The influence of pre-operative Computed Tomography (CT) on surgical approach and fixation for fractures of the tibial plateau

Abstract

Background

It is well known that a computed tomography (CT) scan improves the classification of tibial plateau fractures (TPF) compared with radiographs. However, it is less clear how this translates into clinical practice. The aim of this study is to establish to what extent a pre-operative CT scan alters the approach, setup and fixation choice in TPF compared to radiographs.

Methods

50 consecutive TPF with a preoperative CT and radiographic imaging available, were assessed by 4 consultant surgeons. First, anonymised radiographs were classifying according to the column classification and the planned setup, approach, and fixation technique documented. At a 1-month interval, randomised matched CT scans were assessed and the same data collected. A tibial plateau disruption score (TPDS) was derived for all 4 quadrants (no injury = 0, split = 1, split/depression = 2 and depression = 3). Radiograph and CT TPDS were assessed using an unpaired T-test.

Results

26 female and 24 male patients, mean age 50.3, were included. Mean TPDS on radiographs and CT scans were 2.77 and 3.17 respectively. A significantly higher CT TPDS, of 0.4 (95%CI 0.10–0.71)[P = 0.0093] was observed, demonstrating that radiographs underestimate the extent of injury. The surgical approach changed in 28.5% of cases, thus influencing a change in the patient setup in theatre in 27% of cases. Identification of fractures within a column changed in 34% of cases. A high intra-observer reliability was observed when surgeons were asked to repeat their assessment in a third round at a further one month interval.

Conclusion

A pre-operative CT scan has a significant effect on the approach required to fix TPF. This therefore influences the setup of the patient and can justifiably be requested as part of pre-operative planning.

Highlights

• •The addition of a CT scan, compared to radiographs, changed the setup (27%), approach (28.5%) and location of plates (34%) of cases.
• •A high intra-observer reliability was observed between surgeons and between rounds when CT images were used to form an operative plan.
• •We feel that the addition of a CT scan influences setup, approach and fixation method and is clinically justified prior to surgery in fractures of the tibial plateau.

1. Introduction

Tibial plateau fractures (TPFs) are common injuries managed on the un-selected orthopaedic take. They comprise 1% of all fractures with an incidence of 10.3 per 100,000 people^1,2. Schatzker's widely accepted classification is based on the AP radiograph³. However, the increased availability of a CT scan allows a 3-dimensional (3D) appreciation of these fractures. Luo's 3-column concept, dividing the tibial plateau into medial, lateral, and posterior columns is now seen as the gold standard when planning approach and fixation⁴. A validated, modification by Chang et al. further splits the posterior column into posteromedial and posterolateral columns^5,6. This is an important distinction as it identifies the posterolateral column fracture, a known treatment challenge⁷. Correct identification of fracture morphology allows fracture-specific fixation, thus affecting the approach, setup, and plate choice by the operating surgeon⁸.

A CT scan may require an additional hospital visit, increase the cost of the treatment pathway, and potentially cause a delay in surgery. The Radiation dose for a CT Knee is low, estimated to be 0.16 mSv, equivalent to 24 days of background radiation⁹. For comparison a Chest Radiograph dose is 0.02 mSv, this is equivalent to 3 days of background radiation¹⁰.

We already know that adding a CT scan to orthogonal radiographs improves accuracy when classifying fractures of the tibial plateau, though the use of 3D reconstructions has shown no additional improvement in classification11, 12, 13, 14. However, it is less clear how this actually translates into clinical practice. There is little published in this regard. The purpose of this study is to establish: (1) to what extent does a CT alter the approach, setup, and fixation strategy, (2) if surgeons can accurately assess fracture morphology on radiographs, (3) can surgeons accurately predict which cases require a CT scan pre-operatively.

2. Materials and Methods

All TPFs managed in a busy single trauma unit, in the National Health Service, UK, were screened retrospectively using the department's fracture registry. The medical record numbers were then correlated with the Picture Archiving and Communication System (PACS) to identify patients who had both a preoperative CT scan and subsequent definitive fixation.

Inclusion criteria were all patients over 16 years of age, who had received fixation for their TPF in our unit with either open reduction and internal fixation or ring external fixation. Exclusion criteria were patients under 16 years of age, periprosthetic fractures, and pathological fractures. Open fractures were triaged pre-hospital to the regional Major Trauma Centre, therefore were not included in this study. No exclusions were made based on fracture classification or injury mechanism as any results should be generalizable to the population presenting to a busy Trauma Unit. A consecutive series of fifty patients, backdated from May 2020, were selected and each case was pseudo-anonymised by the author (TF) (not involved in data analysis nor image interpretation).

Four assessors, all consultant trauma and orthopaedic surgeons (CC, DE, HG, and AD), with regular practice in the fixation of tibial plateau fractures and a mean of 29 years post-graduate practice, performed image interpretation. The study was designed around three rounds of imaging assessment, at one-month intervals. Analysis was standardized with the published descriptions of the column classification and quadrant modification provided to each surgeon. In the first round, for each study number, anteroposterior and lateral radiographs only were presented for analysis to each of the assessors. The following data was recorded; identification of the fracture morphology using the quadrant modification of the column classification, including a description of the fracture within each quadrant (no fracture, split, split depression or depression), theatre setup, and patient position (supine, prone, prone then supine or lateral position), the planned approach or approaches and plate location according to the quadrant of the fracture were recorded or whether the surgeon would elect to manage this case non-operatively or refer on for consideration of a circular frame. Finally, a question was asked whether the surgeon would be happy to proceed operatively based on the radiographs alone, without a pre-operative CT scan.

At a one-month interval, randomized, matched CT scans of the TPFs were assessed and the same data collected, in round two. Randomization was performed to minimize any association or recall from the original analysis. All CT scans contained coronal, sagittal, and axial reconstructions only, 3D reconstructions were not analysed. To assess intraobserver reliability, at a further month's interval, a randomized selection of 25 cases had repeat analysis and a preoperative plan made from both radiographs and CT scans. The raw agreement was calculated for each observer (1 = total agreement between rounds, 0 = any change and therefore no agreement). Analysis was performed using R 4.0.1 (R Foundation for Statistical Computing, Vienna), using the packages "tidyverse" and "ggplot".

To assess the complexity of injury on both radiographs and CT scans, a Tibial Plateau Disruption Score (TPDS) was derived for each case. The tibial plateau was split into quadrants; lateral, medial, posterolateral, and posteromedial, and a number assigned to each quadrant based on injury morphology. No quadrant injury = 0, quadrant split = 1, quadrant split + depression = 2, quadrant depression = 3. With four quadrants in total, the maximum potential score was 12. Therefore a higher tibial plateau disruption score (TPDS) indicates an overall greater injury to the tibial plateau. Each radiograph and CT scan was assigned a TPDS for each of the four observers. Radiographs and CT TPDS were assessed using an unpaired T-test. Data were analysed using 2X2 contingency tables between 2 observers, assessing agreement to proceed with or without a CT scan, categorical data was analysed using a 2-way Fisher's Exact Test with an alpha value of <0.05 considered significant. Statistical analysis was performed by an author (AT) independent of image interpretation and case randomization using GraphPad Prism 9.0. ANOVA was used to assess the Interobserver reliability for the TPDS in each modality, radiographs, and CT scans, between all 4 of the assessors. P < 0.05 was considered significant.

3. Results

There were twenty-six females and 24 males included in this study with a mean age of 50.3 years (range 16–76). A higher Tibial Plateau Disruption Score was identified on a CT scan (mean TPDS = 3.17) compared to radiographs (mean TPDS = 2.77), indicating a higher grade of injury, shown in Fig. 1. This score is derived from analysis of the same patient, same surgeon, comparing round one (XR) to round two (CT). Values for each surgeon are shown in Table 1. A higher CT TPDS was seen across all observers with a mean difference of 0.41 (95% CI 0.10–0.71)[P = 0.0093], albeit with just one surgeon reaching statistical significance.

Fig. 1.

Mean radiograph and CT TPDS for each observer showing a higher TPDS on CT scan compared to radiograph for all observers.

Table 1.

Radiograph and CT TPDS were assessed using a two-tailed unpaired T-test for each observer. An alpha value of <0.05 was considered statistically significant.

	Mean Radiograph TPDS	Mean CT TPDS	Net Difference (95% CI)	P Value
Obs 1	2.78	3.66	0.88 (0.23-1.53)	0.0088
Obs 2	2.66	3.04	0.38 (−0.21-0.97)	0.205
Obs 3	3.34	3.48	0.14 (−0.61-0.89)	0.711
Obs 4	2.28	2.50	0.22 (−0.082-0.52)	0.151
All Obs	2.77	3.17	0.41 (0.10-0.71)	0.0093

Overall there was a 34% variation, between the plain film radiograph (round one) and the CT scan (round two), in the anatomical location of the exiting fracture line of the tibial plateau fracture requiring fixation. Interestingly, there was the same number of quadrants containing a fracture by all observers (236/800 [29.5%]) in both rounds. It was the anatomical location, within which quadrant the fracture line breached the cortex, which varied between rounds. Fig. 2 demonstrates an example of a case where the posteromedial fracture is evident on a CT scan but difficult to identify on the radiograph. This difference in fracture morphology correlated with an alteration in the chosen approach to stabilize the TPF. A 28.5% change (57/200 cases, each of the 50 patients analysed by four surgeons) in approach was seen, demonstrated in Table 2, this was mirrored with a 27.4% (43/200 cases) change in patient positioning in theatre. In 30/50 cases, all four surgeons made no change in their planned patient positioning. Only in one case did the CT scan change the planned patient positioning for all surgeons, whilst in 19 cases some but not all of the four surgeons changed their planned patient positioning between the pre-operative radiograph and CT scan.

Fig. 2.

Example imaging showing a posterior shear component that was not immediately apparent on the original radiograph. All surgeons opted for a lateral incision and plate in the supine position in round one. In round two, three surgeons opted to fix the posterior fracture, two altered position and three changed their approach.

Table 2.

Surgical approach based on the XR and subsequently the CT. The total number of approaches is > 50 reflecting the nature for dual incisions to address complex fractures.

	Anterolateral approach	Medial approach	Posteromedial approach
Obs 1 XR	44	3	9
Obs 1 CT	42	9	10
Obs 2 XR	42	13	9
Obs 2 CT	40	7	7
Obs 3 XR	42	14	9
Obs 3 CT	44	9	9
Obs 4 XR	46	6	7
Obs 4 CT	43	7	10

To assess how accurately surgeons can predict which cases can go to the theatre without a CT scan, we looked at the agreement after round one on which cases surgeons felt a CT would not be required. There was agreement from all surgeons in just 11 out of 50 cases that no CT scan was required before surgery. There were no cases in which all four surgeons could agree a pre-operative scan was required. The observer agreement is displayed in Table 3.

Table 3.

The agreement between each observer is shown above, text in bold it statistically significant.

Observer	Agreement (%)	Fisher's Exact Test (2-sided test) P value
Obs 1 vs Obs 2	64	0.309
Obs 1 vs Obs 3	48	0.383
Obs 1 vs Obs 4	38	1
Obs 2 vs Obs 3	60	0.0475
Obs 2 vs Obs 4	50	0.179
Obs 3 vs Obs 4	74	0.0019

When comparing surgeons, there was significant interobserver variability as to the grade and location of tibial plateau fracture on both radiographs (P < 0.0001), (R squared 0.155 and F 5.369) and CT scans (R squared 0.233 and F 7.192). Intraobserver reliability was strongest when observers were asked to repeat their operative plan in round three using CT scans demonstrated in Fig. 2.

4. Discussion

Our study shows that the addition of a CT scan significantly influences the preoperative plan in TPFs, hypothesis (1). We have shown that the surgical approach to address a TPF fracture changed in 28.5% of cases, correlating closely with a change in the setup of the patient in 27.4%. This degree of change directly affects surgical practice and has the potential to affect patient outcomes. A failure to fully appreciate the fracture pre-operatively can result in changing the approach or patient position and plate choice, increasing surgical time and potential complications. With appropriate fracture reduction and adequate fixation, enough stability can be conferred to permit an early functional range of motion and weight bearing can be achieved, permitting better post-operative rehabilitation^15,16.

When assessing how accurately surgeons can assess fracture morphology, our study has shown a 34% change within which quadrant the fracture lines were identified between radiographs and the respective CT scan. In addition, a significantly higher TPDS has been demonstrated on CT scan (3.17) compared to radiographs of the same patient (2.77), hypothesis (2). Fracture pattern in TPFs in the literature is underestimated in up to 43% of cases on radiographs when compared with CT scans¹⁷. This correlates with our findings from this study, albeit to a lesser degree. Depression and displacement are important factors in affecting the outcome, if missed they can result in coronal plane deformity and joint incongruity¹⁸. One study found that a third of OTA-type C-type bicondylar tibial plateau fractures had a posteromedial shear fragment¹⁹. A large post-reduction depression in the articular surface of the tibial plateau has been shown to closely correlate with a reduced range of motion¹⁶.

Importantly, our study showed that in only 11 out of 50 cases, all surgeons could agree that a CT scan was required before surgery, hypothesis (3). There was a large discrepancy between surgeons, highlighting the difficulties in accurately predicting which cases did or did not require a CT scan based on radiographs alone. This demonstrates the variability in assessment on radiographs and highlights the importance of obtaining a pre-operative CT scan for TPFs.

To our knowledge, since the advent of column classification and column-specific fixation, this is the first study that has analysed how a CT scan affects the operative plan in fractures of the tibial plateau. By accurately identifying the location of fracture lines a surgeon can best plan the setup and the approach to optimally reduce and stabilize these fractures. Our high intra-observer reliability on CT scan analysis showed consistent findings when revisiting the same images at a later date.

Results should be interpreted within the limitations of the study. All four surgeons work at the same institution, therefore may exhibit similar patterns of practice. However, each worked through the cases independently in a randomized order, to mitigate any external influence. The study is retrospective in nature and we cannot deduce from this a direct link between operative plan and patient outcome. This study also does not address whether a CT scan has influenced the decision to manage a patient operatively or non-operatively, further work is required in this regard. Therefore the use of a CT scan in assessing un-displaced fractures of the tibial plateau on XR cannot be made from this study. The strengths are that our population is generalizable to those cases that present to a busy Trauma Unit and are consecutive, chosen to reduce selection bias. Over half (26/50) of the cases were complex, involving at least two quadrants, therefore representing a heterogenous range of TPF injuries from simple splits to four quadrant injuries.

5. Conclusion

In conclusion we have demonstrated that a preoperative CT scan not only identifies a higher grade of injury and more accurately determines the location of fracture lines but also alters the approach, setup, and fixation method in tibial plateau fractures. We believe a CT scan should justifiably be requested as part of routine pre-operative workup in these fractures.

Declaration of competing interest

Conflict of interest, none. This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors. Artificial Intelligence was not used in this manuscript. No Ethical committee approval was required.