|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2000 American Society for Clinical Oncology Influence of Biologic Factors and Anatomic Site in Completely Resected LiposarcomaByFrom the Departments of Surgery and BiostatisticsMemorial Sloan-Kettering Cancer Center, New York, NY. Address reprint requests to Murray F. Brennan, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021; email m-brennan{at}mskcc.org
PURPOSE: Soft tissue sarcoma (STS) encompasses a group of neoplasms that are anatomically and biologically diverse. Retroperitoneal/visceral (RP/V) tumors have a poorer prognosis than extremity/trunk (E/T) lesions, and this has been attributed to frequent presentation with tumors of large size and multiorgan involvement that precludes complete resection. The worse prognosis that is associated with RP/V tumors has also been thought to be histopathologically dependent and not necessarily related to anatomic site. The aim of this study was to determine the role of anatomic site and biologic features in prognosis and outcome in patients after complete resection by examining a large cohort of STS patients with a single histopathology, ie, liposarcoma. METHODS: All patients who were treated for liposarcoma from July 1, 1982, through July 1, 1998, were included. Univariate analyses were performed using log-rank test and Kaplan-Meier estimates, and multivariate analyses were performed using Cox regression. The three end points examined were local recurrence (LR), distant recurrence, and disease-specific survival (DSS). RESULTS: Seven hundred twenty patients with liposarcoma were evaluated, and of these, 460 had completely resected primary or completely resected locally recurrent disease. Breakdown of anatomic site was 65% E/T (n = 301) and 35% RP/V (n = 159). The median follow-up period for patients who underwent complete resection was 42 months (range, 1 to 194 months). We found that RP/V site is a poor prognosticator that is independent of patient sex and age; tumor size, grade, and margin; and recurrent presentation. Sixty-nine percent of patients with RP/V tumors who died had local disease only and no distant metastasis at the time of death. CONCLUSION: In liposarcoma, tumor location exerts as strong an influence on prognosis as biology. In contrast to extremity liposarcoma, LR without distant metastasis often results in death for patients with RP/V tumors. For these patients, local control accomplished by complete surgical resection ± adjuvant radiation therapy should impact strongly on DSS.
SOFT TISSUE SARCOMA (STS) is a rare disease that encompasses a histologically and anatomically diverse group of neoplasms, the majority of which share a common embryologic origin from mesodermal tissue. Despite the fact that skeletal and somatic tissues comprise greater than 75% of the human body, STS remains uncommon, representing less than 1% of all human neoplasms.1 Although many histologic subtypes of STS have been identified, many of which can be distinguished by characteristic cytogenetic abnormalities, histologic subtype has not, at present, influenced disease management.2,3 Because of the low incidence of the disease, systematic trials focusing on specific histologic subtypes of STS have not been possible. Liposarcoma is the most common histology seen in STS. Of more than 3,500 cases of STS that have been entered into the Memorial Sloan-Kettering Cancer Center (MSKCC) prospective database from 1982 to 1998, 720 are classified as liposarcoma. Liposarcoma is the most common histologic subtype of both extremity and retroperitoneal STS. It is notable for its frequent presentation with very large tumors that are especially common in the retroperitoneal/visceral (RP/V) location. This tumor most commonly presents in the extremities, with the thigh being the most common site,4 and multivariate analysis of prognostic factors in extremity liposarcoma has been previously reported.5 Clinical behavior is associated with histology. Low-grade lesions show a high incidence of local recurrence but little to no propensity for metastasis, in contrast with high-grade or poorly differentiated tumors, which often manifest clinically aggressive behavior with a high incidence of local recurrence (LR) and distant metastasis. In this study, we analyzed prospective data from a large group of patients with liposarcoma. We focused on the influence of anatomic site and other known clinical and biologic factors in this histologically homogeneous cohort of patients after complete resection of primary or recurrent disease. The outcome for STS is clearly site dependent.2 In the past, this has been attributed to the influence of different histopathologies and large tumor size at presentation that is commonly seen with RP/V liposarcoma. We sought to determine, in a single histopathologic subtype of STS that is found in both the extremity/trunk (E/T) and RP/V site, whether the difference in outcome after complete resection can be explained by known prognostic factors or whether site is in and of itself an independent prognosticator.
A prospective database of adult patients with STS treated at MSKCC was established in July 1982. Patients who underwent treatment for liposarcoma and had complete resection (ie, negative gross margins) from July 1, 1982, through July 31, 1998, were included in this analysis. All patients with E/T and RP/V site of disease were included. Patient, tumor, and treatment variables were analyzed for prognostic significance. The primary end points of the study were LR-free survival, distant recurrence (DR)–free survival, and disease-specific survival (DSS). LR was defined as the first LR after treatment at MSKCC that was not preceded by a distant metastasis. DR was defined as any DR after treatment at MSKCC. For DSS, only deaths that resulted from disease were considered as events; patients who died of other causes were censored at the time of death. Patient variables that were analyzed include the following: age, sex, and presentation status (primary v recurrent). Tumor variables that were analyzed include size, grade, depth, microscopic margin, anatomic site, and histologic variant.
The assessment of microscopic margins was performed with at least 15 sections for most large RP/V tumors. Similarly, for all liposarcomas, the general rule for microscopic margin examination was at least one section per cubic centimeter of tumor. Grade was determined by evaluation of cellularity, mitotic rate, and presence or absence of necrosis. The tumors were classified primarily in a two-grade scale of malignancy (low and high grade). The grade of malignancy was determined by a combined assessment of the following histologic features: tumor differentiation, degree of cellularity, cellular pleomorphism or anaplasia, mitotic activity, and invasive growth. The tumor differentiation was assessed in relation to the degree of histologic resemblance with the normal adult adipose tissue. The mitotic count was divided into two groups: less than 10 mitotic figures/high-powered field and Because the focus of the study was prognostic factors at the time of presentation and because adjuvant treatment (radiation and/or chemotherapy) was not prospectively randomized (but was given either as standard of care at the time of treatment or in experimental protocols), the inclusion of these variables in the statistical analysis was avoided. Univariate analyses were performed using the log-rank test and Kaplan-Meier estimates, and multivariate analysis was performed using Cox regression. The influence of anatomic site (E/T v RP/V) after adjusting for the significant prognostic variables for each end point was then studied by examining the significance of site with the other significant variables in the multivariate Cox model. The results of the Cox model analysis are reported with relative risks (RR) and 95% confidence intervals. P < .05 was considered to be significant in these analyses.
During the time period of July 1982 through July 1998, 720 patients with liposarcoma were treated at MSKCC. Figure 1 shows the anatomic distribution of these tumors, the majority of which (55%) were extremity lesions. Of the 720 patients, 460 (64%) underwent complete resection with grossly negative margins at the time of resection, and these patients are the subject of this analysis. Table 1 lists the distribution of clinical and pathologic characteristics seen in this cohort of patients. In contradistinction to all STS, in which the minority of patients (35%)2,4 present with low-grade lesions, patients with liposarcoma present more often with low-grade lesions (55%). Three hundred one patients (65%) presented with E/T tumors and 159 (35%) had RP/V lesions. With a median follow-up period of 42 months (range, 1 to 197 months), 310 patients (67%) were alive, 109 (24%) had died of disease, and 41 (9%) had died of other causes at the end of the study period. Actuarial overall 5-year DSS for the entire cohort is 75%. LR was seen in 117 patients (25%), and DR was seen in 45 patients (10%) at the time of last follow-up. Between E/T and RP/V, the proportions of patients who developed LR, DR, and death from disease were 16% versus 43%, 12% versus 5%, and 17% versus 36%, respectively.
A large number of patients (52%) had tumors that were greater than 10 cm in size. This high incidence of large tumors was seen especially in patients with RP/V tumors (116 of 159 patients; 73%). Characteristically, because of large tumor size, multiorgan resection is required in many cases to accomplish complete resection. Such multiorgan resection was required in 128 (80%) of 159 patients in this series. Contiguous organs that were resected in this series included kidney, stomach, small bowel, colon, spleen, liver, pancreas, and ovary.
Association Between Site and Clinicopathologic Factors
Incidence and Rate of Developing LR Table 4 lists results relating the prognostic factors to the rates of developing first LR after treatment. The LR-free rates for the RP/V cases were much lower than those for the E/T cases in all categories that we examined. Multivariate analysis identified recurrent disease at presentation as the most important, independent, unfavorable prognostic factor for developing LR in both the E/T (P = .05; RR = 2) and RP/V (P = .03; RR = 1.8) cases. Multivariate analysis also showed high grade to be independently related to a higher rate of developing LR for the RP/V cases (P = .01; RR = 1.9) while approaching significance for the E/T cases (P = .06; RR = 1.7). When analysis was confined to primary cases only, the only independent unfavorable factor that we found to influence LR was high grade (E/T, P = .06, RR = 2; RP/V, P = .05, RR = 2).
Influence of Site and Other Factors on DSS Table 5 lists results relating factors to disease-related mortality in all 460 patients who underwent complete resection. Once again, the 5-year survival rates for the E/T cases were much higher than those for the RP/V cases in each respective category. By multivariate analysis, in both the E/T and RP/V cases, high grade (E/T, P = .001, RR = 3.5; RP/V, P = .001, RR = 2.6) and recurrent disease at presentation (E/T, P = .03, RR = 2.3; RP/V, P = .001, RR = 2.5) were found to be unfavorable prognostic factors.
In addition, tumor size greater than 10 cm (P = .001; RR = 2.7) was an unfavorable prognostic factor for the E/T cases. Size was an unimportant factor for the RP/V cases (P = .8). When adjusting for all of the factors that we considered, RP/V was an independent unfavorable factor (P = .001; RR = 5) for DSS. Interestingly, RP/V site was found to be an adverse factor for LR and DSS but a favorable factor for DR-free survival, which is likely due to the fact that the majority of patients in that group who died of disease (40 of 58) succumbed to an LR in the absence of distant metastasis. Figure 2 shows the Kaplan-Meier survival curves for LR-free survival and DSS stratified by anatomic site.
DSS in Primary, Completely Resected Deep Liposarcoma Our results in Table 5 show that recurrent disease at presentation was an unfavorable prognostic factor for survival in both the E/T and RP/V cases. Because the proportion of RP/V patients who presented with recurrent disease was much higher than that of the E/T patients (Table 3), we would like to eliminate this possible confounder to determine whether RP/V cases are worse than E/T cases, independent of presentation status. Furthermore, because all RP/V cases are, by definition, deep tumors, we also excluded the superficial E/T tumors in the comparison. The result is a group that consisted of 308 patients (n = 210 for E/T and n = 98 for RP/V) with primary, completely resected, deep tumors. The analysis results for these patients are listed in Table 6.
The results are consistent with those listed in Table 5; ie, RP/V cases fare worse than E/T cases based on the 5-year survival rates (Fig 3). The independent unfavorable factors in the E/T cases were found to be high grade (P = .001; RR = 4.5) and tumor size greater than 10 cm (P = .001; RR = 2.6). The only independent unfavorable factor in the RP/V cases was high grade (P = .001; RR = 4.8); size was not significant (P = .5). In this more homogeneous group of patients, RP/V disease was a marginally unfavorable factor (P = .06; RR = 4) after adjusting for all of the factors that we considered. However, the smaller P value may be due to the lack of power in this analysis, in which the number of deaths due to disease was fewer than 50, or one half of that in the analysis in Table 5. Furthermore, the RR of 4 is similar to that of the analysis that included all patients (RR = 5). Therefore, we conclude that RP/V site of disease is an unfavorable prognostic factor for disease-related mortality.
For patients with liposarcoma, tumor location exerts as strong an effect on prognosis as other clinical and biologic factors that are known to influence outcome. Although the worse prognosis that is associated with RP/V tumors has been previously attributed to frequent presentation with large-size tumors and multiorgan involvement that precludes complete resection, anatomic site remains highly significant as a prognosticator, even in this histologically homogeneous cohort of patients in whom complete resection was achieved. Anatomic site was the most significant prognostic factor for all three end points. A comparison of the different prognostic factors revealed that RP/V liposarcomas tended to be associated with age older than 50 years, tumor size greater than 10 cm, positive microscopic margins, and locally recurrent presentation, which are all adverse prognostic factors. Conversely, the proportion of low-grade sarcomas was higher among patients in the RP/V group than in the E/T group. The higher than predicted incidence of high-grade tumors in the E/T group was influenced by the exclusion of some patients with atypical lipomatous tumor, many of whom may have been characterized as having well-differentiated liposarcoma in the first decade of the study period (1982 to 1992). Controlling for all of these factors, anatomic site remains a highly significant, independent prognostic variable. Interestingly, microscopic margin status was not a predictor of LR in either the RP/V group or the E/T group. For the E/T group, this may be a result of the ability of adjuvant radiation therapy to improve local control in this subset of patients. For the RP/V group, the assessment of microscopic margins remains difficult. These tumors are typically quite large and are surrounded by a thin fibrous capsule, and comprehensive microscopic margin assessment is not practical. Even in the setting in which all examined margins are microscopically negative, the presence of microscopic residual disease cannot be ruled out. In fact, LR in the RP/V group occurred just as commonly among patients with microscopically negative margins as those with positive margins. Why should anatomic site alone be an independent prognostic variable for a histologically identical tumor? A plausible explanation for the strong, independent influence of anatomic site can be deduced when the patterns of failure are analyzed for the group of patients with RP/V tumors. The crucial difference can be seen when the differences in cause of death are examined for patients with RP/V versus E/T lesions. Of patients with RP/V liposarcomas who died of disease, 40 (69%) of 58 patients died with LR only and no clinically recognizable distant metastasis. Death occurred in the majority of these patients as a result of extensive growth of their intra-abdominal/retroperitoneal LR, which resulted in progressive cachexia and, in some cases, multifocal bowel obstruction. In contrast, for patients with E/T liposarcoma, the most likely cause of death was pulmonary metastasis, which was seen in more than 60% of patients who presented with an E/T primary tumor and subsequently died of disease. Although unusual metastatic sites for STS are recognized, they still make up a minority of overall metastases, and in this group of patients with E/T liposarcoma, lung metastasis remains the most common site of distant failure. No patients with E/T lesions died of local disease only. The role of radiation therapy in improving local control has been well documented in randomized trials of extremity STS.6,7 The high incidence of positive microscopic margins in patients with RP/V liposarcoma would suggest that a similar benefit with adjuvant radiation therapy would be seen for this group of patients. Radiation, however, is limited by tolerance of adjacent intra-abdominal and retroperitoneal structures, which makes adequate dosing technically difficult, if not impossible. In contrast, for patients with extremity lesions, adequate-dose radiation therapy can be delivered with low morbidity either by brachytherapy or external-beam radiotherapy, with a significant improvement in local control.8,9 Intraoperative radiation therapy with shielding of adjacent organs to minimize toxicity has been used at our institution to improve local control of RP/V tumors, especially for recurrent RP/V liposarcoma. Although this technique improves our ability to administer adequate-dose radiation to the tumor bed, it has not been shown to improve overall survival or DSS, and so it remains an experimental approach.10 In the only randomized trial of adjuvant radiation therapy for patients with RP/V sarcoma, a benefit was not able to be demonstrated.11 Novel methods to improve local control should have an impact on survival of patients with this disease, because many patients die as a result of local disease only. Although presentation with LR is an adverse prognostic variable, these data show that aggressive attempts at re-resection are warranted in certain instances. In this series, negative microscopic margins were achieved in 62% of patients with recurrent RP/V liposarcomas and in 89% of patients with recurrent E/T lesions.12 The rationale for liberal re-exploration for recurrent RP/V liposarcoma is predicated on the finding that median survival after LR of retroperitoneal STS is 60 months for patients who have undergone resection versus 20 months for patients who have not undergone resection.13,14 The need for resection of multiple contiguous organs to obtain complete resection in recurrent RP/V liposarcoma should not deter surgical attempts to render the patient disease-free. The high incidence of low-grade tumors in this group further emphasizes the importance of complete surgical resection, because the majority of patients who die from RP/V liposarcoma die from local disease only. Because recurrent tumors were more common in the RP/V group, we examined the data for primary tumors only and found the same result, ie, that site alone, independent of recurrent presentation, is an important prognostic indicator. As shown in our previous study of 500 patients with retroperitoneal STS, the ability to achieve complete resection decreases precipitously after the first recurrence.14 Because we have not been able to show that incomplete resection impacts on survival, the best chance for local control is at first presentation or first recurrence. The importance of local control, especially in retroperitoneal liposarcoma, is underscored by its direct association with DSS. When feasible, aggressive surgical treatment to include involved contiguous organs is mandated. However, extended resection of contiguous organs that are not involved has not improved survival.15 The direct correlation between LR and survival in patients with RP/V liposarcoma encourages the surgeon to develop novel methods to improve local control in this group of patients for whom toxicity from adequate-dose radiation is prohibitive. Based on what we have learned about the biology of this disease, improvement in local control may correspond with improved overall survival and DSS in patients with RP/V liposarcoma. Conversely, for patients with E/T liposarcoma, ongoing attempts at developing more effective systemic therapy should be pursued.
Supported by National Institutes of Health grant no. P01-CA47179 (M.F.B.).
1. Lewis JJ, Brennan MF: Soft tissue sarcomas. Curr Probl Surg 33:817-872, 1996[Medline] 2. Lewis JJ, Brennan MF: Soft tissue sarcoma, in Cameron JL (ed): Current Surgical Therapy. St. Louis, MS,Mosby, 1998, pp 1033-1038 3. Pisters PWT, Pollock RE: Soft tissue sarcomas, in Winchester DP, Jones RS, Murphy GP (eds): Cancer Surgery for the General Surgeon. Philadelphia, PA,Lippincott Williams and Wilkins, 1999, pp 309-324
4.
Pisters PWT, Leung DH, Woodruff J, et al: Analysis of prognostic factors in 1,041 patients with localized soft tissue sarcomas of the extremities. J Clin Oncol 14:1679-1689, 1996 5. Chang HR, Gaynor JJ, Tan C, et al: Multifactorial analysis of survival in primary extremity liposarcoma. World J Surg 14:610-618, 1990[Medline]
6.
Pisters PWT, Harrison LB, Woodruff JM, et al: A prospective randomized trial of adjuvant brachytherapy in the management of low grade soft tissue sarcomas of the extremity and superficial trunk. J Clin Oncol 12:1150-1150, 1994
7.
Yang JC, Chang AE, Baker AR, et al: A randomized prospective study of the benefits of adjuvant radiation therapy in the treatment of soft tissue sarcoma of the extremity. J Clin Oncol 16:197-203, 1998
8.
Pisters PWT, Harrison LB, Leung DH, et al: Long term results of a prospective randomized trial evaluating the role of adjuvant brachytherapy in soft tissue sarcoma. J Clin Oncol 14:859-868, 1996 9. Harrison LB, Franzese F, Gaynor JJ, et al: Long term results of a prospective trial of adjuvant brachytherapy in the management of completely resected soft tissue sarcomas of the extremity and superficial trunk. Int J Radiat Oncol Biol Phys 27:259-259, 1993[Medline] 10. Sindelar WF, Kinsella TJ, Chen WP: Intraoperative radiotherapy in retroperitoneal sarcomas: Results of a prospective, randomized clinical trial. Arch Surg 128:402-410, 1993[Abstract] 11. Rosenberg SA, Chang AE, Glatstein E: Adjuvant chemotherapy for treatment of extremity soft tissue sarcomas: Review of National Cancer Institute experience. Cancer Treat Symp 3:83-88, 1985
12.
Lewis JJ, Leung D, Heslin M, et al: Association of local recurrence with subsequent survival in extremity soft tissue sarcoma. J Clin Oncol 15:646-652, 1997 13. Jaques DP, Coit DG, Hajdu SI, et al: Management of primary and recurrent soft-tissue sarcoma of the retroperitoneum. Ann Surg 212:51-59, 1990[Medline] 14. Lewis JJ, Leung D, Woodruff JM, et al: Retroperitoneal soft-tissue sarcoma: Analysis of 500 patients treated and followed at a single institution. Ann Surg 228:355-365, 1998[Medline] 15. Bevilacqua RG, Rogatko A, Hajdu SI, et al: Prognostic factors in primary retroperitoneal soft-tissue sarcomas. Arch Surg 126:328-334, 1991[Abstract] Submitted June 18, 1999; accepted January 4, 2000. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2000 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
10 mitotic figures/high-powered field, and the degree of tumor necrosis was quantitated as less than 50% and 
