| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Department of Surgery, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian, 1, 20133, Milan, Italy
2 Department of Medical Oncology, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian, 1, 20133, Milan, Italy
3 Department of Biostatistics, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian, 1, 20133, Milan, Italy
4 Department of Radiation Oncology, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian, 1, 20133, Milan, Italy
5 Department of Pathology, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian, 1, 20133, Milan, Italy
Correspondence: Address correspondence and reprint requests to: Alessandro Gronchi, MD; E-mail: alessandro.gronchi{at}istitutotumori.mi.it.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: A total of 597 consecutive adult patients with primary extremity STS were treated with conservative surgery at our institution over a 20-year time span. A total of 318 patients were referred after unplanned excisions, and the remaining 279 underwent primary resection at our center. The two groups significantly differed in tumor size and depth. The assessed end points were sarcoma-specific mortality, local recurrence, and distant metastasis. Univariable and multivariable analyses, adjusted for other prognostic factors, were performed in the competing risks framework.
Results: The adjusted 10-year cumulative incidences in re-excised and primarily operated patients were, respectively, 18.7% and 16.4% (P = .535) for local relapse, 17.6% and 20.2% (P = .541) for metastasis, and 20.4% and 22.4% (P = .645) for mortality. Among patients who underwent re-excision, evidence of microscopic residual disease on pathologic examination had a significant prognostic effect on multivariable analysis for distant metastases (P = .002). A trend for survival was detected as well.
Conclusions: At a referral center with a liberal policy of re-excisions in adult primary STS of the extremities, the outcome of patients who underwent re-excision was similar to that of patients who had primary resections. Evidence of microscopic residual disease at re-excision was a marker of clinical aggressiveness.
Key Words: Soft tissue sarcoma • Limbs • Surgery • Re-excision • Prognosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The prognostic effect of re-excision after unplanned surgery has been explored in the literature, with contradictory results in terms of survival.11,23–29 Furthermore, the significance of residual disease in the re-resection specimen has not been definitively established.29 We have therefore retrospectively analyzed a large and well-characterized series of consecutive patients with primary extremity STS who were surgically treated at our institution over a 20-year time span.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
From this consecutive series, we selected for the present analysis 597 patients with primary and localized tumors, excluding dermatofibrosarcoma protuberans and aggressive fibromatosis because of their nearly exclusive local prognosis. Patients who had undergone any demolitive surgical procedure (forequarter and hindquarter procedures, major and minor limb amputations) were also excluded.
Two different groups of patients were identified according to their initial presentation at our institution. The first group (re-excision group) consisted of patients who had undergone unplanned resection of a primary sarcoma at a nonspecialized center in the lack of obvious macroscopic residual disease after the first operation and were subsequently referred to our institution for re-excision. The second group (control group) consisted of patients who underwent primary resection at our institution.
The recommendation for radiotherapy was given by both the operating surgeon and the radiation oncologist when a higher risk of recurrence was thought to exist on clinical grounds. However, no prospectively selected criteria were used for this purpose. External beam radiation was used in all such cases, and doses ranged from 45 to 65 Gy (median, 57 Gy).
Chemotherapy was administered at the discretion of the multidisciplinary STS committee or as part of clinical trials. Anthracycline-based regimens were used—in most cases, they were associated with if-osfamide. In any case, the multidisciplinary approach was consistent either with randomized trial results or current standards of care.
Clinical follow-up of the patients was updated to September 2004, with a median follow-up of 97 months (interquartile range, 51–124 months) for the re-excision group and 102 months (interquartile range, 59–127 months) for the control group. A multiple binary logistic model was used for testing the dishomogeneity between the re-excision and control groups with respect to common characteristics. The response variable was assigned the value of 1 for subjects with re-excised disease and 0 for the others. The effects of the different characteristics were tested by two-sided Wald tests.
The prognostic analyses mainly focused on investigating the effect of re-excision. In addition, the effects of residual disease and the time interval (in days) between the previous unplanned sarcoma operation and subsequent re-excision were separately analyzed in the subset of patients with re-excised disease. The end points were local relapse, distant metastasis, and sarcoma-specific death. The time to the occurrence of each event was computed from the date of operation at our institution to the date when the event was first recorded. The time was censored at the date of last follow-up in event-free subjects. All end points were analyzed in a competing risks framework.31 For local relapse (distant metastasis) analysis, death with no evidence of disease and distant metastasis (local relapse), whichever occurred first, were regarded as competing events. In the analysis of sarcoma-specific mortality, death due to conditions unrelated to the disease was regarded as a competing event.
For each of the main prognostic factors investigated (re-excision, residual disease, and time interval between first sarcoma operation and re-excision), univariable analyses of each end point were performed for descriptive purposes by estimating the crude cumulative incidence curves. Comparisons between curves were performed by means of the Gray test.32 Covariate adjustment was necessary in this study because comparison between the groups (in turn, re-excision vs. control and presence vs. absence of residual disease) could be biased by imbalances in the prognostic characteristics. Multivariable analyses were thus performed by using Fine and Gray33 multiple regression models, in which the factor effect was adjusted for other prognostic factors. These models allow making inferences about the factor effect on the cumulative incidence function based on two-sided Wald tests. The following adjustment factors were considered: age at diagnosis, tumor size, tumor depth, Federation Nationale des Centres de Lutte Contre le Cancer (FNCLCC) grade,34 histological subtype, surgical margins, radiotherapy, and chemotherapy. In the models performed on the subset of patients with re-excised disease, only tumor size, tumor depth, FNCLCC grade, and histological subtype were included, because of the low number of events in this subgroup.
The multiple regression model provides estimates of the adjusted cumulative incidence curves in relation to specific values of the prognostic factors included in the model. Thus, ideally, an infinite number of incidence curves can be estimated, because the model contains continuous covariates such as age or tumor size. To allow comparisons between univariable and multivariable incidence estimates for the re-excision and control categories, adjusted incidence curves were calculated for each end point by averaging the curves estimated by the corresponding multiple Fine and Gray model for each of the two categories and the observed combinations of the adjustment covariates. The same calculations were performed for estimating the adjusted curves in both the presence and absence of residual disease categories.
In all analyses, age, tumor size, and the interval between the first sarcoma operation and re-excision were modeled as continuous variables by using three-knot restricted cubic splines,35 whereas the other covariates were modeled as categorical variables, as shown in Table 1
, by using dummy (0/1) variables. We used SAS software (SAS Institute Inc., Cary, NC) and the S-Plus® (MathSoft, Inc., Seattle, WA) design (http://lib.stat.cmu.edu) and cmprsk (http://bio-www.dfci.harvard.edu/~gray/) libraries to perform the modeling and statistical calculations. We considered as significant two-sided P values below the 5% conventional threshold.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. According to the logistic model results, no significant differences between the two study groups were found according to sex (P = .980), tumor site (P = .774), or tumor grade (P = .380). Patients who underwent re-excision and control patients were found to be significantly different as far as age (P = .011), tumor size (P < .001), tumor depth (P = .002), histological subtype (P = .001), margins (P < .001), radiotherapy (P = .008), and chemotherapy (P < .001) were concerned. Thus, in the control group, significantly more patients had deeper and larger tumors. This difference was reflected also in the distribution of multimodality treatment. In the re-excision group, radiotherapy was administered in 26.4% of patients, and only 3.5% of patients received adjuvant chemotherapy. In the control group, 54.1% and 30.0% of the patients received radiotherapy and chemotherapy, respectively.
After definitive surgery, surgical margins were microscopically negative in 97.2% of the re-excision group and in 72.8% of the control group. Among the re-excision group, 75 patients (23.6%) had residual disease in the pathologic specimen, including 17 patients (5.3%) with macroscopic residual disease. This result is slightly lower compared with literature data (Table 2).
|
During the follow-up period, local relapse developed as a first event in 47 (14.8%) patients in the re-excision group and in 51 (18.3%) in the control group. Distant metastasis was the first event in 38 (11.9%) patients in the re-excision group and in 73 (26.2%) patients in the control group.
Forty patients (12.6%) died of their disease in the re-excision group, as did 77 (27.6%) in the control group. A total of 29 patients died for reasons other than sarcoma.
Prognostic Effect of Re-Excision
Univariable analysis results for the different end points in our series of patients with re-excised disease were comparable to the literature data (Table 3). Table 4 shows the results of univariable and multivariable analyses (adjusted for the prognostic factors listed previously).
|
|
). In the re-excision group, the adjusted 10-year local relapse incidence was 18.7%, versus 16.4% in the control group.
|
.
|
.
|
shows the results of multivariable analyses. No difference was found according to the presence or absence of residual disease regarding local relapse (P = .747), as depicted in Fig. 4. The 10-year local relapse incidence was 17.5% in the presence of residual disease and 19.2% if no residual disease was found after re-excision. A significant effect of residual disease was found for distant metastasis, which had an incidence of 20.4% at 5 years and 21.1% at 10 years in patients with residual disease in the re-excision specimen versus 8.7% at 5 years and 9.0% at 10 years in patients with no residual disease (P = .002), as depicted in Fig. 5. For cause-specific mortality, only a trend toward a worse prognosis for patients with residual disease could be detected (P = .143), as depicted in Fig. 6. The mortality incidence was 17.2% and 18.4% at 5 and 10 years, respectively, in the presence of residual disease versus 11.3% and 12.1%, respectively, in the absence of residual disease. For all end points investigated, no significant prognostic effect was found for the time interval between unplanned sarcoma surgery and subsequent re-excision at our institution (data not shown).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
–3). In the subgroup of patients with re-excised disease, the presence of residual disease at pathologic examination was found to be a significant prognostic factor for metastases and, possibly, survival (Figs. 5 and 6).
The rate of re-excisions in this series was high in comparison to those in other published series (Table 2). Evidently, this reflects a liberal policy of re-excisions, which in fact were basically performed whenever the original operation had been unplanned. Indeed, the proportion of patients with residual microscopic disease in this series, 24%, was rather low in comparison with other series (Table 2), in which it ranged from 40% to 63%. Thus, on the basis of these data, it is difficult to state whether re-excisions were beneficial, because the equivalent outcome detected in patients with primarily excised and secondarily re-excised disease might depend either on the efficacy of the procedure as such or on overly liberal patient selection. Available literature is scant, and historical series have reported local failure in 70% to 90% of patients operated on with unplanned excisions,11–14 whereas more recent series have found patients with re-excised disease to have a local outcome similar to that in patients who received primary resection.27,28
No difference in survival has been observed in published series between the two subgroups of patients except for the work by Lewis et al.28 Their analysis of 407 patients with re-excised disease led to the paradoxical conclusion that patients with re-excised disease fared better than those operated on primarily, even after data were adjusted by American Joint Committee on Cancer (AJCC) stage. It is difficult to explain these findings. One possible reason is a selection bias, which may not be compensated for after adjusting comparisons by AJCC stage because of the wide variety of presentations, particularly as far as the primary tumor site is concerned. Furthermore, patients who undergo primary operation at a referral center might carry, on average, bigger tumors than those referred after excision. Again, this might not be evident by stratifying patients by AJCC classification, which is based on a fixed size cutoff (5 cm). Patients with tumors located in a difficult position or with bigger tumors might be referred to specialized centers more frequently than those with easier primary tumor locations and smaller tumors.
Actually, at the univariable analysis, we also found that patients with re-excised disease fared significantly better than those with primarily resected disease regarding distant metastasis and sarcoma-specific death (P < .0001). However, after adjustment for major risk factors (we considered tumor size as a continuous variable), statistical significance was lost. Thus, in our series, the incidence curves of the two groups of patients treated at our institution with a re-excision or a one-time definitive operation are similar for all the end points investigated (Figs. 1–3).
When the re-excised group was analyzed, patients with microscopic residual disease had a worse prognosis—interestingly, not in terms of local control, but in terms of distant metastases and, possibly, survival. This might be explained by assuming that residual disease is a marker of biologic aggressiveness, especially in a selected patient population such as ours. In fact, we labeled as "re-excisions" only those surgical interventions performed in the lack of obvious macroscopic residual disease after the first operation. Indeed, available literature is not consistent with this finding: residual disease was found either not to affect prognosis27,28 or to be a risk factor only for local recurrence.29 However, this could be due to a different starting patient selection: these series mainly included patients with positive margins after primary operation. It is therefore not surprising that they showed residual disease irrespective of the inherent aggressiveness of their disease.
In our series, all patients who had undergone an unplanned resection also underwent re-excision, whether surgical margins were described as positive or not. What was relevant for the decision to re-excise was basically whether the original surgeons were aware that they were operating on a sarcoma. Furthermore, little can be done to determine the presence of residual disease after an unplanned excision, although, because many patients in this series were treated before magnetic resonance imaging was available, the potential benefit of this modality could not be assessed. However, magnetic resonance imaging does not seem to add significantly to this, even in published literature.36 In any case, a lower tumor burden can be expected when patients are selected with these criteria, and residual disease will be less likely to be detected at re-excision. It follows that positive residual disease at re-excision will be more likely to correlate with tumor aggressiveness.
However, an additional factor might be the interval between the primary operation and re-excision in our series. The median time was 66 days, which is almost the double that reported by Lewis et al.28 As a matter of fact, this longer interval might have provided a chance for more aggressive tumors to proliferate by the time of the re-excision.
In conclusion, a practice of liberal surgical re-excision after unplanned surgery for adult STS may be associated with a final outcome for patients with re-excised disease that is similar to the outcomes of those who undergo primary resection at a referral institution. Under this kind of patient selection, residual disease might serve as a marker of tumor aggressiveness. Possibly, it could even serve as a prognostic factor—along with all the well-known prognostic factors of adult STS, i.e., pathologic grade and tumor size and depth—to be considered in the decision regarding, for example, adjuvant chemotherapy.37
Received for publication March 4, 2005. Accepted for publication August 4, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Bonvalot, M. Rivoire, M. Castaing, E. Stoeckle, A. Le Cesne, J. Y. Blay, and A. Laplanche Primary Retroperitoneal Sarcomas: A Multivariate Analysis of Surgical Factors Associated With Local Control J. Clin. Oncol., January 1, 2009; 27(1): 31 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hoshi, M. Ieguchi, M. Takami, M. Aono, S. Taguchi, T. Kuroda, and K. Takaoka Clinical Problems After Initial Unplanned Resection of Sarcoma Jpn. J. Clin. Oncol., October 1, 2008; 38(10): 701 - 709. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R. Chandrasekar, H. Wafa, R. J. Grimer, S. R. Carter, R. M. Tillman, and A. Abudu The effect of an unplanned excision of a soft-tissue sarcoma on prognosis J Bone Joint Surg Br, February 1, 2008; 90-B(2): 203 - 208. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
