Skip to main content
NCBI home page
Clinical Orthopaedics and Related Research logoLink to Clinical Orthopaedics and Related Research
. 2024 May 15;482(10):1896–1908. doi: 10.1097/CORR.0000000000003116

Does Freehand, Patient-specific Instrumentation or Surgical Navigation Perform Better for Allograft Reconstruction After Tumor Resection? A Preclinical Synthetic Bone Study

Harley H L Chan 1,2, Prakash Nayak 3, Ibrahim Alshaygy 4, Kenneth R Gundle 5, Kim Tsoi 6,7,8, Michael J Daly 1,2, Jonathan C Irish 1,2,8,9, Peter C Ferguson 6,7,8, Jay S Wunder 6,7,8,✉
PMCID: PMC11419413  PMID: 38813958

Abstract

Background

Joint-sparing resection of periarticular bone tumors can be challenging because of complex geometry. Successful reconstruction of periarticular bone defects after tumor resection is often performed with structural allografts to allow for joint preservation. However, achieving a size-matched allograft to fill the defect can be challenging because allograft sizes vary, they do not always match a patient’s anatomy, and cutting the allograft to perfectly fit the defect is demanding.

Questions/purposes

(1) Is there a difference in mental workload among the freehand, patient-specific instrumentation, and surgical navigation approaches? (2) Is there a difference in conformance (quantitative measure of deviation from the ideal bone graft), elapsed time during reconstruction, and qualitative assessment of goodness-of-fit of the allograft reconstruction among the approaches?

Methods

Seven surgeons used three modalities in the same order (freehand, patient-specific instrumentation, and surgical navigation) to fashion synthetic bone to reconstruct a standardized bone defect. National Aeronautics and Space Administration (NASA) mental task load index questionnaires and procedure time were captured. Cone-beam CT images of the shaped allografts were used to measure conformance (quantitative measure of deviation from the ideal bone graft) to a computer-generated ideal bone graft model. Six additional (senior) surgeons blinded to modality scored the quality of fit of the allografts into the standardized tumor defect using a 10-point Likert scale. We measured conformance using the root-mean-square metric in mm and used ANOVA for multipaired comparisons (p < 0.05 was significant).

Results

There was no difference in mental NASA total task load scores among the freehand, patient-specific instrumentation, and surgical navigation techniques. We found no difference in conformance root-mean-square values (mean ± SD) between surgical navigation (2 ± 0 mm; mean values have been rounded to whole numbers) and patient-specific instrumentation (2 ± 1 mm), but both showed a small improvement compared with the freehand approach (3 ± 1 mm). For freehand versus surgical navigation, the mean difference was 1 mm (95% confidence interval [CI] 0.5 to 1.1; p = 0.01). For freehand versus patient-specific instrumentation, the mean difference was 1 mm (95% CI -0.1 to 0.9; p = 0.02). For patient-specific instrumentation versus surgical navigation, the mean difference was 0 mm (95% CI -0.5 to 0.2; p = 0.82). In evaluating the goodness of fit of the shaped grafts, we found no clinically important difference between surgical navigation (median [IQR] 7 [6 to 8]) and patient-specific instrumentation (median 6 [5 to 7.8]), although both techniques had higher scores than the freehand technique did (median 3 [2 to 4]). For freehand versus surgical navigation, the difference of medians was 4 (p < 0.001). For freehand versus patient-specific instrumentation, the difference of medians was 3 (p < 0.001). For patient-specific instrumentation versus surgical navigation, the difference of medians was 1 (p = 0.03). The mean ± procedural times for freehand was 16 ± 10 minutes, patient-specific instrumentation was 14 ± 9 minutes, and surgical navigation techniques was 24 ± 8 minutes. We found no differences in procedure times across three shaping modalities (freehand versus patient-specific instrumentation: mean difference 2 minutes [95% CI 0 to 7]; p = 0.92; freehand versus surgical navigation: mean difference 8 minutes [95% CI 0 to 20]; p = 0.23; patient-specific instrumentation versus surgical navigation: mean difference 10 minutes [95% CI 1 to 19]; p = 0.12).

Conclusion

Based on surgical simulation to reconstruct a standardized periarticular bone defect after tumor resection, we found a possible small advantage to surgical navigation over patient-specific instrumentation based on qualitative fit, but both techniques provided slightly better conformance of the shaped graft for fit into the standardized post-tumor resection bone defect than the freehand technique did. To determine whether these differences are clinically meaningful requires further study. The surgical navigation system presented here is a product of laboratory research development, and although not ready to be widely deployed for clinical practice, it is currently being used in a research operating room setting for patient care. This new technology is associated with a learning curve, capital costs, and potential risk. The reported preliminary results are based on a preclinical synthetic bone tumor study, which is not as realistic as actual surgical scenarios.

Clinical Relevance

Surgical navigation systems are an emerging technology in orthopaedic and reconstruction surgery, and understanding their capabilities and limitations is paramount for clinical practice. Given our preliminary findings in a small cohort study with one scenario of standardized synthetic periarticular bone tumor defects, future investigations should include different surgical scenarios using allograft and cadaveric specimens in a more realistic surgical setting.

Introduction

Joint-sparing resection of periarticular bone tumors can be challenging because of complex geometry, their proximity to critical structures, and a narrow margin of error [12, 23, 24]. Precise surgical planning needs a multimodality imaging approach that includes MRI and CT, and fusion of these images can help localize the tumor more accurately. The use of intraoperative navigation or patient-specific instrumentation may facilitate complex osteotomies, helping to compensate for surgical blind spots and minimize surgical exposure without compromising cut accuracy and surgical margins. Successful reconstruction of periarticular bone defects after tumor resection is often performed with structural allografts to preserve the joint. Close contact of host bone and allograft [26] promotes better bone-graft integration, minimizes nonunion and delayed union, and allows for earlier return to function [4, 25-27]. However, achieving a size-matched allograft to fill the defect can be challenging because allograft sizes vary, they do not always match a patient’s anatomy, and cutting the allograft to fit the defect perfectly is demanding.

Tumor surgery requires the ability to produce accurate bone cuts using an oscillating saw or osteotome. Because these are planar instruments, the correct plane of cut is determined by the roll and pitch angle of the plane and the depth of each cut. Real-time tracking and positional feedback of these cutting instruments is an emerging technology that is not yet used by available commercial systems [5-7]. Our previous investigations reported improved accuracy in bone tumor resections because of the real-time feedback provided by our novel navigation system in pelvic tumors [32], joint-sparing resections [33], head and neck surgery, and skull base surgery [2, 17]. Patient-specific instrumentation (also known as custom cutting guides) is an alternative modality which uses three-dimensional (3D) printing technology, and it has been widely reported in the literature [1].

The present study compared the performance of three approaches for allograft shaping: a novel surgical navigation system, patient-specific instrumentation, and freehand osteotomies in accurately shaping allografts.

We asked: (1) Is there a difference in mental workload among the freehand, patient-specific instrumentation, and surgical navigation approaches? (2) Is there a difference in conformance (quantitative measure of deviation from the ideal bone graft), elapsed time during reconstruction, and qualitative assessment of goodness-of-fit of the allograft reconstruction among the approaches?

Methods and Materials

Experimental Overview

Seven orthopaedic surgeons were recruited for our study. Each surgeon was provided with a standard bone defect model as a template to shape three synthetic bones to fill a standardized defect (Fig. 1). The shaping sequence began with the freehand technique, followed by patient-specific instrumentation, and surgical navigation. There was a cool-down period of approximately 15 minutes between each shaping approach, and the National Aeronautics and Space Administration (NASA) mental task load index questionnaire was provided at the end of each shaping session. The procedural time lapse was captured for each approach. All synthetic bone grafts were scanned with cone-beam CT to compute quantitative measurements. Six additional senior surgeons who were blinded to the surgical technique scored the quality of synthetic bone grafts using a Likert scale ranging from 1 (worst) to 10 (best).

Fig. 1.

Fig. 1

Open in a new tab

This schematic diagram shows an overview of this study in which we created a standardized 3D-printed femur defect model for allograft synthetic crafting that was performed with different approaches, including freehand, patient-specific instrumentation, and surgical navigation. Data collection and analyses included NASA mental task load index evaluations, cone-beam CT conformance computation of the shaped grafts, procedural time lapse, and blinded qualitative bone graft evaluations. A color image accompanies the online version of this article.

Study Design, Tumor Model, and Surgical Planning

Three orthopaedic oncology fellows (PN, IA, KRG) and four senior orthopaedic residents (none were study authors) participated in this study, each shaping a simulated allograft using synthetic bone using the three approaches in sequence beginning with the freehand technique, then patient-specific instrumentation, and finally surgical navigation to fill a standardized tumor resection defect in the distal femur. All participating surgeons had 1 to 3 years of surgical experience with freehand surgical resection and reconstruction. One surgeon (PN) had used patient-specific instrumentation several times and had performed 13 procedures using the in-house surgical navigation system during a clinical trial [30]. The remaining participating surgeons had no previous experience with patient-specific instrumentation or the in-house surgical navigation system.

A custom Sawbones® tumor model (Pacific Research Laboratories) made of solid rigid polyurethane was created using CT and MRI images from a patient with a parosteal osteosarcoma of the distal femur with intramedullary involvement who had undergone multiplanar geometric resection and allograft reconstruction to spare the joint (Fig. 2A). A fellowship-trained sarcoma surgeon developed the tumor resection plan using the Insight Toolkit-Snake Automatic Partitioning (ITK-SNAP) [36] software based on preoperative MRI and CT images, which required eight osteotomies to resect the tumor with negative margins (Fig. 2B). The location and orientation of each osteotomy was saved on the in-house surgical navigation platform GTxEyes [11]. A 3D-printed femur model was manufactured, with the tumor resection defect based on the preoperative computerized multiplanar surgical plan [34] (Fig. 2C) and served as a standardized host bone for all participants conducting reconstruction using the three surgical approaches. Synthetic bone that served as a simulated allograft for shaping had a different size and geometry from the host bone used to create the defect model.

Fig. 2.

Fig. 2

Open in a new tab

(A) A distal femur Sawbones model of a parosteal osteosarcoma was based on a patient’s CT and MRI scans [34]. (B) This is a surface rendering of the cone-beam CT tumor bone model overlay; it had multiplanar cutting planes for tumor resection that were developed using in-house surgical navigation and computerized planning software. (C) A 3D-printed femur model with the tumor resection defect based on the preoperative computerized plan was created; the defect model served as a standard template for creating the bone graft by freehand, patient-specific instrumentation, and surgical navigation approaches in this study. A color image accompanies the online version of this article.

Synthetic Bone Shaping

The participant surgeons used all three methods (freehand, patient-specific instrumentation, and surgical navigation) to shape intact synthetic bone of the distal femur that served as the standard simulated allograft, guided by the tumor resection model described previously (Fig. 2), which provided a standardized defect requiring reconstruction with an appropriately shaped synthetic bone. Each surgeon shaped one synthetic bone using each method starting with the freehand technique, patient-specific instrumentation, and surgical navigation. Between shaping methods, the participant had a cool-down period of approximately 15 minutes, after which they completed the NASA task load questionnaire. In the freehand technique (Fig. 3), surgeons measured the anticipated size of the defect from preoperative MRI and CT images, the actual size of the defect in the tumor resection model (Fig. 2C), and the 3D model of the resected tumor, thereby simulating a clinical scenario. Then they used routine surgical instruments including an oscillating saw, ruler, and bone-holding clamps to shape the allograft.

Fig. 3.

Fig. 3

Open in a new tab

A freehand allograft was created simulating the standard approach commonly undertaken in the operating room. The surgeon creating the allograft first drew the dimensions of the resected tumor or the resulting tumor resection defect onto the allograft and then used an oscillating saw to shape the graft based on scalar measurements.

Patient-specific Instrumentation

Patient-specific instrumentation, also known as patient-specific cutting jigs, was designed for the eight osteotomies based on the final computerized tumor resection plan (Fig. 2B). The synthetic bone model was scanned using prototype cone-beam CT with an image volume of 256 x 256 x 192, isotropic voxel of 0.78 mm3, and it was imported into Mimics software (Mimic Innovation Suite, Version 17.0, Materialise). Semiautomated segmentation provided a coarse volume that was refined manually for precise model geometry. The patient-specific instrumentation designed for the study tumor model had two separate pieces for an easier fit around the tumor-bearing bone (Fig. 4A). The patient-specific instrumentation developed for tumor resection was placed on the synthetic bone graft (Fig. 4B). K-wire holes that were incorporated into the patient-specific instrumentation design allowed the construct to be firmly anchored to synthetic bone for allograft shaping with an oscillating saw.

Fig. 4.

Fig. 4

Open in a new tab

Patient-specific instrumentation is designed for tumor resection using computerized surgical planning software. The patient-specific instrumentation incorporates cutting slots that allow placement of the oscillating saw in predefined positions or orientations based on preoperative computerized surgical planning. (A) Shown here is the patient-specific instrumentation design (PSI) and placement for resection of a distal femur surface tumor. (B) This image shows the same 3D-printed patient-specific instrumentation jigs applied to the allograft synthetic bone for shaping for reconstruction. A color image accompanies the online version of this article.

The patient-specific instrumentation was constructed using a fused deposition modeling 3D printer (Dimension SST 1200es, Stratasys). Acrylonitrile butadiene styrene was used with print layer resolution of 0.01 inches and model interior structures set at sparse-high density.

Real-time 3D Model of the Allograft by Structured Light Surface Scan

A 3D model of the synthetic allograft bone was obtained to apply the exact surgical plan used for tumor resection to accurately shape the allograft. To achieve this, we adapted a low-cost, off-the-shelf surface scanner (Carmine 1.09, PrimeSense) (Fig. 5A) to digitize the synthetic bone (Fig. 5B) and create a mesh surface model (Fig. 5C) in real time using a structured light technique with a dual-camera (RGB and IR camera) and near-IR light. The mean ± SD time to scan the synthetic bone that would be used to shape the grafts was 9 ± 1 minutes (n = 7).

Fig. 5.

Fig. 5

Open in a new tab

A navigated allograft crafting approach using an allograft synthetic bone is shown here. (A) A low-cost surface digitizer acquires the geometry of (B) the allograft synthetic bone and produces (C) a point cloud representation of the allograft synthetic bone. (D) The bone tumor model in the surgical navigation system with the overlaid cut planes for tumor resection is co-registered with point cloud allograft representation to allow for precise shaping of the synthetic bone to fit the geometric bone tumor resection defect. The distance, pitch, and roll indicators in our novel navigation system provide visual feedback for the surgeon in terms of tracked saw blade alignment with respect to the preoperative cut plan; RGB = red, green, blue; CMOS = complementary metal-oxide-semiconductor; IR = infrared.

Surgical Navigation

The surgical navigation system included a prototype mobile C-arm intraoperative cone-beam CT (PowerMobile, Siemens), real-time optical tracking system (Polaris, NDI), and an in-house-developed surgical navigation platform (GtxEyes) for visualization. The cone-beam CT imaging system provides 3D images with submillimeter spatial resolution and soft tissue differentiation with low-dose radiation (approximately 0.1 to 0.35 mSv), allowing multiple intraoperative images. Cone-beam CT reconstruction images have an isotropic voxel size of 0.78 mm3 with image volume of 256 x 256 x 192.

The navigation platform GtxEyes [11] is based on open-source imaging libraries such as Image-Guided Surgery Toolkit [13], the Visualization Toolkit [31], and Insight Toolkit [19], which provide a variety of 3D visualization options including a standard triplanar view (axial, sagittal, and coronal planes), virtual rendering models of the bone surface, and dynamic image reslicing. Critical structures such as the arteries, nerves, and articular surfaces can be delineated using ITK-SNAP. A margin of error around critical structures can be built into the surgical plan to provide a warning when the surgeon is close to a critical structure; for example, when using a saw or osteotome.

The rationale for using surgical navigation (Fig. 5D) was to ease the trial-and-error process of shaping the synthetic graft for filling an osseous defect after tumor resection. Taking advantage of the exact preoperative computerized osteotomy plans developed for tumor resection allows the surgeon to reverse-cut an allograft that is equivalent to the resected tumor. The point cloud representation of the synthetic bone graft (Fig. 5C) was registered on a cone-beam CT image of the tumor bone in GtxEyes. Then, the osteotomy planes with their individual pitch, roll, and plane distances were overlayed onto the synthetic graft (Fig. 5D) so the cuts could be navigated, as if applying the same surgical plan to the tumor-bearing bone, but instead providing a matched graft for reconstruction.

NASA Tasks Load Index (NASA-TLX)

After graft shaping, each participant completed the NASA Task Load Index (NASA-TLX) questionnaire [15, 16]. This is a validated, multidimensional, 21-point Likert-scale tool that provides a subjective assessment of mental workload. We anticipated there would be cognitive, interface, time, and effort investments in these three techniques. We measured mental demand, physical demand, temporal demand, performance, effort, and frustration; participants graded each task load on the questionnaire. Scores for each subscale ranged from 0 to 20, with higher scores being worse and representing higher workloads. For the performance subscale, higher scores indicated less successful completion of the task. Total task load (range 0 to 120) was determined by adding all six subscale scores.

Conformance Comparison of Shaped Synthetic Allografts

Conformance is how well something meets a specified standard. In this study, we defined conformance as a quantitative measure of deviation of the shaped synthetic bone from the ideal computer-generated bone graft. We measured conformance using the root-mean-square metric in mm, which computes the square root of the average squared distance between the ideal graft in the reference model and the shaped graft. The root-mean-square provides a meaningful metric by avoiding any canceling-out effect in the measurement associated with negative and positive values. First, all shaped bone allografts were scanned with cone-beam CT. The volumetric image was imported into visual planning software (Mimic Innovation Suite, Version 17.0, Materialise) to create a 3D stereolithography model that was imported into 3D modeling software (3-matic, Materialise). The shaped allografts were registered to the reference ideal graft model to compute the difference in the geometric fit to allow for quantitative measurement. The distances in the geometric fit of the shaped allografts compared with the ideal reference graft were also highlighted by displaying them as color-coded conformance maps. These color-coded maps also served as a qualitative measure to visually demonstrate the goodness-of-fit of the grafts made with each shaping technique to match the computer-generated ideal bone graft model.

Qualitative Bone Graft Evaluation

Across 21 shaped allografts (Fig. 6) that were harvested from the three study arms including freehand, patient-specific instrumentation, and surgical navigation, the shaped synthetic allograft fit was qualitatively scored by six experienced (faculty) sarcoma surgeons (KT, PCF, JSW, JN, RN, PR, the last three were not study authors). There was no overlap between these six senior evaluating surgeons and the seven more-junior surgeons who shaped the grafts. All 21 shaped grafts were deidentified so the evaluating surgeons were blinded to the technique. Surgeons compared each shaped bone graft with the reference 3D-printed tumor resection model for goodness-of-fit. Physical conformance and graft fitting were scored qualitatively on a Likert scale from 1 (worst fit) to 10 (best fit).

Fig. 6.

Fig. 6

Open in a new tab

(A) 3D-printed tumor resection model shows both the standardized defect and resected tumor used for participants whose allografts were shaped by freehand, patient-specific instrumentation, and surgical navigation. (B) The shaped grafts were randomly numbered from 1 to 21 and evaluated qualitatively for goodness-of-fit into the standardized bone tumor defect.

Ethical Approval

Ethical approval was not sought for this present study. A proof-of-principle investigation on which this study is based was also conducted using synthetic bone to simulate clinical practice [33].

Statistical Analyses

Statistical analysis was performed using XLSTAT® (2019.2.2, Addinsoft®). To detect the difference among each approach for allograft shaping across root-mean-square values as a quantitative measure of conformance and procedure time, we performed an ANOVA, performed with Tukey honestly significant difference post hoc testing. We analyzed the NASA-TLX using the nonparametric Kruskal-Wallis test with multiple pairwise comparisons using the Steel-Dwass-Critchlow-Fligner procedure with a two-tailed setting. The level of significance was set at p < 0.05. In terms of minimum clinically important differences, we felt that 1-mm root-mean-square conformance differences or 1 Likert point could be clinically meaningful qualitatively in terms of the ability of a shaped graft to fit well enough into a standardized defect to allow sufficient bone healing and avoid nonunion. To our knowledge, no minimum clinically important difference values have been published for the NASA total score or subscale scores.

Results

NASA-TLX Mental Workload

We found no differences in the total NASA task load scores (median [IQR] for bone graft shaping: 79 (52 to 83) for freehand, 43 (33 to 50) for patient-specific instrumentation, and 61 (58 to 66] for navigation (freehand versus patient-specific instrumentation: difference of medians 36; p = 0.37; patient-specific instrumentation versus navigation: difference of medians 18; p = 0.15; freehand versus navigation: difference of medians 18; p = 0.80). With respect to NASA subscale analyses, we detected no differences in the mental, physical, temporal, effort, and frustration subscales (Fig. 7). Performance scores (median [IQR]) for freehand, patient-specific instrumentation, and surgical navigation were 11 (10 to 12), 6 (6 to 11), and 5 (4 to 7), respectively. Only the comparison of freehand versus surgical navigation showed a difference (difference of medians 6; p = 0.04) (Fig. 7); however, it may be too small to be clinically meaningful.

Fig. 7.

Fig. 7

Open in a new tab

This boxplot shows the median and IQR for each category of the NASA task load index that evaluates mental demand, physical demand, temporal demand, performance, effort, and frustration in three different bone grafting approaches including freehand, patient-specific instrumentation, and surgical navigation; PSI = patient-specific instrumentation.

Conformance, Procedure Time, and Qualitative Allograft Evaluation

Conformance based on the mean ± SD root-mean-square values was 3 ± 1 mm for freehand, 2 ± 1 mm for patient-specific instrumentation, and 2 ± 0 mm for the surgical navigation technique (Fig. 8A). The navigated bone grafts had similar conformance (Table 1) to the grafts shaped by patient-specific instrumentation (mean difference 0 mm [95% CI -0.5 to 0.2]; p = 0.82), and both showed slightly better conformance than the freehand approach (mean difference 1 mm [95% CI 0.5 to 1.1]; p = 0.01 for surgical navigation; mean difference 1 mm [95% CI -0.1 to 0.9]; p = 0.02 for patient-specific instrumentation). The mean ± SD procedural time (Fig. 8A) for freehand was 16 ± 10 minutes, patient-specific instrumentation was 14 ± 9 minutes, and surgical navigation techniques was 24 ± 8 minutes. We found no differences in procedures times across three shaping modalities (freehand versus patient-specific instrumentation: mean difference 2 minutes [95% CI 0 to 7]; p = 0.92; freehand versus surgical navigation: mean difference 8 minutes [95% CI 0 to 20]; p = 0.23; patient-specific instrumentation versus surgical navigation: mean difference 10 minutes [95% CI 1 to 19]; p = 0.12) (Table 1).

Fig. 8.

Fig. 8

Open in a new tab

Conformance results show the differences in geometry or lack of congruence between crafted bone grafts compared with the ideal bone graft generated from computerized planning software. (A) This bar chart shows the mean ± SD of conformance based on the root-mean-square in mm and the required procedure time for each resection approach (minutes and seconds). (B) In these color-coded conformance maps, green and red indicate higher and lower levels of conformance, respectively. The scale to the right measures the actual differences in conformance between the shaped graft and the ideal graft in millimeters; RMS = root-mean-square; PSI = patient-specific instrumentation.

Table 1.

Shape conformance and procedure time comparison among the freehand, PSI, and surgical navigation methods

Parameter PSI (n = 7) Surgical navigation (n = 7)
Freehand (n = 7) Mean ± SD Mean ± SD Mean difference vs freehand (95% CI) p value for PSI vs freehand Mean ± SD Mean difference vs freehand (95% CI) p value for surgical navigation vs freehand Mean difference vs PSI (95% CI) p value for surgical navigation vs PSI
Conformance (RMS) in mm 3 ± 1 2 ± 1 0.86 (-0.06 to 0.93) 0.02 2 ± 0 1.03 (0.46 to 1.08) 0.01 0.17 (-0.5 to 0.23) 0.82
Procedure time in minutes 16 ± 10 14 ± 9 2 (0 to 7) 0.92 24 ± 8 8 (0 to 20) 0.23 10 (1 to 19) 0.12
Open in a new tab

The color-coded conformance maps for the three grafts (Fig. 8B) demonstrated increasing red from bottom to top, suggesting best fit for the navigated graft over the patient-specific instrumented graft, which was better than the freehand-shaped graft. The maximum difference in conformance between the shaped and ideal grafts was approximately 1.2 mm for the navigated graft, 5 mm for the patient-specific instrumented graft, and 7.5 mm for the freehand-shaped graft. Based on qualitative assessment of fit of the shaped grafts into the standardized bone tumor defect (Fig. 9), we found that surgical navigation (median [IQR] 7 [6 to 8]) performed better than patient-specific instrumentation (median 6 [5 to 7.8]), and both scored much higher than the freehand technique (median 3 [2 to 4]). For freehand versus surgical navigation, the difference of medians was 4 (p < 0.001). For freehand versus patient-specific instrumentation, the difference of medians was 3 (p < 0.001). For patient-specific instrumentation versus surgical navigation, the difference of medians was 1 (p = 0.03).

Fig. 9.

Fig. 9

Open in a new tab

To evaluate the blinded bone grafts, six experienced surgeons scored goodness-of-fit from 1 (worst fit) to 10 (best fit) into the standardized bone tumor resection defect. This boxplot depicts minimum, maximum, median, first quartile, and third quartile scores; PSI = patient-specific instrumentation.

Discussion

Joint-sparing resection of periarticular bone tumors can be challenging because of complex anatomic geometry and the goal of obtaining a negative margin. Reconstruction of periarticular bone defects after tumor resection may be performed with structural allografts to preserve the joint. However, achieving a size-matched allograft to fill the defect can be challenging owing to allograft size variations, a mismatch between the patient’s anatomy and the allograft, and the need to sculpt the allograft to fit the defect. Surgical navigation and patient-specific instrumentation have been developed as alternatives to freehand tumor resection and allograft shaping. Here, we evaluated whether these three methods differ regarding mental workload, time to completion, or conformance of the shaped graft to a computer-generated standardized synthetic bone tumor. We found no difference in mental workload and procedural time among the freehand, patient-specific instrumentation, and surgical navigation approaches. In terms of quantitative conformance, we found a difference favoring surgical navigation over patient-specific instrumentation. This is unlikely to have a clinically relevant impact; however, both techniques performed slightly better than freehand (Fig. 8). In the qualitative evaluation of goodness-of-fit, surgical navigation performed better than patient-specific instrumentation and both scored much higher than freehand (Fig. 9). To determine whether these differences in conformance are clinically meaningful will require further study. We do not recommend introducing the current format of this surgical navigation technology into widespread clinical practice until further rigorous validation and testing demonstrates a more clearcut clinical benefit than the currently available alternative techniques.

Limitations

There are several limitations in this study including a small sample, which could affect sensitivity because we did not have a priori estimates to calculate a robust sample size for this pilot study. We also used synthetic bone, which may be inferior to allografts for extrapolation to real patients, and only one surgical scenario. Tumor defect size may play an important role in terms of procedure time and mental workload. For example, a small defect may be more amenable to a freehand approach, while avoiding the expense and lead time required to create cutting jigs. We also used a standardized bone tumor defect model, which is not realistic compared with an actual operating room situation and may occur infrequently even in high-volume sarcoma centers. The periarticular bone defect model we chose was based on one patient. However, in most clinical scenarios, the postresection defect is rarely identical to the initial plan. It has been reported that cuts made around simulated tumors substantially deviate from the preoperative plan, implying that the anticipated defect is rarely the same as the actual defect in real-world application [18, 21]. In our study, participants used a standardized 3D-printed model incorporating a computerized preoperative resection plan. Although this may deviate from the true clinical setting, a standardized 3D-printed bone model can provide a consistent way to assess the three reconstruction modalities. This scenario may favor surgical navigation over the other two approaches. An alternative study design that may be more realistic to clinical practice would be to have each participant individually conduct the tumor resection of the host bone and then reconstruct that defect using each of the three surgical approaches; for example, a host tumor resected freehand would then be reconstructed freehand. In that scenario, the resulting defect and bone graft would both be scanned to determine the accuracy of fit. This alternative study design may be more realistic than the study design we used, and may show different results among the three surgical methods.

The simulation of real-world soft tissue constraints in our prior tumor resection models [2, 33] with the use of putty to cover the tumor was an attempt to make the model more realistic but was still an underrepresentation of real-time anatomy because 3D constraints caused by surrounding soft tissues and critical structures—including major nerves and blood vessels—could not be re-created. In the current synthetic bone-shaping study, the lack of soft tissues should have been advantageous to and overestimated the utility of patient-specific instrumentation. However, navigated grafts were as accurate based on quantitative conformance and scored better on qualitative goodness-of-fit measures, and this may have been due to the real-time feedback associated with that technique. The in-house-designed, patient-specific instrumentation used in this study had two separate components because of the size and location of the tumor in our selected simulation model. However, a two-component cutting jig may have introduced more uncertainty and error than a one-component design. Finally, in the clinical setting, the accuracy of laser scanners to digitize allograft geometry compared with the synthetic bone as used in this study could be compromised because of background noise generated by fluids, soft tissues, and room lighting, which could interfere with registration and affect the surgical navigation accuracy.

NASA-TLX Mental Workload

Although patient-specific instrumentation had a lower total task load score than either the surgical navigation or freehand techniques (Fig. 7), suggesting less mental workload or demand, these differences were not clinically meaningful. Furthermore, of the six NASA-TLX subscale scores, the only small difference identified was in performance. Surgical navigation had a lower performance score (that is, more successful task completion) than patient-specific instrumentation, which was lower than freehand, but the only potential meaningful difference was that surgical navigation performed better than the freehand technique. The cognitive load for patient-specific instrumentation planning exists preoperatively, in a less stressful environment outside the operating room, but that was not accounted for in this study, which addressed the intraoperative component of allograft shaping for each of the three techniques. In comparison, the cognitive demand for surgical navigation occurs intraoperatively, so it was not surprising that the navigated procedures were longer, although the differences compared with the other two techniques were not clinically important.

Patient-specific instrumentation, also known as patient-specific cutting jigs, may be easier to use intraoperatively than surgical navigation because it avoids registration and has no line-of-sight issues. However, this ease of use comes at the cost of flexibility, limiting the degrees of freedom for each osteotomy because it precludes any intraoperative change to surgical plan because of soft tissue constraints, tumor growth (which may require a revised surgical plan), or geometric difficulty in fitting the cutting jig. For pelvic and sacral bone tumors especially, the complex surrounding anatomy can limit the real-time use of patient-specific instrumentation. Patient-specific instrumentation is often planned on CT images that provide higher resolution and are devoid of soft tissue covering the bone. This may cause a misfit or complete lack of ability to fit the cutting jig, leading to inaccuracies because of the need for a soft tissue envelope surrounding most malignant bone tumors to ensure negative resection margins. Patient-specific instrumentation is also not sophisticated in guiding or controlling the depth of the osteotomy and lacks any live visual feedback. In allograft crafting, the patient-specific instrumentation used for tumor resection may not always be adaptable to the allograft bone because of geometric variations.

Conformance, Procedure Time, and Qualitative Allograft Evaluation

Our study showed that patient-specific instrumentation and surgical navigation have slightly better quantitative conformance based on root-mean-square measurement compared with the standard-of-care freehand surgical technique for precise allograft shaping. However, color-coded conformance maps (Fig. 8B) and qualitative bone graft evaluation for goodness-of-fit (Fig. 9) favored navigated synthetic bone shaping. To our knowledge, this synthetic bone allograft study model is the first to compare freehand, patient-specific instrumentation, and surgical navigation techniques. The surgical navigation arm of this study used a novel, real-time surgical navigation tracking system with positional feedback (roll, pitch, and distance relative to planned cuts) of the oscillating saw used for allograft shaping, which has been shown to be effective in orthopaedic oncology as well as head and neck oncology bone tumor resections [2, 17, 30, 32-34]. Another first is the superimposition of the computerized resection plan on the digital synthetic bone graft image to facilitate precise shaping. The turnaround time for this process is rapid (mean ± SD scanning time 9 ± 1 minutes) and can be achieved without breaking operating room sterility. It may also be performed ahead of time during anesthesia induction to minimize intraoperative delays. The surface scanning technology used in this study is commercially available, and its utility for acquiring 3D surface bone geometry has been reported [35]. That study [35] investigated differences in morphometry after four articulated human pelvic bones were scanned using structured light scanners, photogrammetry, and CT. The investigators found that all three scanning methods yielded similar surface representations, with average deviations among the surface models between 0.1 and 0.2 mm.

Patient-specific instrumentation and surgical navigation can improve the accuracy of complex osteotomies to resect bone tumors because they help the surgeon visualize the angle, depth, and roll of each bone cut. This improved accuracy facilitates negative margin tumor resection and enhances safety with respect to surrounding critical structures [14, 20, 21, 22]. Our novel surgical navigation system provides real-time feedback and automatic registration [30], but it can have line-of-sight issues. However, this system can adapt to intraoperative changes the surgeon makes to the surgical plan in real-time to achieve a negative margin. Bosma et al. [3] performed a cadaveric study comparing the accuracy of resecting simulated bone tumors by freehand, computer-assisted surgery, patient-specific instrumentation, and a combination of computer-assisted surgery and patient-specific instrumentation. Interestingly, patient-specific instrumentation-assisted resections had the best overall cut plane accuracy, which is different from the findings in our investigation. There are several differences between the study designs that may have contributed to these discrepancies: (1) a simple circular tumor model versus complex tumor model; (2) a simple surgical resection plan versus complex osteotomies; (3) different surgical navigation system (non-real-time surgical navigation cut versus real-time guided osteotomies); and (4) different quantitative measurements. In previous studies of our surgical navigation system [2, 17, 32, 33], we showed the localization accuracy was between 1.2 and 1.7 mm, which is smaller than 3.6 mm with computer-aided surgery but similar to 1.9 mm for patient-specific instrumentation as reported by Bosma et al. [3]. However, even with patient-specific instrumentation, the authors [3] found increased errors at the exit bone cuts because of the lack of visual feedback with this technique, a problem our surgical navigation system has overcome by introducing real-time system feedback.

There are limited data on the use of patient-specific instrumentation for allograft shaping. One study [2] reported on four pediatric tibial sarcoma resections and reconstructions. In another study [20], the authors reported a technical guideline for using patient-specific instrumentation to perform joint-preserving resections around the knee, and they emphasized the need to shape the undersurface of the guide to match the negative margin of the tumor-bearing bone surface to offer a precise fit. Fan guides (guides for multiple K-wires along a curve) have been used for curvilinear osteotomies [8]; however, these are more feasible in the extremity than in the pelvis because of anatomic constraints. In a systematic review [1], the authors reported on the use of patient-specific instrumentation for corrective osteotomies of the femur and the pelvis, and demonstrated improvement in operating time and accuracy. None of these studies formally compared the two technologies, as we did in our investigation. Our group reported that navigated osteotomies in periarticular pelvic and extremity bone tumor models, pelvic cadaveric studies [32, 33], and head and neck bone tumor resections [2, 17] improved accuracy as benchmarked against freehand surgical techniques. Being able to accurately shape allografts to fit periarticular bone tumor defects implies there is an added flexibility offered by surgical navigation beyond tumor resection. For future research, larger studies should include different surgical scenarios with varied anatomic sites and sizes of the defect including the use of cadaver specimens. Instead of using a homogenized identical defect based on a 3D-printed model of synthetic bone, participants might use different bones for tumor resection and bone graft reconstruction but with the same surgical modality. In terms of technical direction, the in-house-developed surgical navigation system will eventually be incorporated into projection-based augmented reality in the future to facilitate a surgeon’s ability to visualize the cut planes for tumor resection and allograft shaping directly onto the surgical field [9, 10, 29].

Conclusion

In a preclinical synthetic bone model with periarticular bone defect simulation, surgical navigation and patient-specific instrumentation demonstrated better performance in accurate synthetic bone graft shaping than a freehand approach. However, qualitative assessments favored surgical navigation–shaped grafts. This study also demonstrated proof-of-concept feasibility of a novel surgical navigation approach to shaping synthetic bone by using a low-cost surface digitizer to acquire bone geometry and register it to a preoperative computerized tumor resection plan. Despite the limitations, the study model had a validated framework and outcome measures, and its translatable findings form a benchmark for future prospective cadaveric and clinical studies. The surgical navigation system presented here is a product of laboratory research development and is not ready for widespread clinical use, but is being used in a research operating room setting. Introducing new technology is associated with learning curves, capital costs, and potential risk, and further investigation in a range of scenarios will be important to identify whether there are sufficient benefits to patients.

Acknowledgments

We thank Zachary Tran MD, Eyal Ramu MD, Omar Salem Dahduli MD, and Rohit Sheshgir MSc, MD who participated in shaping synthetic bones for this study. We also thank Jennifer Nevin MD; Rosti Novak PhD, MD; and Philip Rowell MD for their qualitative evaluations of the synthetic bone grafts.

Footnotes

The institution of one or more of the authors (KRG) has received, during the study period, funding from Presage Biosciences. This work was supported the Strobele Family Guided Therapeutics Research Fund; the Dorrance Drummond Family Fund; the Garron Foundation; Guided Therapeutics (GTx) Program/Princess Margaret Cancer Centre, University Health Network; and the Princess Margaret Cancer Foundation.

Each author certifies that there are no funding or commercial associations (consultancies, stock ownership, equity interest, patent/licensing arrangements, etc.) that might pose a conflict of interest in connection with the submitted article related to the author or any immediate family members.

All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research® editors and board members are on file with the publication and can be viewed on request.

Clinical Orthopaedics and Related Research® neither advocates nor endorses the use of any treatment, drug, or device. Readers are encouraged to always seek additional information, including FDA approval status, of any drug or device before clinical use.

Ethical approval was not sought for the present study.

This work was performed at Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada.

Contributor Information

Prakash Nayak, Email: nayakprakash@gmail.com.

Ibrahim Alshaygy, Email: ialshaygy@hotmail.com.

Kenneth R. Gundle, Email: gundle@ohsu.edu.

Kim Tsoi, Email: Kim.Tsoi@sinaihealth.ca.

Michael J. Daly, Email: michael.daly@rmp.uhn.ca.

Jonathan C. Irish, Email: Jonathan.Irish@uhn.ca.

Peter C. Ferguson, Email: Peter.Ferguson@sinaihealth.ca.

Jay S. Wunder, Email: Jay.Wunder@sinaihealth.ca.

References

  • 1.Baraza N, Chapman C, Zakani S, Mulpuri K. 3D - printed patient specific instrumentation in corrective osteotomy of the femur and pelvis: a review of the literature. 3D Print Med. 2020;6:34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Bernstein JM, Daly MJ, Chan H, et al. Accuracy and reproducibility of virtual cutting guides and 3D-navigation for osteotomies of the mandible and maxilla. PloS One. 2017;12:e0173111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Bosma SE, Wong KC, Paul L, Gerbers JG, Jutte PC. A cadaveric comparative study on the surgical accuracy of freehand, computer navigation, and patient-specific instruments in joint-preserving bone tumor resections. Sarcoma. 2018;2018:4065846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Brien EW, Terek RM, Healey JH, Lane JM. Allograft reconstruction after proximal tibial resection for bone tumors. An analysis of function and outcome comparing allograft and prosthetic reconstructions. Clin Orthop Relat Res. 1994;303:116-127. [PubMed] [Google Scholar]
  • 5.Cartiaux O, Banse X, Paul L, Francq BG, Aubin CE, Docquier PL. Computer-assisted planning and navigation improves cutting accuracy during simulated bone tumor surgery of the pelvis. Comput Aided Surg. 2013;18:19-26. [DOI] [PubMed] [Google Scholar]
  • 6.Cartiaux O, Paul L, Docquier PL, et al. Accuracy in planar cutting of bones: an ISO-based evaluation. Int J Med Robot. 2009;5:77-84. [DOI] [PubMed] [Google Scholar]
  • 7.Cartiaux O, Paul L, Docquier PL, Raucent B, Dombre E, Banse X. Computer-assisted and robot-assisted technologies to improve bone-cutting accuracy when integrated with a freehand process using an oscillating saw. J Bone Joint Surg Am. 2010;92:2076-2082. [DOI] [PubMed] [Google Scholar]
  • 8.Cernat E, Docquier PL, Paul L, Banse X, Codorean IB. Patient specific instruments for complex tumor resection-reconstruction surgery within the pelvis: a series of 4 cases. Chirurgia (Bucur). 2016;111:439-444. [DOI] [PubMed] [Google Scholar]
  • 9.Chan HHL, Haerle SK, Daly MJ, et al. An integrated augmented reality surgical navigation platform using multi-modality imaging for guidance. PloS One. 2021;16:e0250558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Chan HHL, Sahovaler A, Daly MJ, et al. Projected cutting guides using an augmented reality system to improve surgical margins in maxillectomies: a preclinical study. Oral Oncology. 2022;127:105775. [DOI] [PubMed] [Google Scholar]
  • 11.Daly MJ, Chan H, Nithiananthan S, et al. Clinical implementation of intraoperative cone-beam CT in head and neck surgery. Proc SPIE. 2011;7964:796426-796428. [Google Scholar]
  • 12.Deijkers RL, Bloem RM, Hogendoorn PC, Verlaan JJ, Kroon HM, Taminiau AH. Hemicortical allograft reconstruction after resection of low-grade malignant bone tumours. J Bone Joint Surg Br. 2002;84:1009-1014. [DOI] [PubMed] [Google Scholar]
  • 13.Enquobahrie A, Cheng P, Gary K, et al. The image-guided surgery toolkit IGSTK: an open source C++ software toolkit. J Digit Imaging. 2007;20:21-33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Fujiwara T, Kaneuchi Y, Stevenson J, et al. Navigation-assisted pelvic resections and reconstructions for periacetabular chondrosarcomas. Eur J Surg Oncol. 2021;47:416-423. [DOI] [PubMed] [Google Scholar]
  • 15.Hart SG. NASA-Task Load Index (NASA-TLX); 20 years later. Proc Hum Factors Ergon Soc Annu Meet. 2006;50:904-908. [Google Scholar]
  • 16.Hart SG, Staveland LE. Development of NASA-TLX (Task Load Index): results of empirical and theoretical research. In: Hancock PA, Meshkati N, eds. Advances in Psychology. North-Holland; 1988:139-183. [Google Scholar]
  • 17.Hasan W, Daly MJ, Chan HHL, Qiu J, Irish JC. Intraoperative cone-beam CT-guided osteotomy navigation in mandible and maxilla surgery. Laryngoscope. 2020;130:1166-1172. [DOI] [PubMed] [Google Scholar]
  • 18.He G, Dai AZ, Mustahsan VM, et al. A novel method of light projection and modular jigs to improve accuracy in bone sarcoma resection. J Orthop Res. 2022;40:2522-2536. [DOI] [PubMed] [Google Scholar]
  • 19.Ibanez L, Schroeder W, Ng L, Cates J. The ITK Software Guide. Kitware Inc; 2003. [Google Scholar]
  • 20.Jud L, Muller DA, Furnstahl P, Fucentese SF, Vlachopoulos L. Joint-preserving tumour resection around the knee with allograft reconstruction using three-dimensional preoperative planning and patient-specific instruments. Knee. 2019;26:787-793. [DOI] [PubMed] [Google Scholar]
  • 21.Khan FA, Lipman JD, Pearle AD, Boland PJ, Healey JH. Surgical technique: computer generated custom jigs improve accuracy of wide resection of bone tumors. Clin Orthop Relat Res. 2013;471:2007-2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Laitinen MK, Parry MC, Albergo JI, Grimer RJ, Jeys LM. Is computer navigation when used in the surgery of iliosacral pelvic bone tumours safer for the patient? Bone Joint J. 2017;99:261-266. [DOI] [PubMed] [Google Scholar]
  • 23.Lewis VO, Gebhardt MC, Springfield DS. Parosteal osteosarcoma of the posterior aspect of the distal part of the femur. Oncological and functional results following a new resection technique. J Bone Joint Surg Am. 2000;82:1083-1088. [DOI] [PubMed] [Google Scholar]
  • 24.Liu T, Liu ZY, Zhang Q, Zhang XS. Hemicortical resection and reconstruction using pasteurized autograft for parosteal osteosarcoma of the distal femur. Bone Joint J. 2013;95:1275-1279. [DOI] [PubMed] [Google Scholar]
  • 25.Muller DA, Beltrami G, Scoccianti G, Cuomo P, Capanna R. Allograft-prosthetic composite versus megaprosthesis in the proximal tibia—what works best? Injury. 2016;47(Suppl 4):S124-S130. [DOI] [PubMed] [Google Scholar]
  • 26.Muscolo DL, Ayerza MA, Aponte-Tinao LA, Ranalletta M. Use of distal femoral osteoarticular allografts in limb salvage surgery. J Bone Joint Surg Am. 2005;87:2449-2455. [DOI] [PubMed] [Google Scholar]
  • 27.Pesenti S, Peltier E, Pomero V, et al. Knee function after limb salvage surgery for malignant bone tumor: comparison of megaprosthesis and distal femur allograft with epiphysis sparing. Int Orthop. 2018;42:427-436. [DOI] [PubMed] [Google Scholar]
  • 28.Safir O, Kellett CF, Flint M, Backstein D, Gross AE. Revision of the deficient proximal femur with a proximal femoral allograft. Clin Orthop Relat Res. 2009;467:206-212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Sahovaler A, Chan HHL, Gualtieri T, et al. Augmented reality and intraoperative navigation in sinonasal malignancies: a preclinical study. Front Oncol. 2021;11:723509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Sahovaler A, Daly MJ, Chan HHL, et al. Automatic registration and error color maps to improve accuracy for navigated bone tumor surgery using intraoperative cone-beam CT. JB JS Open Access. 2022;7:e21.00140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Schroeder W, Martin K, Lorensen B. The Visualization Toolkit: An Object Oriented Approach to 3D Graphics. Kitware Inc; 2004. [Google Scholar]
  • 32.Sternheim A, Daly M, Qiu J, et al. Navigated pelvic osteotomy and tumor resection: a study assessing the accuracy and reproducibility of resection planes in Sawbones and cadavers. J Bone Joint Surg Am. 2015;97:40-46. [DOI] [PubMed] [Google Scholar]
  • 33.Sternheim A, Kashigar A, Daly M, et al. Cone-beam computed tomography-guided navigation in complex osteotomies improves accuracy at all competence levels: a study assessing accuracy and reproducibility of joint-sparing bone cuts. J Bone Joint Surg Am. 2018;100:e67. [DOI] [PubMed] [Google Scholar]
  • 34.Sternheim A, Rotman D, Nayak P, et al. Computer-assisted surgical planning of complex bone tumor resections improves negative margin outcomes in a sawbones model. Int J Comput Assist Radiol Surg. 2021;16:695-701. [DOI] [PubMed] [Google Scholar]
  • 35.Waltenberger L, Rebay-Salisbury K, Mitteroecker P. Three-dimensional surface scanning methods in osteology: a topographical and geometric morphometric comparison. Am J Phys Anthropol. 2021;174:846-858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Yushkevich PA, Piven J, Hazlett HC, et al. User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage. 2006;31:1116-1128. [DOI] [PubMed] [Google Scholar]

Articles from Clinical Orthopaedics and Related Research are provided here courtesy of The Association of Bone and Joint Surgeons

RESOURCES