What Are the Complication Rates and Factors Associated With Total Femur Replacement After Tumor Resection? Findings From the Japanese Musculoskeletal Oncology Group

Abstract

Background

Patients undergoing massive tumor resection and total femur replacement (TFR) face a substantial risk of hip dislocation and infection, often resulting in multiple implant revisions or hip disarticulation. These complications can impact their independence and prognosis. Additionally, their shorter life expectancy is influenced by challenges in achieving local radical resection and controlling metastases. Identifying suitable candidates for TFR is vital, necessitating investigations into dislocation, infection, implant failure rates, local recurrence, overall survival, and associated factors.

Questions/purposes

(1) What is the postsurgical complication (hip dislocation and infection) rate and factors associated with postsurgical complications in patients who underwent TFR after tumor resection? (2) What is the local recurrence rate, implant failure rate, overall survival rate, and factors associated with local recurrence and implant failure?

Methods

We retrospectively evaluated 42 patients (median [range] age 47 years [10 to 79 years]) who underwent TFR and tumor resection at the time of the same surgical procedure between 1990 and 2020 at 12 registered institutions that specialized in tumor treatment in Japan. A total of 55% (23) of the patients were men, and 79% (33) had bone sarcoma. The median (range) follow-up period was 36.5 months (2 to 327 months). Of the 42 patients, 12% (5) were lost to follow-up before 2 years without meeting a study endpoint (postsurgical complications, revision, or amputation), and another 19% (8) died before 2 years with implants intact, leaving 69% (29) of the original group who had either follow-up of at least 2 years or met a study endpoint before the minimum surveillance duration. Another 10% (4) had a minimum of 2 years of follow-up but had not been seen in the past 5 years. Infection was defined as deep-seated infection involving soft tissues, bones, joints, and the area around the implant. We did not consider superficial infections. Implant failure was defined when a patient underwent reimplantation or amputation. The complication and implant failure rates were assessed by the cumulative incidence function method, considering competing events. The Kaplan-Meier method was used to estimate the overall survival rate.

Results

The 1-month, 6-month, 1-year, and 2-year dislocation rates were 5%, 12%, 14%, and 14%, respectively. The 1-month, 6-month, 1-year, and 2-year infection rates were 5%, 7%, 10%, and 15%, respectively. Multivariable analyses for hip dislocation and infection revealed that resection of the abductor muscles and large tumor size were positively associated with hip dislocation. The 6-month, 1-year, and 2-year local recurrence rates were 5%, 15%, and 15%, respectively. The 6-month, 1-year, 2-year, and 5-year implant failure rates were 5% (95% confidence interval 1% to 15%), 7% (95% CI 2% to 18%), 16% (95% CI 6% to 29%), and 16% (95% CI 6% to 29%), respectively. Multivariable analyses of local recurrence and implant failure that led to reimplantation or amputation revealed that a positive surgical margin was positively associated with local recurrence. The 1-year, 2-year, and 5-year overall patient survival rates were 95% (95% CI 87% to 102%), 77% (95% CI 64% to 91%), and 64% (95% CI 48% to 81%), respectively.

Conclusion

Hip dislocation, infection, and local recurrence were frequently observed in patients who received massive tumor resection and TFR in our study, eventually leading to reimplantation or amputation. Preserving the abductor muscles and resecting the tumor with a wide margin can prevent postoperative dislocation and local recurrence. Future research should focus on patient selection criteria, prevention of hip dislocation, and innovative treatments.

Level of Evidence

Level IV, therapeutic study.

Introduction

The femur and thigh are common sites of bone and soft tissue sarcomas [4]. Typically, curative surgical treatment is achieved through resection of the soft tissue or bone surrounding these tumors, followed by bone and soft tissue reconstruction. Despite advances in implant techniques, reconstructing the femur after resection of large tumors remains challenging. Various reconstruction methods have been developed to preserve bone stock, such as using allograft bones with hip or knee prostheses [12] and biological reconstruction with heat-treated or irradiated autogenous bone [9]. Another option is the Compliant Pre-Stress Implant (Compress, Zimmer Biomet Corp), which is suitable for patients with limited femur bone stock [2]. Total femur replacement (TFR) remains the last resort for massive bone defects. Introduced in 1977, TFR is used to reconstruct massive bone defects after tumor resection or trauma [14]. Previous studies on TFR with small samples have reported high rates of complications, including hip dislocation and infection, often leading to multiple implant revisions or hip disarticulations and negatively impacting patients' prognoses [1, 5, 10, 11, 13, 16, 20, 21]. Additionally, patients with large tumors may have poor prognoses owing to the challenges of achieving local radical resection and controlling metastases. However, no published reports have addressed which patients undergoing tumor resection and TFR are more prone to complications and how to prevent these complications. We performed a multicenter, retrospective study of patients with bone or soft tissue tumors who underwent tumor resection and TFR in Japanese Musculoskeletal Oncology Group institutions, which consist of surgeons, oncologists, and pathologists specializing in bone and soft tissue sarcoma who provide standard treatment in Japan. We gathered information on 53 patients to evaluate the clinical and functional outcomes of TFR.

We therefore asked: (1) What is the postsurgical complication (hip dislocation and infection) rate and factors associated with postsurgical complications in patients who underwent TFR after tumor resection? (2) What is the local recurrence rate, implant failure rate, overall survival rate, and factors associated with local recurrence and implant failure?

Patients and Methods

Study Design and Setting

This was a multicenter, retrospective study of patients with bone or soft tissue tumors who underwent tumor resection and TFR in Japanese Musculoskeletal Oncology Group institutions between 1990 and 2020. The Japanese Musculoskeletal Oncology Group, established in 1981, consists of 250 orthopaedic oncologists and 83 institutions, including university hospitals and regional and national cancer center hospitals. The group's mission is to investigate the clinical outcomes of rare musculoskeletal tumors. Data for this observational study were collected from the original medical records of 32 participating institutions, with approval from respective institutional review boards. Patient data were tracked in this registry, although patient deaths were confirmed by the registering orthopaedic oncologist and not linked to the national death registry. In patients undergoing revision THA, confirmation of the initial surgery's implants usually requires contacting the facility where the surgery was performed. However, when follow-up observation ceases, the possibility of a patient undergoing revision surgery at another medical institution cannot be ruled out.

Patients

The study included 53 patients treated with TFR at 32 Japanese Musculoskeletal Oncology Group institutions between 1990 and 2020. Eleven patients who underwent TFR without tumor resection as salvage for bone fractures or implant failure after the primary operation were excluded. The study assessed 42 patients who underwent TFR and tumor resection simultaneously. All procedures aimed to achieve local curative treatment and preserve the limb. Of 42 patients, 12% (5) were lost to follow-up before 2 years without meeting a study endpoint (postsurgical complications, revision, or amputation), and another 19% (8) died before 2 years with their implants intact, leaving 69% (29) of the original group who had either follow-up of at least 2 years or met a study endpoint before the minimum surveillance duration. Another 10% (4) had a minimum of 2 years of follow-up but had not been seen in the past 5 years.

Descriptive Data

Clinical features of the patients, such as type of hospital, gender, age at surgery, histologic diagnosis, tumor size, presence of skip lesions and metastasis, and performance status (Eastern Cooperative Oncology Group) were collected. High-volume institutions were defined as hospitals that treated more than eight patients. Tumor size was defined as the maximum diameter of the tumor. Treatment-related factors, reasons for TFR, chemotherapy, radiation therapy, surgical status, operation time, blood loss during surgery, resected muscles and nerves, surgical margin, type of reconstruction, and type of implant were analyzed. The reason for undergoing TFR after tumor resection was classified into three categories: large tumor or skip lesion within the femur, relapse after previous tumor resection, or unplanned surgery in patients with suspected massive contamination in the femur. The surgical margin was classified into three groups: R0 (margin negative on microscopy), R1 (margin positive on microscopy with no evident residual tumor on macroscopic examination), and R2 (remaining tumor on macroscopic examination). A resected muscle was defined as a muscle that was planned to be resected with the tumor to achieve an adequate margin and was resected as planned. Muscles that were only attached to the femur were not included. The number of resected quadriceps muscles was counted and classified into five groups (0 to 4).

All 42 patients who were included in this study were treated at one of 12 institutions, and 45% of the patients (19) were treated in high-volume institutions (Table 1). A total of 55% of the patients were men (23) with a median (range) age of 47 years (10 to 79 years) at the time of surgery. The median (range) follow-up period was 37 months (2 to 327 months). The diagnoses of the primary lesion were bone sarcoma (79% [33]) including osteosarcoma (45% [19]), chondrosarcoma (21% [9]), Ewing sarcoma (10% [4]), and undifferentiated pleomorphic sarcoma of bone (2% [1]); soft tissue sarcoma (14% [6]); and metastatic tumor (7% [3]). The median tumor size was 20 cm (5 to 35 cm), and 17% of the patients (7) had skip lesion in the femur. Among the three patients with tumors smaller than 10 cm, two had skip lesions, while the remaining patient had a tumor contaminated by previous surgery. Twenty-six percent of patients (11) had metastases at the time of surgery. The performance status of the patients was 0 points (38% [16]), 1 point (57% [24]), and 2 points (5% [2]).

Table 1.

Patient demographics

Demographics	Value (n = 42)
High-volume hospitala	45 (19)
Age in years (median 46.5 years) > 50	45 (19)
Male gender	55 (23)
Diagnosis
Bone sarcoma	79 (33)
Osteosarcoma	45 (19)
Chondrosarcoma	21 (9)
Ewing sarcoma	10 (4)
Undifferentiated pleomorphic sarcoma of bone	2 (1)
Soft tissue sarcoma	14 (6)
Metastatic tumor	7 (3)
Tumor size in cm (median 20 cm) > 20	43 (18)
Skip lesion +	17 (7)
Metastasis +	26 (11)
Performance scale
0	38 (16)
1	57 (24)
2	5 (2)

Data presented as % (n).

^aHigh-volume hospitals are facilities that treated more than eight patients.

The reasons for undergoing TFR were as follows: 83% of the patients (35 of 42) underwent TFR because they had a large tumor or a skip lesion in the femur, 12% (5 of 42) underwent TFR because they had relapse after previous tumor resection, and 5% (2 of 42) had unplanned surgery that left suspected massive contamination in the femur (Table 2). The two patients in the third group had sarcomas that were misdiagnosed at another hospital and were initially treated with internal fixation using an intramedullary rod. The mean operation time was 6 hours 34 minutes (range 2 hours 54 minutes to 13 hours 57 minutes). Three patients had a surgical duration of longer than 12 hours, mainly because of soft tissue reconstruction with a free flap or artery reconstruction. The mean intraoperative blood loss was 937 mL (177 to 2520 mL). Information about the surgical margin was available for all patients. R0 surgical margins were achieved in 83% of the patients (35) and R1 was achieved in 17% (7). The muscles surrounding the tumor were resected together with the tumor to achieve R0 or R1 resection. At least one of the hip abductor muscles (gluteus medius, gluteus minimus, or tensor fasciae latae) was resected in nine patients, and the rectus femoris was resected in six. Femoral nerve resection was performed in three patients. Two patients underwent soft tissue reconstruction using a free flap. Artery reconstruction was performed in three patients. Hip reconstruction was performed using mesh in eight patients or a Leeds-Keio artificial ligament (Xiros) in one. No patients received an acetabular cup component. TFR was performed using the Kyocera Modular Limb Salvage System (Kyocera Medical Corp) in 43% of the patients (18), the Howmedica Modular Replacement System (Howmedica Osteonics Corp) in 40% (17), the Orthopedic Salvage System (Zimmer Biomet Corp) in 10% (4), and the Global Modular Replacement System (Stryker Corp) in 7% (3). Five percent of the patients (2) received expandable prostheses. Fourteen percent (6) received radiation therapy after surgery, and 55% of the patients (23) underwent adjuvant chemotherapy. Adjuvant chemotherapy was administered before and after surgery to 14 patients with osteosarcomas, three with Ewing sarcomas, and one with undifferentiated pleomorphic sarcoma of bone. Two patients with osteosarcomas, one with Ewing sarcoma, and one with soft tissue sarcoma received adjuvant chemotherapy only before the surgery, and one patient with soft tissue sarcoma received adjuvant chemotherapy only after the surgery.

Table 2.

Treatment parameters

Treatment parameter	Value (n = 42)
Reason for treating with TFR
Large size or skip lesion	83 (35)
Relapse after the primary surgery	12 (5)
Unplanned primary surgery	5 (2)
Operation time in hours (mean 6:34) > 6 hours	45 (19)
Blood loss in mL (mean 937 mL) > 900 mL	48 (20)
Surgical margin R0	83 (35)
Abductor muscles resection	21 (9)
Rectus femoris resection	14 (6)
Number of quadriceps resection
0	7 (3)
1	17 (7)
2	29 (12)
3	40 (17)
4	7 (3)
Nerve resection	7 (3)
Soft tissue reconstruction	5 (2)
Artery reconstruction	7 (3)
Hip reconstruction with artificial materials	21 (9)
Type of implantation system
KMLS	43 (18)
HMRS	40 (17)
OSS	10 (4)
GMRS	7 (3)
Expandible prosthesis	5 (2)
Radiation therapy	14 (6)
Chemotherapy	55 (23)

Data presented as % (n).

KMLS = Kyocera Modular Limb Salvage System (Kyocera Medical Corporation); HMRS = Howmedica Modular Replacement System (Howmedica Osteonics Corporation); OSS = Orthopedic Salvage System (Zimmer Biomet Corporation); GMRS = Global Modular Replacement System (Stryker Corporation).

Primary and Secondary Study Outcomes

Our primary study goal was to reveal the postoperative complication rate. To achieve this, we used the cumulative incidence function method to analyze occurrences of postoperative hip dislocation, infection, local recurrence, and implant failure. Infection was characterized as infection affecting the deep soft tissue, bone, joint, or areas around the implant. Superficial infections were not included. The time of infection was defined as the moment surgical debridement was performed. We defined implant failure as the need for any revision of the prosthesis or part of the prosthesis (such as aseptic loosening, fracture of the implant, infection, breakage, or acetabular resurfacing for hip dislocation) or failure treated with amputation.

Our secondary study goals were to evaluate overall patient survival and identify factors associated with postoperative complications. Overall patient survival was analyzed using the Kaplan-Meier method. The factors of hip dislocation, infection, local recurrence, and implant failure were analyzed with multivariable analyses. The duration of overall survival was defined as the time from surgery to the date of the patient’s final visit to the clinic or death.

Ethical Approval

Ethical approval for this study was obtained from the National Cancer Center Hospital, Tokyo, Japan (approval number 2017-218).

Statistical Analysis

For the baseline variables, we constructed summary statistics, using frequencies and proportions for categorical data and means and standard deviations for continuous variables. For the time-to-event outcomes of overall survival, we compared the length of time to the first event using the log-rank test, and the Kaplan-Meier method was used to estimate the absolute risk of each event for each group. Hazard ratios and 95% confidence intervals were estimated using a Cox proportional hazards model. Additionally, the cumulative incidence function method was used to estimate the probabilities of dislocation, infection, local recurrence, and implant failure as time-to-event outcomes, considering competing events. Differences between groups were assessed using the Gray test. A competing risk analysis was conducted using the Fine-Gray generalization of the proportional hazards model. Competing risks are defined as events that prevent the occurrence of the outcome of interest, considering death as one of these competing risks. Multivariable analyses were performed using a Cox proportional hazards model and Fine-Gray subdistribution hazard model. Variables that exhibited significant differences in the univariate analysis were included in the multivariable analysis. All comparisons were planned, and the tests were two-sided. We considered p values of less than 0.05 to indicate statistical significance. All statistical analyses were performed using SAS v9.4 (SAS Institute).

Results

Postsurgical Complication Rate and Associated Factors

Postoperative infection occurred in 19% (8 of 42) of patients, and hip dislocation occurred in 17% (7 of 42) (Table 3). A cumulative incidence function curve and Cox proportional hazard models were used to identify the dislocation rate, infection rate, and the potential prognostic factors for hip dislocation and infection. The 1-month, 6-month, 1-year, and 2-year dislocation rates were 5%, 12%, 14%, and 14%, respectively (Fig. 1). The 1-month, 6-month, 1-year, and 2-year infection rates were 5%, 7%, 10%, and 15%, respectively (Fig. 2). Multivariable analyses were performed to determine factors associated with hip dislocation and infection. Resection of the abductor muscles (HR 8.4 [95% CI 1.7 to 42.0]; p = 0.01) and tumor larger than 20 cm (HR 5.8 [95% CI 1.5 to 22.4]; p = 0.01) were associated with hip dislocation in the multivariable analysis (Table 4). A multivariable analysis was also performed for deep infection, but none of the factors showed an association (Supplemental Table 1; http://links.lww.com/CORR/B243).

Table 3.

Treatment outcomes

Treatment outcome	Value (n = 42)
Infection	19 (8)
Hip dislocation	17 (7)
Changing prosthesis	7 (3)
Hip disarticulation	10 (4)
MSTS score (n = 32)
0-10	12 (5)
11-15	21 (9)
16-20	31 (13)
21-30	12 (5)
Local recurrence	17 (7)
Clinical outcome
NED	40 (17)
AWD	10 (4)
DOD	48 (20)
DOOD	2 (1)

Data presented as % (n).

MSTS = Musculoskeletal Tumor Society score; NED = no evidence of disease; AWD = alive with disease; DOD = died of disease; DOOD = died of other disease.

Fig. 1.

The cumulative incidence function curve for hip dislocation is shown.

Fig. 2.

The cumulative incidence function curve for infection is shown.

Table 4.

Multivariable analysis of risk factors for hip dislocation

Variable	Hazard ratio	95% CI	p value
Age older than 50 years	2.9	0.6-12.7	0.17
Tumor size over 20 cm	5.8	1.5-22.4	0.01
Abductor resection +	8.4	1.7-42.0	0.01

Local Recurrence Rate, Implant Failure Rate, Overall Survival Rate, and Associated Factors

A cumulative incidence function curve and Cox proportional hazard models were used to identify the local recurrence and implant failure rates. The 6-month, 1-year, and 2-year local recurrence rates were 5%, 15%, and 15%, respectively (Fig. 3). Seven percent (3 of 42) of patients underwent implant replacement because of hip dislocation, and 10% (4 of 42) of patients underwent hip disarticulation because of local recurrence or infection (Table 3). The 6-month, 1-year, 2-year, and 5-year implant failure rates were 5% (95% CI 1% to 15%), 7% (95% CI 2% to 18%), 16% (95% CI 6% to 29%), and 16% (95% CI 6% to 29%), respectively (Fig. 4). The 1-year, 2-year, and 5-year overall patient survival rates were 95% (95% CI 87% to 102%), 77% (95% CI 64% to 91%), and 64% (95% CI 48% to 81%), respectively (Fig. 5). Multivariable analyses were performed to determine factors associated with local recurrence. Positive surgical margin (HR 6.2 [95% CI 1.5 to 25.6]; p = 0.012) was associated with local recurrence (Table 5). In addition, we performed a multivariable analysis to determine any factors associated with implant failure; however, we identified no such factors.

Fig. 3.

The cumulative incidence function curve for local recurrence is shown.

Fig. 4.

The cumulative incidence function curve for implant failure is shown.

Fig. 5.

The Kaplan-Meier curve for overall survival is shown.

Table 5.

Multivariable analysis of risk factors for local recurrence

Variable	Hazard ratio	95% CI	p value
Reason (Re2 + Re3)a	2.0	0.58-7.2	0.27
Margin (R1)	6.2	1.5-25.6	0.012

^aRe2 = relapse after the primary surgery; Re3 = unplanned primary surgery.

Discussion

Patients who undergo massive tumor resection and TFR are at a high risk of experiencing hip dislocation and infection after surgery, often necessitating multiple implant revisions or hip disarticulation. These complications can lead to a loss of independence for patients and worsen their prognosis. Furthermore, these patients tend to have a shorter life expectancy because of challenges in achieving local radical resection and controlling metastases. Nonetheless, it remains unclear which patients are the most susceptible to implant revisions or amputations, owing to several postsurgical complications and local recurrence and how to mitigate these challenges. Thus, it is important to investigate the dislocation, infection, implant failure (defined as reimplantation or hip disarticulation), local recurrence, and overall survival rates and factors related to hip dislocation, infection, local recurrence, and implant failure. To clarify this, we reviewed the data of 42 patients who underwent TFR after tumor resection in Japanese Musculoskeletal Oncology Group institutions, which consist of surgeons, oncologists, and pathologists specializing in bone and soft tissue sarcoma who provide standard treatment in Japan. We evaluated the postsurgical complication rate and associated factors for each complication, as well as the local recurrence rate, implant failure rate, overall survival rate, and factors associated with local recurrence and implant failure. Our study revealed that resection of abductor muscles and tumor larger than 20 cm were associated with hip dislocation, while positive surgical margins were associated with local recurrence. Efforts to prevent or lessen the occurrence of hip dislocation, deep infection, and local recurrence might decrease the proportion of reimplantations and amputations, leading to better limb prognosis.

Limitations

First, the number of patients was small because of the rarity of these extensive tumors and the infrequent use of TFR. Second, although 59% of the patients were followed for more than 5 years, the duration of follow-up varied greatly among the patients, which could have resulted in a wide margin of error for long-term outcomes, causing transfer bias. Third, this study was a chart review–based, retrospective, observational study, which could have caused selection, assessment, and cotreatment biases. Specifically, factors determining the indication for TFR were not clearly addressed. For example, institutions with experienced plastic surgeons and vascular surgeons may prefer limb salvage surgery over amputation. Additionally, the use of adjuvant chemotherapy in certain patients deviates from the current standard practice because of the 30-year span of the study and the varying characteristics of individual patients. It is unclear whether the reasons for not administering chemotherapy in these patients were because of patient factors or the treatment facility's policy at that time. Furthermore, survival outcomes and complications are influenced by many noncancer factors that were not assessed in this study. Because of variation in the experience of operating surgeons, there are potential differences in surgical duration and blood loss among patients. Four different types of implants were used by various surgeons in different institutions, introducing potential confounding factors.

Postsurgical Complication Rate and Associated Factors

In this study, we demonstrated in multivariable analyses that resection of the abductor muscles and tumor size were strongly associated with hip dislocation. Analyzing the characteristics of seven patients with hip dislocation, we found that in five patients, the abductor muscle was excised, while the remaining two patients had tumor diameters larger than 20 cm. The abductor muscle group is involved in the stability of the hip, and removing these muscles increases the risk of hip dislocation. Although it is unlikely that a large tumor diameter directly causes hip dislocation, tumors with a larger diameter are associated with extensive extracortical involvement, leading to increased removal of the muscles around the femur. Studies [5, 11, 16, 21, 22] have reported the percentage of hip dislocation and infection after TFR (Table 6). Hip dislocation and infection are the two most common postoperative complications, occurring in approximately 10% to 20% of patients undergoing TFR, and the rates of hip dislocation and infection assessed in the present study were comparable to those in past reports. Sevelda et al. [21] pointed out that the loss of soft tissue around the femur may put a patient at risk for stabilizing the implant, and the tumor size can indirectly affect the ease of dislocation. Several techniques to avoid hip dislocation have been attempted. In our study, 21% (9 of 42) of patients underwent mesh or artificial ligament coverage for hip reconstruction, which has been reported to prevent hip dislocation [5]. However, the multivariable analyses revealed that the use of mesh and artificial ligaments was not associated with hip dislocation. In one patient with a massive muscle defect, a soft tissue free flap was used to cover the hip; however, this patient also experienced hip dislocation. Hip dislocation occurred in seven patients; in two, the acetabulum was resurfaced in an additional operation, and this patient did not experience further hip dislocation. Acetabular resurfacing during the primary surgery to lessen the likelihood of dislocation is controversial for patients undergoing endoprosthetic reconstruction after tumor resection. Bickels et al. [3] reported that the important factors for reducing the likelihood of hip dislocation after reconstruction with a proximal femur endoprosthesis were acetabular preservation, capsulorrhaphy, and reconstruction of the abductor muscles, suggesting that acetabular resurfacing will not prevent hip dislocation. However, those authors did not offer a solution to prevent dislocation when massive abductor muscle resection was needed. Further studies should be conducted to identify the most effective approach in order to prevent hip dislocation after abductor muscle resection.

Table 6.

Dislocation, infection, local recurrence, and overall patient survival in previous studies and the current study

Study	Number of patients	Follow-up in months, median (range)	Dislocation, % (n)	Infection, % (n)	Local recurrence, % (n)	Overall survival
Natarajan et al. [16]	17	54 (11-168)	12% (2)	12% (2)	6% (1)	82% in 14 years
Ruggieri et al. [20]	21	48 (1-68)	0% (0)	10% (2)	0% (0)	38% in 16 yearsa
Kalra et al. [11]	26	57 (3-348)	11% (4)	7% (3)	7% (3)	50% in 5 yearsb
Sevelda et al. [21], conT	34	57 (1-280)	24% (8)	18% (6)	3% (1)	59% in 10 years
Sevelda et al. [21], exT	10	172 (43-289)	10% (1)	10% (1)	0% (0)	ND
Du et al. [5]	58	43 (6-129)	21% (12)	14% (8)	ND	ND
Sewell et al. [22]	33	50 (9-197)	18% (6)	3% (1)	9% (3)	32% in 5 years
Present study	42	37 (2-327)	14%c (7)	15%c (8)	15%c (7)	64% in 5 years

Dislocation, infection, and local recurrence are presented as crude percentages (the number of patients with dislocation or infection divided by the number who underwent surgery), except for the present study. Two-year dislocation, infection, and local recurrence rate were calculated using survivorship.

^aDisease-free survival.

^bOverall survival only for patients without metastases.

^c2-year dislocation, infection, and local recurrence rate are calculated using survivorship. conT = study only registered the conventional TFR; exT = study only registered the expandible TFR; ND = no data.

Two studies [6, 15] have identified factors associated with deep infection, such as prolonged operative time, being a man, having soft tissue tumors invading bones, and receiving radiation therapy, which were not associated with infection in our study. Both of these studies indicated that prolonged operative time was associated with postsurgical infection. In our study, two patients with massive muscle defects after tumor resection had a surgical duration of more than 13 hours, mainly because of soft tissue reconstructions with free flaps. Both of these patients experienced postsurgical infection. Although some authors have suggested that soft tissue reconstructions can reduce the likelihood of postsurgical infection [8, 17], our study indicates that prolonged operative time associated with free flap reconstruction might be associated with deep infection. Although not implemented in our study, an alternative approach to soft tissue reconstruction could involve using vacuum-assisted closure for soft tissue regeneration instead of a free flap, followed by skin grafting. This approach could lead to shorter surgical durations. Another technique to avoid infection is using iodine-coated implants [15] or silver-coated implants [18]. Using infection-resistant implants to reduce the infection rate may be a more ideal approach, because it does not extend the surgical time compared with soft tissue reconstruction. However, substantial evidence indicating a reduced infection rate with the use of silver-coated implants is lacking [18], and the challenge stems from the high cost of the implant. In patients who are believed to be at high risk of experiencing postoperative infection, careful intraoperative hemostasis, attempts to shorten the operative time, and careful soft tissue closure or the use of vacuum dressings might be beneficial to reduce the likelihood of infection.

Local Recurrence Rate, Implant Failure Rate, Overall Survival Rate, and Associated Factors

The 2-year local recurrence rate was 15%, which was higher than rates that ranged from 0% to 9% in past reports (Table 6). When evaluating the indications for TFR, it is essential to consider not only the possible postoperative complications but also the patient's overall prognosis. Furthermore, because local recurrence can be as challenging to treat as deep infection and may lead to hip disarticulation, it is necessary to assess factors associated with local recurrence. We could not specify the reason for the high local recurrence rate we observed compared with those in past reports, but the profile of the diagnoses was similar to that of a report that showed a 9% recurrence rate [22]; approximately 30% of patients did not have either osteosarcoma or Ewing sarcoma. Osteosarcoma and Ewing sarcoma are both known as chemotherapy-sensitive sarcomas, and adjuvant chemotherapy before surgery may be helpful in achieving an R0 margin. Other tumors in our study, such as undifferentiated pleomorphic sarcoma of bone or soft tissue sarcoma, may not have the same options, and adjuvant treatments might not help with local control. Moreover, we found that a positive surgical margin was related to local recurrence. Considering that three of four patients who ultimately received amputation after TFR had local recurrence, we believe it is important to aim for an R0 surgical margin when performing tumor resection and TFR to spare the limb.

In our study, the 2-year and 5-year implant failure rate was 16%. Because of small samples, only a few reports have elucidated the implant failure rate after tumor resection. To the best of our knowledge, only one report performed a competing risk analysis of implant failure or implant survival rate [21]. In that report, the authors showed that the 5-year revision-free implant survival rate for conventional TFR was 48%, and the implant failure rate was 50%. Other studies have analyzed the revision-free implant survival using the Kaplan-Meier method, showing good prognosis ranging from 80% to 100% over 10 years [11, 22]. Comparing implant survival or implant failure rates with reports of others is difficult because the definition of implant failure, lengths of follow-up, patients’ backgrounds, and overall survival differ. However, our study, which observed a high recurrence rate, confirmed an implant failure rate consistent with that of previous research.

We could not specify any factors associated with implant failure, possibly owing to numerous confounding factors that are related to implant failure. In our study, three of seven patients underwent implant replacement or amputation because of local recurrence, two because of hip dislocation, one because of infection, and one because of local recurrence and infection, indicating that the implant survival rate reflects the rate of local recurrence and postoperative complications. To avoid reimplantation and amputation after tumor resection and TFR, it is essential to limit the potential for local recurrence and postoperative complications.

Comparing overall survival rates with those in previous reports is also challenging because of differences in patient diagnoses and follow-up periods. However, the 5-year survival rate in this study for 19 patients with osteosarcoma, including patients with metastatic disease, was 80% (95% CI 59% to 100%), indicating a favorable prognosis, even compared with patients with osteosarcoma who undergo wide resection with endoprosthesis reconstruction (other than TFR) or amputation [7, 19]. For patients with conditions such as osteosarcoma, for whom chemotherapy has high efficacy and the long-term prognosis is promising, a favorable implant outcome becomes a crucial factor for patient satisfaction, especially for those with excellent long-term prospects. Additionally, with advancements in systemic chemotherapy, the long-term prognosis for patients who have cancer with bone metastases has improved, leading to an anticipated increase in the indication of TFR for patients with femoral bone metastases. Consequently, controlling local recurrence and preventing complications become even more critical in such patients.

Conclusion

In our study, we observed hip dislocation, infection, and local recurrence rates of 14%, 15%, and 15%, respectively, at the 2-year interval. Factors associated with hip dislocation included abductor muscle resection and tumor size, while positive surgical margins were linked to local recurrence. The 2-year and 5-year implant failure rates were 16%, and the 5-year overall patient survival rate was 64%. Taking measures to reduce the likelihood of hip dislocation, deep infection, and local recurrence may reduce the need for reimplantation procedures and amputations. Future studies should focus on identifying specific patient selection criteria for TFR and determining better approaches to reduce the likelihood of hip dislocation after abductor muscle resection. Additionally, investigating new surgical techniques or infection-resistant implants for high-risk patients might be helpful in reducing the likelihood of postsurgical infections. Long-term follow-up studies comparing different implant types and techniques should be conducted to determine the most effective and reliable TFR method in these patients. Furthermore, exploring multidisciplinary approaches, including advancements in systemic chemotherapy and innovative treatment options, will likely improve the overall prognosis for patients with bone metastases and enhance the outcomes of TFR after tumor resection.

Supplementary Material

abjs-482-702-s001.docx ^{(58.8KB, docx)}