Risk of local recurrence after deltoid-sparing resection for osteosarcoma of the proximal humerus

Abstract

BACKGROUND:

The anatomy of the shoulder poses special challenges with regard to limb-sparing surgery. Resection of the deltoid muscle is considered by some surgeons to be necessary to achieve adequate margins for osteosarcoma of the proximal humerus. However, this can compromise the functional results after reconstruction of the shoulder. The goal of the current study was to determine whether deltoid-sparing resection can be safely performed for osteosarcoma of the proximal humerus.

METHODS:

Between 1978 and 2005, 23 consecutive patients with high-grade nonmetastatic osteosarcoma of the proximal humerus underwent limb-sparing surgery with preservation of the deltoid muscle. All patients received neoadjuvant chemotherapy followed by surgery and postoperative chemotherapy. The mean follow-up was 90 months (range, 7 months-279 months).

RESULTS:

The overall survival at 5 years was 77%. At the time of last follow-up, 14 (61%) of 23 of patients were alive without evidence of disease. Three (13%) patients developed local recurrence. Two of these patients had poor responses to chemotherapy, with tumor necrosis of 50% and 70%. The third patient had a pathologic fracture of the humerus. Positive surgical margins were associated with local recurrence, and 2 of 4 patients with a positive surgical margin developed local recurrence (P = .01).

CONCLUSIONS:

Preservation of the deltoid muscle can be performed for carefully selected patients with osteosarcoma of the proximal humerus. Routine use of the procedure is not justified, because it may be associated with an elevated risk of recurrence. The risk of local recurrence appears to be related to positive surgical margins and possibly the percentage of tumor necrosis. Cancer 2009. © 2009 American Cancer Society.

The proximal humerus is the third most common site of involvement for conventional high-grade intramedullary osteosarcoma. Although it is usually treated with limb-sparing techniques, there is a surgical dilemma in terms of resecting an adequate amount of the tissue for oncologic purposes and preserving enough tissues for functional purposes.

Because the axillary nerve is located on the undersurface of the deltoid muscle, it is particularly vulnerable during resection of tumors.1 The close proximity to the humerus has led certain authors to recommend routine sacrifice of the deltoid muscle and axillary nerve for osteosarcomas of the proximal humerus.2 Such a resection achieves the goal of a wide excision, but it can potentially compromise shoulder function after reconstruction.

Although numerous articles have been written on proximal humeral resections, most have included a mixture of different tumor types, and to our knowledge none have addressed deltoid-sparing surgery specifically for osteosarcomas.3-7 Gebhardt et al made the observation that preservation of the deltoid muscle was important for stability and function in allograft reconstructions.3 The potentially significant role of the deltoid has been underscored in more recent years by the application of the reverse shoulder prosthesis to oncologic reconstructions.8 This implant relies on an intact deltoid muscle.

We hypothesized that sacrifice of the axillary nerve and deltoid muscle is not always necessary for osteosarcomas of the proximal humerus (Fig. 1). We analyzed the oncologic outcome for a consecutive series of 23 patients who underwent deltoid-sparing surgery for nonmetastatic, high-grade osteosarcomas of the proximal humerus. To our knowledge, the current study represents the largest series of conservative resections for osteosarcomas of the proximal humerus published to date. The goal of the study was to examine the factors that could affect local recurrence. The results are important to future work regarding shoulder reconstruction that may depend critically on preservation of the deltoid muscle.

Figure 1. A 21-year-old man presented with osteoblastic osteosarcoma of the proximal humerus. (A) X-ray of the shoulder demonstrated involvement of the anteromedial aspect of the humerus and ossification of the soft tissue mass. (B) A gadolinium contrast-enhanced T1-weighted axial magnetic resonance imaging (MRI) scan demonstrated the tumor mass arising from the medullary canal and extending anteriorly under the biceps muscle. The undersurface of the deltoid muscle was not involved. (C) A coronal T1-weighted MRI image demonstrated the longitudinal extent of the tumor. (D) Three years after surgical resection and reconstruction with an allograft-prosthetic composite, the implant was stable, and the osteotomy had healed. The patient was able to achieve 150° forward flexion at the shoulder.

MATERIALS AND METHODS

Study Design and Patients

Between 1978 and 2005, 26 patients with Musculoskeletal Tumor Society stage IIA or IIB conventional osteosarcoma of the proximal humerus were treated at our institution.9 The patients were identified from the Orthopedic Oncology Surgical Database. Twenty-three of the 26 patients underwent limb-sparing surgery with preservation of the axillary nerve and most or all of the deltoid muscle. The other 3 patients underwent extra-articular resection with sacrifice of the entire deltoid muscle. These 3 patients had massive tumors that invaded a large portion of the deltoid muscle. The group of 23 patients formed the cohort for the current study, which was approved by the institutional review board.

The medical records were reviewed to determine patient demographics, clinical information, and details of treatment. All the imaging studies, surgical reports, and pathology reports were reviewed. The occurrence of local recurrence and metastasis was noted for each patient. The disease status of each patient was determined at the time of last contact. Death of patients was confirmed by the Social Security Death Index.

There were 12 males and 11 females with mean age of 17 years (range, 7 years-31 years) at the time of diagnosis. The minimum follow-up was 24 months, unless patients died of disease before 24 months. The mean follow-up was 90 months (range, 7 months-279 months).

Pathology

All patients had conventional high-grade osteosarcoma. Four patients had stage IIA disease, and 19 patients had stage IIB disease. The histologic subtypes included 12 osteoblastic, 5 telangiectatic, 4 fibroblastic, and 2 chondroblastic osteosarcomas. The histologic response of the tumor to chemotherapy was determined after detailed grid mapping of the resection specimen.10 The percent tumor necrosis was obtained from the pathology report. Positive surgical margins were defined as the presence of tumor at the inked margin on histologic specimens.

Chemotherapy

During the study period, 17 patients received intra-arterial cisplatin (120-160 mg/m²) together with intravenous doxorubicin (90 mg/m²) for neoadjuvant preoperative chemotherapy. The mean number of courses of intra-arterial cisplatin was 3.7 (range, 2-7), and the number of courses are shown for each patient in Table 1. Six patients were treated preoperatively without intra-arterial cisplatin using alternative regimens, including the Children's Cancer Group protocol CCG-7921¹¹ (3 patients) and TIOS-IV¹² (1 patient).

Table 1. Patient Data

Patient No.	Sex	Stage of Disease	OS, mo	CDFS, mo	LRFS, mo	No. of Preoperative IA-CDP Cycles	Necrosis, %	LR	VS	Margins
OS indicates overall survival; CDFS, continuous disease-free survival; LRFS, local recurrence-free survival; IA-CDP, intra-arterial cisplatin; LR, local recurrence; VS, vital status; NA, not available; DOD, died of disease; NED, no evidence of disease; AWD, alive with disease.
1	Female	IIB	29	16	29	0	NA	No	DOD	Negative
2	Female	IIA	219	219	219	2	70	No	NED	Negative
3	Male	IIB	78	78	78	4	95	No	NED	Negative
4	Female	IIB	279	279	279	3	100	No	NED	Negative
5	Male	IIB	145	9	145	4	100	No	NED	Negative
6	Male	IIB	66	51	66	5	100	No	DOD	Negative
7	Male	IIB	176	176	176	4	100	No	NED	Negative
8	Female	IIB	225	225	225	6	75	No	NED	Positive
9	Male	IIA	27	17	27	4	95	No	DOD	Negative
10	Male	IIB	14	7	7	3	70	Yes	DOD	Positive
11	Female	IIB	216	216	216	4	26	No	NED	Positive
12	Female	IIA	180	180	180	7	95	No	NED	Negative
13	Female	IIB	122	122	122	4	100	No	NED	Negative
14	Male	IIA	132	47.	47	0	92	Yes	NED	Negative
15	Female	IIB	22	11	11	0	50	Yes	DOD	Positive
16	Male	IIB	22	9	22	0	99	No	DOD	Negative
17	Male	IIB	44	44	44	0	30	No	NED	Negative
18	Male	IIB	61	19	61	0	99	No	DOD	Negative
19	Male	IIB	36	36	36	2	85	No	NED	Negative
20	Female	IIB	28	15	28	4	80	No	AWD	Negative
21	Male	IIB	39	39	39	2	93	No	NED	Negative
22	Female	IIB	28	9	28	3	96	No	AWD	Negative
23	Female	IIB	30	30	30	2	69	No	NED	Negative

Chemotherapy was reinitiated between 2 and 4 weeks after the surgery. For the patients receiving intra-arterial cisplatin, postoperative chemotherapy was determined by the tumor response. Patients with a good histologic response (>90% tumor necrosis) received additional courses of intravenous doxorubicin (75 mg/m²) together with intravenous cisplatin and/or ifosfamide, whereas those with <90% tumor necrosis were treated with a combination of agents, including high-dose methotrexate (10-12 g/m²) and high-dose ifosfamide (14 g/m²) in multiple alternating courses.

Surgery

Twenty-two patients underwent intra-articular resections, whereas 1 underwent an extra-articular resection. In all 23 patients, the deltoid muscle was preserved together with its innervation by the axillary nerve. The decision to spare the deltoid was made at the discretion of the surgeon on the basis of preoperative radiologic studies. Partial excision of the deltoid (<50%) was performed to remove open biopsy tracts and/or focal areas of tumor involvement. The types of reconstruction varied. Sixteen patients underwent an endoprosthetic reconstruction, 2 patients underwent vascularized fibular bone graft reconstruction, 4 patients underwent an allograft-prosthesis composite reconstruction (Fig. 1), and 1 patient underwent a total humerus replacement.

Statistical Methods

Statistical calculations were performed with the SAS9.1 (SAS Institute Inc, Cary, NC) and SPSS 12.0 (SPSS, Chicago, Ill) statistical computer programs. The Kaplan-Meier method was used to estimate survival distributions, and the log-rank test was used to compare the survival curves for local recurrence-free survival (LRFS) based on different prognostic factors. Comparison of nonparametric distributions was performed with chi-square analysis and the Fisher exact test. Statistical significance was defined as P <.05.

RESULTS

The Kaplan-Meier overall survival rate was 77% at 5 years (Fig. 2). Seven (30%) patients died of disease, and 16 patients (70%) were alive at the time of last follow-up (Table 1). Thirteen (57%) patients had been continuously free of disease at a median follow-up of 162 months. At the time of last follow‒up, 2 patients were alive with disease. One patient was free of disease 85 months after undergoing wide local excision of a recurrence in the deltoid muscle. At last follow-up, 14 (61%) of 23 of the patients were alive and had no evidence of disease.

Figure 2. Overall survival (OS) as determined by Kaplan-Meier analysis is shown for the 23 patients in the study. The OS with 95% confidence intervals (95% CIs) were 0.82 (95% CI, 0.68-0.99), 0.77 (95% CI, 0.60-0.97), and 0.64 (95% CI, 0.45-0.90), respectively, at 3 years, 5 years, and 10 years.

Three (13%) patients developed local recurrence. These developed 7 months, 11 months, and 47 months, respectively, after resection of the primary tumor. In 2 patients, the local recurrence was in the deltoid muscle. In the other patient, the local recurrence was in the subscapularis muscle, away from the deltoid region.

According to Kaplan-Meier analysis, the LRFS at 5 years was 84% (Fig. 3). The factors that were not found to have a significant effect on local recurrence included age (P = .62), sex (P = .58), size of the tumor (P = .80), histologic subtype (P = .70), and subdeltoid location of extraosseous tumor (P = .15).

Figure 3. Local recurrence-free survival (LRFS) as determined by Kaplan-Meier analysis is shown. The LRFS with 95% confidence intervals (95% CIs) were 0.91 (95% CI, 0.81-1), 0.84 (95% CI, 0.69-1), and 0.84 (95% CI, 0.69-1), respectively, at 3 years, 5 years, and 10 years.

A positive surgical margin was associated with local recurrence. Of the 4 patients with positive surgical margins, 2 developed local recurrences, whereas 1 of 19 patients with negative surgical margins developed local recurrence (P = .014).

Using chi-square analysis, tumor necrosis with the conventional 90% cutoff for good chemotherapy response did not appear to correlate with local recurrence. Two of 10 patients with tumor necrosis of <90% developed local recurrence, whereas 1 of 13 patients with tumor necrosis of ≥90% developed local recurrence (P = .23). It was noted, however, that for the 3 patients who developed local recurrence, the tumor necrosis was 50%, 70%, and 92%, respectively (mean, 71%). In comparison, the mean tumor necrosis for patients without local recurrence was 85% (P = .37). The 2 patients with poor responses of 50% and 70%, respectively, developed local recurrence rapidly at 11 months and 7 months, respectively. Both patients developed concurrent distant metastasis and died of disease. The third patient, who had a response rate of 92%, developed local recurrence relatively late at 47 months. This patient initially presented with a pathologic fracture. The patient was alive and free of disease 85 months after undergoing wide local excision of the recurrent tumor.

One of 17 patients receiving intra-arterial cisplatin developed local recurrence, whereas 2 of 6 patients not receiving intra-arterial cisplatin developed local recurrence (P = .074). The mean percentage necrosis was 85% for patients receiving preoperative intra-arterial cisplatin and 74% for patients who were not treated with intra-arterial cisplatin (P = .046).

DISCUSSION

The results of the current study suggest that deltoid-preserving resections for osteosarcomas of the proximal humerus are feasible, but should be performed carefully and selectively. During the study period, the majority of patients (23 of 26) at the study institution were treated with deltoid-sparing procedures. This cohort of patients, although small in number, is valuable because it allows the examination of whether deltoid-sparing surgery can be broadly applied to osteosarcoma patients.

Wittig et al2 published a series of 23 patients with primary osteosarcoma of the proximal humerus. All patients except for 1 were treated with extra-articular resection and sacrifice of the deltoid muscle. There were no local recurrences, and the authors recommended that resections routinely include the deltoid muscle.

To our knowledge, the published experience with deltoid-sparing resections is sparse. Getty and Peabody reported no local recurrences in 4 patients with osteosarcoma who underwent extra-articular deltoid-sparing resection.4 Jensen and Johnston reported 1 local recurrence in 4 patients.5 O'Connor et al reported just 1 local recurrence in 16 cases of stage IIA and IIB osteosarcomas of the proximal humerus and scapula, with preservation of the deltoid muscle in selective cases.6

The rate of local recurrence (13%) for the 23 patients in the current study was relatively high in comparison to the general local recurrence rate of osteosarcoma in large published series, which ranges between 5% and 11%.13-16 It is possible that the elevated local recurrence rate may have contributed to the 5-year continuous disease-free survival rate of 57%, which is low in comparison to the large, recently published Children's Cancer Group trial.11 The rate of local recurrence noted in the current study may reflect the finding that the deltoid-sparing technique was used in a high percentage of patients during the study period. In retrospect, greater selectivity in the use of deltoid-sparing surgery may have resulted in an improvement in the local recurrence rate. Patients who have large, extraosseous tumors on the undersurface of the deltoid, particularly those that do not form a smooth border of ossification after preoperative chemotherapy, are currently not considered good candidates for deltoid-preserving surgery at our institution.

It has been observed by Picci et al that the response to chemotherapy is a significant determinant of local recurrence in osteosarcoma.14 The results of the current study, although limited in numbers, are consistent with this idea. Two of the 3 patients with local recurrence had markedly poor chemotherapy responses of 50% and 70%, respectively. The third patient had a chemotherapy response of 92%, but this patient had a pathologic fracture, which may have contributed to the local recurrence.17

In addition to chemotherapy response being a predictor for local recurrence, we found that positive surgical margin also correlated with local recurrence. The observation is consistent with previous studies. It must be emphasized that although intensive neoadjuvant treatment may diminish the risk of recurrence, adequate surgical margins are still vitally important. The appropriate selection of patients for deltoid-sparing techniques should take into account the extent of extraosseous tumor on the undersurface of the deltoid.

The majority of patients in the current study received intra-arterial cisplatin,18 which is not commonly used in most centers. Although there was better tumor necrosis in patients who received intra-arterial cisplatin, the rate of local recurrence was not found to be significantly less. It ought to be noted that the number of patients in this study is small, and patients were not randomly selected for intra-arterial versus intravenous cisplatin. These data are most likely not adequate to draw any definite conclusions regarding the effectiveness of intra-arterial therapy.

Reconstruction of the shoulder after resection of the proximal humerus is quite difficult, and in most cases, normal shoulder movement cannot be restored. In fact, active motion at the shoulder is generally quite poor after oncologic resection and endoprosthetic replacement.19 In the current series, the reconstructive techniques were varied, and there was quite a large range of functional deficits (data not shown). Although a few patients had excellent function (Fig. 1), many patients had very limited active shoulder motion. The mere fact that the deltoid was preserved did not guarantee good shoulder function. The analysis of shoulder movement in relation to surgical technique is complicated, and we have found that many factors affect the outcome. The scope of the current study is not adequate to address these factors, and further work will be needed in the future to investigate this area.

In summary, the results of the current study indicate that deltoid-sparing resection is feasible for osteosarcoma of the proximal humerus. However, it should not be used indiscriminately, because routine use of the procedure may be associated with an elevated risk of local recurrence. Both chemotherapy response and surgical margins may affect local control. The selection of patients is important. Tumors with extensive infiltration into the deltoid muscle and tumors that do not respond well to preoperative chemotherapy are not appropriate for the procedure. In the future, more accurate prediction of tumor necrosis may help determine which patients may be good candidates for deltoid-sparing techniques.

Conflict of Interest Disclosures

The authors made no disclosures.