AMSTERDAM – Pathological fractures are prognostic of poor outcomes in adults with osteosarcoma, but not in children with osteosarcoma, investigators have found.
A retrospective review of data on consecutive patients treated over the course of 30 years showed that among patients of all ages, both 5-year and 10-year overall survival (OS) rates were worse among patients who had pathological fractures.
But in an analysis stratified by age, the survival difference attributable to fractures was limited entirely to patients who were 18 or older at the time of an osteosarcoma diagnosis, reported Lisa Kelley, a medical student at the Ludwig-Maximilians Universität in Munich.
“It’s important to understand what are the prognostic factors [in osteosarcoma], and while many of the factors have been thoroughly researched, on pathological fractures the data are still fairly inconclusive due to the results of studies contradicting each other, and also having a relatively small patient number due to the fact that osteosarcoma itself is rare, and the pathological fractures occur only in approximately 10% of cases,” she said at an annual congress sponsored by the European Cancer Organisation.
To get a better handle on possible correlations between pathological fractures and prognosis in patients with central high-grade osteosarcoma, Ms. Kelley and her coauthors collected data on consecutive patients treated for localized or metastatic osteosarcoma of the extremities from 1980 through 2010 at one of the member institutions of the Cooperative Osteosarcoma Study Group (COSS).
They identified 2,847 patients, of whom 2,193 (77%) were 18 or younger at the time of diagnosis. Of the entire cohort, 321 patients (11.3%) had a pathological fracture either at presentation or soon after diagnosis.
Comparing patients with and without pathological fractures, the investigators found that factors significantly associated with fracture risk included tumor location, especially the humerus (P less than .001), tumors occurring proximally and in a diaphysis (P less than .001), telangiectatic subtype (P less than .001), the presence of primary metastases (P = .025). and tumors comprising more than one-third of the affected bone (P less than .001).
There were no significant differences in the cohort as a whole between patients with or without fractures in either age in years, sex, body-mass index, history of pain or swelling symptoms, local surgical remission at the main tumor site, total surgical remission including metastases, response to chemotherapy, use of adjuvant chemotherapy rather than neoadjuvant, or type of surgery.
Among adult patients only, however, factors associated with pathological fractures included age (P less than .001). BMI (P = .021), tumor site (P less than .001), histologic subtype (P less than .001), primary metastases (P = .011), relative tumor size (P = .047), and total surgical remission (P = .015).
Among pediatric patients, factors associated with fracture risk were (P less than .001 for all unless otherwise specified) age, BMI (P = .018), history or symptoms (P = .001). tumor site, localization within bone, histologic subtype, and relative tumor size.
In univariate analysis, 5-year OS rates were 70.6% for patients without fractures compared with 63.0% for those with fractures, and respective 10-year OS rates were 64.9% vs. 58.1% (P = .007 for both comparisons).
Among pediatric patients, the Kaplan-Meier survival curves of patients with and without pathological fractures overlapped. But for adult patients, survival of those with fractures was significantly worse, with 5-year OS of 69.3% with no fractures vs. 45.9% with fractures, and respective 10-year OS rates of 62.1% vs. 36.8% (P less than .001 for both comparisons).
Also in univariate analysis, 5-year and 10-year event-free survival (EFS) rates were significantly lower for patients who experienced pathological fractures.
Finally, in multivariable analysis of overall survival by age group, the investigators found that among adults pathological fracture was associated with a nearly twofold risk for death (hazard ratio [HR] 1.893, P = .013). Other factors were primary metastases (HR 2.486, P = .001), response to chemotherapy (P less than .001) and total surgical remission (P less than .001).
As noted before, pathological fracture among pediatric patients was not associated with worse OS. Factors significantly associated with OS in younger patients were primary metastases (HR 2.187, P less than .001), relative tumor size (HR 1.239, P = .024), response to chemotherapy (HR 2.295, P less than .001), total surgical remission (HR 4.253, P less than .001), and type of surgery (HR 1.282, P = .008).
In contrast, pathological fracture was not associated with EFS among either adult or pediatric patients. Factors significantly associated with EFS in both children and adults were primary metastases and response to chemotherapy. Among pediatric patients only, tumor site and relative tumor size were also associated with EFS.
Ms. Kelley acknowledged that the study was limited by the retrospective design.
She said that exploration of the discrepancies between adult and pediatric patients in regard to the influence of pathological fracture on survival and of the role of pathologic fracture as a negative prognostic factor for OS but not EFS in adults is warranted.
The study was funded by the Wilhelm Sander-Stiftung Foundation for cancer research. Ms. Kelley reported having no conflicts of interest.