Functional outcomes and survival analysis of modular endoprostheses following resection of upper extremity tumors: a retrospective clinical study

Abstract

Background

Advances in chemotherapy and radiotherapy have significantly improved patient survival. Furthermore, advancements in implant technology have enabled the successful application of endoprostheses in many cases in which reconstruction via biological methods is not feasible. This study aimed to assess implant survival, overall patient survival, and functional outcomes in patients who underwent upper extremity tumor resection.

Methods

This study had a retrospective design. Patients who underwent resection and endoprosthetic reconstruction between 2011 and 2020 were evaluated. After applying the inclusion and exclusion criteria, 90 patients were included in this study. Factors potentially influencing endoprosthesis and patient survival were identified and analyzed. Functional assessment was performed by measuring the Musculoskeletal Tumor Society (MSTS) scores of patients during outpatient clinic follow-ups.

Results

The mean follow-up period was 36 ± 19.4 months. Implant failure occurred in eight of the 90 patients, resulting in a total endoprosthesis survival rate of 91.1%. Among the eight patients with implant failure, aseptic loosening was observed in two, infection in four, and tumor progression in two cases. During the study period, 37 patients died, resulting in an overall patient survival rate of 58.8%. The mean MSTS score for functional assessment was 22.1 ± 2.564. Studies in the literature suggest that the use of synthetic meshes contributes to improved functional outcomes. The mean MSTS score was 23.2 ± 2.370 for patients with synthetic mesh, compared to 21.5 ± 2.575 for those without, a statistically significant difference (P = 0.005). The overall patient survival rate was 58.8%. Two of the 90 patients underwent amputation, and the limb survival rate was 97.7%. The notably higher implant survival rate compared to patient survival and high limb survival rates indicate favorable outcomes consistent with the intended use of endoprosthetic reconstructions.

Conclusion

Modular endoprostheses are a preferred successful reconstruction method following upper extremity tumor resections owing to their long-term survival and low complication rates.

Level of evidence

Level IV

Background

Until the mid-20th century, amputation was commonly used for the standard treatment for musculoskeletal tumors. Endoprosthetic reconstruction was first implemented by Moore in 1940 [1]. The initial description of proximal humerus resection and endoprosthetic reconstruction in the literature dates back to Marcove in 1977 [2].

Advancements in chemotherapy and radiotherapy in contemporary medicine have significantly increased patient survival rates [1]. The goal of limb-sparing surgery is to provide patients with good function and a complication-free period throughout their lifespan [3–5]. Following tumor resection, two options are available: biological reconstruction or endoprosthetic reconstruction. In the upper extremity, biological reconstruction is preferred for diaphyseal lesions or in children who have not completed their growth period [6, 7]. Endoprosthetic reconstruction is favored in cases where biological reconstruction is not feasible, such as metaphyseal involvement or involvement of the entire humerus [8–10].

Despite the functional benefits achieved with endoprosthetic applications, potential complications and implant failures have been reported in the literature [11–13].

In this study, the survival outcomes of endoprostheses applied after resection of primary malignant and aggressive lesions, as well as metastatic lesions affecting the upper extremity were investigated. Variables potentially influencing survival were assessed, and functional evaluation of patients was conducted using the Musculoskeletal Tumor Society (MSTS) scoring system [14]. The hypothesis was upper extremity endoprostheses would demonstrate a higher implant survival rate relative to overall patient survival, along with low complication rates.

Methods

The study obtained ethical approval from the Ethics Committee of Marmara University Faculty of Medicine (Approval No: 09.2020.760). Patients who underwent resection and endoprosthetic reconstruction for primary or metastatic bone tumors of the upper extremity between 2011 and 2020 were retrospectively screened. After applying the exclusion criteria, a total of 90 patients were included in the study (Table 1). Patients with a follow-up duration of less than 12 months were excluded. This threshold was selected to allow standardized functional evaluation, as MSTS scores were assessed at the 12-month postoperative follow-up, and to ensure consistency with prior survival and functional outcome studies in orthopedic oncology.

Table 1.

Inclusion and exclusion criteria

Inclusion criteria	• Malignant, benign, or metastatic bone tumors of the upper extremity. • Soft tissue tumors requiring bone resection. • Patients undergoing endoprosthetic reconstruction.
Exclusion criteria	• Patients undergoing revision surgery. • Patients who underwent Intercalary prosthesis. • Patients with a follow-up period of less than 12 months. • Patients who underwent reverse shoulder prosthesis.

All patients in the study underwent a multidisciplinary evaluation at the Orthopedic Oncology Council [15, 16]. Radiological assessments of the affected bone, including adjacent joints, were performed using direct radiography and contrast-enhanced magnetic resonance imaging (MRI). Thoracic and abdominal computed tomography (CT) scans were performed for staging purposes. Additionally, scintigraphy was used to detect metastasis to other bones in patients with metastatic disease.

Biopsies were obtained in accordance with orthopedic oncology practice standards [17, 18].Tumor staging for patients with primary malignant bone tumors was performed according to the Enneking staging system, based on histopathological grade, local extent, and presence of metastasis, using preoperative imaging, biopsy results, and multidisciplinary oncologic evaluation. Following biopsy-confirmed diagnoses, staging was completed, and patients requiring neoadjuvant chemotherapy were scheduled for treatment by the council. Chemotherapy was administered according to tumor-specific oncologic protocols following multidisciplinary tumor board recommendations. Patients with primary malignant bone tumors received neoadjuvant and/or adjuvant chemotherapy, whereas chemotherapy was administered postoperatively in selected patients with metastatic disease or multiple myeloma. All radiotherapy treatments in this cohort were administered in the postoperative period as adjuvant therapy. No patients received preoperative radiotherapy prior to tumor resection and endoprosthetic reconstruction. Bone cement was used selectively to improve primary implant stability, particularly in patients with poor bone stock, extensive metaphyseal involvement, or previous radiotherapy, where immediate fixation was considered critical. Cemented fixation was performed using polymethylmethacrylate (PMMA) bone cement (Cemad^®, Turkey). The decision to use cemented or cementless fixation was made intraoperatively based on bone quality and resection characteristics. All surgical procedures were performed by a team of orthopedic oncology specialists.

Following the surgery, all patients were instructed to use an arm sling. Early postoperative rehabilitation included active hand and wrist exercises along with passive elbow exercises. Functional and radiological evaluations of the patients were conducted every three months in the first two years, every six months in the subsequent three years, and annually thereafter.

The start of the follow-up period for implant survival was defined as the time of tumor resection and the implementation of endoprosthetic reconstruction. The endpoint was defined as the complete removal of the prosthesis or any of its components, prosthetic revision, or amputation of the extremity. Patients experiencing implant failure were classified based on the Failure Mode Classification established in 2011 (Type 1: soft tissue failure, Type 2: aseptic loosening, Type 3: structural or mechanical failure, Type 4: infection, Type 5: tumor recurrence) [13].

Patients’ functional evaluation was conducted based on the MSTS scores that were evaluated at the 12-month postoperative follow-up. This time point was chosen to allow standardized evaluation, as functional recovery after extensive upper extremity endoprosthetic reconstruction is generally expected to stabilize by one year. Patients who died before completing 12 months of follow-up were not included in the study cohort, and therefore no MSTS evaluation was performed for these patients. For MSTS comparisons, distributional assumptions were assessed using visual inspection and normality testing. As assumptions were not violated, parametric tests were applied. Effect sizes are reported as mean differences with 95% confidence intervals. Factors potentially affecting functional outcomes were investigated, and significance tests were performed. Implant and patient survivals were analyzed using the Kaplan-Meier test. Kaplan–Meier curves were presented with numbers at risk displayed at predefined time points. Univariable Cox regression analyses were conducted, in which each variable was evaluated separately to explore associations with implant survival. Statistical analysis of the available data was conducted using SPSS 27.0 software.

Results

The study included 90 patients (43 females and 47 males) who underwent endoprosthetic reconstruction following resection of primary and metastatic tumors in the upper extremity. The mean age of the patients was 49 ± 18.52 (range: 14–84) years. The distribution of patients according to diagnosis and resection localization is provided in Table 2. Baseline demographic and clinical characteristics of the full cohort, stratified by diagnostic group and radiotherapy status, are summarized in Table 3. To improve the significance of the statistical analysis, descriptive and outcome analyses were primarily interpreted within major diagnostic categories, including primary malignant tumors, metastatic disease/multiple myeloma, and benign/aggressive tumors. Among patients with primary malignant tumors, disease stages were distributed as follows: Stage I in 7 patients (15,9%), Stage II in 26 patients (59.1%), and Stage III in 11 patients (25%). A patient flow diagram illustrating patient selection, exclusions with reasons, and final inclusion is provided in Fig. 1.

Table 2.

Patient distribution according to diagnoses, diagnostic groups and resection types

Groups	Diagnoses	Resection types			Total numbers	Total numbers of groups
		Proximal Humerus	Distal Humerus	Total Humerus
Primary malignant tumors	Osteosarcoma	5	0	0	5	44
	Ewing sarcoma	3	0	5	8
	Chondrosarcoma	26	0	0	26
	Soft tissue sarcoma	3	2	0	5
Metastasis and Multiple Myeloma	Metastasis	26	4	3	33	40
	Multiple Myeloma	4	3	0	7
Primary Benign Tumors	Desmoid Tumor	2	0	0	2	6
	Giant Cell Tumor	4	0	0	4
Total		73	9	8	90

Table 3.

Baseline demographic and clinical characteristics of the full cohort, stratified by diagnostic group and radiotherapy status. Values are presented as mean ± standard deviation or number (%), as appropriate. Analyses were performed using available data. No missing data were observed for the variables included in this table. Percentages were calculated based on the corresponding column totals. NA indicates not applicable. The proximal humerus subgroup is presented because functional analyses (MSTS) were performed primarily within this subgroup

Variable		Overall (n = 90)	Proximal humerus subgroup (n = 73)	Primary malignant (n = 44)	Metastatic/MM (n = 40)	Radiotherapy (+) (n = 44)	Radiotherapy (–) (n = 46)
Age (years), mean ± SD		49.2 ± 18.23	49.2 ± 17.7	40.4 ± 19.17	59.9 ± 11.91	59.3 ± 11.92	39.5 ± 18.68
Sex (male), n (%)		47(52.2)	37 (50.7)	21 (47.7)	25 (62.5)	27 (61.4)	20 (43.5)
Resection localization, n (%)	─ Proximal humerus	73 (81.1)	73 (100)	37 (84.1)	30 (75.0)	33 (75.0)	40 (87.0)
	─ Distal humerus	9 (10)	NA	2 (4.5)	7 (17.5)	8 (18.2)	1 (2.2)
	─ Total humerus	8 (8.9)	NA	5 (11.4)	3 (7.5)	3 (6.8)	5 (10.9)
Synthetic mesh use, n (%)		42 (46.7)	40 (54.8)	21 (47.7)	18 (45.0)	19 (43.2)	23 (50.0)
Bone cement use, n (%)		26 (28.9)	21 (28.8)	9 (20.5)	15 (37.5)	16 (36.4)	10 (21.7)
Follow-up (months), mean ± SD		36.1 ± 27.28	37.4 ± 27.07	48.1 ± 28.97	20.7 ± 16.22	22.1 ± 18.74	49.7 ± 27.47

Fig. 1.

A patient flow diagram illustrating patient selection, exclusions with reasons, and final inclusion

The mean follow-up duration for patients was 36 ± 19.4 (range: 12–107) months. During the study period, a total of 37 patients died, with 11 in the primary malignant tumor group and 26 in the metastatic tumor group.

Two different prosthesis brands were used in this study. MUTARS (TiA16V4, Implantcast ^®, Buxtehude, Germany) modular humeral prostheses were used between 2011 and 2016, whereas TMTS (CoCrMo, Hipokrat ^®, Izmir, Turkey) modular humeral prostheses were used between 2016 and 2020. Statistical analyses revealed no significant difference between the two brands in terms of prosthesis survival (log-rank test, P = 0.733) and functional outcomes (ANOVA, P = 0.675).

Functional evaluation was conducted using the Musculoskeletal Tumor Society (MSTS) scoring system, which is commonly employed for assessing upper extremity function [19]. Assessment was performed based on scores obtained during the twelfth month outpatient clinic follow-up appointments.

The mean MSTS score was 22.1 ± 2.564. When evaluating MSTS scores based on resection location, the highest average score of 22.4 ± 2.588 was observed in patients who underwent proximal humeral resection and prosthesis application. This was followed by 21.3 ± 2.236 for distal humerus and 20.8 ± 2.295 for total humerus resection and prosthesis application (Fig. 2). An analysis of variance (ANOVA) test was conducted to assess significant differences in MSTS scores among the different endoprosthesis groups applied to various regions. However, no significant difference was observed between the groups (P = 0.139).

Fig. 2.

MSTS scores distribution according to resection types

An assessment of MSTS scores according to diagnostic groups revealed that patients in the primary malignant tumor group had significantly higher scores (23.3 ± 2.309) than those in the metastatic and multiple myeloma groups. (21.1 ± 2.358) (p: 0.001) (Fig. 3).

Fig. 3.

MSTS score distribution according to groups

The use of synthetic mesh and the extent of resection, particularly in proximal humerus resection prostheses, have been extensively investigated in the literature for their impact on functional outcomes [20, 21].Functional outcome analysis related to synthetic mesh use was limited to patients who underwent proximal humerus resection (n = 73). In this subgroup, synthetic mesh was used in 40 patients and was not used in 33 patients. In patients who underwent proximal humerus resection, the use of synthetic mesh was associated with significantly higher functional outcomes. The mean MSTS score was 23.2 ± 2.37 in patients treated with synthetic mesh compared with 21.5 ± 2.58 in those without mesh application. The mean difference in MSTS score was 1.7 points (95% confidence interval [CI]: 0.6–2.8), indicating a clinically meaningful improvement in functional outcome (p = 0.005) (Table 4). When evaluating the relationship between the extent of resection and functional outcome, it was observed that as the extent of resection increased, the MSTS score decreased, although this difference was not statistically significant.

Table 4.

MSTS comparisons and effect sizes

Comparison	Group 1 mean ± SD	Group 2 mean ± SD	Mean difference	95% CI	p-value
Primary malignant vs. Metastatic/MM	23.3 ± 2.309	21.1 ± 2.358	2.2	1.2–3.2	0.001
Mesh vs. No mesh	23.2 ± 2.370	21.5 ± 2.575	1.7	0.6–2.8	0.005

Out of the 90 patients included in the study, eight experienced prosthesis failure. Two patients underwent prosthesis revision due to aseptic loosening, four patients underwent two-stage revision due to infection, and two patients underwent amputation due to tumor progression. Among the patients with prosthesis failure, six had proximal humerus resection prostheses, one had a distal humerus resection prosthesis, and one had a total humeral resection prosthesis. Kaplan-Meier analysis was used to evaluate endoprosthesis survival. The cumulative survival curve is shown in Fig. 4.

Fig. 4.

Kaplan-Meier analysis of implant survival with number at risk

The following factors were identified as potential determinants influencing endoprosthesis survival: synthetic mesh use, extent of resection, diagnostic groups, radiotherapy, bone cement use, localization, and age. These variables were statistically evaluated to assess their impact on endoprosthesis survival rates.

Current study, a synthetic mesh was used for implantation in 42 patients, and implant failure occurred in four of these patients. Kaplan-Meier analysis revealed no significant difference between the use of synthetic mesh and implant survival (P = 0.633).

As the extent of resection is a continuous variable, survival analysis was performed using simple Cox regression. The analysis revealed that as the extent of resection increased, the rate of implant failure also increased, but no statistically significant difference was found (p: 0.317).

When diagnostic groups were examined, it was found that implant failure occurred in four patients with primary malignant tumors, three with metastatic tumors and multiple myeloma, and one with primary benign tumors. However, Kaplan-Meier analysis revealed no significant difference between the diagnostic groups and implant failure (P = 0.439).

Implant failure occurred in three of the 46 patients who did not receive radiotherapy and in five of the 44 patients who received radiotherapy. In the unadjusted Kaplan–Meier analysis, implant failure rates were significantly higher among patients who underwent radiotherapy (P = 0.015). Univariable Cox regression analysis was performed to identify factors associated with implant survival. The following variables were evaluated: synthetic mesh use, extent of resection, diagnostic groups, radiotherapy, bone cement use, resection localization, and age. Among the evaluated variables, radiotherapy was significantly associated with implant failure, with a hazard ratio (HR) of 5.4 (95% confidence interval [CI], 1.12–24.39; P = 0.015). None of the other variables, including synthetic mesh use, extent of resection, diagnostic group, bone cement use, resection localization, or age, demonstrated a statistically significant association with implant survival in univariable analysis. The results of univariable Cox regression analyses for implant survival are presented in Table 5.

Table 5.

Univariable Cox regression analysis for implant survival

Variables		HR	95% CI	P value
Synthetic mesh use (yes vs. no)		1.19	0.55–2.56	0.633
Extent of resection (continuous)		1.21	0.84–1.75	0.317
Diagnostic group (metastatic/MM vs. primary malignant)		1.62	0.41–6.38	0.439
Radiotherapy (yes vs. no)		5.40	1.12–24.39	0.015
Bone cement use (yes vs. no)		1.08	0.24–4.86	0.925
Resection localization	─ Distal humerus vs. proximal humerus	0.88	0.09–8.74	0.912
	─ Total humerus vs. proximal humerus	1.14	0.12–10.61	0.906
Age (per year increase)		1.04	0.99–1.08	0.088

Implant failure occurred in five of the 64 patients who did not receive bone cement during stem fixation, compared with three of the 26 patients who did. Although a higher proportion of failures was observed in patients who received bone cement, the difference was not statistically significant (P = 0.925).

Among the 73 patients who underwent proximal humerus resection, implant failure occurred in six patients, while among the nine patients who underwent distal humerus resection and the eight patients who underwent total humerus resection, implant failure occurred in one patient each. When Kaplan-Meier analysis revealed no significant difference between the type of resection and implant failure (P = 0.561).

As age is a continuous variable, simple Cox regression analysis was applied for the survival analysis. The results indicated that as patient age increased at the time of implantation, the rate of implant failure also increased; however, this difference was not statistically significant (P = 0.088).

In addition, patient survival was evaluated in this study. Out of the 90 patients included, 37 died (osteosarcoma: 2, Ewing sarcoma: 4, chondrosarcoma: 2, metastatic tumor: 23, multiple myeloma: 3, soft tissue sarcoma: 3). Kaplan-Meier analysis was used to evaluate the overall survival (OS) of the patients. The cumulative survival curve is shown in Fig. 5. Univariable Cox proportional hazards regression analysis demonstrated that patients in the metastatic and multiple myeloma diagnostic group had a significantly higher risk of mortality compared with patients with primary malignant tumors (hazard ratio [HR] = 5.5, 95% confidence interval [CI]: 1.23–24.7; p = 0.025). Among patients with primary malignant tumors, advanced disease stage was associated with an increased risk of death (HR = 1.7 per stage increment, 95% CI: 1.15–2.76; p = 0.010). Chemotherapy administration, sex, and age were not significantly associated with overall survival (Table 6).

Fig. 5.

Kaplan-Meier analysis of overall survival with number at risk

Table 6.

Univariable Cox regression analysis for overall survival

Variables	HR	95% CI	P value
Metastatic/MM vs. primary malignant	5.5	1.23–24.7	0.025
Disease stage (per stage increase)	1.7	1.15–2.76	0.010
Chemotherapy (yes vs. no)	1.31	0.60–2.88	0.499
Sex (male vs. female)	1.7	0.91–3.44	0.090
Age (per year increase)	1.10	0.87–2.45	0.081

Complications were evaluated and classified according to the failure classification described by Henderson et al. in 2011 [13]. Among the 90 patients evaluated in present study, complications leading to implant failure occurred in eight patients. Two patients experienced mechanical failure, whereas six experienced non-mechanical failure. Both instances of mechanical failure were due to aseptic loosening, one in a patient with a proximal humerus resection prosthesis and the other in a patient with a distal humerus resection prosthesis. Of the six cases of non-mechanical failure, four were due to infection and two were related to tumor progression. Among the complications related to infection, three occurred with proximal humerus resection prostheses, and one occurred with a total humerus resection prosthesis. Both cases of complications related to tumor progression occurred in patients with proximal humeral resection prostheses.

Additionally, seven patients developed superficial wound infections. These infections were treated with superficial debridement and antibiotic therapy and did not result in prosthesis failure; therefore, superficial infections were not included in the statistical analysis.

Among the 42 patients who received a synthetic mesh in the overall cohort of 90 patients, implant failure due to infection occurred in two patients, while among the 48 patients who did not receive a synthetic mesh, implant failure due to infection occurred in two patients. The difference in infection development rates between these two groups was evaluated using the chi-square test, and no significant difference was found (P = 0.891).

Discussion

In this study, implant and patient survival were evaluated and the factors influencing these survivals in patients undergoing extensive surgical resection and endoprosthetic reconstruction for upper extremity musculoskeletal tumors. Additionally, the functional outcomes of patients were assessed using the MSTS score.

In current series, implant failure occurred in eight of the 90 patients who underwent endoprosthetic reconstruction, resulting in an overall implant survival rate of 91.1%. The total implant survival rates were 91.7%, 88.8%, and 87.5% for proximal humerus resection prostheses, distal humerus resection prostheses, and 87.5% for total humerus resection prostheses, respectively. Prosthetic failure due to aseptic loosening occurred in two patients (2.2%), infection in four patients (4.4%), and tumor progression in two patients (2.2%). Amputation was performed in two patients who developed tumor progression, resulting in an overall limb survival rate of 97.7%. A review of the literature revealed varying survival rates reported for humeral endoprosthetic reconstructions. Although some studies have reported survival rates higher than the general average, the survival rates in present study were similar to those reported in the literature. In a study by Cannon et al. examining long-term survival of proximal humerus resection prostheses with an average follow-up of 30 months, implant failure due to infection occurred in two out of 83 patients, resulting in a total survival rate of 97.6% [22]. Similarly, in a study by Yang et al. focusing on proximal humerus, the total 5-year survival rate was determined to be 88.2% [23]. Additionally, a study by Wafa et al. on the long-term survival of total humerus resection prostheses found 5-year and 10-year survival rates of 90% [24].

In present study, the most common complication was infection, with a rate of 4.4%. When comparing these results with the data obtained from the literature review, both similarities and differences were found. In a multicenter study conducted by Henderson et al., the most common mechanism of failure in 436 humeral endoprostheses was infection, with a rate of 6.1% [13]. In a systematic review evaluating 2721 endoprosthetic reconstructions across forty-nine studies, the mechanism most frequently leading to prosthesis failure was structural failure (Type 3), with a rate of 16%. This was followed by aseptic loosening (12%) and infection (9%) [25]. When comparing current results with those in the literature, lower rates of failure due to infection and other complications were observed.

The use of synthetic meshes to enhance the adherence of surrounding soft tissues to the prosthesis and contribute to postoperative function has raised concerns about their potential role in infection development. In a study by Schmolders et al., the effect of synthetic mesh use on infection development was investigated in 30 patients undergoing proximal humerus endoprosthetic reconstruction, and no significant difference was found [26]. In a study by Marulanda et al., where synthetic aortic grafts were used as synthetic meshes, no infections were observed during mid-term follow-up in 16 patients [27]. In present study, the occurrence of infection was evaluated in patients with and without the use of synthetic mesh. It was observed that two out of 42 patients who received synthetic mesh and two out of 48 patients who did not develop implant failure due to infection. A Chi-square test revealed that the use of synthetic mesh did not significantly affect the development of infection (P = 0.891).

Radiotherapy was found to be associated with higher implant failure rates in unadjusted analyses. However, this finding should not be interpreted as evidence of a direct causal relationship between radiotherapy and implant failure. Radiotherapy was more commonly administered to patients with metastatic disease and multiple myeloma, who represent a subgroup with more advanced disease burden and poorer overall prognosis. Therefore, the observed association likely reflects confounding by indication rather than a direct effect of radiotherapy on implant survival. Although Cox regression analysis demonstrated an increased hazard ratio for implant failure in patients receiving radiotherapy (HR = 5.4, 95% CI: 1.12–24.39), residual confounding cannot be excluded due to the retrospective design of the study and the limited number of failure events. Aseptic loosening has been reported in patients receiving radiotherapy [27]. Cemented endoprosthesis applications have been proposed to prevent aseptic loosening. In present study, cemented and cementless prostheses were compared in patients who received radiotherapy in terms of implant failure and found no significant difference between the two groups. (p:0.302).

Functional assessment was performed using the MSTS score. The mean score at the 12th-month clinic follow-up of the 90 patients was 22.1 ± 2.564. The highest score was observed in patients who underwent proximal humeral resection; however, no significant difference was found among all groups. A systematic review of 10 studies with 141 patients who underwent endoprosthetic reconstruction following proximal humerus resection reported MSTS scores ranging from 61% to 77% (18.3–23.1 points) [28]. Furthermore, in a study evaluating the functional outcomes of proximal humerus endoprostheses, the mean MSTS score was 24 ± 2 in patients with synthetic mesh usage compared to 20 ± 3 in those without synthetic mesh [20]. The contribution of synthetic mesh usage to functional outcomes was evaluated in patients undergoing proximal humerus resection prostheses. It was found that the mean MSTS score was 23.2 for patients with mesh application, while it was 21.5 for those without. When ANOVA was applied, a significant difference between the groups was observed (P = 0.005). In the functional assessment according to diagnostic groups, patients in the primary malignant tumor group demonstrated significantly higher MSTS scores compared with patients in the metastatic and multiple myeloma group. This finding is consistent with the more extensive soft-tissue involvement, systemic disease burden, and adjuvant treatment requirements observed in metastatic and multiple myeloma patients, which may negatively affect postoperative functional outcomes.

The heterogeneity of the study cohort represents an important consideration when interpreting the results. Patients with primary malignant tumors, metastatic disease, multiple myeloma, and benign/aggressive tumors differ substantially in terms of prognosis, systemic treatment requirements, and expected survival. Although this heterogeneity reflects real-world orthopedic oncology practice, it limits the generalizability of the findings across all tumor types. Therefore, the results of the present study should primarily be interpreted within major diagnostic categories, and caution should be exercised when extrapolating these outcomes to specific tumor entities or treatment pathways. The inclusion of patients with metastatic disease and multiple myeloma represents an important source of heterogeneity in the present cohort. These entities differ from primary malignant bone tumors in terms of biological behavior, prognosis, and indications for adjuvant therapy. Consequently, overall survival outcomes were strongly influenced by diagnostic category, with metastatic disease and multiple myeloma demonstrating poorer survival, consistent with existing literature. However, implant survival and functional outcomes primarily reflect the performance of modular endoprosthetic reconstruction and surgical technique, which are largely independent of tumor biology. Therefore, while oncologic survival results should be interpreted within the context of diagnostic heterogeneity, the implant-related findings remain clinically relevant across these groups.

The relationship between the extent of resection and functional outcomes was evaluated for tumors located in the proximal humerus. Although worse functional outcomes were observed with greater resection, no statistically significant difference was detected (P = 0.367). A review of the literature also reveals studies demonstrating that the extent of resection does not adversely affect functional outcomes [9, 29, 30].

Several studies have recommended reverse shoulder arthroplasty to provide better extremity function and reduce complications associated with soft tissue deficiency. In a study by Grosel et al., comparing conventional shoulder prostheses with reverse shoulder prostheses, it was shown that functional outcomes were significantly better with reverse shoulder prostheses, with no difference in complication rates [31, 32]. Limited reverse shoulder prostheses have also been used to prevent soft-tissue deficiency. However, aseptic loosening remains a concern with limited use of reverse shoulder prostheses [33].

The overall patient survival rates were 91.2% at Year 1 (95% CI: 85.4–96.1), 80.2% at Year 2 (95% CI: 72.1–88.3), 75.1% at Year 3 (95% CI: 66.2–84.0), 72.2% at Year 4 (95% CI: 62.5–81.9), and 66.9% at Year 5 (95% CI: 55.8–78.0). A study evaluating the survival of patients undergoing endoprosthetic reconstruction due to osteosarcoma in the humerus showed a 5-year survival rate of 70% [34]. Similarly, in a study evaluating the long-term survival of patients following proximal humerus resection arthroplasty, the survival rate was 77% at the second year and 45% at the 11th year [35]. When patient survival was evaluated according to diagnostic groups, metastatic patients exhibited the worst survival rates, consistent with the literature, and this result was statistically significant (p: 0.001) [36]. In patients with primary malignant tumors, it was also observed that survival decreased as the disease stage advanced (p: 0.01). In the Cox regression analysis, the risk of death was 5.5 times higher in the metastasis and multiple myeloma diagnostic group and 1.7 times higher in stage 3 patients compared to earlier stages. These high-risk ratios were consistent with the literature.

The retrospective design, non-randomized patient selection, and particularly the relatively lower number of benign cases compared to other diagnostic groups, constitute limitations of this study. The exclusion of patients with a follow-up duration of less than 12 months represents a potential source of selection bias, as patients with early mortality were not included. This approach was necessary to allow standardized functional assessment; however, overall survival results should be interpreted with this consideration in mind. Although longer follow-up data were available, MSTS scores beyond 12 months were not analyzed, as later functional assessments may be influenced by disease progression, adjuvant therapies, or deterioration in general health, particularly in patients with metastatic disease. Therefore, 12-month MSTS evaluation was selected to best reflect surgery- and implant-related functional outcomes. The small number of implant failure events represents an important limitation of this study. Although univariable Cox regression analyses were used to explore potential associations with implant survival, the limited number of events precluded robust multivariable modeling and resulted in wide confidence intervals. Accordingly, these hazard ratio estimates should be interpreted as exploratory rather than confirmatory. An important methodological consideration in the present study is the potential impact of competing risks on implant survival analyses. Given the relatively high mortality rate in this cohort, death represents a competing event for implant failure, as patients who die are no longer at risk of experiencing implant-related complications. Consequently, implant survival rates estimated using Kaplan–Meier analysis may be overestimated. However, a key strength is the consistency achieved by having all operations performed using similar techniques by a single surgeon and a team specialized in oncologic orthopedics.

Conclusion

Owing to advances in both neoadjuvant and adjuvant therapies, as well as implant technology, limb-sparing surgery has become the gold standard in the management of upper-extremity tumors in modern medicine.

Patients should be informed of the functional outcomes. Despite optimal repair of the surrounding muscle tissues during surgery, functional limitations, particularly in shoulder abduction, may occur. Based on the data from present study and the available literature, it can be said that the use of synthetic mesh appears to allow for better muscle restoration, leading to improved functional outcomes. In the literature review, reverse shoulder prostheses have been shown to provide better functional outcomes than conventional prostheses. However, potential complications, such as aseptic loosening, should be considered.

In the future, more comprehensive and randomized controlled studies with larger sample sizes comparing the long-term survival and functional outcomes of endoprosthetic reconstructions could provide further valuable contributions to the literature.

Abbreviations

CI

Confidence interval

CT

Computed tomography

HR

Hazard ratio

MRI

Magnetic resonance imaging

MSTS

Musculoskeletal Tumor Society

OS

Overall survival

PMMA

Polymethylmethacrylate

Authors’ contributions

Özer Öztürk and Mert Gündoğdu researched the literature and conceived the study. Bülent Erol was involved in protocol development, ethical approval, patient recruitment, and data analysis. Osman Mert Topkar measured the range of motion of the patients and evaluated the MSTS score. Özer Öztürk and Samed Çatal wrote the first draft of the manuscript. All authors reviewed, edited, and approved the final version of the manuscript.

Funding

The authors received no financial support for the research, authorship, or publication of this article.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the Marmara University Faculty of Medicine, Istanbul, Turkey. The study was approved by the Ethics Committee (approval number: 09.2020.760).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.