Talar OsteoPeriostic Grafting from the Iliac Crest (TOPIC)

Background: Press-fit Talar OsteoPeriostic grafting from the Iliac Crest (TOPIC) is a novel technique for the treatment of large osteochondral lesions of the talus. The purpose of the present study was to prospectively evaluate the 2-year clinical outcomes for patients with medial osteochondral lesions of the talus that were treated with the TOPIC procedure. Methods: Forty-three patients were prospectively assessed before and 24 months after the TOPIC procedure. All procedures were performed through a medial distal tibial osteotomy. Clinical assessment preoperatively and at 24 months of follow-up included determination of the Numeric Rating Scale (NRS) scores for pain during walking (primary outcome), at rest, during running, and during stair-climbing. The Foot and Ankle Outcome Score (FAOS) and the Mental Component Summary (MCS) score and Physical Component Summary (PCS) score of the Short Form-36 (SF-36) were also assessed. A computed tomography (CT) scan was performed 12 weeks postoperatively to assess union of the distal tibial osteotomy site and at 1 and 2 years postoperatively to assess consolidation of the graft as well as cyst development in the graft. Results: All enrolled patients were available for follow-up. The median NRS score for pain during walking improved from 7 points preoperatively to 2 points at 2 years postoperatively (p < 0.001). All other NRS scores for pain improved significantly. All FAOS subscale scores improved significantly, including those for pain (from 53 to 75), symptoms (from 50 to 61), activities of daily living (from 68 to 88), sports (from 30 to 55), and quality of life (from 19 to 38). The SF-36 PCS score improved from 43 to 48 (p < 0.001), and the MCS score changed from 28 to 26 (p > 0.05). There was a 100% rate of union of the osteotomy site at the distal tibia and 100% of the grafts showed consolidation at 1 and 2 years postoperatively. Conclusions: The TOPIC procedure for large osteochondral lesions of the medial talar dome is an effective technique that resulted in significant improvement exceeding the minimal clinically important difference in pain scores (primary outcome) as well as in other outcomes, with 100% consolidation of the grafts. Level of Evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

C urrently, there is no consensus regarding the best treatment for primary or non-primary osteochondral lesions of the talus [1][2][3][4][5] . Large osteochondral lesions of the talus may require strategies that involve autograft or allograft transplantation 6,7 . However, it is known that autologous osteochondral transplantation causes donor-site morbidity in up to 35% of cases when the graft is harvested from the ipsilateral femoral condyle 1,8 . A specific disadvantage of allografts is nonunion and failure of the graft 7,9,10 . An alternative can be the implantation of autologous osteoperiosteal grafts from the iliac crest. The use of cylindrical osteoper-iosteal grafts was described by Hu et al. 8 and Chen et al. 11 , who reported clinically effective outcomes. A disadvantage of cylindrical autografts is that their sizes are static and may not allow for a perfect fit on the talar side 12,13 .
To overcome these disadvantages, a novel joint-sparing technique for the treatment of large osteochondral lesions of the talus was developed 14 . This technique, known as Talar Osteo-Periostic grafting from the Iliac Crest (TOPIC), involves using the iliac crest with the overlying periosteal layer as an autograft [15][16][17] .
The primary aim of the present study was to investigate the 2-year clinical and radiographic outcomes following the use of the TOPIC procedure for the treatment of large medial osteochondral lesions of the talus. The secondary aim was to assess the outcomes of this procedure in terms of safety, reoperations, and donor-site morbidity. We hypothesized that the use of the TOPIC procedure for the treatment of large, complex, medial osteochondral lesions of the talus would be clinically effective with good radiographic and safety outcomes and low donor-site morbidity at 2 years of follow-up.

Materials and Methods
T he present study is a prospective, single-center, nonrandomized, clinical cohort study with a 2-year follow-up. The study was approved by the local Medical Ethics Committee at the University of Amsterdam (reference number MEC 08/326) and was performed in accordance with the current ethical standards (Declaration of Helsinki). Our center is a tertiary referral academic hospital that is nationally and internationally recognized for the diagnosis and treatment of cartilage and osteochondral lesions of the foot and ankle.

Patient Selection
All patients with a medial osteochondral lesion of the talus who were managed with the TOPIC procedure ( Fig. 1) and followed for a minimum of 24 months were included in this study. The indication for the TOPIC procedure was a large (>10 mm in anteroposterior and/or mediolateral diameter and/or depth) symptomatic osteochondral lesion of the talar dome that was treated in either a primary or non-primary (i.e., underwent prior surgical treatment that failed) fashion and was determined to be not fixable 18,19 (i.e., underwent a standardized prior nonoperative treatment protocol for at least 6 months without success) in a patient with a closed epiphyseal growth plate 14 . (Relative) contraindications included rheumatoid arthritis, advanced osteoporosis, advanced osteoarthritis (OA), concomitant painful or disabling disease of the lower limb, infectious pathology, or any kind of malignancy. Lateral and central osteochondral lesions of the talus were excluded from the analysis in order to keep the study group homogeneous.

Preoperative Planning
As part of our clinical workup for patients with a suspected osteochondral lesion of the talus, all patients were evaluated with a dual-energy computed tomography (CT) scan to confirm the diagnosis and to determine the location, morphology, and size of the lesion 5 . After the diagnosis had been confirmed, patients participated in a comprehensive individualized shared decision-making process during which the different treatment options were discussed. Intraoperative visualization of an osteochondral lesion ( Fig. 1-A), en bloc trapezoidal resection of the lesion ( Fig. 1-B), and press-fit insertion of the osteoperiosteal graft matching the curvature of the medial talar dome ( Fig. 1-C). 6,2023 is shown in Figure 1. When there was lateral ankle instability, a lateral stabilization procedure was added in selected cases, and when there was varus or valgus hindfoot malalignment, a supramalleolar osteotomy was added.

Clinical Evaluation
Standard preoperative patient demographics were extracted. Preoperative and postoperative assessment consisted of an electronic questionnaire consisting of the following patient-reported outcome measures (PROMs): the Numeric Rating Scale 20 (NRS) for pain during walking, at rest, during running, and during stairclimbing, the Foot and Ankle Outcome Score (FAOS) 21 including its subscales, and the Short Form-36 (SF-36) Mental Component Summary (MCS) score and Physical Component Summary (PCS) score 22 . Preoperatively and postoperatively, the American Orthopaedic Foot & Ankle Society (AOFAS) 23 ankle-hindfoot score was measured. To minimize the risk of bias, the treating orthopaedic surgeons were not involved in any outcome assessment. Validated Dutch versions of all questionnaires were used 21, 22,24,25 . Other, prospectively assessed clinical outcomes consisted of additional surgical procedures, complications, and donor-site morbidity. At the 2-year follow-up, patients were specifically asked to report on the presence of donor-site morbidity, which was defined as pain, disability, or hindrance related to the donor site or surrounding area affecting activities in daily living.

Radiographic Evaluation
Baseline radiographic assessments included the maximum anteroposterior and mediolateral diameters and depth, surface area (ellipsoid formula: area = abp = coronal length · sagittal length · 0.79), and volume (formula: volume = abc = coronal length · sagittal length · depth) 26 . Morphological description consisted of 3 categories (fragmentary, cystic, or crater) 27 . The lesion location was reported using a 9-grid scheme 28 . The preoperative stage of OA was determined with a CT ankle OA atlas, which includes the presence and number of osteophytes 29 . A CT scan was performed 12 weeks postoperatively to assess union of the distal tibial osteotomy site and consolidation of the graft, and at 1 and 2 years postoperatively to assess consolidation of the graft as well as cyst development in the graft.

Statistical Analysis
The primary outcome was the comparison between the NRS score for pain during walking preoperatively versus postoperatively (2 years). All statistical analyses were performed with SPSS version 28.0 (IBM). Dichotomous and categorical outcome variables were reported as absolute numbers and percentages. Continuous outcome variables were reported as means and standard deviations (SDs) in cases of a normal distribution of the data and as medians and interquartile ranges (IQRs) in cases of non-normally distributed data. Preoperative and postoperative data were compared using a paired t test in cases of normally distributed data and with a Wilcoxon signed rank test in cases of non-normally distributed data. Interrater and intrarater reliability for the lesion size measurements and interrater reliability for the lesion morphology and location assessments were determined with a subset of 20 osteochondral lesions of the talus assessed by 2 independent raters (J.D. and Q.R.). Intrarater reliability was assessed by repeating the measurements at a 2-week interval. Lesion size measurements were assessed with use of intraclass correlation coefficient (ICC) analysis 30 . The Cohen kappa value was used to test interrater reliability for lesion location and morphological classification 31 . A consensus meeting between the 2 independent raters (J.D. and Q.R.) was organized to agree upon lesion localization and morphological category. In cases of disagreement, the senior author (G.M.M.J.K.) made the final decision.
For correlation analyses and subgroup comparisons, the principal dependent variable was the primary outcome. Data were compared between 2 subgroups using a Mann-Whitney U test for non-normally distributed data and a 2-sample unpaired t test for normally distributed data. One-way nonparametric analysis of variance (ANOVA) using a Kruskal-Wallis test was used to compare the dependent variable among the different morphological groups and the OA subgroups. Significance values were adjusted by the Bonferroni correction for multiple tests in case of significance. A Spearman rho test was used for the assessment of correlation of continuous non-normally distributed data with the principal dependent variable, and a Pearson rho test was used for the assessment of correlation of continuous normally distributed data. The level of significance was set at p < 0.05.
A prospective sample size calculation was performed to detect a minimal clinically important difference (MCID) of 1.0 point on the NRS for pain during walking between baseline and follow-up 32 . The calculation indicated that 34 ankles were needed. To correct for a potential loss to follow-up of 15%, the required sample size for the present study was 39 ankles.

Source of Funding
No external funding was received for this study

Results
A total of 44 patients were eligible for inclusion. One patient underwent an arthroscopic ankle arthrodesis before the 2-year follow-up and was excluded in order to not assess the outcomes of the arthrodesis (Fig. 2). Thus, 43 patients were assessed. No patients were lost to follow-up. Patient demographics are summarized in   Table I, and preoperative radiographic characteristics are shown in Table II. An example of a preoperative 3D visualization of a typical lesion is shown in Figure 3.
The change in the NRS score for pain during walking for each patient is shown in Figure 5. The NRS score for pain at rest decreased from a median of 3 (IQR: 2 to 6) preoperatively to a median of 1 (IQR: 0 to 3) postoperatively (p < 0.001). The NRS score for pain during stairclimbing decreased from a median of 7 (IQR: 4 to 7) preoperatively to a median of 2 (IQR: 0.5 to 5) postoperatively (p < 0.001) (Fig. 4). The NRS for pain during running decreased from a median of 9 (IQR: 8 to 9) preoperatively to a median of 4 (IQR: 1 to 6) postoperatively (p < 0.001); this variable was analyzed for the 29 patients in the series who were able to run preoperatively and postoperatively (Fig. 6).

AOFAS
The AOFAS scores are presented in Figure 7.

FAOS and SF-36
The preoperative and postoperative FAOS and SF-36 subscale scores are shown in Table III.

Radiographic Outcomes
One hundred percent of the CT scans were performed at 12 weeks postoperatively and at 2 years postoperatively, although 1 patient did not have a CTscan at 1 year postoperatively due to a pregnancy. Preoperative and postoperative (2-year) Numeric Rating Scale scores for pain during walking, at rest, and during stair-climbing. The IQR is indicated by a box, the median is indicated by a line within the box, and the whiskers indicate the highest and lowest values. Preoperative 3D visualization of a cystic osteochondral lesion (same lesion as in Figure 8) that has a 3D volume measurement of 2,249 mm 3 . There was a 100% rate of union of the distal tibial osteotomy site at 3 months postoperatively. One hundred percent of the grafts showed consolidation at 3 months and at 1 and 2 years postoperatively. The 1-year postoperative CT scans showed cyst formation in the graft in 48% of the patients, and the 2-year postoperative CT scans showed cyst formation in the graft in 67% of the patients. Interobserver reliability and intraobserver reliability related to radiographic characteristics are shown in Appendix I. Examples of preoperative and postoperative CT scans are shown in Figures 8, 9, and 10.
Additional Surgical Procedures, Donor-Site Morbidity, and Complications Additional surgical procedures are described in Table IV. There were 4 complications (9%). Two of these were hypoesthesia: 1 case of temporary hypoesthesia around the iliac crest wound and 1 case of ongoing hypoesthesia of the first web space. One patient developed complex regional pain syndrome (CRPS) postoperatively, and 1 had a fracture within the osteoperiosteal graft 23 months postoperatively. The fracture was treated nonoperatively and clinical symptoms improved. No patients reported symptoms related to donor-site morbidity at the 2-year follow-up.

Subgroup and Correlation Analyses
Subgroup analyses with the primary outcome as dependent variable are shown in Table V and factors correlated with the primary outcome, in Table VI.

Discussion
I n our prospective study, the TOPIC procedure proved to be a clinically effective and safe technique for large medial osteochondral lesions of the talus, with good mid-term functional results and consolidation of all grafts and medial distal tibial osteotomy sites on the postoperative CT scans. The median improvements of 5 points on the NRS for pain during walking and of 35 points on the AOFAS scale are considered highly Fig. 6 Preoperative and postoperative (2-year) Numeric Rating Scale scores for pain during running. The IQR is indicated by a box, the median is indicated by a line within the box, and the whiskers indicate the highest and lowest values. Observed data points (small circles) outside the boundaries of the whiskers are plotted as outliers.  effective, as a mean improvement of 2 on the NRS of 0 to 10 is significantly associated with "much better" and is (far) beyond the MCID 33 . An advantage of the TOPIC procedure is that it is a single-stage technique that requires limited surgical time 14 . Additionally, the geometry of the iliac crest matches the talar geometry and it has a chondrogenic potential due to the periosteal layer 15,17 . As such, the iliac crest autograft provides a scaffold, an appropriate bone block, and growth factors [34][35][36][37] . All transplantation techniques have the potential disadvantage of a less than fully anatomical fit and thus new cyst formation. Cyst formation was observed in our study, but it has not been found to have any correlation with clinical symptoms at 2-year follow-up 8 . Another advantage of the TOPIC procedure is that the defect depth is not a limiting factor when determining whether the procedure is indicated. Lastly, if the TOPIC procedure fails, a repeat TOPIC procedure, an allograft procedure, an ankle arthrodesis, or a total ankle replacement are among the salvage procedures to be considered 14 .
The access route to the talar dome through a fairly large osteotomy of the medial distal tibia may be considered a relative disadvantage of the procedure. In our cohort, all osteotomy sites showed complete union on CT scans performed 3 months postoperatively. The screw was removed in 15 patients (43%) because of concern that it would cause discomfort.
With regard to the indications for the TOPIC procedure as compared with the autologous osteochondral transplantation procedure, autologous osteochondral transplantation is indicated for 100 and 150-mm 2 lesions 38 , reflected by preoperative defect sizes ranging from 78 to 113 mm 2 in the literature [38][39][40] , whereas the median surface area of the defects in our analyzed cohort was 177 mm 2 and in some cases the surface area was twice as large as the area suitable for autologous osteochondral transplantation. Another important comparison in the context of indications is the percentage of prior surgical procedures. In a large study on autologous osteochondral transplantation by Shimozono et al. 38 , the percentage of primarily treated osteochondral lesions of the talus ranged from 75% to 81%, while in our study the percentage was 33%. However, the 2 procedures are similar with regard to reoperation rates and the clinical outcomes [38][39][40][41] . With regard to donor-site morbidity, Shimozono et al. 42 found that the estimated proportion of this complication ranged from 7% to 11% after autologous osteochondral transplantation   Figs. 9-A through 9-D Preoperative and 2-year postoperative CT scans of patient example 2. Fig. 9-A Preoperative coronal CT scan. Fig. 9-B Preoperative sagittal CT scan. Fig. 9-C Postoperative coronal CT scan. Fig. 9-D Postoperative sagittal CT scan.    and a retrospective study 43 showed that donor-site morbidity of the knee after autologous osteochondral transplantation led to functional impairments in patients with previously asymptomatic knees. In our prospective study, the patients were specifically asked to report donor-site morbidity of the iliac crest; none reported functional impairments due to iliac crest donor-site morbidity, which is in line with studies by Wang et al. 44 and Shi et al. 41 , who also used an autograft harvested from the iliac crest. With regard to indications and demographic factors, the application of bulk allograft is comparable with the TOPIC procedure 9,[45][46][47][48][49] . In a recent publication by Fletcher et al. 46 , the preoperative size (volume as based on CT scans) was 1,879 mm 3 , which is 22% smaller than that in our cohort (2,400 mm 3 ). Concerning the reoperation rate, fresh osteochondral allograft transplantation was associated with an additional surgery rate of 48% 46 , which is substantially different from the rate of 37% in our study. It must also be noted that the majority of the additional procedures in our study were implant removal unrelated to the graft, which is considered a minor procedure not requiring access to the joint. A comparative analysis of survival of the TOPIC graft versus the allograft procedure may not be applicable as we did not include long-term follow-up with an appropriate survival analysis. However, the survival rate at 2 years of follow-up in our cohort was 98%, which is higher than the 87% extracted from the literature on osteochondral allograft transplantation 48 .
Our study has limitations. It was limited to a single treatment arm, and only medial osteochondral lesions of the talus were included even though central and lateral TOPIC procedures are also performed at our institution 50 . Another limitation is the low statistical power concerning the subgroup analyses and potentially correlated factors. Strengths of the present study are the prospective nature, the 100% follow-up rate, and the usage of a high number and diverse set of clinical outcome scores. Moreover, 2 independent raters scored the CT scans for both preoperative measurements and postoperative outcomes with high interrater reliability as well as intrarater reliability. Lastly, due to the prospective power calculation, the study was sufficiently powered for the primary outcome.

Conclusions
The novel press-fit TOPIC procedure for large primary and secondary osteochondral lesions of the medial talar dome is an  effective and safe surgical technique with good short-term clinical results (significant decrease in pain, the primary outcome) and with 100% radiographic incorporation of the grafts. As such, the procedure may serve as a clinically effective jointsparing alternative to an autologous osteochondral transplantation procedure, a bulk allograft, an (arthroscopic) ankle arthrodesis, and a total ankle replacement. Future research is necessary to assess long-term results of the TOPIC procedure, effectiveness in other centers, as well as the efficacy for central and lateral talar lesions.