol, as well as the lack of trueness assessment.
No products in the cart.