Fax: (901) 521-9005
Telangiectatic osteosarcoma: The St. Jude Children's Research Hospital's experience
Article first published online: 9 MAR 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 8, pages 1627–1637, 15 April 2007
How to Cite
Weiss, A., Khoury, J. D., Hoffer, F. A., Wu, J., Billups, C. A., Heck, R. K., Quintana, J., Poe, D., Rao, B. N. and Daw, N. C. (2007), Telangiectatic osteosarcoma: The St. Jude Children's Research Hospital's experience. Cancer, 109: 1627–1637. doi: 10.1002/cncr.22574
- Issue published online: 4 APR 2007
- Article first published online: 9 MAR 2007
- Manuscript Accepted: 3 JAN 2007
- Manuscript Revised: 8 DEC 2006
- Manuscript Received: 15 SEP 2006
- Cancer Center Support (CORE). Grant Number: CA21765
- National Cancer Institute. Grant Number: CA23099
- American Lebanese Syrian Associated Charities
- pathologic fracture;
- prognostic factors;
- telangiectatic osteosarcoma
Telangiectatic osteosarcoma (TOS) is a rare subtype of osteosarcoma (OS). The authors reviewed their experience with TOS to characterize its histologic, radiologic, and clinical features.
The authors reviewed records, pathology material, and imaging studies from all patients with TOS who were treated between 1978 and 2005 and compared their outcomes with the outcomes of patients with all other subtypes of high-grade osteosarcoma (OS).
Among 323 patients with OS, 22 patients (6.8%) had TOS. Two additional patients who were treated in Chile on a recent OS trial were included. The median age at diagnosis of the 24 patients was 15.7 years. Four patients (17%) had metastatic disease, and 9 of 21 patients (43%) had pathologic fractures. Only 5 patients (who were treated after 1994) underwent limb-salvage surgery. Estimatesof 5-year event-free survival (58.3% ± 11.9%) and overall survival (66.8% ± 11.6%) were similar to those for patients with other OS subtypes (P ≥ .85). The absence of local disease progression and chemotherapy with ≥3 agents that were active against OS were correlated with improved outcome (P ≤ .005). The presence of a pathologic fracture was not associated with surgery type or patient outcome.
TOS was associated with a high rate of pathologic fracture. With multimodality therapy, the outcome of patients with TOS was similar to that of patients with other high-grade OS subtypes. The absence of local disease progression and chemotherapy with ≥3 active agents were associated with a favorable outcome. Cancer 2007. © 2007 American Cancer Society.
Osteosarcoma (OS) is the most common primary malignant bone tumor of childhood. Several distinct clinicopathologic variants of OS have been recognized. Telangiectatic OS (TOS) is a rare subtype that represents from 2% to 12% of all cases of OS.1–5 TOS is characterized by multiple, aneurysmally dilated, blood-filled cavities with high-grade, sarcomatous cells in the peripheral rim and septae.6 Identification of the unique histologic and radiographic features of TOS is crucial for correct diagnosis. Despite its obvious differences from the other OS subtypes, TOS has been confused with benign entities, such as giant cell tumor, hemangioendothelioma, and aneurysmal bone cyst.7 Many case reports have described misdiagnosed TOS as an aneurysmal bone cyst, and this misdiagnosis often results in a delay in diagnosis and potentially affects the outcome.8–11
Although TOS displays radiographic and histopathologic findings that are distinct from those of other OS variants, the clinical significance of the TOS variant has been unclear; and some have questioned whether its significance is purely academic, because it appears to have no prognostic clinical implications.1 The prognosis for patients with TOS has been debated in the literature for many decades. In the 1970s, an often referenced single-institutional study reported a dismal survival rate for patients with this subtype of OS, and the conclusion was that the prognosis for patients with TOS was far worse than that for patients with conventional OS.2 Since then, other reports, including an updated one from the same institution, have stated that the overall survival estimates are similar for patients with TOS and for patients with other OS subtypes.1, 7, 12 In addition, several studies have demonstrated that TOS, compared with conventional OS, is uniquely sensitive to chemotherapy.3, 4, 13–16 Reports of improved histologic response and disease-free survival estimates for patients with TOS indicate that there may be clinical relevance to this OS variant after all.
Potential predictive risk factors, such as age, sex, tumor size, and location, have not demonstrated consistent clinical significance in patients with nonmetastatic TOS.3, 4, 7, 17 It has been reported that, among patients with OS in general, those whosustain a pathologic fracture have a greater risk of local recurrence and a lower rate of survival than patients who have not sustained a pathologic fracture.18 Although patients with TOS are at a higher risk of pathologic fractures than patients with conventional OS,2, 4, 7 the impact of pathologic fractures on the outcome of patients with TOS is unknown.
In the current study, we reviewed the experience of our institution (St. Jude Children's Research Hospital) with TOS over a period of approximately 3 decades to characterize the histologic, radiographic, and clinical features of this subtype of OS. In addition, we analyzed risk factors that may be predictive of patient outcome, including pathologic fracture, and we compared the outcome of patients with TOS to that of patients with other high-grade OS subtypes who were treated during the same period.
MATERIALS AND METHODS
We conducted a search of our solid tumor database to identify all patients with TOS who were treated at St. Jude Children's Research Hospital between January 1978 and December 2005. In addition, we included all patients with TOS who were treated on our most recent multiinstitutional OS protocol (OS99) at collaborating sites. Treatment on this protocol consisted of aggressive surgery and neoadjuvant and adjuvant chemotherapy comprised of carboplatin, ifosfamide, and doxorubicin for localized OS, with the addition of irinotecan for metastatic or unresectable OS.19 Information regarding the clinical characteristics, treatment, and outcome of TOS patients was collected by a review of the medical records. All available pathology materials and imaging studies for these patients were reviewed retrospectively. Imaging studies that were used to evaluate primary tumors included a combination of plain radiography, bone scintigraphy, computerized tomography (CT), and magnetic resonance (MR) imaging.
Patients who were included in this study fulfilled the histologic and radiographic diagnostic criteria of TOS as defined in the World Health Organization Classification.20 These criteria are summarized as follows: 1) predominantly lytic bone mass with minimal sclerosis on radiographs, 2) grossly cystic medullary mass with no or minimal solid or sclerotic component, and 3) histologic features consisting of bone-forming tumor with notable blood-filled spaces separated by septae lined by and/or containing malignant tumor cells with prominent nuclear atypia and limited osteoid deposition (Fig. 1).
Intracompartmental tumors were defined as tumors that were confined primarily within the cortex or that expanded the cortex in a way similar to that of an aneurysmal bone cyst on imaging studies, whereas extracompartmental tumors were defined as tumors that extended outside the cortex with a soft tissue mass. The presence of a pathologic fracture was determined on the basis of radiologic findings (plain radiography, CT, or MR imaging) and/or histologic findings (Fig. 2, left). Intratumoral bleeding was suggested by the presence of fluid/fluid levels on MR images (Fig. 2, right), and 1 fluid/fluid level was the threshold used.
In terms of treatment response, patients with no significant radiologic change in tumor were considered to have stable disease. Patients whose tumor growth was measurable or patients in whom new lesions developed were considered to have progressive disease. The 4-grade system of Huvos and colleagues21, 22 was used for histologic assessment of the response of the primary tumor to chemotherapy at the time of definitive surgery. Surgical remission was achieved when all detectable tumor foci were removed and no residual disease was evident after surgery. This retrospective review was approved by the Institutional Review Boards of both institutions that contributed the patients who were included in the study.
Survival was defined as the time interval from the date of diagnosis to the date of last follow-up or the date of death from any cause. Event-free survival (EFS) was defined as the time interval from the date of diagnosis to the date of the first event or the date of last follow-up for patients who had no events. An event included recurrent or progressive disease, a second malignancy, or death from any cause. Survival and EFS distributions were estimated by using the method of Kaplan and Meier23; standard errors were calculated by using the method of Peto et al.24 Exact log-rank tests were used to compare survival and EFS distributions according to clinical characteristics.
Of all 323 patients who were treated for OS at our institution between January 1978 and December 2005, 22 patients (6.8%) were identified with TOS. Two more patients who had TOS and were treated on our OS99 protocol in Santiago, Chile, were included in the current analysis. The median age at diagnosis of all 24 patients was 15.7 years (range, 3.2–23 years). Sixteen patients (67%) were male, and 13 patients (54%) were white. The demographic and clinical characteristics of the 24 patients are summarized in Table 1.
|Characteristic||No. of patients (%)|
|Age at diagnosis, y|
|Primary tumor site|
|Tumor compartmental status|
|Not evaluable||3 (12)|
|Not evaluable||3 (12)|
Most patients had localized disease (n = 20 patients; 83%). Four patients (17%) had metastatic disease at diagnosis: in 3 patients, the disease had metastasized to the lung; and, in 1 patient, the disease had metastasized to the skull. The most common primary site was the femur (n = 12 patients; 50%). Available for retrospective review were diagnostic imaging studies from 21 patients, including 1 patient who had only images from bone scintigraphy available. Seventeen of those 21 patients (81%) had extracompartmental tumors at diagnosis, and 4 patients had intracompartmental tumors. Records for 21 patients were evaluable for analysis of pathologic fracture, including 20 patients for whom imaging studies were available for review (excluding the patient with bone scintigraphs only) and 1 patient for whom no imaging studies were available but for whom the report of the plain radiograph and the histologic review of the resected specimen showed a pathologic fracture. Nine of the 21 evaluable patients (43%) had a pathologic fracture. Fractures were apparent on plain radiography in all 9 patients and in 3 of 5 patients who had CT scans, in 6 of 7 patients who had MR images, and in 4 of 9 patients by histologic review. Pathologic fractures affected the femur (n = 5 patients), the tibia (n = 2 patients), or the humerus (n = 2 patients). Pathologic fractures occurred prior to the biopsy date in 4 patients and within 2 months after the biopsy date in the remaining 5 patients. Two patients had skip metastases noted on radiologic review but not confirmed by histologic review. Intratumoral bleeding, as suggested by the presence of fluid/fluid levels, was observed on MR images from 11 of 14 patients (79%) for whom such images were available; this finding was present in patients with or without fracture.
Patients were treated with different chemotherapy regimens depending on the treatment era (Table 2). All but 2 patients underwent surgery for tumor removal. Two patients did not undergo surgery because of unresectable primary tumors that arose in the sacrum and pubis. The latter patient also had disease that had metastasized to the skull. Most patients who underwent surgery for tumor removal underwent amputation (n = 17 patients; 77%), and the median time from the date of diagnosis to the date of surgery was 7.5 weeks (range, 0–20 weeks). Only 5 patients underwent limb-salvage surgery, and all 5 were treated between 1995 and 2005 (in total, 9 patients were treated during this period). Limb-salvage surgery was performed first at our institution in December 1980 and initially was restricted to selected patients with localized disease. One of our patients underwent amputation before presentation to our institution. Among the remaining 16 patients, amputation was performed because of treatment in 1980 or earlier (5 patients), presence of metastatic disease (2 patients), local disease progression (3 patients), primary tumor arising in the distal tibia (1 patient), presence of skip metastasis on imaging studies (1 patient), large intratumoral bleeding (1 patient), local tumor extension (1 patient), patient/family preference (1 patient), or undocumented reason (1 patient). Six of 17 patients (35%) who underwent amputation had a pathologic fracture, and 3 of 5 patients (60%) who underwent limb-salvage surgery had a pathologic fracture. Only 1 patient experienced significant intraoperative blood loss; this complication occurred at the time of above-the-knee amputation to remove distal femoral OS. Twenty of 24 patients achieved surgical remission; and, of the remaining 4 patients, 2 patients had unresectable primary tumors, 1 patient had gross pulmonary disease, and 1 patient had both gross pulmonary disease and gross residual disease at the primary site.
|Characteristic||No. of patients (%)|
|CDDP, DOX||1 (4)|
|CYC, DOX, MTX||5 (21)|
|ACD, BLEO, CDDP, CYC, DOX, MTX||2 (8)|
|CDDP, DOX, IFOS, MTX||5 (21)|
|CBP, CDDP, DOX, IFOS||4 (17)|
|CBP, DOX, IFOS||7 (29)|
|Type of definitive surgery|
|Limb salvage||5 (21)|
|No surgery||2 (8)|
|No. of active chemotherapeutic agents|
|Clinical response to neoadjuvant chemotherapy, n = 14|
|Stable disease/response||8 (57)|
|Local progressive disease||5 (36)|
|Not evaluable||1 (7)|
|Histologic response to neoadjuvant chemotherapy, n = 14|
|Rosen grade I/II||8 (57)|
|Rosen grade III/IV||4 (29)|
|Unresectable tumor||2 (14)|
Thirteen of 24 patients (54%) received ≥3 chemotherapeutic agents with activity against OS (cisplatin, doxorubicin, methotrexate, ifosfamide, or carboplatin in combination with ifosfamide) as part of their primary chemotherapy (neoadjuvant and adjuvant). Fourteen patients received neoadjuvant chemotherapy that consisted of single-agent ifosfamide; combined ifosfamide, high-dose methotrexate, and doxorubicin; combined carboplatin and ifosfamide; or combined carboplatin, ifosfamide, and doxorubicin. Six of those 14 patients received ≥3 neoadjuvant agents. Of those 14 patients, 8 patients (57%) had clinically stable disease or had a response to chemotherapy, 5 patients (36%) had local progressive disease, and 1 patient who had very extensive intratumoral bleeding during neoadjuvant chemotherapy was not considered evaluable for clinical response. Of the 5 patients who had local disease progression (3 of whom had a pathologic fracture), 1 patient experienced disease progression after treatment with single-agent ifosfamide, and 4 patients experienced disease progression after treatment with carboplatin in combination with ifosfamide. Two of the 14 patients did not undergo surgery for tumor removal; hence, for these patients, there were no data on the histologic response to neoadjuvant chemotherapy. In 4 of the 12 remaining patients (33%) with data on histologic response, >90% tumor necrosis (Rosen grade III/IV) was observed. Three of those 4 patients received combined carboplatin, ifosfamide, and doxorubicin, and 1 patient received carboplatin in combination with ifosfamide.
Fourteen patients (58%) survived with a median follow-up of 7.1 years (range, 1–23.7 years) from the time of diagnosis. At the time of the current analyses, 10 patients (42%) had died, including 9 patients who died of disease and 1 patient who died of doxorubicin-induced cardiomyopathy. The median time from the date of diagnosis to death for these 10 patients was 1.7 years (range, 0.9–20.5 years).
First events included recurrent or progressive disease (10 patients), second malignancy (1 patient), and toxic death (1 patient). For the 10 patients who developed recurrent or progressive disease, the median time to recurrence or progression was 3 months (range, 1–10.4 months) after diagnosis. Eight of the 10 patients who experienced disease recurrence or progression died, whereas 2 patients remained alive without evidence of disease. The latter 2 patients experienced local disease progression after treatment with carboplatin in combination with ifosfamide and survived after surgical removal of the tumor and additional chemotherapy (which included cisplatin, doxorubicin, and high-dose methotrexate) for 1.1 years and 2.7 years after diagnosis. The patient who developed a second malignancy was diagnosed with colon cancer 7.2 years after the diagnosis of OS. Six years later, that patient developed a third malignancy (OS) in the radiation field that was used during therapy for colon cancer and died of progressive disease. With respect to the 4 patients who had metastasis at diagnosis, 1 patient with pulmonary disease experienced resolution of the pulmonary nodules after 2 courses of ifosfamide and remained alive without evidence of disease 15.4 years after diagnosis, 1 patient with pulmonary disease died of disease within 1 year of diagnosis, 1 patient with a single pulmonary metastasis that was resected was cured of TOS but died of a third malignancy that occurred 20.5 years after the initial diagnosis, and 1 patient with skull metastasis died of disease 6 years after diagnosis. Of the 5 patients who had local disease progression after neoadjuvant chemotherapy, 2 patients remained alive without evidence of disease after surgical removal of the primary tumor and adjuvant chemotherapy, and 3 patients died of progressive disease.
The 5-year estimates of EFS and survival for the 24 patients were 58.3 ± 11.9% and 66.8 ± 11.6%, respectively (Fig. 3). The median survival and EFS estimate was 15.3 years.
Potential Predictors of Survival and EFS
Exact log-rank tests were used to examine potential prognostic factors for survival and EFS. Because of the small samples, each factor was examined in a univariate manner only. The results are summarized in Table 3.
|No. of patients||5-Year estimate (1 SE), %||Exact P||5-Year estimate (1 SE), %||Exact P|
|<16||14||71.4 (14.4)||74.3 (14.2)|
|≥16||10||40 (15.5)||.37||60 (17)||.54|
|Girl||8||75.0 (15.3)||75 (15.3)|
|Boy||16||50.0 (15.8)||.23||60.3 (15.5)||.55|
|Black||6||66.7 (22.2)||80 (17.9)|
|Caucasian/Hispanic||18||55.6 (13.1)||.53||61.5 (13.5)||.41|
|Localized||20||60.0 (13.4)||64.1 (13.6)|
|Metastatic||4||50.0 (20.4)||.48||75 (18.8)||.55|
|Pathologic fracture, n = 21|
|No||12||66.7 (14.5)||82.5 (12.2)|
|Yes||9||55.6 (21.4)||.74||59.3 (21.8)||.71|
|Femur||12||58.3 (16.8)||64.8 (17.2)|
|Other bones||12||58.3 (15.4)||.95||70 (14.5)||.85|
|Type of surgery, n = 22|
|Limb salvage||5||100 (0)||100 (0)|
|Amputation||17||47.1 (12.1)||.043||55.1 (13.1)||.22|
|No||10||50 (14.4)||50 (14.4)|
|Yes||14||64.3 (17.2)||.91||79.5 (14.7)||.50|
|Clinical response to neoadjuvant chemotherapy|
|Stable disease/response||8||100 (0)||100 (0)|
|Local progressive disease||5||0 (0)||<.001||37.5 (21)||.005|
|Histologic response to neoadjuvant chemotherapy|
|Rosen grade I/II||8||64.3 (22.2)||64.3 (22.2)|
|Rosen grade III/IV||4||100 (0)||.081||100 (0)||.29|
|No. of active agents|
|<3||11||9.1 (6.1)||27.3 (13.4)|
|≥3||13||100 (0)||<.001||100 (0)||.001|
|Tumor compartmental status, n = 21|
|Intracompartmental||4||50 (25)||100 (0)|
|Extracompartmental||17||64.7 (13.6)||.23||65.7 (13.6)||.82|
|No||4||50 (20.4)||75.0 (18.8)|
|Yes||20||60 (13.4)||.99||64.1 (13.6)||.84|
Only 2 factors were significantly predictive of patient outcome. Patients who did not have local disease progression during neoadjuvant chemotherapy had a better overall survival estimate (P = .005) than patients who experienced local disease progression. Patients who received ≥3 chemotherapeutic agents that were active against OS had survival and EFS estimates greater than those of patients who received ≤2 active agents (Fig. 4) (P ≤ .001). Although the degree of tumor necrosis after neoadjuvant chemotherapy in our cohort did not correlate with survival (P = .29), there was a trend toward statistical significance with regard to EFS (P = .081).
Of the 9 patients who had pathologic fractures, 6 patients underwent amputation, and 3 patients underwent limb-salvage surgery. Of the 12 patients who were without fractures, 8 patients underwent amputation, 2 patients underwent limb-salvage surgery, and 2 patients did not undergo surgery. The presence of a pathologic fracture was not associated with the type of surgery performed (P = .63) or patient outcome (P > .70). Patients who underwent amputation had a worse EFS estimate than patients who underwent limb-salvage surgery (P = .043). A repeat analysis that was restricted only to the subset of patients who had tumors of the extremity did not indicate an association between the presence of a pathologic fracture and patient outcome (P > .73).
Comparison of the Outcome of TOS and of Other Subtypes of High-Grade OS
Overall survival and EFS estimates for our 24 patients with TOS were compared with those for patients with other subtypes of high-grade OS who were treated at our institution and on our OS99 protocol at the Chilean institution during the same period. There was no evidence of a significant difference in EFS (P = .85) or survival (P = 1.0) estimates between the groups. EFS estimates at 5 years (±standard error) were 58.3 ± 11.9% for patients with TOS and 44.4 ± 3.2% for patients with all other OS subtypes (Fig. 5, top). EFS estimates at 10 years were 51.9 ± 12.7% for patients with TOS and 42.1 ± 3.8% for patients with all other OS subtypes. Survival estimates at 5 years were 66.8 ± 11.6% for patients with TOS and 57.7 ± 3.1% for patients with all other OS subtypes (Fig. 5, bottom). Survival estimates at 10 years were 60.1 ± 12.7% for patients with TOS and 53.5 ± 3.6% for patients with all other OS subtypes.
TOS is a rare variant of OS with well-defined histologic, radiographic, and clinical characteristics that clearly differentiate it from other OS subtypes. The incidence of TOS among patients with OS in our study was approximately 7%, a finding consistent with the reports of others.1–4 Although it was believed previously that patients with TOS fared much worse than patients with conventional OS,2 this belief no longer appears to be true. We observed 5-year EFS and overall survival estimates of 58 ± 12% and 67 ± 12%, respectively. These estimates compare favorably with the 5-year EFS and overall survival estimates of patients with other subtypes of OS who were treated at our institution during the same period (44 ± 3% and 58 ± 3%, respectively). EFS and overall survival estimates for patients with TOS that are at least as good as those for patients with conventional OS have been reported in other large institutional studies (Table 4).1, 7 In addition, in a large study of 570 patients with osteosarcoma, survival did not differ significantly according to histologic subtype. In that study, patients with TOS were combined with patients who had other rare histologic subtypes.25
|Study||Study period||No. of TOS patients (%)||Disease stage||Rate of pathologic fracture, %||DFS/EFS, %||Overall Survival, %|
|TOS||All other OS subtypes||TOS||All other OS subtypes|
|Farr et al., 19741||1929–1971||28 (2)||Localized and metastatic||—||—||—||18||18|
|Matsuno et al., 19762||1929–1974||25 (2.5)||Localized and metastatic||24||—||—||8|
|Huvos et al., 19827||1921–1979||124 (11)||Localized and metastatic||29||—||—||27 (5-y)||25 (5-y)|
|Rosen et al., 19863||1973–1980||25 (12)||Localized||32||68 (5-y DFS)||—||—|
|Vanel et al., 19879||1969–1983||14 (4.5)||Localized and metastatic||29||—||—||36|
|Mervak et al., 199112||1975–1983||17 (4)||Localized and metastatic||—||—||—||47|
|Bacci et al., 199416||1983–1990||28 (10)||Localized||14||82 (5-y DFS)||61 (5-y DFS)||—||—|
|Bacci et al., 20014||1990–1994||24 (8)||Localized||17||83 (5-y DFS)||55 (5-y DFS)||87 (5-y)||69 (5-y)|
|Current study||1978–2005||24 (6.8)||Localized and metastatic||43||58 (5-y EFS)||44 (5-y EFS)||67 (5-y)||58 (5-y)|
A combined-modality approach, including the use of strict pathologic and radiologic criteria, is crucial in diagnosing TOS. There are many reports in the literature about patients with TOS that was diagnosed incorrectly; and it is important to note that delaying the correct diagnosis and initiation of proper therapy may have a negative effect on survival.8–11 An aneurysmal bone cyst in particular is almost indistinguishable from TOS on radiographs, but the correct diagnosis can be established by careful microscopic examination.26–28 Accordingly, the radiologic and pathologic inclusion criteria used by us and others to make the diagnosis of TOS seem most appropriate.2, 3, 7, 14
Most of the histologic, radiologic, and clinical features in our TOS cohort were consistent with those described previously by others.1, 3, 6, 12, 13 The rate of metastatic disease in our patients with TOS (17%) was similar to that reported for patients with OS. The 43% rate of pathologic fracture among our patients with TOS was consistent with previous findings of significantly higher rates of this complication in TOS (17–32%) than in conventional OS (6–13%).2–4, 7, 9, 18, 29, 30 Our retrospective review of the diagnostic imaging studies may have led to a higher rate of detection of pathologic fracture in our study than in previous studies. A very high rate of pathologic fracture (61%) also was noted in a recent retrospective review of radiologic features in 36 patients with TOS.6 One possible explanation for the high rate of pathologic fractures in patients with TOS is the extensively lytic and cystic nature of the tumor, which may make the bone more prone to fracture.3, 9, 10
One purpose of our retrospective review was to uncover any potential risk factors that hold predictive clinical significance for this OS subtype. In agreement with results of previous studies, the results of our univariate analyses of survival did not reveal statistically significant differences among most variables analyzed (eg, age, sex, race, and primary tumor site).7, 12, 16 We observed similar 5-year EFS and overall survival rates for patients with localized and metastatic disease. Of the 4 patients who had metastatic disease at diagnosis, 1 patient had complete resolution of bilateral pulmonary metastases after 2 courses of ifosfamide therapy and survived without evidence of disease, 1 patient died of a third malignancy 20.5 years after the initial diagnosis, and 1 patient died of progressive disease 6 years after diagnosis.
It is most noteworthy that the strongest predictor of EFS and overall survival for our TOS cohort was the number of active chemotherapeutic agents. Patients with TOS who received treatment with ≥3 agents that were active against OS had an outcome that was significantly better than patients who received ≤2 active agents. No previous study specifically has compared the number of active agents and its effect on survival in patients with TOS. Bacci et al.13 reported that patients with OS in general had a better histologic response after they received ≥3 active agents rather than ≤2; in addition, those investigators reported that the use of >3 agents did not further improve the histologic response.
In 1982, Huvos et al.7 first made the observation that the telangiectatic subtype of OS appeared to be particularly sensitive to chemotherapy. When patients with TOS received 3 active chemotherapeutic agents, they had a better histologic response (>90% tumor necrosis) than patients with all other histologic variants (82% vs 49%, respectively). The addition of active, multiagent chemotherapy negated the survival differential once believed to exist between the 2 groups.2 Those authors hypothesized that the increased vascularity of telangiectatic tumors may increase drug delivery to the tumor site; therefore, TOS may be more chemoresponsive than other OS subtypes. Since then, other groups have confirmed this same favorable histologic response in TOS.3, 14, 15
It has been reported that a pathologic fracture in patients with OS in general is a risk factor for local recurrence and decreased overall survival.18, 30 In addition, tumor response to chemotherapy is predictive of fracture union, improved overall survival, and local disease control.18, 29 Despite the occurrence of pathologic fractures in 9 of our patients, no patient experienced a local recurrence of disease, and the presence of a pathologic fracture had no statistically significant impact on EFS or overall survival of patients with TOS. The negative prognostic significance associated with an increased incidence of pathologic fractures in patients with TOS may be compensated by the extreme chemosensitivity of the telangiectatic tumors.3
Historically, the surgical treatment of patients with OS and pathologic fractures was immediate amputation.29 With the more recent advent of preoperative chemotherapy, limb-salvage surgery has become another option for these patients. Such an approach, although it is associated with a small increased risk of local recurrence, has not been found to compromise the overall survival of patients.14, 18, 30, 31 The presence of a pathologic fracture in our study did not predispose patients to local recurrence, and none of our patients developed a local recurrence. However, it is difficult to determine whether the presence of a pathologic fracture contributed to local disease progression. Three of our 5 patients with local disease progression had a pathologic fracture. The presence of a pathologic fracture did not influence the type of surgery performed, and the outcome of our patients who underwent limb-salvage surgery was at least as good as that of patients who underwent amputation. This result demonstrates that the presence of a pathologic fracture does not negate the feasibility of limb-salvage surgery and that this type of surgery, when performed in carefully selected patients, does not appear to compromise outcome. It is important to realize that the excellent outcome of patients who underwent limb-salvage surgery in our study may reflect an effect of tumor size and/or advances in treatment, because all patients who underwent limb-salvage surgery were treated after 1994.
The current study was limited by its retrospective nature, the heterogeneity of treatment over the years, and the small number of patients. However, the rarity of TOS renders a prospective clinical trial of this OS subtype unfeasible. Although it was comparable in size to the cohorts in previous large, institutional studies, our cohort was small, and its small size may not have permitted significant power to detect differences among variables. This possibility may explain the few prognostically significant variables that were identified in our study and by others.
In conclusion, TOS is a rare yet distinct subtype of osteosarcoma that is associated with a high rate of pathologic fracture. In our cohort of patients with TOS, the presence of a pathologic fracture was not associated significantly with the type of surgery or patient outcome. With the use of multimodality therapy, the outcome of patients with TOS was similar to that of patients with other subtypes of high-grade OS. In particular, an improved outcome was associated with the absence of local disease progression and the use of ≥3 chemotherapeutic agents with activity against OS. The increased chemosensitivity of TOS may compensate for the potentially negative effects of the increased rate of pathologic fracture. Its extreme vascularity makes it an intriguing target for alternative drug-delivery systems or antiangiogenic agents. Patients with TOS should continue to be treated on conventional OS protocols with the expectation of outcomes that are at least similar if not better than those of patients with other OS subtypes.
We thank Julia Cay Jones, PhD, ELS, for editing the article.
- 5Neoadjuvant chemotherapy with high-dose ifosfamide, high-dose methotrexate, cisplatin, and doxorubicin for patients with localized osteosarcoma of the extremity: a joint study by the Italian and Scandinavian Sarcoma Groups. J Clin Oncol. 2005; 23: 8845–8852., , , et al.
- 13Neoadjuvant chemotherapy for high-grade central osteosarcoma of the extremity. Histologic response to preoperative chemotherapy correlates with histologic subtype of the tumor. Cancer 2003; 97: 3068–3075., , , et al.
- 20Telangiectatic osteosarcoma. In: FletcherC, UnniK, MertensF, eds. Tumors of Soft Tissue and Bone. Lyon, France: IARC Press; 2000: 271–272., , .
- 25Does the histological subtype of high-grade central osteosarcoma influence the response to treatment with chemotherapy and does it affect overall survival? A study on 570 patients of two consecutive trials of the European Osteosarcoma Intergroup. Eur J Cancer. 2002; 38: 1218–1225., , , , .
- 30The surgical treatment of patients with osteosarcoma who sustain a pathologic fracture. Clin Orthop Relat Res. 1996; 234: 227–232., , , , .