1. Keivan Abtahi, DO – Orthopaedic & Trauma Service of Oklahoma, Tulsa, Oklahoma
2. Jake Checketts, DO – Oklahoma State University Medical Center, Department of Orthopaedic Surgery, Tulsa, Oklahoma
3. Conner Howard, BS – Oklahoma State University Center for Health Sciences, Tulsa, Oklahoma
4. Byron Detweiler, DO – Oklahoma State University Medical Center, Department of Orthopaedic Surgery, Tulsa, Oklahoma
5. Zachary Nollin, DO – Integris Baptist Medical Center, Oklahoma City Oklahoma, Oklahoma
6. Nasir Mushtaq, PhD – Office of Research Development at the University of Oklahoma – Tulsa, Oklahoma
7. Brent Norris, MD – Orthopaedic & Trauma Service of Oklahoma, Tulsa, Oklahoma

Abstract

Introduction
Bone graft harvesting from the intramedullary (IM) femoral cavity using the Reamer Irrigator Aspirator (RIA) has grown in popularity over the past two decades. Harvesting the bone from the IM canal requires the surgeon to slightly oversize the femoral isthmus by 1-1.5mm. The RIA ruler is used with C-arm fluoroscopy to measure the femoral canal during the operation. Our study aims to compare the accuracy of the RIA ruler with two other measurements: CT derived femoral diameter and direct anatomic femur measurements.

Methods
Six fresh frozen human cadavers were used in this study. C-Arm fluoroscopy was used to visualize the femoral isthmus and measure the diameter using the RIA ruler in the AP and lateral planes. The measurements were recorded for estimated size. CT scans of each femur were performed, followed by transection and direct caliper measurements in the AP and lateral planes.

Results
The RIA ruler measurements consistently overestimated the femoral canal diameter by approximately 1mm in the AP plane, with poor to moderate reliability between surgeons. There was a significant difference between the RIA ruler measurement and CT canal measurement in the AP plane (p=0.0107) and between the RIA ruler and canal caliper in both lateral plane and mean measurements (p<0.0001). Inter-rater reliability for RIA ruler measurements by investigators was also found to be poor to moderate in both the AP and lateral planes.

Conclusion
To prevent exceeding the suggested limits of reaming, surgeons using the RIA ruler should reduce their reamer head selection to 1mm over the fluoroscopic measurement and use a +/- 1.5 mm tolerance to stay within an adequate range of bone harvesting.However, staying within a 1.5 mm tolerance is difficult for surgeons of all experience levels, highlighting the necessity of developing a more precise tool for measuring the femoral canal intraoperatively. Until a new tool is developed, caution should be exercised when using the RIA ruler, and measurements should be interpreted with care.

Keywords: Hip; Femur; Fracture; Intramedullary; Autogenous; Bone Graft; Femoral Canal Measurement

Introduction
Autogenous bone graft is a common surgical technique used for the treatment of nonunion bone defects, and arthrodesis. While iliac bone graft harvesting has traditionally been used for autogenous bone grafts, this technique has been associated with significant complications, such as fracture, blood loss, and persistent pain.¹ A newer option is the Reamer Irrigator Aspirator (RIA) system, which generates a large volume of morselized cortical/cancellous bone graft material in a single-step reaming process. The RIA system has demonstrated advantages such as abundant bone graft and potentially less pain and morbidity;^2–4 however, has also been associated with iatrogenic femur fractures and blood loss.^5,6

When using RIA, the initial step is to determine the size of reamer head. An intraoperative measurement of the femoral canal is taken using the RIA ruler in conjunction with C-arm fluoroscopy. The measurement is used to determine the size of the reamer head, which is the part of the device that cuts the intramedullary bone for harvesting. Selecting an undersized reamer head may lead to limited bone graft harvest, while an overestimation of size increases the risk of incarceration and femur fracture.^2–6

The accuracy of the RIA ruler in determining the femoral canal size is unclear. The authors believe it provides a rough estimate of the femoral canal, at best. Our study aims to compare femoral canal measurements taken from the RIA ruler to those taken from CT scans and direct anatomic caliper measurements to evaluate the effectiveness of the RIA ruler as a measurement tool.

Materials and Methods
We first obtained IRB approval for this study on fresh frozen anatomic lower extremity specimens. The St. John Research Institute helped us with storing and maintaining the specimens during the course of the data collection. Six fresh frozen human lower extremity cadaver specimens were obtained from the University of Oklahoma (OU) Anatomy Board. The specimens had no prior orthopaedic intervention to the femur. Each cadaveric femur was then thawed and taken to the OR for placement of three (2.0 mm) percutaneous reference wires in the region of the isthmus as was determined by AP plane fluoroscopy. Each wire was separated by approximately 3 cm. Each reference point was then measured with the RIA ruler on the distal aspect of the reference wire on both the AP and lateral image with the fluoroscope. The femurs were then taken to the CT scanner and a MSK radiologist measured the femoral canal just below each reference wire in the AP and lateral planes. The specimens were then brought back to the OR where each femur was then dissected down to the region of the reference wires. The thickness of the soft tissue was then measured in the AP and lateral plane using a standard ruler (to assess possible C-arm magnification error given variations in human thigh thickness). The femurs were cut with an oscillating saw just distal to each reference wire and a direct digital caliper measurement was taken at each reference point in the AP and lateral plane.

Statistics
Statistical evaluation of RIA ruler measurements was compared to direct anatomic caliper and CT measurements to assess RIA ruler accuracy. CT measurements were considered gold standard for comparison. Thigh thickness measurements were assessed for possible RIA ruler magnification measurement error with fluoroscopy. RIA ruler measurements were considered correct if they were within 2.5mm of the actual measurement (manufacturer’s upper limit for safe use of reamer head). Exploratory analysis was performed as follows: means and standard deviations were calculated for different measurements, frequency and proportion were calculated for correct RIA ruler measurements (i.e., within 2.5mm and 1.5mm of the actual measurement). Differences between RIA ruler measurements by investigators and actual CT measurements and direct digital caliper (AP and Lateral planes) were evaluated by mixed effects models and paired tests. Bland Altman’s limits of agreement (LOA) were used to determine whether measurement differences were clinically significant. Interclass correlation (ICC) and 95% confidence intervals (95%CI) were calculated to evaluate level of agreement between RIA ruler measurements and actual measurements in the AP and Lateral planes. Association between investigators and proportion of correct measurements (within 2.5mm and 1.5mm of actual measurement) was examined by Fisher’s exact test. ICC was calculated using SPSS version 24, whereas all other analyses were performed using SAS version 9.4, and statistical significance was determined at P < 0.05.

Results
RIA Ruler measurements consistently overestimate femoral canal diameter in the AP plane by approximately 1mm, and reliability between surgeons is largely poor to moderate at best.

We measured femoral canal diameter using different measurement systems (Table 1). Our measurements using CT canal in the AP plane measured at a mean of 12.17mm (+/- 1.62), in the lateral plane measured at a mean of 14.61mm (+/-1.53), and our mean measurement (AP and Lateral) was 13.37mm (+/- 1.56). Our measurements using canal calipers in the AP plane measured at a mean of 12.44mm (+/- 1.72), in the lateral plane measured at a mean of 11.95mm (+/- 1.78), and our mean measurement (AP and Lateral) was 12.19mm (+/- 1.56). Lastly, our measurements using the RIA ruler in the AP plane measured at a mean of 13.00mm (+/- 1.65), in the lateral plane measured at a mean of 14.64mm (+/- 1.45), and our mean measurement (AP and Lateral) was 13.82mm (+/- 1.39). There was a significant difference between RIA ruler measurement and CT canal measurement in the AP plane (p=0.0107). Whereas, Lateral plane and mean measurements were significantly different between RIA ruler and canal caliper (p<0.0001). 

Similarly, all investigators’ RIA ruler measurements significantly differ from canal caliper measurements in the Lateral plane, whereas, in the AP plane, RIA ruler measurements by Junior Attendings were not significantly different from CT canal measurements. (Table 2 and 3). Evaluation of clinical significance of these differences in measurements was performed by Bland Altman’s limits of agreement (95%LOA). The results showed that except for the Lateral measurement by CT and RIA ruler measurements by Fellow, all measurements obtained by RIA ruler were significantly different as the 95%LOA included + 2.5 which was considered as a clinically significant difference (Table 4). 

The results of Interclass correlation evaluating reliability of RIA ruler measurements by each investigators and actual measurement obtained through CT and canal caliper indicated poor reliability for the Lateral plane. However, the reliability of RIA ruler measurements in the AP plane ranged from poor to moderate (Table 4). Inter-rater reliability for RIA ruler measurements by investigators was also found to be poor to moderate for the AP and Lateral planes (ICC 0.66, 95%CI: 0.46, 0.83 and ICC 0.68, 95%CI: 0.49, 0.84, respectively).

We also calculated the percentage of correct measurements (+/- 2.5mm) by 5 separate surgeons at differing levels of training as compared to the CT canal measurement (Table 5). There was no significant difference between investigators and correct measurement. Our Senior Attending measured correctly in the at a rate of 94.44% in the AP plane, 77.78% in the lateral plane, and his overall mean rate of correct measurement was 94.44%. Similarly, both Junior Attending 1 and Junior Attending 2 were correct in their overall mean measurements at a rate of 94.44%. Junior Attending 1 measured in the AP plane correctly at a rate of 94.44%, and was correct in the lateral plane 88.89% of the time. Junior Attending 2 measured in the AP plane correctly at a rate of 88.89%, and was correct in the lateral plane 88.89% of the time. Our Fellow measured correctly in the AP plane at a rate of 88.89%, measured correctly in the lateral plane at a rate of 88.89%, and had an overall mean accuracy of 100.00%. Lastly, our resident measured correctly in the AP plane at a rate of 88.89%, measured correctly in the lateral plane at a rate of 88.89%, and had an overall mean accuracy of 88.89%. When using a +/- 1.5mm tolerance, the percentage of correct measurements fell drastically by all investigators (Table 5). 

Discussion
We appear to overestimate the femoral canal size by approximately 1 mm on the AP image when using the RIA ruler compared to CT scan. Furthermore, RIA ruler measurements are not very accurate compared to direct caliper measurements, especially on the lateral measurement. Using a +/-2.5 mm tolerance (the manufacturer’s upper limit for safe use of the reamer head), we accurately predicted the canal size nearly 90% of the time in the AP plane and slightly less (87%) in the lateral plane. However, when using a +/-1.5mm tolerance our accuracy fell drastically (Table 5), which suggests our measurements even when within the +/- 2.5mm tolerance are just barely so. Additionally, reliability between investigators was poor to moderate. 

Since surgeons often choose a reamer head 1.5 mm above the ruler measurement to assure adequate bone harvest, an overestimation of 1mm (as found in the AP plane of our study) suggests surgeons using the RIA ruler consistently approach the manufacturer’s upper limit of safe reaming of 2.5mm over the isthmus of the femoral canal. These results are significant because many surgeons only attempt to measure the femoral canal via AP imaging as lateral imaging is often difficult to obtain intraoperatively, and recent studies have found the AP view to be more accurate in measuring the femoral canal.⁷ If this is the case, RIA ruler measurements should be interpreted cautiously, and a new more accurate tool for assessing femoral canal width is needed. 

Accurate femoral canal measurement is essential to avoid costly outcomes. Underestimating the canal leads to poor bone harvest, requiring additional reamer heads, increased blood loss, and higher costs.⁵ Overestimating the canal, as demonstrated in our study, while following the manufacturer’s 2.5 mm limit, may result in reamer head incarceration, RIA device failure, or femur shaft fracture.⁵ Magnification errors during measurements taken with the RIA ruler are likely due to the ruler not being fixed to the bone, the variable distance between the ruler and the X-ray beam, and the measuring gauge relying on the surgeon’s subjective impression.⁸ These factors can lead to measurement interpretation errors, increasing the margin of error and adding to variability in accuracy. A new measurement tool is needed to eliminate the inadequacies of the current RIA ruler, reduce the risk of complications such as femur fracture and excessive blood loss, and lower costs associated with multiple reamer head use.

Strengths
Strengths of this paper include our use of three separate methods for measurement of femoral canal. This allows direct comparisons between these methods, and discussion as to the implication of these findings. Furthermore, because both CT and RIA ruler overestimated femoral canal width when compared to direct caliper measurements, our paper supports the need of a more robust method or tool for measuring femoral canal width. In addition, because we used surgeons at differing levels of training, our results are highly generalizable as many levels of experience was evaluated and had similar results. 

Limitations
Limitations of this study include possible magnification error occurring during our measurements due to variation in tissue thickness surrounding the bone. Though we took efforts to prevent this, it is possible this could have affected our measurements. Furthermore, due to the position in which each specimen was set during measurement, true laterals were difficult to asses which could have contributed to variations when using the RIA ruler and CT findings depending on view. Because of this, our mean AP measurements are likely the most accurate. Finally, the direct caliper measurement was made using a digital caliper that made contact with the endosteum of the femur. In thawed specimens this endosteum was fragile and ‘indentable’ possibly adding to some measurement error.

Conclusion
Using the RIA ruler, femoral canal size was commonly overestimated by approximately 1 mm in the AP plane compared to CT measurements. Additionally, RIA ruler measurements were not very accurate compared to direct caliper measurements in the lateral plane. The reliability between investigators was also found to be poor to moderate. The implications of inaccurate measurements can be costly, resulting in poor bone harvest, additional reamer heads, increased blood loss, and higher costs. To prevent exceeding the manufacturer’s suggested upper limit of reaming, surgeons using the RIA ruler should reduce their reamer head selection to 1mm over the fluoroscopic measurement. The authors propose that surgeons use a +/- 1.5mm tolerance instead of the manufacturer’s +/- 2.5mm tolerance to stay within an adequate range of bone harvesting. However, staying within a 1.5mm tolerance is difficult for surgeons of all experience levels, highlighting the necessity of developing a more precise tool for measuring the femoral canal intraoperatively. Until a new tool is developed, caution should be exercised when using the RIA ruler, and measurements should be interpreted with care.

Table 1 | Table 2 | Table 3 | Table 4 | Table 5

References

1. Fowler BL, Dall BE, Rowe DE. Complications associated with harvesting autogenous iliac bone graft. Am J Orthop . 1995;24(12):895-903.
2. Dawson J, Kiner D, Gardner W 2nd, Swafford R, Nowotarski PJ. The reamer-irrigator-aspirator as a device for harvesting bone graft compared with iliac crest bone graft: union rates and complications. J Orthop Trauma. 2014;28(10):584-590.
3. Conway JD, Shabtai L, Specht SC, Herzenberg JE. Sequential harvesting of bone graft from the intramedullary canal of the femur. Orthopedics. 2014;37(9):e796-e803.
4. Cox G, Jones E, McGonagle D, Giannoudis PV. Reamer-irrigator-aspirator indications and clinical results: a systematic review. Int Orthop. 2011;35(7):951-956.
5. Marchand LS, Rothberg DL, Kubiak EN, Higgins TF. Is This Autograft Worth It?: The Blood Loss and Transfusion Rates Associated With Reamer Irrigator Aspirator Bone Graft Harvest. J Orthop Trauma. 2017;31(4):205-209.
6. Dimitriou R, Mataliotakis GI, Angoules AG, Kanakaris NK, Giannoudis PV. Complications following autologous bone graft harvesting from the iliac crest and using the RIA: a systematic review. Injury. 2011;42 Suppl 2:S3-S15.
7. Walsh KP, Fowler JR, Chen O, et al. The validity and reliability of preoperative radiographic canal diameter measurements of the femur. HSS J. 2013;9(2):150-156.
8. Rosenstein DS. IMAGE MAGNIFICATION IN DIGITAL FLUOROSCOPY. Veterinary Radiology Ultrasound. 1998;39(4):304-308. doi:10.1111/j.1740-8261.1998.tb01610.x

Required Disclosures and Declaration

• Copyright Information: No Copyright Information Added
• IRB Approval Information: Not applicable
• Disclosure Information: The authors have no conflicts of interest to report. An AO North America (AONA) Grant was used to fund this work.