Abstract
Background
The management of primary malignant bone tumors in patients with metastatic disease at presentation remains a challenge. Although surgical resection has been a mainstay in the management of nonmetastatic malignant bone tumors, there is a lack of large-scale evidence-based guidance on whether surgery of the primary site/tumor improves overall survival in malignant bone tumors with metastatic disease at presentation.
Questions/purposes
(1) Is surgical resection of the primary tumor associated with improved overall survival in patients with primary malignant bone tumors who have metastatic disease at presentation? (2) What other factors are associated with improved and/or poor overall survival?
Methods
The 2004 to 2016 National Cancer Database (NCDB), a national registry containing data from more than 34 million cancer patients in the United States, was queried using International Classification of Diseases, 3rd Edition, topographical codes to identify patients with primary malignant bone tumors of the extremities (C40.0-C40.3, C40.8, and C40.9) and/or pelvis (C41.4). The NCDB was preferred over other national cancer registries (that is, the Surveillance, Epidemiology, and End Results database) because it includes a specific variable that codes for patients who received additional surgeries at metastatic sites. Patients with malignant bone tumors of the head or skull, trunk, and spinal column were excluded because these patients are not routinely encountered and treated by orthopaedic oncologists. Histologic codes were used to categorize the tumors into the following groups: osteosarcomas, chondrosarcomas, and Ewing sarcomas. Patients whose tumors were classified as Stage 1, 2, or 3 based on American Joint Commission of Cancer guidelines were excluded. Only patients who presented with metastatic disease were included in the final study sample. The study sample was divided into two distinct groups: those who underwent surgical resection of the primary tumor and those who did not receive any operation for the primary tumor. A total of 2288 patients with primary malignant bone tumors (1121 osteosarcomas, 345 chondrosarcomas, and 822 Ewing sarcomas) with metastatic disease at presentation were included, of whom 46% (1053 of 2288) underwent surgical resection of the primary site. Thirty-three percent (348 of 1053) of patients undergoing surgical resection of the primary site also underwent additional resection of metastases. Patients undergoing surgical resection of the primary site typically were younger than 18 years, lived further from a facility, had tumors involving the upper or lower extremity, had a diagnosis of osteosarcoma or chondrosarcoma, and had a greater tumor size and higher tumor grade at presentation. To account for baseline differences within the patient population and to adjust for additional confounding variables, multivariate Cox regression analyses were used to assess whether undergoing surgical resection of the primary tumor was associated with improved overall survival, after controlling for differences in baseline demographics, tumor characteristics (grade, location, histologic type, and tumor size), and treatment patterns (resection of distant or regional metastatic sites, positive or negative surgical margins, and use of radiation therapy or chemotherapy). Additional sensitivity analyses, stratified by histologic type for osteosarcomas, chondrosarcomas, and Ewing sarcomas, were used to assess factors associated with overall survival for each tumor type.
Results
After controlling for differences in baseline demographics, tumor characteristics, and treatment patterns, we found that surgical resection of the primary site was associated with reduced overall mortality compared with those who did not have a resection of the primary site (hazard ratio 0.42 [95% confidence interval 0.36 to 0.49]; p < 0.001). Among other factors, in the stratified analysis, radiation therapy was associated with improved overall survival for patients with Ewing sarcoma (HR 0.71 [95% CI 0.57 to 0.88]; p = 0.002) but not for those with osteosarcoma (HR 1.14 [95% CI 0.91 to 1.43]; p = 0.643) or chondrosarcoma (HR 1.0 [95 % CI 0.78 to 1.50]; p = 0.643). Chemotherapy was associated with improved overall survival for those with osteosarcoma (HR 0.50 [95% CI 0.39 to 0.64]; p < 0.001) and those with chondrosarcoma (HR 0.62 [95% CI 0.45 to 0.85]; p = 0.003) but not those with Ewing sarcoma (HR 0.7 [95% CI 0.46 to 1.35]; p = 0.385).
Conclusions
Surgical resection of the primary site was associated with an overall survival advantage in patients with primary malignant bone tumors who presented with metastatic disease. Further research, using more detailed data on metastatic sites (such as, size, location, number, and treatment), chemotherapy regimen and location of radiation (primary or metastatic site) is warranted to better understand which patients will have improved overall survival and/or a benefit in the quality of life from resecting their primary malignant tumor if they present with metastatic disease at diagnosis.
Level of Evidence
Level III, therapeutic study.
Introduction
Primary malignant bone sarcomas, which constitute less than 2% of all newly diagnosed cancers in the United States, are associated with substantial morbidity and mortality [21, 33]. For individuals presenting with localized and/or early-stage disease, surgical resection combined with adjunct treatments (chemotherapy or radiation therapy, depending on histology) has been successful in achieving 5-year survival rates up to 70% [9, 19, 24, 29]. However, 20% to 30% of patients with malignant bone tumors present with metastatic disease at the time of diagnosis, with 5-year survival rates ranging from 8% to 35%, depending on the tumor’s histology [10, 19, 22, 27].
Because evidence from small studies is limited by generalizability and the presentation of metastatic disease is diverse, the strategy for the management of the primary tumor in patients with metastatic disease remains a matter of debate [16, 18]. Although prior studies have advocated for aggressive resection of tumor sites, systemic chemotherapy, and adjunct radiation (for radiosensitive tumors) for local control in patients with metastatic disease [1, 3, 4, 6, 7, 11, 18], there is limited evidence on whether resection of the primary tumor/site is associated with improved survival, irrespective of whether or not metastatic sites are resected. Understanding the role of surgical resection of the primary site alone in improving overall survival would allow surgeons to implement shared-decision making when choosing whether or not to operate on these challenging patients.
Considering the latter observations, we used a national cancer dataset to answer our primary research questions: (1) Is surgical resection of the primary tumor associated with improved overall survival in patients with primary malignant bone tumors who have metastatic disease at presentation? (2) What other factors are associated with improved and/or poor overall survival?
Patients and Methods
Database and Patient Selection
This retrospective, comparative study was performed using the National Cancer Database (NCDB) [2]. Jointly owned by the American College of Surgeons and the American Cancer Society, the NCDB is a comprehensive national cancer dataset that contains patient records from more than 1500 Commission on Cancer-accredited facilities across the United States. The database contains the records of more than 34 million patients with cancer and represents more than 70% of newly diagnosed cancers nationally. We preferred to use the NCDB this study over the Surveillance, Epidemiology, and End Results (SEER) database [21] because the NCDB contains specific surgical information on whether or not the primary tumor was resected, as well as whether an individual received a resection of metastases, which must be considered in subsequent analyses. The database, however, does not provide details of other treatments for metastases such as radiotherapy to the metastatic site(s). In addition, there is no information on the number of metastatic sites for each patient or whether they lung metastases versus bone, lymph node, other organs. This is important since metastases are treated differently between tumor sites. For instance, pulmonary metastatic lesions are often resected with thoracotomy/thoracoscopy in osteosarcoma, whereas bone sites are less commonly resected and might get radiation. In contrast, for patients with Ewing sarcoma, pulmonary radiation is used more often than resection for pulmonary metastases. The data provided by the NCDB are deidentified, thus, the study was exempt from institutional review board approval.
We queried the 2004 to 2016 NCDB files using International Classification of Diseases in Oncology, 3rd Edition, topographical codes to identify pediatric (younger than 18 years) and adult (18 years or older) patients with primary malignant bone tumors of the extremities (C40.0-C40.3, C40.8, and C40.9) and/or pelvis (C41.4). Only patients with metastatic disease at presentation were included in the study, and those with American Joint Commission of Cancer Stage 1 to 3 tumors were excluded. Patients with malignant bone tumors of the head or scalp, trunk, and vertebral column were excluded because these patients are not routinely treated by orthopaedic oncologists. Histologic codes were used to group tumors into osteosarcomas, chondrosarcomas, and Ewing sarcomas. Patients with metastatic chordomas were excluded from the study because of a relatively small sample size (n = 32); this would have prevented us from performing an adequately powered analysis. We used the NCDB variable “RX_SUMM_SURG_PRIM” to identify whether a patient received a surgical resection (local excision/partial resection, radical resection/limb-salvage, amputation and unspecified type) at the primary tumor site. Patients with missing data with regard to surgery (n = 5) at the primary site were excluded from the study. We used the variable “RX_SUMM_SURG_OTHREGDIS” to identify whether patients received a resection of metastases to a regional and/or distant site. After application of the inclusion and exclusion criteria, 2288 patients with a malignant primary bone sarcoma and metastatic disease at presentation were included in the study (Fig. 1). The study cohort was divided into two distinct groups: those who underwent surgical resection of the primary site and those who did not receive any surgical resection of the primary site. Of the 2288 patients, 46% (1053) underwent surgery of the primary site. Thirty-three percent (348 of 1053) of patients undergoing surgical resection of the primary site also underwent additional resection of metastases. The mean follow-up of all patients was 28.5 months (SD 31.6).
Fig. 1.
This flowchart diagram depicts how the study sample was derived.
Variables Included
Variables included in the study were patient demographics, tumor characteristics, and treatments. Patient demographics included age (stratified into 18 years and younger, 19 to 50 years, 51 to 75 years, and older than 75 years), sex, race (based on what was recorded in medical record documentation), insurance status (private, Medicare, other government, Medicaid, uninsured, and unknown), median household income, proportion of individuals without a high school education residing in the patient’s ZIP code, Charlson comorbidity index, and distance from the facility (divided into three equal tertiles: 0 to 13 miles, 13 to 49 miles, and more than 49 miles).
Tumor characteristics included location (upper extremity, lower extremity, pelvic, and/or unspecified part of the upper or lower extremity), histology (osteosarcoma, chondrosarcoma, and Ewing sarcoma), tumor size (stratified into 0 to 8 cm, larger than 8 cm, or unknown), and tumor grade (1: well-differentiated or low-grade; 2: moderately differentiated or intermediate-grade; 3: poorly differentiated or high-grade; and 4: undifferentiated or high-grade and unknown). Treatment factors included whether or not patients underwent surgical resection (yes or no), additional resection of metastases of regional or distant sites, surgical margins (negative: R0; positive: R1 or R2; and unknown), and whether individuals received additional radiation therapy (adjuvant or neoadjuvant to primary or metastatic sites) and chemotherapy. Readers should note that the NCDB does not provide detailed information on metastatic lesions (size, location, number and whether these nonprimary sites received radiation). Our purpose of including “resection of metastases” as a variable in the study was to ensure that any confounding being imposed by surgery on nonprimary sites on the overall survival was accounted for in subsequent analyses. It is important to understand that due to the way data is coded in the NCDB, we are unable to identify if the patient received radiation to the primary site only, metastatic site only, or both primary and metastatic sites. When recording the radiation treatment location of an individual receiving treatment at both primary and metastatic sites, the database only captures and records radiation at the primary site. Therefore, we chose not to attempt to delineate, report, and analyze the effect of treatment location, as our answers might be imprecise (that is, it might show inflated improved overall survival for patients with a reported radiation to the primary site only versus a metastatic site without knowing if patients in the primary site group received additional radiation at the metastatic site or not). Similarly, the database does not provide information on the specific chemotherapy regimen used (such as, the type of medication, when it was administered, or duration of treatment).
Baseline Clinical Characteristics of the Study Population
Individuals undergoing surgical resection of the primary site were more likely to be younger and live further away from the treating facility than those who did not undergo surgical resection of the primary site (Table 1). Patients undergoing surgical resection of the primary site typically had tumors involving the upper or lower extremity and a diagnosis of osteosarcoma or chondrosarcoma, as well as a greater tumor size and higher tumor grade at presentation (Table 2). Readers should note that, based on the aforementioned differences, it is likely that clinical judgment and/or patient selection influenced a physician’s choice of doing a resection on the primary tumor. Therefore, we included and adjusted for all previously mentioned variables in a multivariate Cox regression model to ensure that results were independent of any confounding effects being imposed by baseline differences (for example, larger tumor sizes, higher grade, location of tumor) between both cohorts.
Table 1.
Baseline characteristics of those who underwent surgery of the primary site and those who did not
| Baseline characteristics | No surgery of primary site (n = 1235) | Surgery of primary site (n = 1053) | p value |
| Age (years) | < 0.001 | ||
| ≤ 18 | 34% (417) | 44% (467) | |
| 19-50 | 44% (540) | 33% (345) | |
| 51-75 | 16% (202) | 18% (188) | |
| > 75 | 6% (76) | 5% (53) | |
| Sex | 0.455 | ||
| Male | 59% (734) | 61% (642) | |
| Female | 41% (501) | 39% (411) | |
| Racea | 0.526 | ||
| White | 83% (1020) | 81% (857) | |
| Black | 10% (120) | 11% (114) | |
| Asian | 3% (41) | 3% (34) | |
| American Indian/Alaska Native | 1% (7) | 1% (9) | |
| Other | 2% (27) | 3% (29) | |
| Unknown | 2% (20) | <1% (10) | |
| Insurance status | 0.630 | ||
| Private | 53% (655) | 55% (574) | |
| Medicare | 11% (139) | 12% (125) | |
| Other government | 2% (24) | 2% (19) | |
| Medicaid | 23% (283) | 23% (242) | |
| Uninsured | 6% (69) | 5% (53) | |
| Unknown | 5% (65) | 4% (40) | |
| Median household income | 0.874 | ||
| ≥ USD 63,000 | 32% (397) | 31% (323) | |
| USD 48,000 to USD 62,999 | 26% (319) | 27% (284) | |
| USD 38,000 to USD 47,999 | 25% (312) | 25% (259) | |
| < USD 38,000 | 16% (199) | 17% (181) | |
| Unknown | < 1% (8) | < 1% (6) | |
| Proportion without high school education in defined ZIP code | 0.426 | ||
| < 7% | 24% (297) | 22% (233) | |
| 7% to 12.9% | 28% (351) | 32% (336) | |
| 13% to 20.9% | 26% (325) | 25% (264) | |
| ≥ 21% | 21% (254) | 20% (215) | |
| Unknown | < 1% (8) | < 1% (5) | |
| Charlson comorbidity score | 0.163 | ||
| 0 | 91% (1121) | 91% (961) | |
| 1 | 7% (86) | 8% (79) | |
| ≥ 2 | 2% (28) | 1% (13) | |
| Distance from facility (miles) | < 0.001 | ||
| 0 to 13 | 37% (456) | 30% (311) | |
| 13 to 49 | 33% (407) | 33% (349) | |
| ≥ 49 | 30% (364) | 37% (388) | |
| Unknown | < 1% (8) | < 1% (5) |
Race was self-reported, based off medical record documentation
Table 2.
Tumor characteristics of those who underwent surgery of the primary site and those who did not
| Tumor characteristics | No surgery of primary site (n = 1235) | Surgery of primary site (n = 1053) | p value |
| Location | < 0.001 | ||
| Upper | 13% (160) | 17% (177) | |
| Lower | 36% (449) | 69% (729) | |
| Pelvic | 48% (596) | 11% (116) | |
| Unspecified part of limb | 2% (30) | 3% (31) | |
| Histologic result | < 0.001 | ||
| Osteosarcoma | 38% (469) | 62% (652) | |
| Chondrosarcoma | 13% (158) | 18% (187) | |
| Ewing sarcoma | 49% (608) | 20% (214) | |
| Tumor size (cm) | < 0.001 | ||
| 0-8 | 18% (221) | 20% (210) | |
| > 8 | 44% (541) | 57% (600) | |
| Unknown | 38% (473) | 23% (243) | |
| Grade | < 0.001 | ||
| 1 | 1% (10) | 2% (20) | |
| 2 | 2% (29) | 5% (51) | |
| 3 | 15% (188) | 32% (336) | |
| 4 | 20% (246) | 27% (281) | |
| Unknown | 62% (762) | 35% (365) |
The most common type of procedure at the primary site was radical resection (database designation) with limb salvage (59%, 623 of 1053), followed by amputation (24%, 249 of 1053), local excision or partial resection (15%, 155 of 1053), and unspecified type of resections (2%, 26 of 1053) (Table 3). Due to the database’s coding methodology of lumping resections together, it is hard to identify whether these resections were truly “radical” or “wide” excisions. Given that several patients also had positive margins, some of the resections in this cohort may have been wide that were grouped under radical resection and some were marginal and intralesional. Thirty-three percent (348 of 1053) of patients undergoing surgical resection of the primary site also underwent resection of distant or regional metastases. Radiation therapy was used more often in patients who did not undergo surgery of the primary site than in those who did (44% versus 18%; p < 0.001), although the NCDB does not report the site of radiation treatment, which limited our ability to differentiate between radiation for metastatic versus primary site disease.
Table 3.
Treatment characteristics of those who underwent surgery of the primary site and those who did not
| Treatment characteristics | No surgery of primary site (n = 1235) | Surgery of primary site (n = 1053) | p value |
| Type of surgery | < 0.001 | ||
| Local excision or partial resection | 0% (0) | 15% (155) | |
| Radical resection or limb salvage | 0% (0) | 59% (623) | |
| Amputation | 0% (0) | 24% (249) | |
| Surgery, not specified | 0% (0) | 2% (26) | |
| Underwent resection of distant or regional metastases | 20% (245) | 33% (348) | < 0.001 |
| Surgical margins | < 0.001 | ||
| Negative | 0% (0) | 71% (746) | |
| Positive | 0% (0) | 13% (141) | |
| Unknown | 0% (0) | 16% (166) | |
| Radiation therapy | 44% (547) | 18% (192) | < 0.001 |
| Chemotherapy | 81% (998) | 83% (869) | 0.077 |
Statistical Analysis
As a means of exploratory analysis, we used Pearson’s chi-square tests to assess for baseline differences in patient demographics, tumor characteristics, and treatment factors between the two cohorts (surgical resection and no surgical resection of the primary site). A multivariate Cox regression analysis was used for the entire sample to assess whether surgery of the primary site was associated with overall survival, after adjusting for differences in patient demographics, tumor characteristics, and treatment patterns between the two cohorts. Additional sensitivity analyses using Cox regression models, stratified by histologic type for osteosarcomas, chondrosarcomas, and Ewing sarcomas, were used to assess prognostic factors for overall survival for each histologic type. The results of the multivariate Cox regression models are reported as adjusted hazard ratios, along with their 95% confidence intervals and respective p values. For all statistical analyses, a p value less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 24 (IBM Corp, Armonk, NY, USA).
Results
Is Surgical Resection of the Primary Tumor Site Associated with Improved Overall Survival?
After controlling for differences in patient demographics, tumor characteristics, and treatment factors, we found that surgical resection of the primary site was associated with reduced overall mortality (HR 0.42 [95% CI 0.36 to 0.49]; p < 0.001) for the entire cohort. After stratifying by histologic type, we found that surgical resection of the primary site was associated with improved overall survival for patients with osteosarcomas (HR 0.43 [95% CI 0.36 to 0.52] ; p < 0.001), chondrosarcomas (HR 0.45 [95% CI 0.30 to 0.68]; p < 0.001), and Ewing sarcomas (HR 0.36 [95% CI 0.27 to 0.51]; p < 0.001) (Table 4).
Table 4.
Multivariate Cox regression analyses showing prognostic factors associated with survival, after stratifying based on histology into osteosarcoma, chondrosarcoma, and Ewing sarcoma
| Osteosarcoma | ||
| Associated factors | Hazard ratio (95% CI) | p value |
| Surgical resection of primary site | 0.43 (0.36 to 0.52) | < 0.001 |
| Age (years) | ||
| ≤ 18 | Ref. | |
| 19-50 | 1.38 (1.15 to 1.66) | 0.001 |
| 51-75 | 2.19 (1.69 to 2.82) | < 0.001 |
| > 75 | 3.06 (2.04 to 4.59) | < 0.001 |
| Race | ||
| White | Ref. | |
| Black | 0.78 (0.62 to 0.97) | 0.027 |
| Asian | 0.86 (0.51 to 1.44) | 0.561 |
| American Indian/Alaska Native | 0.84 (0.41 to 1.71) | 0.615 |
| Other | 1.09 (0.64 to 1.87) | 0.749 |
| Unknown | 0.98 (0.47 to 2.01) | 0.946 |
| Insurance | ||
| Private | Ref. | |
| Medicare | 1.55 (1.16 to 2.07) | 0.003 |
| Other government | 0.68 (0.37 to 1.24) | 0.208 |
| Medicaid | 0.94 (0.76 to 1.15) | 0.526 |
| Uninsured | 1.12 (0.81 to 1.56) | 0.493 |
| Unknown | 0.89 (0.54 to 1.54) | 0.632 |
| Resection of regional or distant metastases | 0.97 (0.81 to 1.16) | 0.725 |
| Radiation therapy | 1.14 (0.91 to 1.43) | 0.261 |
| Chemotherapy | 0.50 (0.39 to 0.64) | < 0.001 |
| Positive surgical margins | 1.55 (1.13 to 2.14) | 0.007 |
| Chondrosarcoma | ||
| Associated factors | Hazard ratio (95% CI) | p value |
| Surgical resection of primary site | 0.45 (0.30 to 0.68) | < 0.001 |
| Age (years) | ||
| ≤ 18 | Ref. | |
| 19-50 | 3.48 (0.76 to 16.07) | 0.110 |
| 51-75 | 5.36 (1.14 to 25.13) | 0.033 |
| > 75 | 5.81 (1.18 to 28.78) | 0.031 |
| Race | ||
| White | Ref. | |
| Black | 0.48 (0.26 to 0.89) | 0.020 |
| Asian | 1.10 (0.47 to 2.61) | 0.826 |
| American Indian/Alaska Native | 1.75 (0.20 to 15.42) | 0.615 |
| Other | 1.87 (0.25 to 13.90) | 0.541 |
| Unknown | 0.74 (0.17 to 3.23) | 0.689 |
| Median household income | ||
| ≥ USD 63,000 | Ref. | |
| USD 48,000-USD 62,999 | 1.21 (0.81 to 1.82) | 0.356 |
| USD 38,000-USD 47,999 | 1.41 (0.89 to 2.25) | 0.146 |
| < USD 38,000 | 1.90 (1.10 to 3.28) | 0.021 |
| Unknown | 2.07 (0.22 to 19.51) | 0.524 |
| Location | ||
| Upper | Ref. | |
| Lower | 0.68 (0.44 to 1.06) | 0.155 |
| Pelvic | 0.50 (0.31 to 0.80) | 0.004 |
| Unspecified part of limb | 0.58 (0.27 to 1.27) | 0.175 |
| Grade | ||
| 1 | Ref. | |
| 2 | 1.84 (0.72 to 4.68) | 0.203 |
| 3 | 3.50 (1.41 to 8.66) | 0.007 |
| 4 | 4.19 (1.66 to 10.56) | 0.002 |
| Unknown | 2.01 (0.82 to 4.97) | 0.129 |
| Resection of regional or distant metastases | 0.74 (0.51 to 1.06) | 0.101 |
| Radiation therapy | 1.08 (0.78 to 1.50) | 0.643 |
| Chemotherapy | 0.62 (0.45 to 0.85) | 0.003 |
| Positive surgical margins | 1.38 (0.87 to 2.17) | 0.169 |
| Ewing sarcoma | ||
| Associated factors | Hazard ratio (95% CI) | p value |
| Surgical resection of primary site | 0.36 (0.27 to 0.51) | < 0.001 |
| Age (years) | ||
| ≤ 18 | Ref. | |
| 19-50 | 2.07 (1.67 to 2.57) | < 0.001 |
| 51-75 | 4.76 (2.76 to 8.23) | < 0.001 |
| > 75 | 87.27 (8.21 to 927.35) | < 0.001 |
| Race | ||
| White | Ref. | |
| Black | 1.93 (1.20 to 3.09) | 0.007 |
| Asian | 2.64 (1.48 to 4.70) | 0.001 |
| American Indian/Alaska Native | 6.71 (2.04 to 22.01) | 0.002 |
| Other | 0.67 (0.34 to 1.34) | 0.257 |
| Unknown | 0.54 (0.17 to 1.71) | 0.295 |
| Location | ||
| Upper | Ref. | |
| Lower | 1.27 (0.91 to 1.78) | 0.155 |
| Pelvic | 1.22 (0.88 to 1.69) | 0.243 |
| Unspecified part of limb | 2.26 (1.15 to 4.44) | 0.018 |
| Resection of regional or distant metastases | 0.85 (0.65 to 1.11) | 0.235 |
| Radiation therapy | 0.71 (0.57 to 0.88) | 0.002 |
| Chemotherapy | 0.79 (0.46 to 1.35) | 0.385 |
| Positive surgical margins | 1.79 (0.98 to 3.29) | 0.058 |
Adjustment was performed for all demographics, tumor characteristics, and treatment patterns.
What Factors Are Associated with Improved and/or Poor Overall Survival?
After controlling for baseline demographics, tumor characteristics, and treatment patterns (including whether the patient received surgical resection of the primary site), we found that resection of metastases was not associated with overall survival for the entire cohort (HR 0.92 [95% CI 0.81 to 1.05]; p = 0.235) and nor was it associated with overall survival in the histology-stratified subsets of osteosarcomas (Table 4), chondrosarcomas (Table 4) and Ewing sarcomas (Table 4). Radiation therapy was associated with improved overall survival for patients with Ewing sarcoma (HR 0.71 [95% CI 0.57 to 0.88]; p = 0.002) but not for those with osteosarcoma (HR 1.14 [95% CI 0.91 to 1.43] ; p = 0.643) or chondrosarcoma (HR 1.0 [95 % CI 0.78 to 1.50]; p = 0.643). Chemotherapy was associated with improved overall survival for patients with osteosarcoma (HR 0.50 [95% CI 0.39 to 0.64]; p < 0.001) and those with chondrosarcoma (HR 0.62 [95% CI 0.45 to 0.85]; p = 0.003) but not for those with Ewing sarcoma (HR 0.79 [95% CI 0.46 to 1.35]; p = 0.385). Among other factors, increasing age, low median household income, higher tumor grade (3 or 4) at presentation, having Medicare insurance, and presence of positive surgical margins were predictors of poor prognosis.
Discussion
Even with substantial advances in orthopaedic oncology and improvements in insurance coverage of patients with cancer nationwide [8, 20, 23, 25], up to one-third of all individuals with malignant primary bone sarcomas present with metastatic disease at the time of diagnosis. Evidence guiding the management of patients with metastatic primary bone tumors remains conflicted largely because of limitations inherent to single-institutional experiences and/or small sample sizes. Our results showed that, regardless of whether or not metastatic sites were resected, surgical resection of the primary tumor was associated with improved survival among patients with metastatic primary bone tumors in our study sample. However, given the lack of detailed data on metastatic sites and the absence of quality-of-life scores in the database, conclusions on the value of extra surgery for these vulnerable patients should be derived with caution.
Limitations
The most important limitation of the study is the lack of detailed data on metastatic sites. Even though the NCDB records whether or not patients underwent resection of regional and/or distant metastases, it does not define the number (one to two metastatic lesions/nodules versus diffuse metastatic disease), size of metastases, location of metastases (lung, other bone, and other locations) and whether radiation was received at metastatic sites. Although the American Joint Commission of Cancer staging system subclassifies Stage IV into IVa (presence of lung metastases) and IVb (metastases to other sites), the NCDB does not consistently rely on subclassification coding systems and mostly uses sequential staging numbers (1, 2, 3, and 4) to identify patients. The variable for resection of metastases “RX_SUMM_SURG_OTHREGDIS” does not report the number of resections performed (for example, resection for one or two solitary pulmonary nodules) or the size of the resected lung lesions, which are known to influence the odds of complete resection and subsequently, overall survival [13, 15, 31, 32]. Due to these limitations, readers should be cautious about deriving conclusions regarding our finding that resection of metastases are not associated with an improvement in overall survival. Our purpose of including “resection of metastases” in the study was to simply use the variable as a covariate in our analyses, to ensure that any confounding being imposed by resection at nonprimary sites was accounted for in our final regression model. We also did not have access to data on specific chemotherapy regimens that were employed. The database also does not report quality-of-life data, which would have proven useful in understanding whether the survival benefit associated with surgery was associated with an improvement or worsening in patient quality-of-life. Although this may be a confounding factor, such clinically relevant data is not recorded in national cancer registries and databases. Although single-center studies might be able to replicate the idea and adjust for these factors, the results of these studies may ultimately be limited by small, underpowered sample sizes and a lack of generalizability. The availability of a large sample size coupled with a robust cancer dataset currently available through the NCDB are strong supporting factors in our study. Future multicenter studies/trials may be required to readdress the impact of metastases location on overall survival.
Another important limitation of the database is its inability to record detailed data on the site undergoing radiation therapy. Specifically, when a patient receiving radiation to both the primary and metastatic sites, the database only captures and records the treatment as radiation received at the primary site. Thus, we chose not to attempt to delineate, report, and analyze the effect of treatment location, as this may introduce bias (that is, it might show prolonged overall survival for patients with a reported radiation to the primary site only versus the metastatic site, without knowing if patients in the primary site group received radiation to additional metastatic sites). However, this does not entirely discount the findings of the study, as we chose to adjust for the impact of overall radiation therapy on the overall survival as a means of controlling for any confounding when assessing the impact of surgical resection of primary tumor on overall survival. Regardless, there is a need for more accurate and detailed clinical data that may provide clarity regarding the effectiveness of radiation therapy to the primary tumor as well as distant metastases in patients with metastatic primary malignant bone tumors.
It should also be understood that there is no information on the decision-making process for which patients had a resection of the primary and those who did not, thereby including a possible selection bias in the cohort. It is likely that those operated on had factors that the treating team decided made them more favorable patients, such as tumors in easily resectable sites, a small number of pulmonary-only metastases, or issues such as concern about local pain and skin integrity at the primary site. Although differences between the patients receiving surgical resection and those who did not (higher tumor grade, extremity versus pelvic location, younger age among the surgical resection group) existed, it is important to mention that some of these baseline differences were accounted for in our multivariate Cox regression analyses. Thus statistically, we were able to account for some of this patient selection bias, however, it is difficult to discount other confounding variables (such as, socioeconomic status and the availability of a caregiver) that may have an influenced a physician’s or patient’s choice of whether or not to undergo surgery.
We were unable to adjust or account for the effect of facility volume on overall survival, as the sample of patients with metastatic disease (n = 2288) was too small to identify appropriate volume cutoffs. The NCDB also does not contain data on individual surgeon volume, which may have a confounding effect on indications for surgical resection and its effect on overall survival. We also did not account for facility type (academic or national cancer institute versus community cancer programs) because the database does not provide information on this variable for individuals younger than 40 years, and including this variable may have led to the introduction of bias in our findings. We chose not to investigate whether the type of surgery at the primary site (amputation versus radical resection with limb salvage) influenced overall survival, as this was beyond the scope of the study. The NCDB does not report the patients’ causes of death; therefore, we were unable to calculate disease-specific survival rates. The database does not provide clinical information on whether the patient was rendered disease-free after surgical resection. It is possible that the survival of patients with a complete surgical resection was still poor but better than patients with residual disease. We also did not have access to billing or coding data to estimate the overall costs of care, which would have been critical for discussing the “value” of surgery for late-stage primary bone sarcoma. We did not differentiate between neoadjuvant and/or adjuvant radiation or chemotherapy and their impact on overall survival because this was also beyond the objective of the study. The NCDB does not provide data on the type of chemotherapy regimens administered to patients, which would have been useful to include in the analysis. Lastly, the database only records the definitive surgical procedure, which prevented us from completely differentiating individuals who may have received more than one surgical procedure at the primary site (specifically, an individual who underwent amputation for a failed radical resection with limb salvage is recorded as an amputation in the dataset, making it impossible to differentiate from a primary amputation). Lastly, even though the NCDB is a national cancer registry, follow-up quality of the database is unknown. However, the NCDB does rely on stringent process-of-measures, such as audits and physician querying, to ensure that the data contained within has high validity and accuracy.
Surgical Resection of the Primary Tumor and Overall Survival
In line with past research and recent unpublished data and conference abstracts, we observed that surgical resection of the primary tumor was associated with an overall survival benefit for patients presenting with metastatic primary malignant bone tumors. Using the national SEER database, Ren et al. [28] found that surgical resection of primary tumors in patients with metastatic Ewing sarcoma was associated with substantially improved overall survival. Two additional SEER-based abstracts presented at the 2019 Musculoskeletal Tumor Society (MSTS) meeting found that surgical resection of the primary tumor site improves survival for both patients with metastatic osteosarcoma and those with soft-tissue sarcomas. However, because these were only abstracts, we will wait for the publication of both as full studies before drawing conclusions. Another SEER-based study by Song et al. [30] found resection of the primary tumor to be associated with better overall survival. These prior studies [28, 30] were limited because the SEER database does not record whether the patient underwent additional resection of metastases, thereby preventing the adjustment of this possible confounding factor in respective analyses. We believe the two SEER-based MSTS abstracts were limited for the same reason, but again, we will withhold drawing conclusions on them until both abstracts are published as full studies. As mentioned previously, the NCDB, however, allows researchers to identify whether patients received some additional surgical procedure of nonprimary or metastatic locations, thereby allowing us to control and adjust for the possible confounding effect of resection of metastases on overall survival.
Factors Associated with Improved and/or Poor Overall Survival
Unlike prior evidence that found resection of metastatic lesions (in addition to resection of primary site) to be associated with improved overall survival in metastatic osteosarcomas [14], we were unable to observe an independent survival benefit for resection of metastases in our study. However, as mentioned previously, given the lack of detailed data regarding the number, location, size, and adjunct treatment (such as radiation therapy) in our analysis, readers should be cautious against drawing clinical correlations regarding this finding. It is possible that the lack of an association may be due to underlying selection bias, and more detailed information characterizing metastatic lesions is necessary to identify a subset of patients who might benefit from resection of metastatic disease in addition to resection of the primary site. Given these limitations, we are unable to provide a clinical recommendation for or against the independent role of resection of metastases on overall survival of patients with primary malignant bone tumors who have metastatic disease. With regard to other treatment options, the survival benefit associated with the use of chemotherapy and/or radiation therapy was histology-dependent. Specifically, chemotherapy in patients with osteosarcoma and chondrosarcoma was associated with a 50% and 38% lower risk-adjusted hazard of overall mortality, respectively. Similar to our findings, a prior single-institution report by Bacci et al. [5] showed that in patients with osteosarcomas and lung metastases at presentation, using a combination treatment of aggressive chemotherapy and surgical resection of primary and metastatic tumor sites was associated with substantially improved overall survival. In another study, Meyers et al. [17] analyzed 62 patients with metastatic osteosarcomas and concluded that aggressive surgical resection of the tumor combined with chemotherapy is necessary for prolonged survival. As expected, radiation therapy was associated with nearly a 30% reduction in the risk-adjusted hazard rate of overall mortality among patients with metastatic Ewing sarcomas. Although this finding is in line with prior studies showing that radiation therapy of both primary and metastatic sites in patients with Ewing sarcoma was associated with improved overall survival [12, 26], readers should interpret this observation considering the database’s limitation/inability to differentiate whether radiation was given to the primary or metastatic site. Although our results showing chemotherapy to be associated with an improved overall survival in chondrosarcomas was statistically significant, this finding may not be biologically appropriate given that this specific tumor type is traditionally considered to be chemoresistant. Most patients with chondrosarcomas who underwent chemotherapy had high-grade tumors (52% of chemotherapy patients were 3 or 4, and 36% were unknown), and it is possible that a fair number of these patients had dedifferentiated or mesenchymal chondrosarcomas, where chemotherapy does appear to have some role [14]. Unfortunately, the lack of detailed histological data prevents us further stratifying our analysis to understand the role of chemotherapy in these patients. As mentioned previously, the NCDB also does not report detailed information on the chemotherapy regimen and treatment duration, which may confound our findings. Although studying the role of chemotherapy in the management of metastatic and advanced chondrosarcomas is beyond the scope of this study, future research exploring this specific topic is warranted.
Using a national validated surgical dataset with relatively robust data compared with other conventional cancer registries, our study showed that surgical resection of the primary site is associated with improved overall survival for patients presenting with metastatic primary malignant bone tumors. Although these findings may be helpful when making decisions on whether or not to surgically resect the primary tumor in these patients, physicians should be aware that patients who underwent surgical resection of the primary tumor/site in this study are likely carefully selected. The findings should be interpreted with caution given the lack of data on metastatic lesions and presence of possible cotreatment bias (that is, adjunct systemic and/or radiation therapy may have led to improvement for reasons other than surgical treatment alone). Future prospective studies that collect detailed data on metastatic lesions, radiation specifications, and treatment regimens are required to establish the most appropriate treatment protocol to improve the overall survival of this vulnerable patient population.
Footnotes
Each author certifies that neither he, nor any member of his immediate family, has funding or commercial associations (consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article.
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research® editors and board members are on file with the publication and can be viewed on request.
Each author certifies that his institution waived approval for the human protocol for this investigation and that all investigations were conducted in conformity with ethical principles of research.
References
- 1.Aljubran AH, Griffin A, Pintilie M, Blackstein M. Osteosarcoma in adolescents and adults: survival analysis with and without lung metastases. Ann Oncol. 2009;20:1136-1141. [DOI] [PubMed] [Google Scholar]
- 2.American. American College of Surgeons. National Cancer Database. 2019. Available at: https://www.facs.org/quality-programs/cancer/ncdb. Accessed September 3.
- 3.Bacci G, Briccoli A, Ferrari S, Ruggieri P, Avella M, Casadei R, Battistini A, Picci P. [Osteosarcoma of the extremities metastatic at presentation. Results obtained with primary chemotherapy followed by simultaneous resection of the primary and metastatic lesion]. Minerva Chir. 1993;48:35-45. [PubMed] [Google Scholar]
- 4.Bacci G, Briccoli A, Ferrari S, Saeter G, Donati D, Longhi A, Manfrini M, Bertoni F, Rimondini S, Monti C, Forni C. Neoadjuvant chemotherapy for osteosarcoma of the extremities with synchronous lung metastases: treatment with cisplatin, adriamycin and high dose of methotrexate and ifosfamide. Oncol Rep. 2000;7:339-346. [DOI] [PubMed] [Google Scholar]
- 5.Bacci G, Mercuri M, Briccoli A, Ferrari S, Bertoni F, Donati D, Monti C, Zanoni A, Forni C, Manfrini M. Osteogenic sarcoma of the extremity with detectable lung metastases at presentation. Results of treatment of 23 patients with chemotherapy followed by simultaneous resection of primary and metastatic lesions. Cancer. 1997;79:245-254. [PubMed] [Google Scholar]
- 6.Bacci G, Rocca M, Salone M, Balladelli A, Ferrari S, Palmerini E, Forni C, Briccoli A. High grade osteosarcoma of the extremities with lung metastases at presentation: treatment with neoadjuvant chemotherapy and simultaneous resection of primary and metastatic lesions. J Surg Oncol. 2008;98:415-420. [DOI] [PubMed] [Google Scholar]
- 7.Bielack SS, Kempf-Bielack B, Delling G, Exner GU, Flege S, Helmke K, Kotz R, Salzer-Kuntschik M, Werner M, Winkelmann W, Zoubek A, Jurgens H, Winkler K. Prognostic factors in high-grade osteosarcoma of the extremities or trunk: an analysis of 1,702 patients treated on neoadjuvant cooperative osteosarcoma study group protocols. J Clin Oncol. 2002;20:776-790. [DOI] [PubMed] [Google Scholar]
- 8.Cannon RB, Shepherd HM, McCrary H, Carpenter PS, Buchmann LO, Hunt JP, Houlton JJ, Monroe MM. Association of the Patient Protection and Affordable Care Act With Insurance Coverage for Head and Neck Cancer in the SEER Database. JAMA Otolaryngol Head Neck Surg. 2018;144:1052-1057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Damron TA, Ward WG, Stewart A. Osteosarcoma, chondrosarcoma, and Ewing's sarcoma: National Cancer Data Base Report. Clin Orthop Relat Res. 2007;459:40-47. [DOI] [PubMed] [Google Scholar]
- 10.Duchman KR, Gao Y, Miller BJ. Prognostic factors for survival in patients with Ewing's sarcoma using the surveillance, epidemiology, and end results (SEER) program database. Cancer Epidemiology. 2015;39:189-195. [DOI] [PubMed] [Google Scholar]
- 11.Goorin AM, Delorey MJ, Lack EE, Gelber RD, Price K, Cassady JR, Levey R, Tapper D, Jaffe N, Link M, et al. Prognostic significance of complete surgical resection of pulmonary metastases in patients with osteogenic sarcoma: analysis of 32 patients. J Clin Oncol. 1984;2:425-431. [DOI] [PubMed] [Google Scholar]
- 12.Haeusler J, Ranft A, Boelling T, Gosheger G, Braun-Munzinger G, Vieth V, Burdach S, van den Berg H, Juergens H, Dirksen U. The value of local treatment in patients with primary, disseminated, multifocal Ewing sarcoma (PDMES). Cancer. 2010;116:443-450. [DOI] [PubMed] [Google Scholar]
- 13.Handy JR, Bremner RM, Crocenzi TS, Detterbeck FC, Fernando HC, Fidias PM, Firestone S, Johnstone CA, Lanuti M, Litle VR, Kesler KA, Mitchell JD, Pass HI, Ross HJ, Varghese TK. Expert Consensus Document on Pulmonary Metastasectomy. Ann Thorac Surg. 2019;107:631-649. [DOI] [PubMed] [Google Scholar]
- 14.Italiano A, Mir O, Cioffi A, Palmerini E, Piperno-Neumann S, Perrin C, Chaigneau L, Penel N, Duffaud F, Kurtz JE, Collard O, Bertucci F, Bompas E, Le Cesne A, Maki RG, Ray Coquard I, Blay JY. Advanced chondrosarcomas: role of chemotherapy and survival. Ann Oncol. 2013;24:2916-2922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lindner LH, Litiere S, Sleijfer S, Benson C, Italiano A, Kasper B, Messiou C, Gelderblom H, Wardelmann E, Le Cesne A, Blay JY, Marreaud S, Hindi N, Desar IME, Gronchi A, van der Graaf WTA. Prognostic factors for soft tissue sarcoma patients with lung metastases only who are receiving first-line chemotherapy: An exploratory, retrospective analysis of the European Organization for Research and Treatment of Cancer-Soft Tissue and Bone Sarcoma Group (EORTC-STBSG). Int J Cancer. 2018;142:2610-2620. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Meazza C, Scanagatta P. Metastatic osteosarcoma: a challenging multidisciplinary treatment. Expert Rev Anticancer Ther. 2016;16:543-556. [DOI] [PubMed] [Google Scholar]
- 17.Meyers PA, Heller G, Healey JH, Huvos A, Applewhite A, Sun M, LaQuaglia M. Osteogenic sarcoma with clinically detectable metastasis at initial presentation. J Clin Oncol. 1993;11:449-453. [DOI] [PubMed] [Google Scholar]
- 18.Mialou V, Philip T, Kalifa C, Perol D, Gentet JC, Marec-Berard P, Pacquement H, Chastagner P, Defaschelles AS, Hartmann O. Metastatic osteosarcoma at diagnosis: prognostic factors and long-term outcome--the French pediatric experience. Cancer. 2005;104:1100-1109. [DOI] [PubMed] [Google Scholar]
- 19.Miller BJ, Cram P, Lynch CF, Buckwalter JA. Risk factors for metastatic disease at presentation with osteosarcoma: an analysis of the SEER database. J Bone Joint Surg Am. 2013;95:e89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Moss HA, Havrilesky LJ, Chino J. Insurance coverage among women diagnosed with a gynecologic malignancy before and after implementation of the Affordable Care Act. Gynecol Oncol. 2017;146:457-464. [DOI] [PubMed] [Google Scholar]
- 21.National Institutes of Health. Surveillance, Epidemiology and End Results - Cancer Statistics. 2016. Available at: https://seer.cancer.gov/statfacts/html/bones.html. Accessed 8th October, 2019.
- 22.Nie Z, Lu Q, Peng H. Prognostic factors for patients with chondrosarcoma: A survival analysis based on the Surveillance, Epidemiology, and End Results (SEER) database (1973–2012). Journal of Bone Oncology. 2018;13:55-61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Nogueira LM, Chawla N, Han X, Jemal A, Yabroff KR. Patterns of Coverage Gains Among Young Adult Cancer Patients Following the Affordable Care Act. JNCI Cancer Spectr. 2019;3:pkz001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ottaviani G, Jaffe N. The epidemiology of osteosarcoma. Cancer Treat Res. 2009;152:3-13. [DOI] [PubMed] [Google Scholar]
- 25.Parsons HM, Schmidt S, Tenner LL, Bang H, Keegan TH. Early impact of the Patient Protection and Affordable Care Act on insurance among young adults with cancer: Analysis of the dependent insurance provision. Cancer. 2016;122:1766-1773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Paulino AC, Mai WY, Teh BS. Radiotherapy in metastatic ewing sarcoma. Am J Clin Oncol. 2013;36:283-286. [DOI] [PubMed] [Google Scholar]
- 27.Ramkumar DB, Ramkumar N, Miller BJ, Henderson ER. Risk factors for detectable metastatic disease at presentation in Ewing sarcoma - An analysis of the SEER registry. Cancer Epidemiol. 2018;57:134-139. [DOI] [PubMed] [Google Scholar]
- 28.Ren Y, Zhang Z, Shang L, You X. Surgical Resection of Primary Ewing's Sarcoma of Bone Improves Overall Survival in Patients Presenting with Metastasis. Med Sci Monit. 2019;25:1254-1262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Smeland S, Bielack SS, Whelan J, Bernstein M, Hogendoorn P, Krailo MD, Gorlick R, Janeway KA, Ingleby FC, Anninga J, Antal I, Arndt C, Brown KLB, Butterfass-Bahloul T, Calaminus G, Capra M, Dhooge C, Eriksson M, Flanagan AM, Friedel G, Gebhardt MC, Gelderblom H, Goldsby R, Grier HE, Grimer R, Hawkins DS, Hecker-Nolting S, Sundby Hall K, Isakoff MS, Jovic G, Kuhne T, Kager L, von Kalle T, Kabickova E, Lang S, Lau CC, Leavey PJ, Lessnick SL, Mascarenhas L, Mayer-Steinacker R, Meyers PA, Nagarajan R, Randall RL, Reichardt P, Renard M, Rechnitzer C, Schwartz CL, Strauss S, Teot L, Timmermann B, Sydes MR, Marina N. Survival and prognosis with osteosarcoma: outcomes in more than 2000 patients in the EURAMOS-1 (European and American Osteosarcoma Study) cohort. Eur J Cancer. 2019;109:36-50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Song K, Song J, Chen F, Lin K, Ma X, Jiang J. Does Resection of the Primary Tumor Improve Survival in Patients With Metastatic Chondrosarcoma? Clin Orthop Relat Res. 2019;477:573-583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Suzuki M, Iwata T, Ando S, Iida T, Nakajima T, Ishii T, Yonemoto T, Tatezaki S, Fujisawa T, Kimura H. Predictors of long-term survival with pulmonary metastasectomy for osteosarcomas and soft tissue sarcomas. J Cardiovasc Surg (Torino). 2006;47:603-608. [PubMed] [Google Scholar]
- 32.Todd TR. The surgical treatment of pulmonary metastases. Chest. 1997;112:287S-290S. [DOI] [PubMed] [Google Scholar]
- 33.Ward WG, Corey RM. The Burden of Musculoskeletal Diseases in the United States. 2019. Available at: https://www.boneandjointburden.org/2013-report/tumors/ixa Accessed September 4.