Abstract
Background
Heterotopic ossification (HO) is common after total joint arthroplasty and usually does not cause diagnostic problems. However, the occurrence of HO after oncologic prostheses implantation can be troublesome as it may mimic a locally recurrent tumor. Because this distinction could have a profound impact on the surgeon and patient, it is important to distinguish the two entities; to our knowledge, no study has evaluated this after oncologic endoprosthetic reconstruction around the knee after tumor resection.
Questions/purposes
(1) How common is the occurrence of HO compared with local recurrence (LR) after resection of bone sarcoma and the use of an oncologic knee prosthesis? (2) Are there any factors associated with the development of HO after limb salvage procedures with an endoprosthesis? (3) What features allow the surgeon to differentiate HO from a locally recurrent tumor in this setting?
Methods
Between 2002 and 2018, we performed 409 resections of primary bone tumors followed by reconstructions with oncologic endoprostheses. Of these, 17% (71 of 409) died before 2 years and did not have HO at that time, 2% (8 of 409) were lost to follow-up before 2 years, and another 2% (10 of 409) did not have radiographs available at a minimum of 2 years after surgery (and had not developed HO before then), and so could not be analyzed, leaving 320 patients for analysis in this retrospective study. Forty-two patients were excluded; 2% (5 of 320) for a history of failed allograft reconstruction, 3% (8 of 320) for pathologic fracture at presentation, 2% (6 of 320) for inadequate or complicated biopsy, 1% (2 of 320) for stem fractures, 2% (7 of 320) for stem loosening, and 4% (14 of 320) for extracortical bone bridging, leaving 278 patients for final evaluation. Two observers analyzed AP and lateral radiographs for signs of HO at a mean follow-up of 63 ± 33 months after surgery. We defined HO as extraskeletal bone formation in soft tissues. The primary study endpoint was survivorship free from HO, as ascertained by a competing-risks estimator. To identify factors associated with HO appearance, the demographic, radiographic, clinical, pathologic, and surgical characteristics were compared between patients with HO and those who had no lesion. Characteristic features were also compared between patients with HO and those with LR to help their differentiation. Univariate analysis was used for all statistical evaluations.
Results
HO developed in 8% (21 of 278) of patients in whom oncologic knee prosthesis was implanted. LR developed in 10% (28 of 278) of the patients. According to survivorship estimates, the HO-free survival rate was not different from the LR-free survival rate at 2 years after oncologic knee reconstruction (76 ± 5% [95% CI 63 to 87] versus 74 ± 5% [95% CI 62 to 88]; p = 0.19). History of infection was more common in patients with HO than in patients with no lesion (19% [4 of 21] versus 5% [12 of 229], Odds ratio [OR] 6 [95% CI 2 to 17]; p < 0.001). The male sex was more common in the HO group as well (76% [16 of 21] versus 55% [128 of 229], OR 2 [95% CI 1 to 5]; p = 0.03). The Modular Universal Tumor and Revision System prosthesis was more frequently used in patients with HO (67% [14 of 21]) compared to those with no lesions (40% [92 of 229]; OR 2 [95% CI 1 to 5]; p = 0.02). The lesion border in radiographs was ill-defined in 19% (4 of 21) of patients with HO and 100% (28 of 28) of patients with LR (OR 8 [95% CI 3 to 20]; p < 0.001). The median time to the appearance of HO was shorter than the time to LR (8 months [3 to 13] versus 16 months [11 to 21], [95% CI 10 to 13]; p < 0.001). Pain at presentation was more frequent in patients with LR than in those with HO (86% [24 of 28] versus 14% [3 of 21], OR 36 [95% CI 7 to 181]; p < 0.001).
Conclusion
HO may occur after the use of oncologic knee prostheses for reconstruction after tumor resection. In most patients, HO could be differentiated from local recurrence through identifying a well-defined border on radiographs. Otherwise, factors such as an earlier time of presentation and absence of pain could suggest an HO, rather than an LR.
Level of Evidence
Level III, therapeutic study.
Introduction
The distal femur is the most common site of primary bone sarcoma, followed by the proximal tibia [5, 19]. An oncologic prosthesis is a popular option for the reconstruction of the defect of these areas after resection of a bone sarcoma because of its availability, durability, and permitting early weightbearing [10, 13]. Even so, local recurrence (LR) remains a devastating oncologic outcome after the resection of musculoskeletal tumors [11, 12, 23] with morbidity and mortality, despite adjuvant therapy and obtaining adequate margin.
Heterotopic ossification (HO) is a common complication after total joint arthroplasty and does not generally pose a diagnostic difficulty [6, 9]. However, its presentation after knee reconstruction with an oncologic prosthesis could be troublesome because the lesion might mimic a locally recurrent tumor. To our knowledge, no study has examined HO incidence after oncologic knee prosthesis implantation, the factors associated with its occurrence, and its differentiation from LR.
Regarding the above-mentioned gaps in knowledge, we performed this retrospective study answer the following questions: (1) How common is the occurrence of HO compared with LR after resection of bone sarcoma and the use of an oncologic knee prosthesis? (2) Are there any factors that predispose patients to develop HO after limb salvage procedures with an endoprosthesis? (3) What features allow the surgeon to differentiate HO from a locally recurrent tumor in this setting?
Patients and Methods
This study was approved by our institutional review board of our institute under the code of IR.IUMS.REC.1398.1151. The patients or their families provided written consent to use their medical data for publication in this retrospective study.
Patients
We reviewed the medical records of 453 patients who were treated for a primary malignant bone tumor in the knee (distal femur or proximal tibia) between November 2002 and December 2018. Patients were included if they were treated with wide resection and oncologic knee prosthesis reconstruction, had a follow-up duration of more than 2 years, and had available radiographic records. Forty-two patients were treated by either an amputation or wide resection and reconstruction with osteoarticular allograft. The remaining 409 patients were treated with wide resections and reconstructions with oncologic endoprostheses. Of these, 17% (71 of 409) died before 2 years and did not have HO at that time, 2% (8 of 409) were lost to follow-up prior to 2 years, and another 2% (10 of 409) did not have radiographs available at a minimum of 2 years after surgery (and had not developed HO before then), and so could not be analyzed, leaving 320 patients. Forty-two patients were excluded from the study for the following reasons: 2% (5 of 320) for a history of failed allograft reconstruction, 3% (8 of 320) for a pathologic fracture at presentation, 2% (6 of 320) for an inadequate or complicated biopsy, 1% (2 of 320) for stem fractures, 2% (7 of 320) for stem loosening, and 4% (14 of 320) for extracortical bone bridging. Finally, 278 patients met the eligibility criteria for the study (Fig. 1). The medical records of these patients were extracted, and patients were categorized into three groups: HO, LR, and no lesion. We first calculated the prevalence of HO versus LR. To identify factors associated with the incidence of HO, we compared characteristic features between the patients with HO and those with no lesion. To provide evidence to differentiate HO from an LR, we compared the characteristic features of patients with HO and those with LR patients.
Fig. 1.
This flow diagram displays the study inclusion and exclusion criteria.
Treatments, Implants, and Surgical Approaches
In all, 263 patients underwent preoperative and postoperative chemotherapy, but the remaining 15 patients, who had either chondrosarcoma (six patients) or parosteal osteosarcoma (nine patients), did not. Radiation therapy was only used after the detection of LR in a patient with Ewing sarcoma. No other radiation therapy was performed either before or after HO and LR detection. As the HO in the knee is not debilitating, we did not use radiotherapy or NSAIDs as routine prophylaxis for HO.
We used three types of the prosthesis to reconstruct the knee, including the Global Modular Reconstruction System (GMRS, Stryker Orthopaedics, Mahwah, NJ, USA; 16% [44 of 278]), Howmedica Modular Reconstruction System (HMRS, Howmedica GmbH, Kiel, Germany; 42% [117 of 278]), and the Modular Universal Tumor and Revision System (MUTARS, Implantcast, Buxtehude, Germany; 42% [117 of 278]). All prostheses were uncemented.
The lesion was resected using an anteromedial approach for tumors with medial prominence and anterolateral incisions for those with lateral masses. We elevated the limb for 5 minutes and applied a pneumatic tourniquet. When there was limited space in the proximal thigh to apply a pneumatic tourniquet, we used an Esmarch bandage without exsanguination. The average tourniquet time was 90 ± 30 minutes. Osteotomies were performed with a power saw. After wound closure, the limb was immobilized for 2 weeks for patients with distal femur tumors and 4 weeks for those with proximal tibial tumors. The patients attended postoperative visits at 2 weeks and 6 weeks. Follow-up was continued every 3 months for the first 2 years, every 6 months up to 5 years, and yearly up to 10 years. At each follow-up visit, the patients were evaluated clinically. Moreover, AP and lateral radiographs of the femur or tibia were obtained for radiographic evaluations of outcome. A chest CT image was also obtained in each follow-up to evaluate for the presence of chest metastasis.
Primary and Secondary Study Outcomes
Our primary study outcome was survivorship free from HO, as ascertained by a competing-risks estimator [25]. We assessed this on AP and lateral radiographs of the patients. The diagnosis of HO was based on the identification of a well-defined cortical border around the lesion (Fig. 2A). In patients in whom a well-defined border was not seen on radiographs (Fig. 2B-C), a core needle biopsy was performed to make a definitive diagnosis. The border of lesions on radiographs (ill or well-defined) was reviewed by two observers (KJ, AB) who were not involved in patients’ care. An inter- and intraobserver reliability of 0.95 and 0.96 was obtained in this respect.
Fig. 2.
A-C (A) This 9-month postoperative radiograph shows a calcified lesion with a well-defined border that was diagnosed as HO. (B) This 6-month postoperative radiograph shows a calcified lesion with an ill-defined border and unrecognizable etiology that was diagnosed as HO during a pathologic examination. (C) This 18-month postoperative radiograph shows a calcified lesion with an ill-defined border that was diagnosed as a recurrent tumor during a pathologic examination.
Our secondary outcomes were survivorship free from LR, potential factors associated with HO incidence, and potential factors helping the differentiation of HO from LR. We used AP and lateral radiographs to identify recurrence based on the presence of a mineralized mass and bone change with ill-defined radiographic border. In patients who presented with clinical symptoms of recurrence (pain with or without mass) and no radiographic lesion, a nonossifying LR was suspected and was initially inspected with ultrasonography. If a mass was detected in ultrasonography, an ultrasound-guided core needle biopsy was performed to diagnose the LR. We did not use cross-sectional imaging such as MRI or CT to detect LR because of metal artifacts.
To evaluate the factors associated with HO or helping its differentiation from LR, the demographic, clinical, histologic, and surgical characteristics of the patients were extracted from the patients’ medical profile. The clinical evaluation of the patients included the assessment of postoperative complications such as pain and mass.
The radiographic records of patients were re-reviewed, and lesion characteristics such as size, time of presentation, and location were assessed. The size of HO was measured by multiplying the maximum length and width of the lesion on plain radiographs. All radiographs were obtained using a picture archiving and communication system (PACS), which provides actual dimensions on the image. According to the classification proposed by Dalury and Jiranek [6], HO was classified as Grade I (less than 2 cm2), Grade II (2 to 5 cm2), or Grade III (larger than 5 cm2) (Fig. 3A-C).
Fig. 3.
A-C HO was graded based on the size of the lesion; these radiographs show (A) a Grade 1 lesion, (B) a Grade 2 lesion, and (C) a Grade 3 lesion.
Statistical Analysis
Statistical analyses were mainly performed using SPSS for Windows, version 16 (Chicago, IL, USA). All the associations were assessed in univariate statistics as the number of events was not adequate to perform a multivariate analysis. Descriptive data were presented as percentages (with numerator and denominator) and odds ratio with 95% CI for categorical variables. Numerical variables were presented with mean ± SD and mean difference with 95% CI if normally distributed and using median and range with 95% CI if non-normally distributed.
To answer the first question about survivorship free from HO or LR, the competing-risks analysis was done using Stata Statistical Software (Release 13 StataCorp LP, College Station, TX, USA). To answer the second and third questions about the factors associated with the occurrence of HO and factors helping the differentiation of HO from LR, we used an independent t-test or its nonparametric counterpart (Mann-Whitney U test) for the comparison of mean between numeric variables and a chi-square test to compare qualitative variables. A median split approach was used to suggest a time point helping the differentiation of HO from LR. A p value < 0.05 was considered significant.
Results
Frequency of HO Compared with LR after Sarcoma Resection and Reconstruction with an Oncologic Knee
HO prevalence was not different from LR prevalence after the implication of oncologic knee prosthesis. In this respect, HO was detected in 8% (21 of 278) of patients who were treated with an oncologic knee prosthesis. Local recurrence occurred in 10% (28 of 278) of these patients. According to the survival analysis in the presence of competing risks, the HO-free survival rate was not different from the LR-free survival rate at 2 years after oncologic knee reconstruction (76 ± 5% [95% CI 63 to 87] versus 74 ± 5% [95% CI 62 to 88]; p = 0.19). Forty percent (11 of 28) of patients with LR were alive in the last follow-up.
Factors Associated with HO after Limb Salvage Procedures with an Endoprosthesis
HO was associated with sex, history of postoperative infection, and prosthesis type. The sex distribution was different between the two study groups; there were more men in the HO group than in the no-lesion group (76% [16 of 21] versus 56% [128 of 229], odds ratio [OR] 2 [95% CI 1 to 5]; p = 0.03). Moreover, the proportion of patients with a history of postoperative infection was higher in the HO group than in the no-lesion group (19% [4 of 21] versus 5% [12 of 229], OR 6 [95% CI 2 to 17]; p < 0.001). The incidence of HO was associated with prosthesis type; The MUTARS prosthesis was more frequently used in patients with HO (67% [14 of 21]) compared to those with no lesions (40% [92 of 229]; OR 2 [95% CI 1 to 5]; p = 0.02). The mean follow-up duration was not different between patients with HO and those without a lesion (62 ± 32 months versus 63 ± 34 months mean difference 1 months [95% CI - 12 to 8]; p = 0.59). The mean age of the patients was not different between those with HO and those without a lesion (26 ± 14 years versus 27 ± 12 years, mean difference 1 year [95% CI - 6 to - 4]; p = 0.65). The number of patients who received chemotherapy was not different between the HO and no-lesion group (95% [20 of 21] versus 96% [219 of 229], OR 1.2 [95% CI 1 to 2]; p = 0.97, either. There was no association between the incidence of HO and other patients’ demographic, clinical, and surgical characteristics, as well (Table 1).
Table 1.
Comparison of the characteristics of patients with HO and those with no lesion after the implantation of knee oncologic prostheses (n = 278)a
| Variable | With no lesion, % (n) | With HO, % (n) | p value |
| 82 (229) | 8 (21) | ||
| Age in years, mean ± SD | 27 ± 12 | 26 ± 14 | 0.65 |
| Gender, % (n) | |||
| Male | 56 (128) | 76 (16) | 0.03b |
| Female | 44 (101) | 24 (5) | |
| BMI in kg/m2, mean ± SD | 25 ± 3 | 24 ± 2 | 0.89 |
| Follow-up in months, mean ± SD | 63 ± 34 | 62 ± 32 | 0.59 |
| Tumor type, % (n) | |||
| Osteosarcoma | 85 (195) | 81 (17) | |
| Ewing sarcoma | 3 (6) | 9 (2) | |
| Parosteal osteosarcoma | 3 (6) | 5 (1) | |
| Periosteal osteosarcoma | 2 (5) | 0.78 | |
| Chondrosarcoma | 3 (6) | ||
| MFH | 4 (8) | 5 (1) | |
| Leiomyosarcoma | 1 (2) | ||
| Chemotherapy, % (n) | |||
| Yes | 95 (217) | 95 (20) | 0.97 |
| No | 5 (12) | 5 (1) | |
| Tumor stage, % (n) | |||
| IIA | 4 (10) | 5 (1) | 0.92 |
| IIB | 96 (219) | 95 (20) | |
| Tumor location, % (n) | |||
| Distal femur | 80 (182) | 71 (15) | 0.68 |
| Proximal tibia | 20 (47) | 29 (6) | |
| Laterality, % (n) | |||
| Right | 40 (92) | 43 (9) | 0.95 |
| Left | 60 (137) | 57 (12) | |
| Prosthesis type, % (n) | |||
| MUTARS | 40 (92) | 67 (14) | 0.02a |
| HMRS | 44 (101) | 19 (4) | |
| GMRS | 16 (36) | 14 (3) | |
| Infection, % (n) | |||
| No | 95 (217) | 81 (17) | <0.001a |
| Yes | 5 (12) | 19 (4) | |
| Size of resection in cm, mean ± SD | 18 ± 3 | 18 ± 3 | 0.71 |
| Surgical approach, % (n) | |||
| Medial | 45 (104) | 38 (8) | 0.61 |
| Lateral | 55 (125) | 62 (13) |
Patients with local recurrence are not included in this table.
Statistically significant (p value < 0.05); HO = heterotopic ossification; MHF = malignant fibrous histiocytoma; MUTARS = Modular Universal Tumor and Revision System; HMRS = Howmedica Modular Reconstruction System; GMRS = Global Modular Reconstruction System.
Features That Allow the Surgeon to Differentiate HO from Locally Recurrent Tumors
Radiographic border, presence or absence of pain, and time of the presentation could be used in the differentiation of HO from LR. The lesion border was ill-defined on radiographs in 19% (4 of 21) of patients with HO and 100% (28 of 28) of patients with LR (OR 8 [95% CI 3 to 20]; p < 0.001). Accordingly, four patients with suspected LR proved to have HO, but no patients in the HO group proved to have LR until the last follow-up (false-positive rate for LR: 13%, false-negative rate: 0%). The median time to the appearance of HO was shorter than the median (range) time to LR (8 months [3 to 13] versus 16 months [11 to 21], [95% CI 10 to 13]; p < 0.001). HO was symptomatic in 19% (4 of 21) of patients with HO compared with 100% (28 of 28) of patients with LR (OR 8 [95% CI 3 to 20]; p < 0.001). The main HO symptoms included pain (in three patients), limited ROM (in two patients), and swelling (in two patients). Some patients had more than one symptom. Pain at presentation was more frequent in patients with LR than in those with HO (86% [24 of 28] versus 14% [3 of 21], OR 36 [95% CI 7 to 181]; p < 0.001).
The lesion was ossified in 79% (22 of 28) of patients with LR and nonossified in the remaining 21% (6 of 28). There was no difference between the mean ages of the patients in the HO group (26 ± 14 years) and the LR group (12 ± 26 years; mean difference 0.3 years [95% CI ‐ 6 to 5]; p = 0.85). The mean follow-up of the patients was not different between the HO and LR groups (62 ± 32 months versus 65 ± 36, mean difference 3 months [95% CI -23 to 16]; p = 0.23).
With respect to size, 43% (9 of 21) of patients had Grade 1 HO, 33% (7 of 21) of patients had Grade 2, and 24% (5 of 21) of patients had Grade 3 HO. The mean size of the lesion was not different between the HO and LR groups (4 ± 3 cm2 versus 4 ± 3 cm2, mean difference 0.2 [95% CI - 2 to 1.7]; p = 0.83). Moreover, there was no difference between other demographic, clinical, and surgical characteristics of the HO and LR groups (Table 2).
Table 2.
Comparison of characteristics between patients with HO and patients with LR (n = 278)
| Variable | With HO, % (n) | With LR, % (n) | p value |
| 8 (21) | 10 (28) | ||
| Age in years, mean ± SD | 26 ± 14 | 26 ± 12 | 0.74 |
| Gender, % (n) | |||
| Male | 76 (16) | 61 (17) | 0.21 |
| Female | 24 (5) | 39 (11) | |
| Follow-up in months, mean ± SD | 62 ± 32 | 65 ± 36 | 0.29 |
| Mean time of occurrence postoperatively in months, mean ± SD | 6 ± 4 | 18 ± 8 | <0.001a |
| Tumor type, % (n) | |||
| Osteosarcoma | 81 (17) | 89 (25) | |
| Ewing sarcoma | 9 (2) | 4 (1) | 0.42 |
| Parosteal osteosarcoma | 5 (1) | 4 (1) | |
| Periosteal osteosarcoma | |||
| Chondrosarcoma | |||
| MFH | 5 (1) | 4 (1) | |
| Leiomyosarcoma | |||
| Pain at presentation of HO or LR, % (n) | |||
| Yes | 14 (3) | 8 (24) | <0.001a |
| No | 86 (18) | 14 (4) | |
| Chemotherapy, % (n) | |||
| Yes | 95 (20) | 96 (27) | 0.83 |
| No | 5 (1) | 4 (1) | |
| Tumor stage, % (n) | |||
| IIA | 5 (1) | 7 (2) | 0.39 |
| IIB | 95 (20) | 93 (26) | |
| Tumor location, % (n) | |||
| Distal femur | 71 (15) | 75 (21) | 0.52 |
| Proximal tibia | 29 (6) | 25 (7) | |
| Length of resection in cm, mean ± SD | 18 ± 3 | 18 ± 3 | 0.33 |
| Size of lesion in cm2, mean ± SD | 4 ± 3 | 4 ± 3 | 0.21 |
Statistically significant (p value < 0.05); HO = heterotopic ossification; LR = local recurrence; MHF = malignant fibrous histiocytoma; MUTARS = Modular Universal Tumor and Revision System; HMRS = Howmedica Modular Reconstruction System; GMRS = Global Modular Reconstruction System.
Discussion
The appearance of HO after tumor resection and reconstruction with an oncologic knee prosthesis could pose a serious challenge, as it might mimic a locally recurrent tumor. There is no data available on the prevalence of this presentation, its predisposing factors, and its differentiation from LR. In this study, we evaluated the proportion of patients in whom HO developed after the implantation of oncologic knee prosthesis. We also assessed the factors that might potentiate the occurrence of HO in this setting and provided some preliminary evidence to noninvasively differentiate HO from LR. According to our study, the prevalence of HO after the implication of oncologic knee prosthesis was not different from LR. Male gender, history of postoperative infection, and using MUTARS prosthesis were associated with the appearance of HO. We also found there were differences in the radiographic appearance, symptoms, and time to presentation between heterotopic bone and locally recurrent tumors in patients who had a bone sarcoma that was resected and reconstructed with an endoprosthesis.
Limitations
Eight patients were lost to follow-up and another 10 did not have adequate radiographs for analysis. Thus, our estimates of the frequency of HO (and of LR) are conservative because patients with problems are more likely to be lost. The actual percentages almost certainly are higher. Therefore, one may use the high end of that 95% CI as the most liberal estimate. This study was performed retrospectively. In retrospective studies, selection bias is a matter of concern, as treatments are not always applied consistently, and variation in surgical indication, surgical approach, and adjuvant/neoadjuvant therapy might have influenced the study results. This study was not large enough to perform a multivariable analysis and thus there may be some confounding of estimates of association. It is possible that some of the factors that were found to be associated with HO might have been influenced by unmeasured confounding. Although a multivariable analysis would have been preferable, we did not have sufficient events to do this. There may have been other factors associated with HO, but small sample size of the present study did not have sufficient power to identify them.
Frequency of HO Compared with LR after Sarcoma Resection and Reconstruction with an Oncologic Knee
The proportion of patients in whom HO developed after the implantation of oncologic knee prosthesis was 8% (21 of 278). We also found that 10% (28 of 278) of the patients had a local recurrence. Based on survivorship estimates, HO prevalence was not different from LR at two years after oncologic knee reconstruction (76% versus 74%). According to the evidence, LR is observed in 5% to 17% of patients presenting with nonmetastatic, malignant primary bone tumors that are managed with an adequate margin, and adjuvant therapy [7]. LR was estimated 10% in the present series, which was comparable with that cited in the evidence. The overall incidence of HO after TKA ranges from 1% to 42% [2]. However, no earlier investigation has evaluated the incidence of HO after the implantation of oncologic knee prosthesis. The similar incidence of HO and LR in the present study suggests more awareness regarding the differentiation of these entities.
Factors Associated with HO after Limb Salvage Procedures with an Endoprosthesis
In this study, postoperative infection, the use of a MUTARS prosthesis, and male sex were associated with the incidence of HO. Postoperative infection was also associated with a higher incidence of HO in earlier investigations. In one study, HO was present in 76% of the patients who had an infection and in 47% of patients who did not have an infection [3]. The association of postoperative infection with HO appearance has been reported in other investigations [16, 21]. Similarly, in this study, the postoperative infection was associated with HO incidence after the implantation of oncologic knee prostheses. Male sex was associated with HO presentation in earlier investigations [1, 14] and with the incidence of HO in the present study. The association of prosthesis type with the presentation of HO has also been reported earlier; a study that investigated the prevalence of HO in patients who underwent cervical artificial disc replacement with three different prosthesis types showed that the prosthesis type had a definite effect on the occurrence of HO [26]. The present study also revealed that the prosthesis type could be associated with the incidence of HO; the proportion of HO was higher in patients with the MUTARS prosthesis (67%) than in those with the GMRS (12%) or HMRS (19%) prostheses. One explanation could be that the type of material used for the prosthesis manufacturing has been associated with the incidence of HO [15, 17].
The median time to the appearance of HO was 8 months in the present study, while HO generally appears 3 weeks to 4 weeks after THA and TKA [9]. The delayed appearance of HO in this study might be attributed to the suppressive effects of chemotherapy on bone turnover [4], although we could not find any publications to support the chemotherapy effect on postponing HO presentation.
Features That Allow the Surgeon to Differentiate HO from Locally Recurrent Tumors
The median time to the appearance of HO was shorter than the median time to the appearance of LR (8 months [3-13] versus 16 months [11-21]; p < 0.001). The median time to the initial LR of osteosarcoma was reported to be 14 months in an earlier study [20]. Accordingly, the time of presentation might help the differentiation of HO from LR whenever the diagnosis is complicated. In this respect, a lesion that appears within 12 months is more likely to be HO than LR, because only one HO in the current study appeared at more than 12 months after a knee oncologic prosthesis was implanted, and only one LR occurred at less than 12 months. The frequency of pain at presentation was greater in patients with LR than in those with HO (86% [24 of 28] versus 14% [3 of 21]; p < 0.001). Hence, the presence or absence of pain may also be useful for differentiating HO from LR. Pain in the affected extremity is the most common symptom of primary or recurrent bone tumors [8, 24]. According to the evidence, almost 85% of patients with osteosarcoma and 64% of patients with Ewing sarcoma present with symptoms of pain [24]. HO is asymptomatic in most patients [6]. The lesion border in radiographs was ill-defined in 19% (4 of 21) of HO patients and 100% (28 of 28) of LR patients (p < 0.001). An insight into the ossification process of HO might more help differentiation of HO from LR. The ossification process of HO consists of three steps: early-stage, mineralization, and maturation. The early-stage generally takes 10 days, and no mineralization findings are usually seen on radiographs. The mineralization stage generally occurs as early as 10 to 14 days after onset. After approximately 6 weeks, a well-defined cortical border can be observed around the HO on radiographs [18, 22]. Differentiation of HO at the mineralization stage from a locally recurrent tumor is the main challenge after implantation of oncologic knee prostheses. The lesion border is initially ill-defined on radiographs, and therefore, a simple HO might be confused with LR. In most patients of this study (17 of 21), the definitive diagnosis of HO was made radiographically by identifying a well-defined border around the lesion. In the remaining four patients, a well-defined border was not detected on radiographs, and a biopsy was done to identify the lesion etiology. According to the findings of our study, patients with a painless, mineralized mass around the prosthesis which has a well-defined border and is present in the first 12 months after surgery are more likely to have HO, whereas painful masses that have poorer or no mature mineralization occurring 12 or more months after the procedure are more likely to have LR. Obviously, larger studies will be needed to confirm this approach, but in the meantime, these findings might be useful to the clinician to guide decisions about whether or not to biopsy a specific lesion (Fig. 4).
Fig. 4.
This diagnostic algorithm may help differentiate HO at the mineralization stage from a locally recurrent tumor.
Conclusion
HO was observed in a small proportion of patients after an oncologic knee prosthesis was implanted after resection of a bone sarcoma. Postoperative infection, male sex, and the MUTARS prosthesis might be associated with the development of HO. A well-defined border on radiographs is reassuring for distinguishing HO from a locally recurrent tumor. In the mineralization stage of HO, in which a well-defined border is not traceable on radiographs, a biopsy may be warranted. Presentation within 1 year postoperatively and the absence of pain may help differentiate HO from LR.
Footnotes
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.
The authors certify that neither they, nor any members of their immediate families, have 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.
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.
Each author certifies that his institution approved the human protocol for this investigation and that all investigations were conducted in conformity with ethical principles of research.
This work was performed at the Bone and Joint Reconstruction Research Center, Shafa Orthopedic Hospital, Iran University of Medical Sciences, Tehran, Iran.
References
- 1.Ahrengart L, Lindgren U. Functional significance of heterotopic bone formation after total hip arthroplasty. J Arthroplasty. 1989;4:125-131. [DOI] [PubMed] [Google Scholar]
- 2.Baroudi M, Derome P, Malo M. Severe heterotopic ossification and stiffness after revision knee surgery for a periprosthetic fracture. Arthroplast Today. 2017;3:147-150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Barrack RL, Brumfield CS, Rorabeck CH, Cleland D, Myers L. Heterotopic ossification after revision total knee arthroplasty. Clin Orthop Relat Res. 2002;404:208-213. [DOI] [PubMed] [Google Scholar]
- 4.Bath LE, Crofton PM, Evans AE, et al. Bone turnover and growth during and after chemotherapy in children with solid tumors. Pediatr Res. 2004;55:224-230. [DOI] [PubMed] [Google Scholar]
- 5.Dahlin DC, Unni KK. Bone Tumors: General Aaspects and Data on 8,547 Cases. Oak Ridge, TN: U.S. Department of Energy Office of Scientific and Technical Information; 1986. [Google Scholar]
- 6.Dalury DF, Jiranek WA. The incidence of heterotopic ossification after total knee arthroplasty. J Arthroplasty. 2004;19:447-452. [DOI] [PubMed] [Google Scholar]
- 7.Davies A, Vanel D. Follow-Up of musculoskeletal tumors I. Local recurrence. Eur Radiol. 1998;8:791-799. [DOI] [PubMed] [Google Scholar]
- 8.Igarashi K Yamamoto N Shirai T, et al. Late recurrence of osteosarcoma: a report of two cases. J Orthop Surg. 2014;22:415-419. [DOI] [PubMed] [Google Scholar]
- 9.Iorio R, Healy WL. Heterotopic ossification after hip and knee arthroplasty: risk factors, prevention, and treatment. J Am Acad Orthop Surg. 2002;10:409-416. [DOI] [PubMed] [Google Scholar]
- 10.Jamshidi K, Bahrabadi M, Mirzaei A. Long-term results of osteoarticular allograft reconstruction in children with distal femoral bone tumors. Arch Bone Jt Surg. 2017;5:296-301. [PMC free article] [PubMed] [Google Scholar]
- 11.Jamshidi K, Najd-Mazhar F, Ahangar FA. The effect of cement augmentation and anteromedial plating on proximal humerus allograft reconstruction. J Orthop Sci. 2017;22:69-74. [DOI] [PubMed] [Google Scholar]
- 12.Jamshidi K, Shirazi MR, Bagherifard A, Mirzaei A. Curettage, phenolization, and cementation in paediatric Ewing’s sarcoma with a complete radiological response to neoadjuvant chemotherapy. Int Orthop. 2019;43:467-473. [DOI] [PubMed] [Google Scholar]
- 13.Kotz RI. Progress in musculoskeletal oncology from 1922 - 2012. Int Orthop. 2014;38:1113-1122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Łęgosz P, Otworowski M, Sibilska A, et al. Heterotopic ossification: a challenging complication of total hip Arthroplasty: risk factors, diagnosis, prophylaxis, and treatment. Biomed Res Int. 2019;16:3860142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lindholm T, Viljakka T, Vankka E, Popov L. Development of heterotopic ossification around the hip. Arch Orthop Trauma Surg. 1986;105:263-267. [DOI] [PubMed] [Google Scholar]
- 16.Manrique J, Alijanipour P, Heller S, Dove M, Parvizi J. Increased Risk of Heterotopic Ossification Following Revision Hip Arthroplasty for Periprosthetic Joint Infection. Arch Bone Jt Surg. 2018;6:486-491. [PMC free article] [PubMed] [Google Scholar]
- 17.Ohlmeier M, Krenn V, Thiesen DM, Sandiford NA, Gehrke T, Citak M. Heterotopic Ossification in Orthopaedic and Trauma surgery: A Histopathological Ossification Score. Sci Rep. 2019;9:18401-18401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Orzel JA, Rudd TG. Heterotopic bone formation: clinical, laboratory, and imaging correlation. J Nucl Med. 1985;26:125-132. [PubMed] [Google Scholar]
- 19.Ottaviani G, Jaffe N. The epidemiology of osteosarcoma. Cancer Treat Res. 2009;152:3-13. [DOI] [PubMed] [Google Scholar]
- 20.Rodriguez‐Galindo C, Shah N, McCarville MB, et al. Outcome after local recurrence of osteosarcoma: the St. Jude Children's Research Hospital experience (1970–2000). Cancer. 2004;100:1928-1935. [DOI] [PubMed] [Google Scholar]
- 21.Rosteius T, Rausch V, Pätzholz S, Lotzien S, Baecker H, Schildhauer TA, Geßmann J. Incidence and risk factors for heterotopic ossification following periprosthetic joint infection of the hip. Archives Orthop Trauma Surg. 2019;139:1307-1314. [DOI] [PubMed] [Google Scholar]
- 22.Subhawong TK, Fishman EK, Swart JE, Carrino JA, Attar S, Fayad LM. Soft-tissue masses and masslike conditions: what does CT add to diagnosis and management? AJR Am J Roentgenol. 2010;194:1559-1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Verma N, Cowperthwaite MC, Burnett MG, Markey MK. Differentiating tumor recurrence from treatment necrosis: a review of neuro-oncologic imaging strategies. Neuro Oncol. 2013;15:515-534. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Widhe B, Widhe T. Initial symptoms and clinical features in osteosarcoma and Ewing sarcoma. J Bone Joint Surg Am. 2000;82:667-674. [DOI] [PubMed] [Google Scholar]
- 25.Wongworawat MD Dobbs MB Gebhardt MC, et al. Editorial: Estimating survivorship in the face of competing risks. Clin Orthop Relat Res. 2015;473:1173-1176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Yi S Kim KN Yang MS, et al. Difference in occurrence of heterotopic ossification according to prosthesis type in the cervical artificial disc replacement. Spine (Phila Pa 1976). 2010;35:1556-1561. [DOI] [PubMed] [Google Scholar]