Treatment of malignant tumors of the proximal humerus with allograft-prosthesis composite reconstruction

Treatment of malignant tumors of the proximal humerus with allograft-prosthesis composite reconstruction Amy Williams Black, MD,a Robert M. Szabo, MD, MPH,b and Robert M. Titelman, MD,c Davis, CA, and Atlanta, GA

Malignant tumors of the proximal humerus are challenging to treat. Reconstruction with a metallic implant or allograft is the most common method, but each has known risks and frequent complications. Allograftprosthesis composite reconstruction has not been widely used and may avoid problems posed by metal prostheses or allografts used alone. Six patients with malignant tumors of the proximal humerus were treated with allograft-prosthesis composite reconstruction after excision of the intra-articular tumor. Outcomes were assessed by use of the Disabilities of the Arm, Shoulder and Hand questionnaire; the Short Form 36 (SF-36) Health Survey; and the American Shoulder and Elbow Surgeons Shoulder Assessment Form. Preoperative and postoperative scores at a mean of 55 months were compared. Disability increased after surgery based on the Disabilities of the Arm, Shoulder and Hand questionnaire and SF-36, although disability appeared to decrease with time. The mean mental component score on the SF-36 showed continued improvement with time after surgery. One asymptomatic nonunion was repaired, and painful loosening developed in one patient, requiring revision at 45 months. Allograft-prosthesis composite reconstruction is a safe method for treating some malignant tumors of the proximal humerus, providing stable reconstruction and preserving function of the shoulder joint. (J Shoulder Elbow Surg 2007;16:525-533.)

T he proximal humerus is the third most common site

of osteosarcoma, the second most common site of all bony sarcomas, and a frequent location of metastatic disease.1 Historically, primary malignant tumors of

From aSpecialized Orthopedic Surgeons and bDepartment of Orthopaedic Surgery, University of California, Davis, and cResurgens Orthopaedics, Atlanta. Reprint requests: Robert M. Szabo, MD, MPH, Department of Orthopaedic Surgery, University of California, Davis, 4860 Y St, Suite 3800, Sacramento, CA 95817. Copyright © 2007 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2007/$32.00 doi:10.1016/j.jse.2006.12.006

the proximal humerus were treated with forequarter amputation. Limb salvage treatment has included the Tikhoff-Linberg interscapulothoracic resection; shoulder arthrodesis with allograft, autograft, or vascularized autograft; and reconstruction with allograft or metallic prostheses. Of the many reconstructive procedures that have been described for treating intraarticular malignant tumors of the proximal humerus, the most common appears to be prosthetic or allograft reconstruction, depending on the surgeon’s preference. Both allograft reconstruction and prosthetic reconstruction have long-term problems. Osteoarticular allograft reconstruction has been frequently complicated by fracture, subchondral collapse, and infection. Reconstruction with metallic prostheses is often complicated by shoulder instability, limitation of range of motion, and failure of fixation. Combined allograft-prosthesis reconstruction has previously been described9,15 and potentially circumvents some of the problems seen in reconstruction with either a metallic prosthesis or allograft alone. The purpose of this study is to review and report outcomes, by use of validated instruments, in a series of 6 patients with malignant tumors of the proximal humerus treated with excision and combined allograftprosthesis reconstruction. This information will update expectations for limb salvage outcomes. MATERIALS AND METHODS The study group consists of 6 consecutive patients with a malignant tumor of the proximal humerus who presented to the orthopaedic upper extremity service of a single institution between 1999 and 2001. Patients were initially evaluated with plain radiographs, magnetic resonance imaging, and computed tomography of the involved extremity, as well as total-body bone scans and chest computed tomography scans. Tissue diagnosis was obtained through an incisional biopsy of the lesions. After preoperative evaluation, as well as preoperative chemotherapy for osteosarcoma in 1 case, all patients underwent a type 1A intraarticular proximal humeral resection11 and combined allograft-prosthesis hemiarthroplasty reconstruction by a single surgeon. Soft tissues were resected to ensure negative margins with intraoperative biopsies. An effort was made to preserve as much of the abductor mechanism as possible. The Biomet ATLAS Modular Shoulder System (Warsaw, IN)

525

526

Williams Black, Szabo, and Titelman

J Shoulder Elbow Surg September/October 2007

was used in all patients, and a fresh-frozen proximal humeral allograft with intact soft-tissue attachments had been selected before surgery based on a radiograph of the contralateral normal humerus.

Surgical technique Although no two resections were exactly the same due to the nature of the tumor and whether previous surgery had been done other than a biopsy, the following description is the fundamental approach that we used in our resection and reconstruction. The patient was placed in a modified beach-chair position with the involved upper extremity hanging over the edge of the table. The head was secured in position and turned toward the operative arm to prevent stretch on the brachial plexus. A longitudinal incision, 20 cm in length, extending from just distal to the coracoid down the anterolateral arm was used, elliptically around the entire 3-cm biopsy site island of tissue. A standard deltopectoral approach was used with the following modifications: the pectoralis insertion was identified and the superior portion of it taken off the humerus for improved exposure. The anterior proximal deltoid was also taken off the distal clavicle and the anterior portion of the acromion with a thick cuff of periosteum. As the deltoid was elevated, the biopsy island including skin, subcutaneous tissue, and deltoid was isolated; divided from the surrounding normal tissue; and maintained with the tumor mass on the proximal humerus. Deeper dissection then included partially releasing the conjoined tendon from the coracoid and dividing the coracoacromial ligament. The circumflex scapular vessels were identified and cauterized anteriorly before the subscapularis was sharply divided from its insertion on the proximal humerus, tagged, and allowed to retract medially. The latissimus dorsi and teres major were identified, tagged, and released from the proximal humerus. This exposed the humerus down to a level above the deltoid insertion. Further dissection was done distally to arrive at a point 3 cm distal to the most distal extent of the tumor based on magnetic resonance imaging studies. This junction was considered to be a safe zone for osteotomy. Before osteotomy, the remainder of the rotator cuff muscles were divided approximately 1 cm off from their insertion onto the proximal humerus. The long head of the biceps tendon was divided at that same level. The proximal humerus was then well mobilized. At that time, subperiosteal dissection around the humerus at the level of the proposed osteotomy was done, fully exposing it. A chevron osteotomy was made at that level with an oscillating saw and osteotomes (Figure 1). This allowed removal of the proximal humerus in its entirety with the tumor envelope uninterrupted. Curettage specimens from the distal margin were obtained for frozen section analysis. Soft tissue from around the rotator cuff and the inferior capsule was also obtained for frozen section analysis. The specimen was used as a template to make a similar chevron cut in the proximal humerus allograft at the same level and orientation. A starter hole was drilled in the top of the humeral head of the allograft, and serial broaching of the allograft was done to the appropriate size. Similar broaching to the same size was performed in the remainder of the humeral shaft. With the original broach in place in

Figure 1 Chondrosarcoma resection of proximal humerus in a 45-year-old man’s dominant extremity illustrating Chevron osteotomy.

the allograft, the humeral head was resected via the resection guide, and the template was placed over the humeral head. Osteotomes were used to make starter slots for the fins, and the fin template was then driven into the proximal humerus with a secure fit. With the template removed, a half batch of cement was mixed and injected into the proximal humeral allograft, and the final component, which had been tested for fit, was inserted under pressure into the allograft (Figure 2). This component routinely had a 175-mm stem. The remaining humeral shaft of the patient was then lavaged by use of a canal brush. A cement restrictor was placed to the appropriate depth. Another batch of cement was made and injected into the intramedullary canal. The allograft and prosthesis were then seated and impacted into the humeral shaft. Interposing cement was removed at the osteotomy site, and final seating was achieved (Figure 3). Cement was allowed to cure. There was some bone graft obtained from the reaming of the distal aspect of the humerus, and this was placed along the interface between the allograft and the native bone. Trial humeral heads were tried, with the best fit being selected, and were impacted securely onto the prosthesis. The native capsule was then repaired to the allograft capsule by use of horizontal mattress, No. 2 braided, Teflon-coated Dacron suture. After circumferential repair of

Williams Black, Szabo, and Titelman

J Shoulder Elbow Surg Volume 16, Number 5

527

Figure 2 Resection specimen of a chondrosarcoma of proximal humerus in a 43-year-old woman alongside allograft-prosthesis construct used to reconstruct defect. The extent of soft tissue available on the allograft for future attachment of the host tissues should be noted.

the joint capsule, the long head of the biceps tendon was brought up into the allograft and underwent tenodesis into the proximal allograft. The patient’s remaining rotator cuff tendons were repaired to the allograft tissue via No. 2 braided, Teflon-coated Dacron suture. The humeral head was then reduced, and the sutures were tied in a horizontal mattress fashion into the allograft rotator cuff. This established a secure repair of the rotator cuff and seating of the humeral head into the glenoid. The teres major and latissimus dorsi were likewise repaired to the allograft. The deltoid was repaired back to the clavicle and acromion, both to tissue and through bony holes. Fluoroscopy was used to place cross-locking screws through the distal holes in the prosthetic stem, securing it to the humeral diaphysis. The wound was then closed. Postoperatively, the arms were kept in a shoulder immobilizer and sling for 2 weeks. Passive range-of-motion exercises were then started and gradually advanced as pain and healing allowed. Patients were seen at regular intervals after surgery and were enrolled in the study as approved by our institutional review board during their postoperative course. Assessments were performed at the time of enrollment into the study and included the American Shoulder and

Figure 3 Composite allograft-prosthesis after fixation to distal humerus.

Elbow Surgeons (ASES) Shoulder Assessment Form; the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire; and the Short Form 36 (SF-36) Health Survey. The Functional Rating System of the Musculoskeletal Tumor Society was also used based on the patients’ questionnaire answers, clinical interviews, and physical examination findings.

RESULTS Demographic data are displayed in Table I. The mean age at the time of surgery was 40.7 years (range, 15-73 years). The mean initial follow-up time was 25.2 months (range, 13-39 months), and the mean second follow-up time was 55 months. There were 4 female patients and 2 male patients. Three patients had involvement of their dominant arm. Four had a diagnosis of chondrosarcoma, one had osteosarcoma (Figure 4), and one had metastatic thyroid cancer. One of the patients with chondrosarcoma had undergone previous resection and recon-

528

Williams Black, Szabo, and Titelman

J Shoulder Elbow Surg September/October 2007

Table I Demographic data Patient no.

Age (y)

Sex

Diagnosis

1 2

15 73

F F

3 4

43 49

F F

5 6

45 19

M M

Osteoblastic osteosarcoma Metastatic follicular cell carcinoma of thyroid Chondrosarcoma Chondrosarcoma, failed osteoarticular allograft Chondrosarcoma Chondrosarcoma

struction with an osteoarticular allograft by a different surgeon. In this patient, the allograft had fractured and then failed to heal with nonoperative treatment (Figure 5). The failed allograft and fixation device were removed, and a new allograft and prosthesis were used to reconstruct the shoulder in the same manner as in the other 5 patients. The data gathered from the ASES Shoulder Assessment Form14 allow the computation of a shoulder score index, which is shown in Table II. The shoulder score index is computed from a visual analog scale score for pain and from the score on the 10-item survey of activities of daily living. The mean shoulder score index was 61 (out of 100 possible points) at the early follow-up and 59 at the most recent follow-up. Strength and range-of-motion data were also collected. Two of six patients had at least one finding associated with instability. Three patients reported feeling that their operative shoulder was unstable using a visual analog scale to indicate the degree of instability. There have been no dislocations. The DASH questionnaire allows computation of a single score representing the degree of upper extremity disability experienced by the patient. The higher the DASH score, the greater the magnitude of disability. The mean preoperative score was 63.5, and the mean early postoperative score was 76.4 (Table III). The mean postoperative score from the recent follow-up was 68.5, suggesting a mean decrease in disability when compared with the first follow-up (Figure 6). The normative population mean is 50, with an SD of 10. We elected to forego statistical methods, such as the Student t test, as the number of patients was low and the power would be insufficient. The SF-36 Health Survey is a measure of overall quality of life. There are multiple subscales within the SF-36, and we used the two major scales: physical component summary (PCS), which is a measure of a patient’s physical functioning, and mental component summary (MCS), which is a measure of a patient’s psychological well-being. As with the DASH questionnaire, the normative population mean score is 50 and the SD is 10. However, the higher the score on this

Initial follow-up (mo)

Latest follow-up (mo)

Dominant arm

39 13

76 53

No Yes

29 19

58 59

No Yes

25 26

58 26

Yes No

survey, the better the quality of life. The mean preoperative score on the PCS was 46.8, the mean early postoperative score was 40.6, and the mean later postoperative score was 41.5. The mean preoperative score on the MCS was 48.3, with the mean early postoperative score improving to 53.1 and the mean later postoperative score continuing to improve to 57.5 (Table IV and Figure 7). Patient scores derived from the Functional Rating System of the Musculoskeletal Tumor Society4 are shown in Table V. The mean total score was 74% of the maximum (range, 57%-90%) at early follow-up and decreased to 69% at later follow-up. A score of greater than 60% has been considered to indicate a good or excellent outcome in other studies.7,12,13 Complications

One patient had a radiographic nonunion. Although this was painless, the nonunion persisted, and therefore, the patient underwent revision surgery with bone grafting and cable plating of the nonunion. She has had an uneventful recovery, and union was achieved after this second procedure. One patient returned to construction work after being advised by his surgeon not to do so. After 6 months back on the job, shoulder pain developed. He had moved to another region and ultimately underwent revision by another surgeon to a prosthesis-only reconstruction 45 months after his initial surgery. He has now returned to work as a building inspector and reports decreased range of motion and significant instability. There were no disease recurrences, fractures, dislocations, or infections in our study group. DISCUSSION Malignant tumors of the proximal humerus are challenging problems, and numerous reconstructive methods have been described. No single reconstructive procedure has been shown to be clearly superior to other available methods. This case series reports good early and intermediate results in 6 consecutive

J Shoulder Elbow Surg Volume 16, Number 5

Williams Black, Szabo, and Titelman

Figure 4 Images in a 15-year-old female patient with an osteoblastic sarcoma of nondominant left proximal humerus. Clinical photographs were taken at 39 months postoperatively. A, Preoperative radiograph after chemotherapy. B, Forward elevation. C, Maximum active external rotation. D, Maximum active internal rotation. E, Postoperative anterior-posterior radiograph at 39 months demonstrating healing of allograft at osteotomy site. It should be noted how distal this resection had to be to obtain adequate tumor resection margins.

529

530

Williams Black, Szabo, and Titelman

J Shoulder Elbow Surg September/October 2007

Figure 6 DASH scores at preoperative and postoperative intervals. Table IV SF-36 results: Mean normative scores

Figure 5 Images in a 49-year-old woman who had a chondrosarcoma of the dominant humerus initially treated with an osteoarticular allograft. A, Failure of the allograft is demonstrated by the proximal humeral fracture. B, Image obtained after conversion to a new allograft-prosthesis reconstruction. Table II Shoulder scores according to ASES Shoulder Assessment Form Shoulder score

Patient No. 1 2 3 4 5 6 Mean

Initial follow-up

Latest follow-up

68 62 70 67 58 43 61

87 68 43 67 28 — 59

Table III DASH results: Mean normative scores

Mean DASH score

Preoperative

First follow-up

Second follow-up

63.5

76.4

68.5

patients treated with combined allograft and prosthesis reconstruction. Reconstructive options after intra-articular resection of a proximal humeral tumor include arthrodesis, which is uncommonly used because of the substantial bone deficit and difficulty of achieving fusion. Other

Mean SF-36 score PCS MCS

Preoperative

First follow-up

Second follow-up

46.8 48.3

40.6 53.1

41.5 57.5

options include prosthetic implant, osteoarticular allograft, allograft-prosthesis composite reconstruction, and fibular autogenous grafts (vascularized or nonvascularized).5 The advantages of an osteoarticular allograft are the restoration of the glenohumeral joint and the preservation of the abductor mechanism. Reattachment of the rotator cuff tendons and deltoid muscle to their insertions potentially results in better stability and active range of motion of the shoulder, which should lead to better overall function and higher patient satisfaction. The complications associated with osteoarticular allografts, however, are substantial and include fracture, subchondral collapse, and infection. In 1990, Gebhardt et al6 reported on their experience with osteoarticular allografts. Functional results in 20 patients were reported by use of a system developed by one of the authors of that study. Approximately 70% of their patients had a satisfactory result (minimum pain and a return to normal activity), and the authors concluded that despite a high complication rate, the overall success rate compared favorably with those of other reconstructive procedures. Complications included 3 infections, 7 fractures, 1 nonunion, and 1 case of significant instability. They noted a lack of studies directly comparing reconstructive options and no consistent use of a grading system for function. After their study, the Functional Rating System of the Musculoskeletal Tumor Society became commonly used as an outcome measure in tumor patients.

Williams Black, Szabo, and Titelman

J Shoulder Elbow Surg Volume 16, Number 5

Figure 7 SF-36 PCS (A) and MCS (B) scores at preoperative (preop) and postoperative intervals.

Table V Scores on Functional Rating System of Musculoskeletal Tumor Society

Mean total score

First follow-up

Second follow-up

74%

69%

In 1999, Getty and Peabody7 reported on their experience in 16 patients undergoing osteoarticular allograft reconstruction. An unacceptably high complication rate and function levels that decreased with time led them to abandon osteoarticular allografts as a primary method of reconstruction. In the 16 patients, there were 4 fractures, 1 deep infection, 3 dislocations, and 3 cases of anterior subluxation. Whereas the mean score on the Musculoskeletal Tumor Society scale was 81% at 14 months, it had decreased to 70% at a mean follow-up of 34 months. Metal (or ceramic) prosthetic reconstruction has the advantage of modular constructs to facilitate the restoration of limb length; however, problems have been well documented. The most common complication is shoulder instability, with reported subluxation and dislocation rates of 33% to 80%. Failure of fixation is common, as is the lack of active range of motion without reconstruction of the abductor mechanism.

531

There is also concern about the longevity of the implants, as oncology patients are often considerably younger than their degenerative joint counterparts. Ross et al16 reported on 25 proximal humeral prosthetic reconstructions done between 1950 and 1982, 20 of which were done for tumors. Sixteen had subluxation and three had dislocations; however, these were not counted among the complications. Range of motion was noted to be less than 30° in all directions after surgery, and only 1 patient required reoperation. O’Connor et al13 reported on 11 patients who underwent prosthetic reconstruction. Complications included subluxation (6 patients), loosening of the prosthesis (2), and deep infection (1). Four patients required revision to arthrodesis. Infections developed in two, requiring additional surgery. Several reports have been published in the last few years about the reverse shoulder prosthesis in the treatment of proximal humeral tumors. De Wilde et al3 reported on two series of 6 and 7 patients, one group of which had a reverse shoulder prosthesis reconstruction and the other had autoclaved autograft augmentation of the same implant. The authors concluded that both groups had significant functional improvement over other reconstructive methods’ results despite their early follow-up. Similarly, De Wilde et al2 reported on 4 cases of full functional recovery after reverse prosthesis reconstruction in patients requiring sacrifice of the rotator cuff tendons. Indeed, the use of the reverse shoulder prosthesis is an intriguing idea for patients with rotator cuff deficiency and intact deltoid function after proximal humeral tumor resection, but intermediate- and long-term follow-up data are lacking. Allograft-prosthesis composite arthroplasty was first used in hip and knee surgery to address the problems of resorption, fracture and cartilage degeneration of the allograft, and instability and lack of soft-tissue attachments of the prosthesis. Although there are several reported series of allograft-prosthesis composite reconstructions in the hip and knee, long-term follow-up remains scant.8,10 The first report of bone graft– prosthesis composite reconstruction in the proximal humerus was in 1978. Imbriglia et al11 reported on the use of a Neer prosthesis combined with fibular autograft in a 13-year-old patient. He was able to abduct and externally rotate 45° after 30 months. Rock15 reported the first series of allograft-prosthesis reconstructions in the proximal humerus. Four of the five patients had excellent results and no fractures, infections, or loosening. We are unaware of any long-term follow-up in this series. Harrington et al9 reported on a single case of proximal humeral allograft-prosthesis composite reconstruction in a study of 42 patients undergoing limb salvage with prosthetic joint reconstruction. After 41 months, the patient whose allograft was actually autoclaved autograft

532

Williams Black, Szabo, and Titelman

after removal of the tumor was pain-free; could abduct 130°, flex 160°, and externally rotate to 80°; and had returned to work as a plumber. Jensen and Johnston12 reported results in 19 patients undergoing proximal humerus reconstruction between 1978 and 1991. Eleven had intercalary autoclaved autograft with a metal prosthesis, and four had intercalary allograft (and the remaining four had prosthetic reconstruction). Complications included 2 with local recurrence, 1 dislocation, and 2 with subluxation, as well as 2 superficial wound infections. Functional outcome evaluated by use of the Musculoskeletal Tumor Society scale was excellent in 17 patients and good in 2. The authors concluded that both autoclaved autograft and allograft combined with a cemented Neer prosthesis were excellent options for proximal humeral reconstruction. Although the outcome instruments used in this study have not been previously applied to limb salvage patients, we believe that it is appropriate to use current validated outcome instruments to facilitate better communication among investigators and comparison between patients. Ablative surgery clearly is no longer the gold standard treatment for many tumors. Functional outcomes in patients who have undergone limb salvage procedures may be good enough to be compared with patients who have undergone reconstructive procedures for nononcologic problems. Other investigators have remarked that the most frequently used outcome measure in orthopaedic oncology, the Functional Rating System of the Musculoskeletal Tumor Society, seems to produce good or excellent functional outcome scores in patients who demonstrate substantial disability. Getty and Peabody7 remarked that “a more discerning and critical measure of function is necessary.” We have used the ASES, DASH, and SF-36 to provide baseline outcome data for ourselves as well as for other investigators. The difference between preoperative and postoperative DASH scores was anticipated; we expected the degree of upper extremity disability to increase after such an extensive reconstructive procedure. We find the decrease in DASH scores from early postoperative scores to longer-term follow-up interesting in that it suggests that disability level decreases with time after surgery (Figure 6). Whether this phenomenon reflects patients adapting to their functional limitations or truly improving function with time and whether this improvement will continue with longer-term follow-up are unanswered questions. Also interesting are the differences between the mean preoperative and postoperative scores on the MCS of the SF-36 (Figure 7). Perhaps, from a psychological standpoint, patients improve after surgery. This idea contrasts sharply with the accepted fact that patients have a decrease in functional ability after surgery. This mental improvement could account for the consistently high rates of

J Shoulder Elbow Surg September/October 2007

good or excellent outcome scores in patients who clearly have substantial functional deficits after surgery. This study has several weaknesses. It is a retrospective study: patients were asked to complete preoperative evaluations after their operations, thinking back to how they would have responded before the surgery. As with many orthopaedic oncology studies, the number of patients is small. We believed that it was important to report the early postoperative outcomes, as well as intermediate- or longer-term results, to learn about changes that patients experience over time as well as to track complications. Equally important will be reporting long-term results in the future. Combining allograft and prosthetic reconstructions to treat tumors of the proximal humerus allows for a stable reconstruction of the shoulder joint, preservation of the abductor mechanism, and active range of motion of the shoulder joint. At a mean of 55 months after surgery, this procedure appears safe from an oncologic standpoint, the incidence of complications is low (provided that the patient is compliant with postoperative activity restrictions), and functional outcome is comparable to existing studies for alternative procedures. REFERENCES

1. Dahlin D. Bone tumors: general aspects and data on 6221 cases. Springfield (IL): Charles C. Thomas; 1978. 2. De Wilde L, Plasschaert F, Audenaert E, Verdonk R. Functional recovery after a reverse prosthesis for reconstruction of the proximal humerus in tumor surgery. Clin Orthop Relat Res 2005: 156-62. 3. De Wilde L, Sys G, Julien Y, Van Ovost E, Poffyn B, Trouilloud P. The reversed Delta shoulder prosthesis in reconstruction of the proximal humerus after tumour resection. Acta Orthop Belg 2003;69:495-500. 4. Enneking W, Dunham W, Gebhardt M, Malawer M, Pritchard D. A system for the functional evaluation of reconstructive procedures after surgical treatment of tumors of the musculoskeletal system. Clin Orthop Relat Res 1993:241-6. 5. Enneking W, Eady J, Burchardt H. Autogenous cortical bone grafts in the reconstruction of segmental skeletal defects. J Bone Joint Surg Am 1980;62:1039-58. 6. Gebhardt M, Roth Y, Mankin H. Osteoarticular allografts for reconstruction in the proximal part of the humerus after excision of a musculoskeletal tumor. J Bone Joint Surg Am 1990;72:334-45. 7. Getty P, Peabody T. Complications and functional outcomes of reconstruction with an osteoarticular allograft after intra-articular resection of the proximal aspect of the humerus. J Bone Joint Surg Am 1999;81:1138-46. 8. Gitelis S, Piasecki P. Allograft prosthetic composite arthroplasty for osteosarcoma and other aggressive bone tumors. Clin Orthop Relat Res 1990:197-201. 9. Harrington K, Johnston J, Kaufer H, Luck J, Moore T. Limb salvage and prosthetic joint reconstruction for low-grade and selected high-grade sarcomas of bone after wide resection and replacement by autoclaved autogeneic grafts. Clin Orthop Relat Res 1986:180-214. 10. Hejna M, Gitelis S. Allograft prosthetic composite replacement for bone tumors. Semin Surg Oncol 1997;13:18-24.

J Shoulder Elbow Surg Volume 16, Number 5

11. Imbriglia J, Neer C, Dick H. Resection of the proximal one-half of the humerus in a child for chondrosarcoma. J Bone Joint Surg Am 1978;60:262-4. 12. Jensen K, Johnston J. Proximal humeral reconstruction after excision of a primary sarcoma. Clin Orthop Relat Res 1995: 164-75. 13. O’Connor M, Sim F, Chao E. Limb salvage for neoplasms of the shoulder girdle. J Bone Joint Surg Am 1996;78: 1872-88.

Williams Black, Szabo, and Titelman

533

14. Richards RR, An KN, Bigliani LU, Friedman RJ, Gartsman GM, Gristina AG, et al. A standardized method for the assessment of shoulder function. J Shoulder Elbow Surg 1994;3:347-52. 15. Rock M. Intercalary allograft and custom Neer prosthesis after en bloc resection of the proximal humerus. In: Enneking WF, editor. Limb salvage in musculoskeletal oncology. New York: Churchill Livingstone; 1987. p. 586-97. 16. Ross A, Wilson J, Scales J. Endoprosthetic replacement of the proximal humerus. J Bone Joint Surg Br 1987;69:656-61.