Outcome of Patients with Soft‐Tissue Sarcomas: An Age‐Specific Conditional Survival Analysis

Prognostic outcomes of patients with soft‐tissue sarcomas are usually established at the time of the patient's initial disease presentation, based on baseline patient and tumor characteristics. This study focused on conditional survival and prognostic factors in patients according to age at diagnosis.

Abstract

Background.

Soft‐tissue sarcomas (STSs) are a group of rare cancers that can occur at any age. Prognostic outcomes of patients with STS are usually established at the time of the patient's initial disease presentation. Conditional survival affords a dynamic prediction of prognosis for patients surviving a given period after diagnosis. Estimates of conditional survival can provide crucial prognostic information for patients and caregivers, guide subsequent cancer follow‐up schedules, and impact decisions regarding management. This study aims to estimate conditional survival and prognostic factors in patients with STS according to age at diagnosis (≤75 years and ≥75 years).

Subjects, Materials, and Methods.

A total of 6,043 patients with nonmetastatic STS at first diagnosis who underwent complete surgical resection (R0 or R1) were assessed. Cox proportional hazards regression was used to establish prognostic factors of conditional metastasis‐free survival and overall survival at 1, 2, and 5 years after diagnosis.

Results.

Elderly patients have more adverse prognostic features at presentation and tend to receive less aggressive treatment than do younger patients. However, at baseline as well as at each conditional survival time point, the 5‐year estimated probability of metastatic relapse decreases in both young and elderly patients and is almost identical in both groups at 2 years and 5 years after initial diagnosis. Prognostic factors for metastatic relapse and death change as patient survival time increases in both young and elderly patients. Grade, the strongest prognostic factor for metastatic relapse and death at baseline, is no longer predictive of metastatic relapse in patients surviving 5 years after initial diagnosis. Leiomyosarcoma is the histological subtype associated with the highest risk of metastatic relapse and death in young patients surviving 5 years after initial diagnosis. The positive impact on the outcome of peri‐operative treatments tends to decrease and disappears in patients surviving 5 years after initial diagnosis.

Conclusion.

Conditional survival estimates show clinically relevant variations according to time since first diagnosis in both young and elderly patients with STS. These results can help STS survivors adjust their view of the future and STS care providers plan patient follow‐up.

Implications for Practice.

For patients with sarcoma who are followed up years after being treated for their disease, a common scenario is for the patient and caregivers to ask practitioners what the longer‐term prognosis may be. The question posed to practitioners may be, “Doc, am I now cured? It's been 5 years since we finished treatment.” Survival probability changes for patients who survive a given period of time after diagnosis, and their prognosis is more accurately described using conditional survival. By analyzing more than 6,000 sarcoma patients, an overall improvement was found in the risk of relapse as patients conditionally survive. Prognostic factors for metastatic relapse and death change as patient survival time increases in both young and elderly patients.

Introduction

Soft‐tissue sarcomas (STSs) represent 1% of cancers in adults. Several studies have investigated prognostic factors of patients with STSs [1], [2], [3], [4], [5]. However, all of them have focused on the probability that a patient will still be alive x years after diagnosis by estimating overall survival (OS) at the time of diagnosis based on baseline patient and tumor characteristics. Therefore, such estimated OS does not reflect how prognosis can change over time. For instance, a patient, still alive n years after an initial diagnosis of STS, would be much more interested in having data on the conditional probability of surviving x years further [6]. Moreover, even though more than 20% of STSs are diagnosed in patients ≥75 years of age [7], there is no large study investigating the outcome of this specific population in comparison with that of younger patients.

Subjects, Materials, and Methods

Patients

The French Sarcoma Group (FSG) database comprises 6,043 patients 18 years and older who were surgically treated for primary extremity or trunk wall nonmetastatic STS from 1991 to 2006. Patients with Kaposi sarcoma, dermatofibrosarcoma protuberans, Ewing sarcoma, embryonal or alveolar rhabdomyosarcomas, and retroperitoneal or visceral sarcomas were excluded. For all patients, histological review was performed by the members of the pathological subcommittee of the FSG. The histological diagnosis was established according to the World Health Organization Classification of Tumors [8]. The histological grade was determined after central review as previously described according to the grading system of the Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC) [9].

Statistical Analysis

The statistical analysis of baseline demographics and clinical outcomes is based on all data available up to the cutoff date of October 30, 2008. Descriptive statistics were used to show the distribution of variables in the population. OS was defined as the interval between histological diagnosis and the time of death or last follow‐up. Metastasis‐free survival (MFS) was defined as the interval between histological diagnosis and the time of distant recurrence or the last follow‐up. Follow‐up times were described as medians by using the inverse Kaplan‐Meier estimator [10]. Survival rates were estimated with the Kaplan‐Meier method and were compared using the log‐rank test. Five‐year estimates were calculated using Kaplan‐Meier estimators in strata with corresponding 95% confidence interval (CI). Multivariate analyses were carried out by the Cox regression model based on combination of stepwise regression, Akaike information criteria, and the best subset selection as suggested by Shtatland et al [11]. Univariate and multivariate analyses included the following variables: sex, anatomic site, tumor size, tumor location (superficial or deep), margin status, histological subtype, FNCLCC grade, peri‐operative radiotherapy, and peri‐operative chemotherapy. Age was not included in the model given that this variable was indirectly assessed by analyzing elderly patients separately. Variables were included in the multivariate regression as previously described [11]. Analyses were carried out using SAS 9.4 statistical software (SAS Institute, Cary, NC). All statistical tests were two sided, and p < .05 indicated statistical significance.

Results

Patients

As shown in Table 1, 859 patients (14.21%) were 75 years of age or older. There were significant differences in the distribution of clinical characteristics at diagnosis in young (<75 years) versus elderly (≥75 years) patients (p < .001). Sarcomas with complex genomic profiles (leiomyosarcomas [LMSs], undifferentiated pleomorphic sarcomas [UPSs]) were diagnosed more often in elderly patients, whereas translocation‐related sarcomas were almost exclusively diagnosed in younger patients. Elderly patients also presented with large (>10 cm) tumors and high‐grade disease more often.

Table 1. Patient characteristics (n = 6,043 patients).

Abbreviations: MPNST, malignant peripheral nerve sheath tumors; UPS, undifferentiated pleomorphic sarcoma.

Treatment patterns also significantly differed between elderly and younger patients (Table 2). Patients ≥75 years of age were more likely to receive R1 surgery (31.9 vs. 19.08, p < .0001). Moreover, peri‐operative radiotherapy and peri‐operative chemotherapy use were significantly more prevalent in patients <75 years of age than in elderly patients: 59.5% versus 54.4%, p = .007; 35.9% versus 7.7%, p < .0001 (Table 2).

Table 2. Patterns of treatment according to age (n = 6,043).

Prognostic Factors at Diagnosis in Elderly and Younger Patients

Multivariate analysis revealed that gender, histological subtype, tumor size, tumor site, grade, margin status, peri‐operative radiotherapy, and peri‐operative chemotherapy were significantly associated with metastasis‐free survival in patients <75 years of age (supplemental online Table 1). In elderly patients, only grade, tumor site, and tumor size remained independently associated with MFS (supplemental online Table 1).

Gender, histological subtype, tumor size, tumor site, tumor depth, tumor grade, margin status, peri‐operative radiotherapy, and peri‐operative chemotherapy were significantly associated with overall survival in patients <75 years of age (supplemental online Table 5). In elderly patients, all these prognostic factors, except tumor depth and peri‐operative chemotherapy, were statistically significant independent prognostic factors for OS (supplemental online Table 5).

Conditional Metastasis‐Free Survival

At baseline as well as at each conditional survival time point, the 5‐year estimated probability of metastatic relapse decreases in both young and elderly patients and is almost identical in both groups at 2 years (16.6%, 95% CI [15.2–18.1] vs. 18.8%, 95% CI [14.5–24.2]) and 5 years (9.0%, 95% CI [7.6–10.6] vs. 10.1%, 95% CI [5.4–18.5]) after initial diagnosis (Fig. 1A). Multivariate analyses showed that prognostic factors for metastatic relapse change as patient survival time increases in both young and elderly patients (supplemental online Tables 2–4). In patients <75 years of age, at 1 year after diagnosis, gender, histological subtype, tumor grade, tumor depth, margin status, and tumor size were important predictors of metastasis‐free survival. After surviving 2 years, MFS was no longer different in males and females. After 5 years, only histological subtype was significantly associated with MFS with patients, with leiomyosarcoma and “other histology” having the highest probabilities of metastatic relapse (supplemental online Tables 2–4; supplemental online Fig. 1). Grade, the most important predictor of metastatic relapse at diagnosis, no longer had prognostic value in patients surviving 5 years after initial diagnosis (supplemental online Tables 2–4; supplemental online Fig. 1). At baseline, a patient <75 years of age with grade 3 STS had an estimated 41.9% (95% CI [39.5–44.0]) probability of metastatic relapse in 5 years that reduced to 20.2% (95% CI [17.8–22.8]) after surviving 2 years and 8.9% (95% CI [6.7–11.8]) after surviving 5 years.

Figure 1.

Conditional survival of soft‐tissue sarcomas patients. Five‐year conditional metastases‐free survival (A) and overall survival (B) as a function of prediction times, by age category.

In patients ≥75 years of age, at 1 year after diagnosis, histological subtype, tumor grade, tumor site, and tumor size were important predictors of metastasis‐free survival. After surviving 2 years and 5 years, no patient or tumor characteristics were predictive of MFS (supplemental online Tables 2–4; supplemental online Fig. 1).

Conditional Overall Survival

At baseline as well as at each 2‐year and 5‐year conditional survival time point, the 5‐year estimated probability of death decreased slightly in young patients (23.2%, 95% CI [21.9–24.5], 21.2%, 95% CI [19.8–22.8], and 15.4%, 95% CI [13.8–17.2]) and remained almost stable in patients ≥75 years of age (45.4%, 95% CI [41.2–49.8], 45.0%, 95% CI [39.2–51.3], and 47.6%, 95% CI [38.3–57.9]). Multivariate analyses showed that prognostic factors for metastatic relapse change as patient survival time increases in both young and elderly patients (supplemental online Tables 6–8; supplemental online Fig. 2).

In patients <75 years of age, at 1 year after diagnosis, gender, histological subtype, tumor depth, tumor grade, tumor site, margin status, tumor size, peri‐operative radiotherapy, and peri‐operative chemotherapy were important predictors of overall survival. After surviving 2 years, OS was no longer significantly different in patients having received peri‐operative radiotherapy or not. After 5 years, only histological subtype, grade, and tumor size were significantly associated with OS, and patients with leiomyosarcoma, grade 3, and tumor size >10 cm had the highest probability of death (supplemental online Tables 6–8; Fig. 2).

In patients ≥75 years of age, at 1 year after diagnosis, histological subtype, tumor site, margin status, and tumor size were important predictors of overall survival. After surviving 2 years, only tumor site and tumor size were predictive of OS. After surviving 5 years, no patient or tumor characteristics were predictive of OS (supplemental online Tables 6, 7; supplemental online Fig. 2).

Discussion

The results of this study provide information to patients with STS and care providers. Fear of cancer recurrence is a well‐documented distressful concern for survivors [12], [13], [14], [15]. A recent study focusing on patients with STS demonstrated that up to 25% of them had anxiety or depression symptoms [16]. Conditional survival estimates may therefore represent a very useful tool to alleviate the anxiety of patients and of their caregivers. Indeed, studies have suggested a mutual influence between survivors’ and caregivers’ fears and, frequently, a higher fear of recurrence in caregivers than in survivors [17]. Finally, conditional estimates may also have an impact on the professional and financial decisions of survivors and their families and on their capacities to set up new projects.

One important finding of our study is that the impact of prognostic factors varies over time. It is interesting to note that, for younger patients surviving 5 years, exposure to peri‐operative chemotherapy or radiotherapy and having an R1 surgical margin status did not appear to individually increase the risk of death. Of note, we observed that the use of peri‐operative chemotherapy is associated with time‐varying clinical effects and loses its statistically significant beneficial effect in patients having survived 5 years after the initial diagnosis. This result was somewhat expected because 70% of the metastatic events in patients occurred within 2 years after the initial diagnosis. Interestingly, we found that histological subtype remained a significant predictor of metastatic relapse and death at each time point, with patients diagnosed with LMS having the worst outcomes. This result agrees with a previous study showing a substantial incidence of late metastatic recurrence (up to 9% of patients) and late disease‐specific mortality in patients with LMS (up to 6% of patients dying more than 8 years after initial diagnosis) [18]. These data suggest that long‐term follow‐up of patients with LMS is recommended whereas such an approach may be useless in other histological subtypes such as UPS. It is, however, still important to note that no studies to date have convincingly shown that an earlier diagnosis of metastatic relapse of STS can lead to better overall survival. Therefore, the modality and frequency of surveillance of patients with STS has yet to be defined.

In agreement with several other studies, we found worse overall survival outcomes among elderly patients with STS [19], [20], [21], [22]. Differences in presentation and management have been reported as potential explanations for this reduced survival [20], [21], [23], [24]. Indeed, as reported in other series, we found that patients ≥75 years of age had more frequent pleomorphic undifferentiated sarcomas as well as larger and higher‐grade tumors [19], [20], [21]. In our series, almost 31.9% of patients ≥75 years of age had R1 surgery versus 19% in younger patients. This probably reflects the reluctance of most surgeons to perform potentially morbid surgeries on elderly patients, leading to suboptimal surgical treatment [20], [25]. However, our results also demonstrate the importance of considering conditional survival rates in elderly patients because those who continue to survive after their initial diagnosis actually have rates very similar to the 5‐year probability of metastasis‐free survival of younger patients, increasing over time and exceeding 90%. This may have important implications for decision making about treatment options for older patients with STS, supporting a role for optimal loco‐regional treatment including the potential use of peri‐operative therapies. Of note, we were not able to identify any prognostic factor on multivariate analysis in elderly patients surviving at 2 years and 5 years after initial diagnosis because of the limited sample size and number of events.

Conclusion

A limitation of the current study is the lack of details regarding the ultimate cause of death. However, we believe this very large database study provides a strong benchmark of prognosis, which we expect will improve the management of patients with STS, particularly elderly ones. The magnitude of observed improvement in the risk of distant relapse and death as patients conditionally survive their STS diagnosis is important information for patients and their treating clinicians and is indeed good news for surviving patients.

See http://www.TheOncologist.com for supplemental material available online.

Author Contributions

Conception/design: Antoine Italiano

Provision of study material or patients: Kevin Bourcier, Derek Dinart, Axel Le Cesne, Charles Honoré, Pierre Meeus, Jean‐Yves Blay, Audrey Michot, François Le Loarer, Antoine Italiano

Collection and/or assembly of data: Kevin Bourcier, Derek Dinart, Axel Le Cesne, Charles Honoré, Pierre Meeus, Jean‐Yves Blay, Audrey Michot, François Le Loarer, Antoine Italiano

Data analysis and interpretation: Kevin Bourcier, Derek Dinart, Axel Le Cesne, Charles Honoré, Pierre Meeus, Jean‐Yves Blay, Audrey Michot, François Le Loarer, Antoine Italiano

Manuscript writing: Kevin Bourcier, Derek Dinart, Axel Le Cesne, Charles Honoré, Pierre Meeus, Jean‐Yves Blay, Audrey Michot, François Le Loarer, Antoine Italiano

Final approval of manuscript: Kevin Bourcier, Derek Dinart, Axel Le Cesne, Charles Honoré, Pierre Meeus, Jean‐Yves Blay, Audrey Michot, François Le Loarer, Antoine Italiano

Disclosures

Axel Le Cesne: Eli Lilly and Company, Pharmamar, Novartis (C/A). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board