Postoperative Extremity Tomosynthesis—A Superimposition-Free Alternative to Standard Radiography?

Abstract

Rationale and Objectives

This study investigates the performance of tomosynthesis in the presence of osteosynthetic implants, aiming to overcome superimposition-induced limitations in conventional radiograms.

Materials and Methods

After surgical fracture induction and subsequent osteosynthesis, 8 cadaveric fracture models (wrist, metacarpus, ankle, metatarsus) were scanned with the prototypical tomosynthesis mode of a multiuse x-ray system. Tomosynthesis protocols at 60, 80, and 116 kV (sweep angle 10°, 13 FPS) were compared with standard radiograms. Five radiologists independently rated diagnostic assessability based on an equidistant 7-point scale focusing on fracture delineation, intra-articular screw placement, and implant positioning. The intraclass correlation coefficient (ICC) was calculated to analyze interrater agreement.

Results

Radiation dose in radiography was 0.48 ± 0.26 dGy·cm² versus 0.12 ± 0.01, 0.36 ± 0.02, and 1.95 ± 0.11 dGy·cm² for tomosynthesis scans at 60, 80, and 116 kV. Delineation of fracture lines was superior for 80/116 kV tomosynthesis compared with radiograms (P ≤ 0.003). Assessability of intra-articular screw placement was deemed favorable for all tomosynthesis protocols (P ≤ 0.004), whereas superiority for evaluation of implant positioning could not be ascertained (all P's ≥ 0.599). Diagnostic confidence was higher for 80/116 kV tomosynthesis versus radiograms and 60 kV tomosynthesis (P ≤ 0.002). Interrater agreement was good for fracture delineation (ICC, 0.803; 95% confidence interval [CI], 0.598–0.904), intra-articular screw placement (ICC, 0.802; 95% CI, 0.599–0.903), implant positioning (ICC, 0.855; 95% CI, 0.729–0.926), and diagnostic confidence (ICC, 0.842; 95% CI, 0.556–0.934).

Conclusions

In the postoperative workup of extremity fractures, tomosynthesis allows for superior assessment of fracture lines and intra-articular screw positioning with greater diagnostic confidence at radiation doses comparable to conventional radiograms.

Digital tomosynthesis is based on serial x-ray images providing high in-plane resolution. After acquiring a limited number of low-dose projections at different angles in a single sweep, these are merged into 3-dimensional datasets designed to deliver superior in-depth resolution, improve local tissue separation, and offset summation effects typical to conventional radiography.¹ Tomosynthesis constitutes a standard technique with widespread application in the evaluation of breast tissue^2–4 yet not in musculoskeletal imaging. To date, promising approaches have been made for tomosynthesis-based imaging of the appendicular skeleton, for example, for fracture diagnosis and healing^5–7 as well as for rheumatoid arthritis.^8–10 Furthermore, encouraging results have been reported for weight-bearing spine imaging¹¹ with general advantages over conventional radiography in terms of bone structure delineation.^12,13

With high cost-efficiency and ubiquitous availability at relatively low radiation dose, conventional radiography represents the primary imaging modality for postoperative evaluation of the appendicular skeleton regarding bone structure and implant placement. Typically, diagnostic value of summation images is restricted due to superimposition of metal hardware, bone structures, and thus limited spatial depth.¹⁴ In particular, it is often difficult to assess the exact implant placement and potential postoperative complications, such as intra-articular screw positioning or implant loosening. To address these diagnostic uncertainties, computed tomography (CT) examinations are usually the method of choice for further evaluation, although incurring a substantially higher radiation dose.^15,16 Offering an alternative to gantry-based multidetector CT, cone-beam CT has attracted increasing attention in recent years with excellent 3-dimensional visualization of the appendicular skeleton even in presence of metal implants, albeit still at the cost of a much higher dose burden than standard radiography.^17–20

Whether low-dose tomosynthesis can offer similar advantages in the workup of patients with osteosynthetic implants has not been thoroughly investigated thus far. Therefore, the current study aims to assess the best compromise between radiation dose reduction and image quality for optimized evaluation of osteosynthetic implants of the appendicular skeleton by means of tomosynthesis.

MATERIALS AND METHODS

Cadaveric Specimens

After permission for this experimental study was granted by the local ethics committee, 2 formalin-fixed cadaveric specimens were obtained from the local institute of anatomy. No further written informed consent was required as donors had previously consented to posthumous scientific research. All examinations were carried out in accordance with institutional regulations, as well as with state and federal laws. Fractures of the distal tibia and fibula, distal radius, metatarsal, and metacarpal bones were induced by a board-certified trauma surgeon with subsequently performed osteosynthetic treatment thereof. Fractures of the distal tibia and fibula were fixed with designated plates (Contour Plate II; Argomedical, Braunschweig, Germany) and screws (3.5-mm cortical screws; Argomedical) after K-wire fixation. Volar locking plate fixation was performed on distal radius fractures (Aptus 2.5 radial plate; Medartis, Basel, Switzerland). Metatarsal and likewise metacarpal fractures were treated with plate fixation (Aptus 2.8-mm T-plate and 1.5-mm 6-hole trapezoidal plate; Medartis). Figure 1 provides a radiographic overview of the different fracture regions and osteosynthetic implants.

FIGURE 1.

Cadaveric fracture models. Conventional x-ray scans depict the different fracture regions in the distal tibia and fibula (A), distal radius (B), fifth metatarsal (C), and third metacarpal bones (D) as well as the employed osteosynthetic implants. Corresponding images of osteosynthetic treatment after fracture induction (E to H).

Image Acquisition and Reconstruction

A total of 8 osteosynthetically treated fracture regions were examined employing a gantry-free multipurpose x-ray system (Multitom Rax; Siemens Healthineers, Forchheim, Germany). All scans were performed in supine position. Each body region was examined with 3 prototypical tomosynthesis protocols in lateral and anterior-posterior orientation using preselected tube voltages (60 kV, 80 kV, 116 kV) and a tube current–exposure time product of 0.49 mAs (60 kV, 80 kV)/0.55 (116 kV) per image. A 0.3-mm copper filter was employed for all studies, and the sweep angle was set to 10° with a fixed frame rate of 13 FPS. Tomosynthesis protocols were compared with conventional biplanar radiographs with equivalent image volumes as per clinical standard. The acquisition parameters for x-ray scans were set to a tube voltage of 50 kV (anterior-posterior) or 52 kV (lateral), with a tube current–exposure time product of 1.63 mAs/2.04 mAs per image.

Dose-area products (DAPs) were obtained from the automatically generated scan report. Tomosynthesis datasets were reconstructed employing a prototypical reconstruction software. Depending on projection orientation, reconstruction parameters were selected with a slice thickness of 1.25 mm, a field of view of 213 × 189 mm (anterior-posterior)/219 × 195 mm (lateral), and an image matrix of 1440 × 1275 (anterior-posterior)/1478 × 1316 (lateral) pixels.

Image Quality Analysis

Five independent radiologists with 4 to 10 years of expertise in musculoskeletal imaging performed tomosynthesis and radiographic image analyses in randomized and blinded fashion using commercially available picture archiving and communication software (Merlin; Phoenix-PACS, Freiburg, Germany). Prior to image quality analyses, readers underwent a structured briefing, explaining the focus of each category on the basis of a prereading dataset. Based on an equidistant 7-point scale, readers rated image quality by assessing the delineation of fracture lines and the positioning of the osteosynthesis implants, questioning whether intra-articular screw placement could be excluded (7 = excellent, 6 = very good, 5 = good, 4 = satisfactory, 3 = fair, 2 = poor, 1 = very poor). Readers were permitted to alter viewing settings as personally required.

Statistics

Dedicated software (SPSS Statistics 29.0.1; IBM, Armonk, NY) was used for statistical analysis. Reporting of data comprises median values and interquartile ranges (IQRs). For data comparison, Friedman rank-based analysis of variance was performed with Bonferroni-corrected pairwise post hoc tests. Interrater agreement was established via intraclass correlation coefficient (ICC) analyses in a 2-way random-effects model. Interpretation of ICC values was performed as follows: ICC >0.90 = excellent; 0.75–0.90 = good; 0.50–0.75 = moderate; and <0.50 = poor reliability.²¹ Statistical significance is indicated by an α level of 0.05.

RESULTS

Radiation Dose

The DAP of the conventional radiographs amounted to 0.48 ± 0.26 dGy·cm², whereas tomosynthesis protocols ranged from 0.12 ± 0.01 dGy·cm² (60 kV) to 1.95 ± 0.11 dGy·cm² (116 kV). Although the 80 kV protocol resulted in a 3-fold dose increase compared with tomosynthesis at 60 kV, the prior does constitute a 25.0% reduction compared with radiographs. Of note, the 116 kV protocol induces more than a 4-fold increase in dose over 80 kV. Table 1 summarizes acquisition parameters and resulting radiation exposures. Figure 2 provides an overview of the different tomosynthesis protocols compared with standard radiography.

TABLE 1.

Scan Parameters and Radiation Dose

	Voltage (kV)	Current-Time Product Per Projection (mAs)	Sweep Angle (°)	Framerate (FPS)	No. Projections	Copper Filter (mm)	DAP (dGy·cm2)	Relative Dose (%)
T1	60	0.49	10	13	15	0.3	0.12 ± 0.01	25%
T2	80	0.49	10	13	15	0.3	0.36 ± 0.02	75%
T3	116	0.55	10	13	15	0.3	1.95 ± 0.11	406%
X-ray	50–52	1.63	-	-	-	-	0.48 ± 0.26	100%

DAP, dose-area product; T1–3, tomosynthesis scan protocols 1–3.

FIGURE 2.

Illustration of image quality. Anterior-posterior radiograph (A) of the distal tibia and fibula after osteosynthesis compared with tomosynthesis scans at 60 kV (B), 80 kV (C), and 116 kV (D). Focused on the distal fibula fracture, delineation of the fracture line is superior in 80 kV and 116 kV scans compared with the corresponding radiograph.

Delineation of Fracture Lines

Subjective analyses revealed best the delineation of fracture lines for 80 kV and 116 kV tomosynthesis protocols (median rating of 6 [IQR 5–7]) with superior ratings compared with x-ray scans (P ≤ 0.003). Furthermore, 80 kV and 116 kV protocols were determined to be superior to 60 kV tomosynthesis scans in this respect (P ≤ 0.001). Comparisons of 80 kV and 116 kV protocols, as well as 60 kV tomosynthesis and x-ray scans, revealed no significant difference (all P's > 0.999). Interreader reliability for delineation of fracture lines was good, indicated by an ICC value of 0.803 (95% confidence interval [CI], 0.598–0.904; P < 0.001). Figure 3 displays how tomosynthesis resolves the superimposition effect limiting conventional radiography.

FIGURE 3.

Illustration of spatial depth. Lateral radiograph (A) of a midcarpal fracture after osteosynthesis compared with tomosynthesis scans at 80 kV (B–E). Since each tomosynthesis slice depicts a different portion of scan volume in focus, the delineation of structures varies depending on its location within the plane.

Assessability of Osteosynthesis Material Positioning

With regards to the evaluation of intra-articular screw placement, all tomosynthesis protocols were considered superior to x-ray scans, irrespective of radiation dose (P ≤ 0.004). All tomosynthesis scan protocols yielded at least very good assessability (6 [5–7]), bearing no significant difference in direct comparison (P > 0.999). Focusing on the evaluation of implant positioning, no significant difference was ascertained among the different tomosynthesis protocols (6 [5–7]) or between tomosynthesis and standard radiographs (6 [4–7]) (all P's ≥ 0.599). Interrater agreement for intra-articular screw placement (ICC, 0.802; 95% CI, 0.599–0.903; P < 0.001) and implant position (ICC, 0.855; 95% CI, 0.729–0.926; P < 0.001) was both good. Cumulative results for subjective image quality assessment are provided in Table 2, whereas Supplemental Table S1, http://links.lww.com/RLI/A933, contains the individual frequencies of designated ratings.

TABLE 2.

Diagnostic Assessability

	Fracture	Articular Screw Placement	Implant Positioning	Diagnostic Confidence
	Median rating of 5 radiologists (interquartile range)
X-ray	5 (3–6)	5 (4–6)	6 (4–7)	5 (4–6)
60 kV	5 (4–6)	6 (5–7)	6 (5–7)	5 (4–6)
80 kV	6 (5–7)	6 (5–7)	6 (5–7)	6 (5–7)
117 kV	6 (5–7)	6 (5–7)	6 (5–7)	6 (5–7)
	P
X-ray vs 60 kV	>0.999	0.004	>0.999	>0.999
X-ray vs 80 kV	0.003	<0.001	0.846	0.02
X-ray vs 117 kV	<0.001	<0.001	0.599	0.02
60 kV vs 80 kV	0.001	>0.999	>0.999	0.02
60 kV vs 117 kV	<0.001	>0.999	>0.999	0.02
80 kV vs 117 kV	>0.999	>0.999	>0.999	>0.999
ICC (95% CI)	0.803 (0.598–0.904)	0.802 (0.599–0.903)	0.855 (0.729–0.926)	0.842 (0.556–0.934)

Note: P values indicating statistical significance highlighted in bold.

ICC, intraclass correlation coefficient (2-way random-effects model based on absolute agreement); CI, confidence interval.

Diagnostic Confidence

Readers stated superior diagnostic confidence in tomosynthesis scans at 80 kV and 116 kV (6 [5–7]) compared with conventional x-ray examinations as well as tomosynthesis scans at 60 kV (P ≤ 0.002). Comparisons between 80 kV and 116 kV tomosynthesis protocols as well as between 60 kV tomosynthesis and x-ray scans revealed no statistical difference (all P's ≥ 0.999). Confidence ratings displayed good interrater agreement, indicated by an ICC of 0.842 (95% CI, 0.556–0.934; P < 0.001). Table 2 comprises the results of diagnostic confidence. In addition, detailed diagnostic confidence ratings are included in Table 3.

TABLE 3.

Detailed Reader Analysis for Each Dimension of Image Assessability

	Fracture				Articular Screw Placement				Implant Positioning				Diagnostic Confidence
	60 kV	80 kV	117 kV	X-ray	60 kV	80 kV	117 kV	X-ray	60 kV	80 kV	117 kV	X-ray	60 kV	80 kV	117 kV	X-ray
Excellent (7)	9 (22.5%)	11 (27.5%)	11 (27.5%)	9 (22.5%)	14 (35.0%)	15 (37.5%)	15 (37.5%)	2 (5.0%)	17 (42.5%)	16 (40.0%)	17 (42.5%)	16 (40.0%)	8 (20.0%)	11 (27.5%)	11 (27.5%)	8 (20.0%)
Very good (6)	2 (5.0%)	10 (25.0%)	14 (35.0%)	8 (20.0%)	9 (22.5%)	9 (22.5%)	9 (22.5%)	13 (32.5%)	5 (12.5%)	9 (22.5%)	8 (20.0%)	7 (17.5%)	7 (17.5%)	13 (32.5%)	13 (32.5%)	9 (22.5%)
Good (5)	15 (37.5%)	12 (30.0%)	10 (25.0%)	8 (20.0%)	8 (20.0%)	9 (22.5%)	11 (27.5%)	12 (30.0%)	9 (22.5%)	7 (17.5%)	7 (17.5%)	6 (15.0%)	13 (32.5%)	9 (22.5%)	9 (22.5%)	8 (20.0%)
Satisfactory (4)	7 (17.5%)	5 (12.5%)	4 (10.0%)	4 (10.0%)	7 (17.5%)	5 (12.5%)	3 (7.5%)	6 (15.0%)	4 (10.0%)	4 (10.0%)	4 (10.0%)	4 (10.0%)	6 (15.0%)	6 (15.0%)	6 (15.0%)	8 (20.0%)
Fair (3)	6 (15.0%)	2 (5.0%)	1 (2.5%)	5 (12.5%)	2 (5.0%)	2 (5.0%)	2 (5.0%)	5 (12.5%)	3 (7.5%)	2 (5.0%)	2 (5.0%)	5 (12.5%)	5 (12.5%)	1 (2.5%)	1 (2.5%)	5 (12.5%)
Poor (2)	1 (2.5%)	0	0	2 (5.0%)	0	0	0	0	2 (5.0%)	2 (5.0%)	2 (5.0%)	0	1 (2.5%)	0	0	1 (2.5%)
Very poor (1)	0	0	0	4 (10.0%)	0	0	0	2 (5.0%)	0	0	0	2 (5.0%)	0	0	0	1 (2.5%)
Median (IQR)	5 (4–6)	6 (5–7)	6 (5–7)	5 (3–6)	6 (5–7)	6 (5–7)	6 (5–7)	5 (4–6)	6 (5–7)	6 (5–7)	6 (5–7)	6 (4–7)	5 (4–6)	6 (5–7)	6 (5–7)	5 (4–6)

Note: Ratings of 5 radiologists are reported as absolute numbers (frequencies).

IQR, interquartile range.

DISCUSSION

This experimental study investigated the potential of postoperative extremity tomosynthesis employing a twin-robotic x-ray system in cadaveric fracture models. Compared with conventional x-ray scans, both diagnostic confidence and the delineation of fracture lines were superior for tomosynthesis scans at 80 kV and 116 kV. Meanwhile, assessability of intra-articular screw placement was considered higher even with an ultra-low-dose tomosynthesis protocol at 60 kV. No substantial difference was ascertained between conventional x-ray scans and either tomosynthesis protocol regarding implant positioning.

Our results are in line with the findings of De Silvestro et al,²² who concluded that the overlap of cortical bone and osteosynthetic implants can be reduced and the delineation of fracture margins improved in tomosynthesis compared with radiography. This also concurs with Göthlin and Geijer²³ who evaluated the added clinical value of tomosynthesis at 75 kV/90 kV for the detection of implant loosening following total hip arthroplasty. The authors found tomosynthesis to be superior to radiography for visualizing demineralization and osteolysis adjacent to implants. Blum et al¹² postulates that tomosynthesis in the presence of metal implants may even supersede CT without dedicated metal artifact reduction algorithms. Indeed, this is supported by other studies in postoperative settings reporting fewer metal artifacts in tomosynthesis compared with CT scans.^24–26 Furthermore, Tang et al,²⁷ who investigated the potential of metal artifact reduction approaches in tomosynthesis after hip arthroplasty, asserted significantly higher diagnostic accuracy regarding implant loosening in tomosynthesis with metal artifact reduction compared with CT and radiography. Although De Silvestro et al²² postulated a predominance of metal artifacts in tomosynthesis compared with conventional radiography, the authors negate a negative impact on diagnostic assessability. However, it should be borne in mind that optimal scan parameters may vary depending on the implant size, position, and material.¹ Of note, no relevant metal artifacts were detected at the bone-metal interface or in adjacent tissue within the present study. With that being said, the aforementioned authors did not take into account more recent technical developments, particularly the possibilities presented by introducing photon-counting technology and associated metal artifact reduction capabilities.^28–31

Scan parameters of this study were set to a sweep angle of 10° and a frame rate of 13 FPS in order to compare dose-reduced tomosynthesis protocols at different voltage levels (60, 80, 116 kV) to conventional radiographs. Aiming to find the best compromise between radiation dose and image quality, we determined that tomosynthesis images at 80 kV offer superior diagnostic confidence as well as considerably improved delineation of fracture lines and intra-articular screw placement. With a 25% reduction in DAP compared with conventional radiography, tomosynthesis at 80 kV appears to be suitable for optimized assessment after osteosynthesis of the appendicular skeleton. To the authors' best knowledge, this study represents the first investigation using the unique scanner architecture of a multipurpose twin robotic x-ray system for postoperative evaluation of the appendicular skeleton based on a low-dose tomosynthesis protocol. For context, the DAP for an anterior-posterior projection tomosynthesis applied by Becker et al⁵ ranged between 10 and 40 dGy·cm², whereas the mean DAP in the study by De Silvestro et al²² amounted to 9 dGy·cm². Noël et al³² demonstrated a 25% dose reduction in biplane tomosynthesis of the wrist compared with a 5-view radiographic examination (0.72 vs 0.96 mGy), reportedly amounting to a 28-fold radiation exposure reduction to a CT scan (19.8 mGy). Given the relatively low radiosensitivity of the appendicular skeleton, high radiation exposures have a significantly lower radiobiological impact on the extremities compared with dose-sensitive visceral organs. However, even if dose reduction only plays a subordinate role in imaging of the peripheral skeletal system, in light of radiation protection efforts, reducing the applied radiation dose to a minimum ought to be maintained as a major goal. On the other hand, tomosynthesis will not achieve the same depth, contrast, and spatial resolution as CT,²⁷ rendering the latter irreplaceable in complex cases. Thus, tomosynthesis should not primarily be seen as an alternative to CT examinations but rather be understood as a supplement or even alternative to conventional radiography, as it can be subsequently performed in terms of a one-stop-shop approach when employing adequate scanner hardware.^12,22 In terms of cost-effectiveness, reduced radiation exposure, improved spatial resolution beyond standard radiographs, and artifact reduction, tomosynthesis can be classified as a potent alternative to conventional radiography in postoperative imaging.³³ Beyond that, tomosynthesis may offset the need for additional magnetic resonance imaging or CT examinations and reduce diagnostic imaging costs while accelerating patient care management and workflows.³⁴ This is of particular interest in low-resource settings as well as in peripheral urgent care units, where CT scanner capacities are limited.

Some limitations should be discussed regarding this cadaveric study. First, tomosynthesis scans were not compared with CT due to the latter's significantly higher radiation exposure. Second, only peripheral appendicular body regions were investigated. Third, precise information regarding the donors' age and medical history were not available. Furthermore, the extent of bone demineralization due to formalin fixation remained unknown, theoretically hampering the detailed visualization of bone microarchitecture and fine fracture lines, thus potentially representing confounders or impairing reproducibility of results. Fourth, due to the chosen cadaveric study design, movement artifacts, which may occur in a clinical routine, do not play a role that could not be assessed in the current investigation. Therefore, drawing conclusions from the current cadaveric study with regards to a real-world population requires a certain degree of extrapolation. Future studies are warranted to evaluate the potential of the investigated tomosynthesis settings to further improve patient care. Fifth, it lies in the nature of the study design that no hardware complications were present, preventing an analysis regarding the detection rate of implant failure or loosening. Lastly, no intervendor comparison was performed as the prototypical tomosynthesis mode of only 1 gantry-free multipurpose x-ray system was examined.

CONCLUSIONS

In the postoperative workup of extremity fractures, tomosynthesis allows for superior delineation of fracture lines and improved evaluation of intra-articular screw placement, as well as greater diagnostic confidence assessing the appendicular skeleton after performed osteosynthesis at radiation doses comparable to conventional radiographs.