Accuracy of biplanar linear radiography versus conventional radiographs when used for lower limb and implant measurements

Towards a better understanding of knee angular deformities: discrepancies between clinical examination and 2D/3D assessments

INTRODUCTION

Discrepancy between the clinical examination and the 2D/3D radiographs is a common concern in patients with angular or rotational deformities of the lower limbs, as it may alter clinical judgment and subsequent treatment. The aim was to identify such discrepancies and assess determinants that may contribute to their existence.

MATERIALS AND METHODS

A retrospective chart review was conducted on 329 consecutive patients (658 lower limbs) who underwent physical examination and long-leg biplanar radiographs in our institution between 2013 and 2018 for limb length discrepancy or angular deformity of the knees (varus/valgus). Eleven parameters were measured on 2D and 3D images. 3D measurements were based on standing biplanar X-rays and their 3D reconstructions and were considered the gold standard. Contingency tables and multiple linear regression were used to assess discrepancies between the three modalities and their determinants respectively.

RESULTS

Significant mismatches were found between physical examination and 2D images (1% in varus and 1% in valgus), between physical examination and 3D assessment (1% in varus and 4.6% in valgus) as well as between 2 and 3D assessments (1.9% in varus and 7.6% in valgus). The significant determinants of the mismatch between 2 and 3D modalities were frontal pelvic obliquity, neck shaft angle, knee flexion, femoral torsion, and tibial mechanical angle.

CONCLUSION

In the presence of positional and/or morphological deformities, physical examination and 2D assessment of knee alignment could be biased due to axes projection errors. A better understanding of 3D alignment of the knee as part of the entire lower limb from pelvis to toes, may lead to a better diagnosis and subsequently a better treatment of knee angular deformities.

Normative Femoral and Tibial Lengths in a Modern Population of Twenty-First-Century U.S. Children

BACKGROUND

The Green-Anderson (GA) leg-length data remain the gold standard for the age-based assessment of leg lengths in children despite their methodologic weaknesses. We aimed to summarize current growth trends among a cross-sectional cohort of modern U.S. children using quantile regression methods and to compare the median femoral and tibial lengths of the modern U.S. children with those of the GA cohort.

METHODS

A retrospective review of scanograms and upright slot-scanning radiographs obtained in otherwise healthy children between 2008 and 2020 was completed. A search of a radiology registry revealed 3,508 unique patients between the ages of 2 and 18 years for whom a standard-of-care scanogram or slot-scanning radiograph had been made. All patients with systemic illness, genetic conditions, or generalized diseases that may affect height were excluded. Measurements from a single leg at a single time point per subject were included, and the latest available time point was used for children who had multiple scanograms made. Quantile regression analysis was used to fit the lengths of the tibia and femur and overall leg length separately for male patients and female patients.

RESULTS

Seven hundred patients (328 female and 372 male) met the inclusion criteria. On average, the reported 50th percentile tibial lengths from the GA study at each time point were shorter than the lengths in this study by 2.2 cm (range, 1.4 to 3.3 cm) for boys and 2 cm (range, 1.1 to 3.1 cm) for girls. The reported 50th percentile femoral lengths from the GA study at each time point were shorter than the lengths in this study by 1.8 cm (range, 1.1 to 2.5 cm) for boys and 1.7 cm (range, 0.8 to 2.3 cm) shorter for girls.

CONCLUSIONS

This study developed new growth charts for femoral and tibial lengths in a modern U.S. population of children. The new femoral and tibial lengths at nearly all time points are 1 to 3 cm longer than traditional GA data. The use of GA data for epiphysiodesis could result in underestimation of expected childhood growth.

LEVEL OF EVIDENCE

Prognostic Level IV . See Instructions for Authors for a complete description of levels of evidence.

Accuracy, inter- and intrarater reliability, and user-experience of high tibial osteotomy angle measurements for preoperative planning: manual planning PACS versus semi-automatic software programs

Purpose

To compare the accuracy, inter- and intrarater reliability, and user-experience of manual and semi-automatic preoperative leg-alignment measurement planning software for high tibial osteotomy (HTO).

Methods

Thirty patients (31 lower limbs) who underwent a medial opening wedge HTO between 2017 and 2019 were retrospectively included. The mechanical lateral distal femur angle (mLDFA), mechanical medial proximal tibial angle (mMPTA), and planned correction angle were measured on preoperative long-leg full weight-bearing radiographs utilising PACS Jivex Review® v5.2 manual and TraumaCad® v2.4 semi-automatic planning software. Independent measurements were performed by four raters. Two raters repeated the measurements. Accuracy in the standard error of measurement (SEM), inter- and intrarater reliability, and user-experience were analysed. Additionally, measurements errors of more than 3° were remeasured and reanalysed.

Results

The SEMs of all measured varus malalignment angles and planned correction angle were within 0.8° of accuracy for both software programs. Measurements utilising the manual software demonstrated moderate interrater intraclass correlation coefficient (ICC)-values for the mLDFA and mMPTA, and an excellent interrater ICC-value for the correction angle (0.810, 0.779, and 0.981, respectively). Measurements utilising the semi-automatic software indicated excellent interrater ICC-values for the mLDFA, mMPTA, and correction angle (0.980, 0.909, and 0.989, respectively). The intrarater reliability varied substantially per angle, presenting excellent intrarater agreements by both raters (ICC >  0.900) for the correction angle in each software program as well as poor-to-excellent ICC-values for the mLDFA (0.282–0.951 and 0.316–0.926) and mMPTA (0.893–0.934 and 0.594–0.941) in both the manual planning and semi-automatic software. Regarding user-experience, semi-automatic software was preferred by two raters, while the other two raters had no distinctive preference. After remeasurement of five outliers, excellent interrater ICC-values were found for the mLDFA (0.913) and mMPTA (0.957).

Conclusions

Semi-automatic software outperforms the manual software when user-experience and outliers are considered. However, both software programs provide similar performance after remeasurement of the human-related erroneous outliers. For clinical practice, both programs can be utilised for HTO planning.

Level of evidence

Diagnostic study, Level III.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40634-022-00475-x.

Applications and Error Ratios of Calibration Techniques in EOS, Orthoroentgenogram, and Teleoroentgenogram for Length Measurement: A Comparative Study

BACKGROUND

Accurate length measurements of extremity bones are essential in treating limb deformities and length discrepancies in children.

OBJECTIVE

This study aimed to determine errors in common techniques used to measure lower limb lengths in children.

METHODS

Precision and instrument errors in length measurements were studied utilizing electro-optical system (EOS), orthoroentgenogram, and teleoroentgenogram The goal was to measure a 70-cm metallic rod phantom (average length of the lower extremity of a 10-year-old boy in the 50th percentile) in 3 phases. In Phase 1, the length measurements were performed in an EOS unit with internal calibrations, a magball/magstrip in various scan positions, and measurement with TraumaCAD software. In Phase 2, the measurements were repeated utilizing a single radiation "shot" teleoroentgenogram. In Phase 3, an orthoroentgenogram was utilized with a radiopaque ruler reference. The reliability and validity of measurements were calibrated by 4 physicians (a radiologist, senior orthopaedic attending, and 2 orthopaedic fellows).

RESULTS

EOS measurements utilizing internal references had excellent accuracy (for a 700-mm real length, magnification error (ME)] of 0.09%. Teleoroentgenogram with a magball reference and measurements performed with automatic calibration by TraumaCAD program results in ME of 1.83% with insignificant intraobserver/interobserver difference. Teleoroentgenogram with a magball or magstrip reference measured manually showed that the magball has higher intraobserver/interobserver variance than magstrip, with a 6.60 and 0.33-mm SD, respectively. The length by manual measurement utilizing the magstrip has ME of 2.21%. Orthoroentgenogram is accurate with ME of 0.26%, but does not allow anatomical analysis and is also radiation-costly.

CONCLUSION

EOS and orthoroentgenogram are very accurate for length measurements. Teleoroentgenogram is less accurate in measuring length; however, addition of an external reference (magball, magstrip) placed lateral to the target improves accuracy. Automatic calibration with computer-based analysis of the external reference improves the accuracy more than manual calibration. If manual calibration is utilized, the length measurement is less accurate with the magball than the magstrip.

LEVEL OF EVIDENCE

Level II-comparative in vitro study.

Influence of axial limb rotation on radiographic lower limb alignment: a systematic review

INTRODUCTION

The influence of limb malrotation on long-leg radiographs (LLR) is frequently discussed in literature. This systematic review aimed to describe the influence of limb rotation on alignment measurements alone and in combination with knee flexion, and determine its clinical impact.

MATERIALS AND METHODS

A literature search was conducted in June 2021 using the databases MEDLINE, Cochrane, Web of Science (Clarivate Analytics), and Embase. The search term ((radiograph OR X-ray) AND (position OR rotation) AND limb alignment) was used. Database query, record screening, and study inclusion and exclusion were performed by two reviewers independently. Experimental studies (using either specimens or synthetic bones) or clinical studies (prospective or retrospective using radiographs of patients) analyzing the influence of limb rotation on anatomic and mechanical limb alignment measurements were included. Characteristics and results of the included studies were summarized, simplified, and grouped for comparison to answer the research question. Studies were compared descriptively, and no meta-analysis was performed.

RESULTS

A total of 22 studies were included showing large heterogeneity, comprising studies with cadavers, patients, and synthetic bones. Most studies (7 out of 8) reported that external rotation (ER) causes less apparent valgus and leads to more varus and internal rotation (IR) causes more valgus and leads to less varus. However, there is no consensus on the extent of rotation influencing alignment measures. Studies reported about an average change of > 2° (n = 4) and < 2° (n = 4) hip-knee-ankle angle (HKA) between 15°IR and 15°ER. There is a consensus that the impact of rotation on mechanical alignment is higher if additional sagittal knee angulation, such as knee flexion, is present. All five studies analyzing the influence of rotation combined with knee flexion (5°-15°) showed an HKA change of > 2° between 15°IR and 15°ER.

CONCLUSION

Malrotation is frequently present on LLR, possibly influencing the measured alignment especially in knees with extension deficit. Surgeons must consider this when measuring and treating deformities (high tibial osteotomy or total knee arthroplasties), and analyzing surgical outcomes. Especially in patients with osteoarthritis with knee extension deficits or postoperative swelling, the effect of malrotation is significantly greater.

Distance from the magnification device contributes to differences in lower leg length measured in patients with TSF correction

INTRODUCTION

In absence of deformity or injury of the contralateral leg, the contralateral leg length is used to plan limb lengthening. Length variability on long-leg weight-bearing radiographs (LLR) can lead to inaccurate deformity correction. The aim of the study was to (1) examine the variability of the measured limb length on LLR and (2) to examine the influence of the position of the magnification device.

MATERIALS AND METHODS

The limb lengths of 38 patients during deformity correction with a taylor-spatial-frame were measured retrospectively on 7.3 ± 2.6 (4-13) LLR per patient. The measured length of the untreated limb between LLR were used to determine length variability between LLR in each patient. To answer the secondary aim, we took LLR from a 90 cm validation distance. A magnification device was placed in different positions: at the middle of the 90 cm distance (z-position), 5 cm anterior and 5 cm posterior from the z-position, at the bottom and top of the validation distance as well as 5 cm medial and 15 cm lateral from the z-position.

RESULTS

The measured length variability ranged within a patient from 10 to 50 mm. 76% of patients had a measured limb length difference of ≥ 2 cm between taken LLR. Compared to length measurement of the 90 cm test object with the magnification device in the z-position (90.1 cm), positioning the device 5 cm anterior led to smaller (88.6 cm) and 5 cm posterior led to larger measurements (91.7 cm). The measured length with the magnification device at the bottom, top, medial or lateral (90.4; 89.9; 90.2; 89.8 cm) to the object differed not relevantly.

CONCLUSIONS

High variability of limb length between different LLR within one patient was observed. This can result from different positions of the magnification device in the sagittal plane. These small changes in positioning the device should be avoided to achieve accurate deformity correction and bone lengthening. This should be considered for all length and size measurements on radiographs.