JBJS | Pathologic Fracture in Osteosarcoma : Prognostic Importance and Treatment Implications

The Journal of Bone & Joint Surgery, Volume 84, Issue 1

Scientific Article | January 01, 2002

Pathologic Fracture in Osteosarcoma : Prognostic Importance and Treatment Implications

Sean P. Scully, MD, PhD; Michelle A. Ghert, MD; David Zurakowski, PhD; Roby C. Thompson, Jr.MD; Mark C. Gebhardt, MD

Investigation performed at the Musculoskeletal Oncology Division, Department of Orthopaedic Surgery, Duke University Medical Center, Durham, North Carolina; Massachusetts General Hospital, Boston, Massachusetts; Oregon Health Sciences University, Portland, Oregon; Primary Children’s Hospital, Salt Lake City, Utah; University of California at Los Angeles, Los Angeles, California; University of Florida, Gainesville, Florida; University of Miami Medical Center, Miami, Florida; and University of Minnesota, Minneapolis, Minnesota

Sean P. Scully, MD, PhD
Department of Orthopaedic Surgery, Mayo Clinic, 200 First Street S.W., Rochester, MN 55905. E-mail address: scully.sean@mayo.edu

Michelle A. Ghert, MD
Division of Orthopaedics, Department of Surgery, Duke University Medical Center, Durham, NC 27710

David Zurakowski, PhD
Departments of Orthopaedic Surgery and Biostatistics, Children’s Hospital, Harvard Medical School, 300 Longwood Avenue, Boston, MA 02115

Roby C. Thompson Jr., MD
Department of Orthopaedic Surgery, University of Minnesota Medical Center, 420 Delaware Street S.E., Minneapolis, MN 55455
Mark C. Gebhardt, MD
Department of Orthopaedic Surgery, Massachusetts General Hospital, 32 Fruit Street, Boston, MA 02114

The authors did not receive grants or outside funding in support of their research or preparation of this manuscript. They did not receive payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.

A commentary is available with the electronic versions of this article, on our web site (www.jbjs.org) and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM).

The Journal of Bone & Joint Surgery. 2002; 84:49-57

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

text A A A

Abstract

Background: The presence of a pathologic fracture in an osteosarcoma has been considered a poor prognostic factor and an indication for immediate amputation. The purpose of the present study was to determine, in the current era of neoadjuvant chemotherapy, whether a pathologic fracture in an osteosarcoma has prognostic importance and whether limb salvage can be safely performed in such patients without compromising clinical outcome.

Methods: In a cooperative effort of the Musculoskeletal Tumor Society, members from eight institutions provided retrospective data on fifty-two patients with osteosarcoma who had a pathologic fracture and on fifty-five patients with osteosarcoma who had not had a pathologic fracture and had been followed for at least two years or until disease recurrence, metastasis, or death. The two groups were matched for patient age and tumor location. Outcomes examined were survival and local recurrence. A subgroup analysis was performed to assess differences in outcome within the group with the pathologic fracture.

Results: The five-year estimated survival rates were 55% for the group with a pathologic fracture and 77% for the group without a fracture (p = 0.02). The rate of survival without a local recurrence at five years was 75% for the group with a fracture and 96% for the group without a fracture (p = 0.007). In the group with a fracture, seven (23%) of the thirty patients managed with limb salvage and four (18%) of the twenty-two managed with an amputation had a local recurrence (p = 0.75). Eleven (37%) of the thirty patients with a fracture who were managed with limb salvage and ten (45%) of the twenty-two patients with a fracture who were managed with an amputation died of the disease (p = 0.50). Five patients underwent open reduction and internal fixation followed by limb-salvage surgery. Two of them had a local recurrence and died at an average of eight months postoperatively. The remaining three patients were alive at an average of 6.1 years postoperatively. Local disease control and the survival of these patients were not significantly different from those for the thirty-three patients who were treated with nonoperative immobilization of the fracture followed by limb-salvage surgery.

Conclusions: Patients with osteosarcoma who present with a pathologic fracture or sustain one during preoperative chemotherapy have an increased risk of local recurrence and a decreased rate of survival compared with patients who have not sustained a pathologic fracture. The performance of a limb-salvage procedure in carefully selected patients with a pathologic fracture does not significantly increase the risk of local recurrence or death. Factors predictive of improved outcome, such as the response to chemotherapy and union of the fracture, should be taken into account when limb salvage is being considered.

Figures in this Article

Introduction

Limb-salvage has become a popular alternative to amputation for the local control of osteosarcoma. Nonrandomized studies have demonstrated that limb salvage has no adverse effect on survival¹. When complete tumor resection is anatomically possible and neoadjuvant and/or adjuvant chemotherapy is used, limb salvage can improve functional outcome without sacrificing local disease control^1,2. One risk factor that is thought to be a contraindication to limb salvage and that has been historically associated with a poor outcome in patients with osteosarcoma is a pathologic fracture³. The fracture results in a local hematoma with the possible (theoretical) dissemination of tumor cells into adjacent tissue and perhaps into the vasculature and adjacent joints⁴. It has also been hypothesized that damage to the microcirculation may facilitate metastasis⁴. The prevalence of pathologic fracture, either at diagnosis or during preoperative chemotherapy, is between 5% and 10%^5,6.

Some surgeons believe that immediate and aggressive removal of the tumor may halt fracture-induced disease progression and that early amputation is the treatment of choice in many patients with a pathologic fracture in an osteosarcoma^4,5. However, few recent studies have specifically compared the outcome of limb salvage with that of amputation in patients with a pathologic fracture in an osteosarcoma^4,7. The purpose of the present study was to determine whether a pathologic fracture in an osteosarcoma has prognostic importance and whether limb salvage can be safely performed in patients with such a fracture. In addition, we sought to identify any independent variables that may be associated with a poor outcome for patients with a pathologic fracture in an osteosarcoma. Therefore, we performed a multi-institutional, case-matched retrospective review of a large population of patients who had osteosarcoma with or without a pathologic fracture.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 1:: Kaplan-Meier estimated survivorship curves comparing survival rates between patients who had osteosarcoma with a pathologic fracture and those without a fracture. The estimated rates were significantly lower for patients with a pathologic fracture (log-rank test = 5.19, p = 0.02). The error bars around the survivorship curves represent 95% confidence intervals derived by Greenwood’s formula. The numbers of patients on whom the estimates were based are shown in parentheses.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2:: Kaplan-Meier estimated survivorship curves comparing rates of freedom from local recurrence between patients who had osteosarcoma with a pathologic fracture and those without a fracture. The estimated rates were significantly lower for patients with a pathologic fracture (log-rank test = 7.38, p = 0.007). The error bars around the local recurrence curves represent 95% confidence intervals derived by Greenwood’s formula. The numbers of patients on whom the estimates were based are shown in parentheses.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 3:: Kaplan-Meier estimated rates of survival for patients with a pathologic fracture according to tumor management. The estimated rates were not significantly different for patients who had limb salvage compared with those who had an amputation (p = 0.50). The error bars around the survivorship curves represent 95% confidence intervals derived by Greenwood"s formula. The numbers of patients on whom the estimates were based are shown in parentheses.

View Larger

TABLE I: Patient and Tumor Characteristics

*Comparisons were based on the two-sample Student t test, Pearson chi-square test, or Fisher exact test. †The values are given as the mean and the standard deviation. ‡The values are given as the number of patients, with the percentage in parentheses. §Enneking stage8. #Data are missing for one patient who did not have a fracture. **Results are based on forty-five patients with a fracture and thirty-five patients without a fracture for whom pathologist had data available.

Characteristic	Patients with Pathologic Fracture (N = 52)	Patients without Pathologic Fracture (N = 55)	P Value*
Age at surgery† (yr)	23.5 ± 17.4	22.0 ± 13.6	0.63
Gender‡			0.99
Male	28 (54)	30 (55)
Female	24 (46)	25 (45)
Year of surgery‡			0.84
1977-1989	32 (62)	35 (64)
1990-1996	20 (38)	20 (36)
Anatomic location‡			0.82
Proximal part of humerus	13 (25)	13 (24)
Proximal part of femur	?8 (15)	?7 (13)
Distal part of femur	23 (44)	28 (51)
Femoral diaphysis	?3 (6)	?1 (2)
Proximal part of tibia	?4 (8)	?3 (6)
Distal part of tibia	?0 (0)	?1 (2)
Other	?1 (2)	?2 (4)
Tumor size on anteroposterior radiograph‡			0.55
4-8 cm	21 (40)	27 (49)
9-12 cm	21 (40)	21 (38)
>12 cm	10 (19)	?7 (13)
Tumor grade‡§
High (IIB)	52 (100)	55 (100)	1.00
Type of resection‡#
Marginal	?2 (4)	?6 (11)	0.13
Wide	38 (73)	42 (78)
Radical	12 (23)	?6 (11)
Percentage of tumor necrosis estimated by pathologist‡**			0.95
£80	24 (53)	16 (47)
81-90	?5 (11)	?5 (12)
91-95	?4 (9)	?4 (12)
>95	12 (27)	10 (29)

Add to My JBJS

TABLE I: Patient and Tumor Characteristics

Characteristic	Patients with Pathologic Fracture (N = 52)	Patients without Pathologic Fracture (N = 55)	P Value*
Age at surgery† (yr)	23.5 ± 17.4	22.0 ± 13.6	0.63
Gender‡			0.99
Male	28 (54)	30 (55)
Female	24 (46)	25 (45)
Year of surgery‡			0.84
1977-1989	32 (62)	35 (64)
1990-1996	20 (38)	20 (36)
Anatomic location‡			0.82
Proximal part of humerus	13 (25)	13 (24)
Proximal part of femur	?8 (15)	?7 (13)
Distal part of femur	23 (44)	28 (51)
Femoral diaphysis	?3 (6)	?1 (2)
Proximal part of tibia	?4 (8)	?3 (6)
Distal part of tibia	?0 (0)	?1 (2)
Other	?1 (2)	?2 (4)
Tumor size on anteroposterior radiograph‡			0.55
4-8 cm	21 (40)	27 (49)
9-12 cm	21 (40)	21 (38)
>12 cm	10 (19)	?7 (13)
Tumor grade‡§
High (IIB)	52 (100)	55 (100)	1.00
Type of resection‡#
Marginal	?2 (4)	?6 (11)	0.13
Wide	38 (73)	42 (78)
Radical	12 (23)	?6 (11)
Percentage of tumor necrosis estimated by pathologist‡**			0.95
£80	24 (53)	16 (47)
81-90	?5 (11)	?5 (12)
91-95	?4 (9)	?4 (12)
>95	12 (27)	10 (29)

View Larger

TABLE II: Management and Outcome of Patients with a Pathologic Fracture

*P values were determined with use of the Fisher exact test for comparing proportions of local recurrence and death between subgroups. †Fourteen patients who were treated with amputation were excluded. ‡Data available for thirty-three patients.

Characteristic	Total No. of Patients	Local Recurrence		Death
No. (%) of Patients	P Value*	No. (%) of Patients	P Value*
Fracture			0.74		0.57
At presentation	29	?7 (24)		13 (45)
After induction of chemotherapy	23	?4 (17)		?8 (35)
Stabilization of fracture†			0.64		0.99
Open reduction and internal fixation	?5	?2 (40)		?2 (40)
Closed immobilization	33	?8 (24)		13 (39)
Fracture union‡			0.07		0.13
Yes	18	1 (6)		?3 (17)
No	15	?5 (33)		?7 (47)
Fracture displacement			0.72		0.98
Yes	16	?4 (25)		?6 (38)
No	36	?7 (19)		15 (42)
Tumor management			0.75		0.58
Limb salvage	30	?7 (23)		11 (37)
Amputation	22	?4 (18)		10 (45)

Add to My JBJS

TABLE II: Management and Outcome of Patients with a Pathologic Fracture

Characteristic	Total No. of Patients	Local Recurrence		Death
No. (%) of Patients	P Value*	No. (%) of Patients	P Value*
Fracture			0.74		0.57
At presentation	29	?7 (24)		13 (45)
After induction of chemotherapy	23	?4 (17)		?8 (35)
Stabilization of fracture†			0.64		0.99
Open reduction and internal fixation	?5	?2 (40)		?2 (40)
Closed immobilization	33	?8 (24)		13 (39)
Fracture union‡			0.07		0.13
Yes	18	1 (6)		?3 (17)
No	15	?5 (33)		?7 (47)
Fracture displacement			0.72		0.98
Yes	16	?4 (25)		?6 (38)
No	36	?7 (19)		15 (42)
Tumor management			0.75		0.58
Limb salvage	30	?7 (23)		11 (37)
Amputation	22	?4 (18)		10 (45)

View Larger

TABLE III: Predictors of Outcome

*Based on the log-rank test according to Kaplan-Meier survivorship analysis. †Based on results of the final Cox proportional-hazards regression model. ‡Comparison of outcomes in patients managed with amputation and those managed with limb salvage.

Outcome	Variable	Univariate Analysis	Multivariate Analysis
P Value*	Hazard Ratio	?95% Confidence Interval	P Value†
Death	Pathologic fracture	?0.02			?0.80
	Fracture union	?0.03			?0.50
	Tumor size	?0.03			?0.33
	Tumor management‡	?0.006			?0.08
	Percentage necrosis	?0.002	?0.6	0.4-0.9	?0.02
	Local recurrence	<0.0001	10.3	?3.7-25.0	<0.0001
Local recurrence	Pathologic fracture	?0.007	?6.2	?2.6-28.1	?0.005
	Fracture union	?0.03			?0.18
	Fracture displacement	?0.06			?0.54

Add to My JBJS

TABLE III: Predictors of Outcome

Outcome	Variable	Univariate Analysis	Multivariate Analysis
P Value*	Hazard Ratio	?95% Confidence Interval	P Value†
Death	Pathologic fracture	?0.02			?0.80
	Fracture union	?0.03			?0.50
	Tumor size	?0.03			?0.33
	Tumor management‡	?0.006			?0.08
	Percentage necrosis	?0.002	?0.6	0.4-0.9	?0.02
	Local recurrence	<0.0001	10.3	?3.7-25.0	<0.0001
Local recurrence	Pathologic fracture	?0.007	?6.2	?2.6-28.1	?0.005
	Fracture union	?0.03			?0.18
	Fracture displacement	?0.06			?0.54

Materials and Methods

In a cooperative effort of the Musculoskeletal Tumor Society, members from eight institutions provided retrospective data on fifty-two patients who were managed because of an osteosarcoma with a pathologic fracture and on fifty-five patients who were managed because of an osteosarcoma without a pathologic fracture between December 1977 and May 1996. The two groups were matched at each institution with respect to patient age and tumor location. Data were collected with use of a standard data sheet (see Appendix). The patients were observed for at least twenty-four months or until local recurrence, metastasis, or death.

The following data were collected and analyzed for each patient: age, gender, Enneking stage of the disease⁸, anatomic location of the tumor, size of the tumor (determined by the best radiographic method available to the clinician at the time of presentation), presence of a pathologic fracture, evidence of fracture union, extent of fracture displacement, type of fracture stabilization (operative or closed treatment), type of tumor management (limb salvage or amputation), use of neoadjuvant chemotherapy, extent of surgical margins, and percentage of tumor necrosis as determined by the pathologist’s analysis of the resected specimen following complete induction chemotherapy at each institution. Fracture displacement was assessed on plain radiographs and categorized as impacted, displaced less than one-third of the bone diameter, displaced less than two-thirds of the bone diameter, or completely displaced. For the purposes of statistical analysis, the patients were divided into one of two fracture displacement groups: nondisplaced (impacted) or displaced. Fracture angulation was not examined as a variable. Clinicians were asked to identify the chemotherapeutic protocol and the agents used to treat each patient. Chemotherapeutic regimens varied with each institution and era of treatment.

Survival analysis was performed with the Kaplan-Meier product-limit method, and variables that had a p value of <0.10 in the univariate analysis (log-rank test) were entered as candidates into the multivariate Cox proportional-hazards regression model^9,10. Outcomes that were examined included overall survival, disease-free survival, and local recurrence. Survivorship curves were compared with use of the log-rank test. The rates of survival and freedom from local recurrence at five years for patients with or without a pathologic fracture were estimated with use of the Kaplan-Meier product-limit method with 95% confidence intervals derived by Greenwood’s formula, and the curves were compared with use of the log-rank test⁹.

A power analysis revealed that a sample size of 107 patients (fifty-two in the group with the pathologic fracture and fifty-five in the group without a fracture) would provide an 84% power to detect a difference in each outcome based on the log-rank test, assuming a constant hazard ratio of 2.0 during the follow-up period and a two-sided alpha of 0.05. The hazard ratio of 2.0 means that there is an 84% chance of detecting a twofold increase in the risk of death or local recurrence. Variables with a p value of <0.10 in the univariate analysis were entered as covariates into the stepwise Cox proportional-hazards regression model to identify multivariate predictors of each outcome¹⁰. The hazard ratio with a 95% confidence interval was used to measure the strength of the relationship between each multivariate predictor and the corresponding outcome. A subgroup analysis of patients with a pathologic fracture was performed with use of the Fisher exact test or Pearson chi-square test to compare proportions. All reported p values are two-tailed. Power calculations were determined with use of the nQuery Advisor software (version 4.0; Statistical Solutions, Boston, Massachusetts). Statistical analysis was performed with use of the SAS software package (version 6.12; SAS Institute, Cary, North Carolina) and the SPSS software package (version 10.0; SPSS, Chicago, Illinois).

Results

Demographic Data and Patient Management

The median age of the fifty-eight male and forty-nine female patients was 16.5 years (range, two to sixty-nine years). Demographic data for the two matched groups of patients are summarized in Table I. The patients were comparable with respect to age, gender, tumor location, chemotherapy, tumor grade, tumor size, type of resection, and response to chemotherapy. The patients were followed for an average of 4.5 years (range, six months to 12.7 years). All 107 patients presented with Enneking stage-IIB disease. Distant metastasis developed in forty-one patients (38%). The most common primary tumor locations were the distal aspect of the femur (fifty-one patients, 48%), the proximal part of the humerus (twenty-six patients, 24%), and the proximal part of the femur (fifteen patients, 14%). Chemotherapy protocols were the standard ones used at the time of tumor presentation.

Limb salvage was performed in sixty-five patients (thirty with a pathologic fracture and thirty-five without a fracture), and amputation was performed in forty-two patients (twenty-two with a pathologic fracture and twenty without a fracture). Surgical margins were wide in eighty patients (75%), marginal in eight (7%), and radical in eighteen (17%). Pathologic examination of resected specimens from eighty patients for whom data were available revealed that tumor necrosis was £80% in forty patients (50%), 81% to 90% in ten patients (13%), 91% to 95% in eight patients (10%), and >95% in twenty-two (28%).

Data on the presentation and management of the patients who had a pathologic fracture are shown in Table II. Twenty-nine (56%) of the fifty-two patients sustained a pathologic fracture prior to the initial presentation, and twenty-three (44%) sustained a pathologic fracture during the treatment period. As determined by a review of preoperative radiographs, thirty-six fractures were nondisplaced and sixteen were displaced by at least one-third of the bone diameter. With respect to fracture stabilization, thirty-three patients with a pathologic fracture were managed with nonoperative immobilization and five were managed with internal fixation. Of the thirty-three patients for whom data with regard to fracture union were available (data were not available for those who had immediate or early amputation or limb salvage), eighteen (55%) had union and fifteen patients (45%) did not.

Of the twenty-two patients managed with an amputation, nine (41%) had the procedure on presentation and thirteen (59%) had it either during the period when chemotherapy was being carried out or following the completion of induction chemotherapy.

Outcomes in the Entire Study Population

Results of univariate and multivariate analyses of potential prognostic factors are summarized in Table III. With respect to univariate analysis, the presence of a pathologic fracture was found to be significantly associated with a decreased rate of overall survival and an increased rate of local recurrence. The disease-free survival rate at five years was 55% (95% confidence interval, 34% to 67%) in the group that had a pathologic fracture and 77% (95% confidence interval, 68% to 86%) in the group without a fracture (log-rank test = 5.19, p = 0.02) (Fig. 1). The rate of surival without a local recurrence at five years was estimated to be 75% (95% confidence interval, 63% to 87%) in the group that had a pathologic fracture and 96% (95% confidence interval, 92% to 100%) in the group without a fracture (log-rank test = 7.38, p = 0.007) (Fig. 2). With the numbers available, no significant association was found between surgical margins and overall survival or local recurrence. There was no detectable difference in outcome between patients who had a marginal resection and those who had a wide resection (survival, p = 0.49; local recurrence, p = 0.69).

Multivariate analysis with use of the Cox proportional-hazards regression model revealed that a pathologic fracture was a significant independent risk factor for local recurrence (p = 0.005) but not for death (p = 0.80). The most powerful multivariate risk factor for death was local recurrence (p < 0.0001).

Outcomes in the Group with a Pathologic Fracture

We did not find a significant difference with respect to outcome between the patients with a pathologic fracture who underwent an amputation and those who had limb salvage. Seven (23%) of the thirty patients who were managed with limb salvage and four (18%) of the twenty-two who were managed with an amputation had a local recurrence (p = 0.75). Similarly, eleven (37%) of the thirty patients who were managed with limb salvage and ten (45%) of the twenty-two patients who were managed with an amputation died of the disease (p = 0.50) (Fig. 3). We could not detect a significant difference with respect to overall survival between the patients who had an amputation on presentation (four of nine patients died) and those who had an amputation during or after the period of treatment with chemotherapy (six of thirteen patients died) (p = 0.96, Pearson chi-square test with Yate’s correction).

Five patients underwent open reduction and internal fixation followed by limb salvage surgery. Two of them had a local recurrence and died at an average of eight months postoperatively. The remaining three patients were alive at the time of the latest follow-up (average, 6.1 years). With the numbers available, no significant difference with respect to local disease control and survival was detected between these five patients and the thirty-three patients who were managed with nonoperative immobilization of the fracture followed by limb-salvage surgery (p = 0.64 and 0.99, respectively). For the twenty-eight patients for whom data were available, a significant association was found between fracture-healing and the percentage of tumor necrosis (p < 0.03). Of the fifteen patients in whom the fracture united, two had tumor necrosis of £80% and eight had tumor necrosis of >95%. In contrast, twelve of the thirteen patients in whom the fracture did not unite had necrosis of <80% and only one had necrosis of >95%. The Fisher exact test indicated that the fractures that united were associated with a significantly greater percentage of tumor necrosis than were those that did not unite (p < 0.001).

On the basis of simple proportions with use of the Fisher exact test, neither the timing of the fracture (at presentation or after induction of chemotherapy) nor the type of fracture management (open reduction and internal fixation or closed treatment) were significantly associated with any outcome measure (Table II). Of the twenty-nine patients who presented with a pathologic fracture, thirteen (45%) died of the disease, whereas eight (35%) of the twenty-three patients who sustained a fracture during the chemotherapy period died (p = 0.57). Similarly, seven (24%) of the twenty-nine patients who had a fracture on presentation had a local recurrence, whereas four (17%) of the twenty-three patients who sustained a fracture during the chemotherapy period had a local recurrence (p = 0.74).

Fracture displacement was not significantly associated with local recurrence or with overall survival (p = 0.72 and 0.98, respectively) (Table II). Although fracture union was a significant univariate factor for both a decreased rate of local recurrence and an increased rate of survival, it was not a significant multivariate predictor of these outcomes (p = 0.18 and 0.50, respectively) (Table III). However, with use of simple proportions, the associations between fracture union and a decreased risk of local recurrence (p = 0.07) and an increased rate of overall survival (p = 0.13) approached significance (Table II). The fact that fracture union was not found to have independent prognostic significance may be due to the overwhelming effect of other, stronger factors (i.e., the response to chemotherapy and local recurrence).

Discussion

The fifty-two patients with a pathologic fracture in the present study had an overall five-year mortality rate of 32% and a five-year local recurrence rate of 14%. These data are consistent with those reported in other studies on the treatment of osteosarcoma^2,11-13. The risk factors associated with a poorer prognosis in the study population as a whole were a poor response to chemotherapy (a low percentage of tumor necrosis), recurrent local disease, and a pathologic fracture. These findings are also consistent with those of similar large studies^2,3,13-17. To our knowledge, this is the first case-matched evaluation of outcome in patients with a pathologic fracture in an osteosarcoma.

One characteristic of our patient population that does not reflect contemporary cohorts of patients with osteosarcoma is the tumor location. The most common tumor locations in both the patients with a fracture and the patients without a fracture were the distal part of the femur and the proximal aspect of the humerus. In contrast, the proximal part of the tibia has been found to be a more common location for osteosarcoma than the proximal part of the humerus in the majority of recent studies^15,18. However, the patients in our two groups were comparable with respect to age, gender, treatment period, anatomic location of the tumor, tumor size, type of resection, and response to chemotherapy. Therefore, the anatomic discrepancy with respect to other study cohorts should not affect this case-matched analysis. In addition, in the two largest studies on pathologic fracture in osteosarcoma of which we are aware, the proximal part of the humerus was reported to be a more common location than the proximal aspect of the tibia^4,7. These data suggest that osteosarcomas of the proximal part of the humerus may have a relatively increased risk for fracture compared with those in other anatomic locations.

There are several possible reasons why few studies in the orthopaedic literature specifically address the outcome of patients with a pathologic fracture in an osteosarcoma. First, osteosarcoma is a relatively rare malignant lesion, with an annual incidence of two per million¹⁹. Second, the occurrence of a pathologic fracture in these tumors is relatively uncommon (5% to 10%)⁵. Finally, there is controversy with regard to the management of this problem. Although some believe that immediate amputation is indicated^1,17, others think that neoadjuvant chemotherapy can allow limb salvage in selected patients5,20. Through a cooperative effort of the Musculoskeletal Tumor Society, we performed a case-matched study with sufficient numbers and statistical power for definitive analysis of this issue.

Smaller studies have suggested that, in children, the use of neoadjuvant chemotherapy provides the opportunity for fracture-healing and limb salvage. Thompson et al. reviewed the cases of eleven patients who had an osteosarcoma in which a pathologic fracture had occurred²⁰. In that study, all fractures in patients who were less than seventeen years old healed during the period of administration of induction chemotherapy consisting of ifosfamide, Adriamycin (doxorubicin), and methotrexate. The authors concluded that, in children, an effective chemotherapy protocol can retard tumor progression and promote fracture-healing. Jaffe et al. reviewed the records of thirteen similar pediatric patients at M.D. Anderson Hospital and Tumor Institute⁵. Prior to the chemotherapeutic regimen, the tumors were treated with intra-arterial cis-diamminedichloroplatinum-II. Eleven patients had fracture-healing. Although only seven of these patients went on to have operative resection, the authors concluded that a pathologic fracture in an osteosarcoma may heal with therapy and therefore such patients should not be eliminated as candidates for a limb-salvage procedure.

The association between the response to chemotherapy and fracture union reported in the above studies is consistent with the result in the present study. Despite the fact that chemotherapy in and of itself retards the healing process, we found that the fractures that healed were associated with a significantly higher percentage of tumor necrosis than were those that did not. Fracture union in these patients can be attributed to the tumor-cell response to chemotherapy, specifically tumor necrosis, a decrease in or disappearance of a soft-tissue mass, and periosteal bone formation⁵. The local tumor response to chemotherapy was indicative of a similar response of subclinical disseminated disease, as fracture union was associated with an increase in overall and disease-free survival rates and with a decrease in the prevalence of local recurrence.

The findings in the present study with regard to the poorer prognosis associated with a pathologic fracture are consistent with those reported by others in the literature^3,16. However, certain characteristics of fracture presentation may more accurately predict outcome. As in the study by Abudu et al.⁴, we found no significant difference with respect to overall survival or local recurrence between the patients who had a pathologic fracture on presentation and those who sustained a fracture during the period of treatment with chemotherapy. Therefore, the presence of a pathologic fracture on presentation should not be an absolute indication for immediate amputation, as it does not portend a worse prognosis than that portended by a fracture that develops later in the treatment process.

Fracture displacement may also be important in predicting outcome. In our study, univariate analysis revealed that fracture displacement did not predict an increased risk of local recurrence or death. Although the extent of the local tumor and hematoma dissemination was not evaluated in this study, these results suggest that fracture displacement does not necessarily predict an increased risk of local tumor dissemination or distant tumor spread.

Because of the possibility of tumor-cell dissemination during operative treatment of a pathologic fracture in an osteosarcoma, cast immobilization has been the standard initial management during the period of chemotherapy treatment⁴. In contrast, in a study by two of us (S.P.S. and M.C.G.) and colleagues, in which six of eighteen patients underwent operative stabilization for treatment of a pathologic fracture in an osteosarcoma⁷, it was concluded that the method of preoperative fracture treatment did not influence local disease control or overall survival. Similarly, in the present study, the type of fracture management was not found to be significantly associated with outcome. The small number of patients managed with operative stabilization, however, may not have been sufficient to allow detection of a difference between the two groups.

The final issue regarding management of pathologic fractures in osteosarcoma is the surgical approach to local disease control. In three recent studies, the outcomes in patients with a pathologic fracture were evaluated and were stratified on the basis of the type of operative management (limb salvage or amputation).

In the study by two of us and colleagues, the cases of eighteen patients with osteosarcoma who had sustained a pathologic fracture were retrospectively reviewed⁷. The difference in local tumor control and survival between the ten patients who had limb salvage and the eight patients who had an amputation was not found to be significant. It was suggested that low statistical power due to the small number of patients might have been the reason that a difference was not detected.

Dorey et al., in a review of the cases of twenty-four patients who had limb salvage for the treatment of a sarcoma in which a pathologic fracture had occurred, found that these patients had a threefold increased risk of local recurrence compared with that in a similar cohort of patients in whom a pathologic fracture had not occurred (33% compared with 9%)²¹. The difference between the two groups with respect to survival was not significant. It is of note that the patients in that study were treated for all types of sarcoma, not specifically osteosarcoma, and the number of patients with osteosarcoma was not reported.

Abudu et al. reviewed the surgical treatment and oncological results in forty patients with a pathologic fracture in an osteosarcoma⁴. Twenty-seven patients underwent limb salvage, and thirteen had an amputation. Five (19%) of the patients managed with limb salvage and none of those treated with an amputation had a local recurrence. It is of note that all five recurrences occurred in tumor beds where an intralesional or contaminated resection had been performed. Despite the difference in local recurrence between the two groups, the overall survival rate was not significantly different. The authors concluded that if wide margins cannot be achieved, the risk of local recurrence following limb salvage is substantial.

Our results with regard to the relationship between the type of operative management and overall survival and local disease control were similar to the results in these studies. We did not find a difference with respect to outcome between the patients with a pathologic fracture who underwent limb salvage and those who underwent an amputation. It is possible that there was a selection bias in that patients with a tumor that was responsive to chemotherapy and who had fracture union were chosen to undergo limb salvage, leading to an improved outcome in the group managed with limb salvage. The multi-institutional nature of this study is one of its strengths as well as one of its limitations. Because this was a retrospective study, the treatment of these patients was not standardized. The choice of chemotherapeutic regimens, fracture management, and tumor management were specific to the institution and to the surgeon and were not matched for the two study groups. In addition, the chemotherapeutic regimens changed over the course of the study, with advances in medical oncology protocols. As a result, there was a lack of uniformity in the medical and surgical approaches used for these patients. However, the large number of patients and the case-matched format provided a useful tool with which to evaluate the outcome of patients with this rare condition.

In summary, a pathologic fracture in an osteosarcoma is an independent multivariate risk factor for local recurrence and a univariate risk factor for decreased overall survival. With the numbers available, the extent of fracture displacement did not portend a poorer prognosis and the type of fracture stabilization (open reduction and internal fixation compared with closed treatment) did not significantly affect the outcome. The response of the tumor to chemotherapy was predictive of fracture union and improved overall survival and local disease control. Our data indicated that the performance of a limb-salvage procedure in a patient with a pathologic fracture, particularly one that unites following chemotherapy, does not significantly increase the risk of local recurrence or death. It should be stressed that the patients in this study were carefully selected for limb salvage. Factors such as response to chemotherapy and fracture union should be taken into account when limb salvage is being considered.

Appendix

An appendix showing the standard data sheet used to collect patient information is available with the electronic versions of this article, on our web site at www.jbjs.org (go to the article citation and click on "Supplementary Material") and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM).

References

Simon MA. Current concepts review. Limb salvage in osteosarcoma. J Bone Joint Surg Am,1988;70: 307-10. 70307 1988 [PubMed]

Simon MA, Aschilman MA, Thomas N,Mankin HJ. Limb-salvage treatment versus amputation for osteosarcoma of the distal end of the femur. J Bone Joint Surg Am,1986;68: 1331-7. 681331 1986 [PubMed]

Coley BL,Pool JL. Factors influencing the prognosis in osteogenic sarcoma. Ann Surg,1940;112: 1114-28. 1121114 1940 [PubMed]

Abudu A, Sferopoulos NK, Tillman RM, Carter SR,Grimer RJ. The surgical treatment and outcome of pathological fractures in localised osteosarcoma. J Bone Joint Surg Br,1996;78: 694-8. 78694 1996 [PubMed]

Jaffe N, Spears R, Eftekhari F, Robertson R, Cangir A, Takaue Y, Carrasco H, Wallace S, Ayala A, Raymond K,et al. Pathologic fracture in osteosarcoma. Impact of chemotherapy on primary tumor and survival. Cancer,1987;59: 701-9. 59701 1987 [PubMed]

Mulder JO, Schutte HE, Kroon HM, Taconis WK. Radiologic atlas of bone tumors. Amsterdam: Elsevier Science; 1993. Intraosseous osteosarcoma: conventional type; p 51-5

Scully SP, Temple HT, O’Keefe RJ, Mankin HJ,Gebhardt M. The surgical treatment of patients with osteosarcoma who sustain a pathologic fracture. Clin Orthop,1996;324: 227-32. 324227 1996 [PubMed]

Enneking WF. A system of staging musculoskeletal neoplasms. Clin Orthop,1986;204: 9-24. 2049 1986 [PubMed]

Kaplan EL,Meier P. Nonparametric estimation from incomplete observations. J Am Statist Assn,1958;53: 457-81. 53457 1958

Cox DR. Regression models and life-tables. J Roy Statist Soc, Series B,1972;34: 187-202. 34187 1972

Link MP, Goorin AM, Miser AW, Green AA, Pratt CB, Belasco JB, Pritchard J, Malpas JS, Baker AR, Kirkpatrick JA,et al. The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremity. N Engl J Med,1986;314: 1600-6. 3141600 1986 [PubMed]

Bramwell VH, Burgers M, Sneath R, Souhami C, van Oosterom AT, Voute PA, Rouess J, Spooner D, Craft AW, Somers R,et al. A comparison of two short intensive adjuvant chemotherapy regimens in operable osteosarcoma of limbs in children and young adults: the first study of the European Osteosarcoma Intergroup. J Clin Orthop,1992;10: 1579-91. 101579 1992

Meyers PA, Heller G, Healey J, Huvos A, Lane J, Marcove R, Applewhite A, Vlamis V,Rosen G. Chemotherapy for nonmetastatic osteogenic sarcoma: the Memorial Sloan-Kettering experience. J Clin Oncol,1992;10: 5-15. 105 1992 [PubMed]

Taylor WF, Ivins JC, Unni KK, Beabout JW, Golenzer HJ,Black LE. Prognostic variables in osteosarcoma: a multi-institutional study. J Natl Cancer Inst,1989;81: 21-30. 8121 1989 [PubMed]

Glasser DB, Lane JM, Huvos AG, Marcove RC,Rosen G. Survival, prognosis, and therapeutic response in osteogenic sarcoma. The Memorial Hospital experience. Cancer,1992;69: 698-708. 69698 1992 [PubMed]

Dahlin DC. Osteosarcoma of bone and a consideration of prognostic variables. Cancer Treat Rep,1978;62: 189-92. 62189 1978 [PubMed]

Finn HA,Simon MA. Limb-salvage surgery in the treatment of osteosarcoma in skeletally immature individuals. Clin Orthop,1991;262: 108-18. 262108 1991 [PubMed]

Springfield DS, Schmidt R, Graham-Pole J, Marcus RB Jr, Spanier SS,Enneking WF. Surgical treatment for osteosarcoma. J Bone Joint Surg Am,1988;70: 1124-30. 701124 1988 [PubMed]

Widhe B,Widhe T. Initial symptoms and clinical features in osteosarcoma and Ewing sarcoma. J Bone Joint Surg Am,2000;82: 667-74. 82667 2000 [PubMed]

Thompson RC Jr, Pritchard DJ,Nelson TE. Pathologic fractures in osteosarcoma. J Bone Joint Surg Br,1992;74 Suppl III: 277. 74 Suppl III277 1992

Dorey F, Smith S, Eckardt J. Are limb salvage pathological fracture patients at an increased risk of local recurrence but not at increased risk of death? Read at the Combined Meeting of The American and European Musculoskeletal Tumor Societies; 1998 May 6-10; Washington, DC

Topics

View Large | Download Slide(.ppt)
Add to My JBJS

View Larger

TABLE I: Patient and Tumor Characteristics

Characteristic	Patients with Pathologic Fracture (N = 52)	Patients without Pathologic Fracture (N = 55)	P Value*
Age at surgery† (yr)	23.5 ± 17.4	22.0 ± 13.6	0.63
Gender‡			0.99
Male	28 (54)	30 (55)
Female	24 (46)	25 (45)
Year of surgery‡			0.84
1977-1989	32 (62)	35 (64)
1990-1996	20 (38)	20 (36)
Anatomic location‡			0.82
Proximal part of humerus	13 (25)	13 (24)
Proximal part of femur	?8 (15)	?7 (13)
Distal part of femur	23 (44)	28 (51)
Femoral diaphysis	?3 (6)	?1 (2)
Proximal part of tibia	?4 (8)	?3 (6)
Distal part of tibia	?0 (0)	?1 (2)
Other	?1 (2)	?2 (4)
Tumor size on anteroposterior radiograph‡			0.55
4-8 cm	21 (40)	27 (49)
9-12 cm	21 (40)	21 (38)
>12 cm	10 (19)	?7 (13)
Tumor grade‡§
High (IIB)	52 (100)	55 (100)	1.00
Type of resection‡#
Marginal	?2 (4)	?6 (11)	0.13
Wide	38 (73)	42 (78)
Radical	12 (23)	?6 (11)
Percentage of tumor necrosis estimated by pathologist‡**			0.95
£80	24 (53)	16 (47)
81-90	?5 (11)	?5 (12)
91-95	?4 (9)	?4 (12)
>95	12 (27)	10 (29)

Add to My JBJS

TABLE I: Patient and Tumor Characteristics

Characteristic	Patients with Pathologic Fracture (N = 52)	Patients without Pathologic Fracture (N = 55)	P Value*
Age at surgery† (yr)	23.5 ± 17.4	22.0 ± 13.6	0.63
Gender‡			0.99
Male	28 (54)	30 (55)
Female	24 (46)	25 (45)
Year of surgery‡			0.84
1977-1989	32 (62)	35 (64)
1990-1996	20 (38)	20 (36)
Anatomic location‡			0.82
Proximal part of humerus	13 (25)	13 (24)
Proximal part of femur	?8 (15)	?7 (13)
Distal part of femur	23 (44)	28 (51)
Femoral diaphysis	?3 (6)	?1 (2)
Proximal part of tibia	?4 (8)	?3 (6)
Distal part of tibia	?0 (0)	?1 (2)
Other	?1 (2)	?2 (4)
Tumor size on anteroposterior radiograph‡			0.55
4-8 cm	21 (40)	27 (49)
9-12 cm	21 (40)	21 (38)
>12 cm	10 (19)	?7 (13)
Tumor grade‡§
High (IIB)	52 (100)	55 (100)	1.00
Type of resection‡#
Marginal	?2 (4)	?6 (11)	0.13
Wide	38 (73)	42 (78)
Radical	12 (23)	?6 (11)
Percentage of tumor necrosis estimated by pathologist‡**			0.95
£80	24 (53)	16 (47)
81-90	?5 (11)	?5 (12)
91-95	?4 (9)	?4 (12)
>95	12 (27)	10 (29)

View Larger

TABLE II: Management and Outcome of Patients with a Pathologic Fracture

Characteristic	Total No. of Patients	Local Recurrence		Death
No. (%) of Patients	P Value*	No. (%) of Patients	P Value*
Fracture			0.74		0.57
At presentation	29	?7 (24)		13 (45)
After induction of chemotherapy	23	?4 (17)		?8 (35)
Stabilization of fracture†			0.64		0.99
Open reduction and internal fixation	?5	?2 (40)		?2 (40)
Closed immobilization	33	?8 (24)		13 (39)
Fracture union‡			0.07		0.13
Yes	18	1 (6)		?3 (17)
No	15	?5 (33)		?7 (47)
Fracture displacement			0.72		0.98
Yes	16	?4 (25)		?6 (38)
No	36	?7 (19)		15 (42)
Tumor management			0.75		0.58
Limb salvage	30	?7 (23)		11 (37)
Amputation	22	?4 (18)		10 (45)

Add to My JBJS

TABLE II: Management and Outcome of Patients with a Pathologic Fracture

Characteristic	Total No. of Patients	Local Recurrence		Death
No. (%) of Patients	P Value*	No. (%) of Patients	P Value*
Fracture			0.74		0.57
At presentation	29	?7 (24)		13 (45)
After induction of chemotherapy	23	?4 (17)		?8 (35)
Stabilization of fracture†			0.64		0.99
Open reduction and internal fixation	?5	?2 (40)		?2 (40)
Closed immobilization	33	?8 (24)		13 (39)
Fracture union‡			0.07		0.13
Yes	18	1 (6)		?3 (17)
No	15	?5 (33)		?7 (47)
Fracture displacement			0.72		0.98
Yes	16	?4 (25)		?6 (38)
No	36	?7 (19)		15 (42)
Tumor management			0.75		0.58
Limb salvage	30	?7 (23)		11 (37)
Amputation	22	?4 (18)		10 (45)

View Larger

TABLE III: Predictors of Outcome

Outcome	Variable	Univariate Analysis	Multivariate Analysis
P Value*	Hazard Ratio	?95% Confidence Interval	P Value†
Death	Pathologic fracture	?0.02			?0.80
	Fracture union	?0.03			?0.50
	Tumor size	?0.03			?0.33
	Tumor management‡	?0.006			?0.08
	Percentage necrosis	?0.002	?0.6	0.4-0.9	?0.02
	Local recurrence	<0.0001	10.3	?3.7-25.0	<0.0001
Local recurrence	Pathologic fracture	?0.007	?6.2	?2.6-28.1	?0.005
	Fracture union	?0.03			?0.18
	Fracture displacement	?0.06			?0.54

Add to My JBJS

TABLE III: Predictors of Outcome

Outcome	Variable	Univariate Analysis	Multivariate Analysis
P Value*	Hazard Ratio	?95% Confidence Interval	P Value†
Death	Pathologic fracture	?0.02			?0.80
	Fracture union	?0.03			?0.50
	Tumor size	?0.03			?0.33
	Tumor management‡	?0.006			?0.08
	Percentage necrosis	?0.002	?0.6	0.4-0.9	?0.02
	Local recurrence	<0.0001	10.3	?3.7-25.0	<0.0001
Local recurrence	Pathologic fracture	?0.007	?6.2	?2.6-28.1	?0.005
	Fracture union	?0.03			?0.18
	Fracture displacement	?0.06			?0.54