Evaluation of neoadjuvant therapy and histopathologic response in primary, high-grade retroperitoneal sarcomas using the sarcoma nomogram

Authors

  • Timothy R. Donahue MD,

    1. Department of Surgery, Division of Surgical Oncology, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Michael W. Kattan PhD,

    1. Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, Ohio
    Search for more papers by this author
  • Scott D. Nelson MD,

    1. Department of Pathology, University of California at Los Angeles, Los Angeles, California
    2. University of California at Los Angeles Sarcoma Program, Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • William D. Tap MD,

    1. University of California at Los Angeles Sarcoma Program, Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, California
    2. Department of Medicine, Division of Hematology-Oncology, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Frederick R. Eilber MD,

    1. Department of Surgery, Division of Surgical Oncology, University of California at Los Angeles, Los Angeles, California
    2. University of California at Los Angeles Sarcoma Program, Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Fritz C. Eilber MD

    Corresponding author
    1. Department of Surgery, Division of Surgical Oncology, University of California at Los Angeles, Los Angeles, California
    2. University of California at Los Angeles Sarcoma Program, Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, California
    3. Department of Molecular and Medical Pharmacology, University of California at Los Angeles, Los Angeles, California
    • Department of Surgery, Division of Surgical Oncology, University of California at Los Angeles, 10833 LeConte Avenue, Room 54-140 CHS, Los Angeles, CA 90095-1782
    Search for more papers by this author
    • Fax: (310) 825-7575


  • Presented in part at the American Society of Clinical Oncology Annual Meeting, Sarcoma Parallel Session, Chicago, Illinois, May 31, 2008.

Abstract

BACKGROUND:

Patients with primary high-grade retroperitoneal soft tissue sarcomas have a 5-year disease-specific survival (DSS) of <40%. The impact of neoadjuvant therapy on histopathologic response and DSS are unknown.

METHODS:

From 1987 to 2007, 55 patients with primary high-grade retroperitoneal sarcoma received neoadjuvant therapy. All patients underwent surgical resection, and response was assessed histopathologically. Patients with ≥95% pathologic necrosis were classified as responders. Clinicopathologic variables were analyzed for association with DSS. Observed DSS was then compared with the Memorial Sloan-Kettering Cancer Center Sarcoma Nomogram predicted DSS.

RESULTS:

The median tumor size was 15 cm, and the median follow-up time for survivors was 68 months. The 5-year DSS for all 55 patients was 47% and did not significantly differ from the 37% predicted by the sarcoma nomogram for such patients (P = .44). Fourteen (25%) of the patients had ≥95% pathologic necrosis and were defined as responders; 41 (75%) were nonresponders. The 5-year DSS for responders was 83%. This was significantly better than the 5-year DSS of 34% for nonresponders (P = .002) and the 39% predicted by the sarcoma nomogram for responders (P = .018). The 34% 5-year DSS for nonresponders did not significantly differ from the 35% predicted by the sarcoma nomogram (P = .51).

CONCLUSIONS:

Neoadjuvant therapy was not associated with an overall improvement in DSS in patients with primary high-grade retroperitoneal sarcoma compared with the sarcoma nomogram prediction. Histopathologic response to neoadjuvant therapy was associated with a significantly improved DSS compared with nonresponders and with the sarcoma nomogram prediction for such patients. Cancer 2010. © 2010 American Cancer Society.

Although retroperitoneal sarcoma are rare, they account for up to 15% to 20% of soft tissue sarcomas (STSs).1, 2 Compared with other sites of STSs, retroperitoneal sarcomas constitute a difficult management problem and carry a particularly poor prognosis, with an overall 5-year disease-specific survival (DSS) of about 50%.3 High-grade retroperitoneal sarcomas do significantly worse than low-grade tumors, and most patients die from local or regional disease recurrences.3-8 The current standard of care remains complete surgical resection.9 The improvements in outcome seen in extremity STS with the addition of radiation therapy ± chemotherapy provide a benchmark for retroperitoneal sarcomas. Unfortunately, multimodality treatment approaches have not yielded similar improvements in either local control or survival for retroperitoneal sarcomas. Radiation therapy is limited by toxicity to adjacent organs, and adjuvant chemotherapy is viewed as marginally effective.10, 11

Neoadjuvant therapy for solid malignancies has several potential advantages, including cytoreduction of tumor burden, which can lead to a less radical surgical resection, immediate treatment of small clinically undetectable micrometastasis, and an in vivo test of chemoradiosensitivity. Methods to accurately measure response to therapy remain an area of active investigation. Our group has previously illustrated that neoadjuvant treatment-induced histopathologic response in the resected specimen is a significant predictor of local recurrence and overall survival for high-grade extremity STSs.12 Histopathologic response to neoadjuvant therapy has also correlated with recurrence and survival in many other malignancies, including bone sarcomas, metastatic colorectal cancer to the liver, and breast and lung cancer.13-18

Nomograms are prognostic tools that are designed from established databases to predict patient outcomes. Rather than scoring the presence or absence of pathologic variables as in the American Joint Committee on Cancer staging system, they integrate individual patient data to predict outcome. In addition, there has been recent interest in using them to assess the impact of treatment when a randomized clinical trial is not available. Novel instances where disease-specific nomograms have been used to assess the impact of treatment include sarcoma, prostate cancer, and pancreatic cancer.19-25

In patients who received neoadjuvant therapy for primary high-grade retroperitoneal sarcomas, the purposes of this study were to 1) determine whether histopathologic response to neoadjuvant therapy was associated with an improved DSS and 2) compare the observed survival of these treated patients with the sarcoma nomogram prediction.

MATERIALS AND METHODS

From 1987 to 2007, 522 patients with retroperitoneal sarcomas underwent surgical resection at University of California, Los Angeles (UCLA). One hundred twenty-nine (25%) patients had low- or intermediate-grade tumors, and 393 (75%) had high-grade tumors. Of the 393 patients with high-grade tumors, 179 (46%) presented with primary disease and 214 (54%) with recurrent disease. Fifty-five (31%) of the 179 patients with high-grade, primary retroperitoneal sarcomas underwent neoadjuvant chemotherapy and represent the study population. The other 124 (69%) patients with high-grade, primary retroperitoneal lesions did not receive preoperative therapy because of age >80 years, patient preference, logistics (distance of patient residence, referring medical groups, etc), or medical (bowel obstruction, ureteral obstruction, severe pain, etc) reasons. There were no gastrointestinal stromal tumors in the study population.

The neoadjuvant therapy protocols (chemotherapy ± radiation) were administered in a nonrandomized fashion and reflect the evolution of treatment at UCLA during the past 20 years. The regimens for retroperitoneal sarcomas were similar to those used for the extremities and other anatomic sites. Before 1990, patients received doxorubicin (Adriamycin)-based chemotherapy. After 1990, patients received either dacarbazine or gemcitabine/docetaxel (Gemzar/Taxotere)-based chemotherapy for leiomyosarcomas and ifosfamide-based chemotherapy for all other histologies. All patients completed the neoadjuvant chemotherapy and received adjuvant chemotherapy. Radiation therapy was administered as external-beam radiation to a total dose of 5000 centigrays. All patients were treated with radiation therapy. Patients who did not receive neoadjuvant radiation therapy were treated in an adjuvant manner.

Specialized sarcoma pathologists prospectively evaluated specimens in a standard fashion. Each specimen was bisected along its greatest diameter, and the perimeter of the tumor was macroscopically defined. An entire slice of this greatest diameter of tumor was partitioned into blocks of tissue en face and processed for histologic examination. The extent of necrosis was assessed relative to the percentage of residual viable tumor and done in a similar manner to that established for bone tumors.26, 27 The percentage of necrosis ranged from 0% to 100%. For this study, the results were grouped into 2 categories: histopathologic responders were defined as having ≥95% pathologic necrosis and nonresponders as <95% pathologic necrosis. Ninety-five percent was chosen as the cutoff value based on a prior published report of extremity STSs.12 Tumor grade was classified as high, intermediate, or low based on established criteria, including degree of differentiation, nuclear pleomorphism, and number of mitoses per high-powered field.28 For each tumor specimen, microscopic margins were also recorded. A positive microscopic margin was defined as microscopic tumor cells present at the surgical margin.

A tumor was considered primary if there was no evidence of metastatic disease and it was untreated at presentation. Recurrent disease was defined as tumor recurrence at the site of previous surgery (local) or at a distant (metastatic) site. Tumor size was defined as maximum tumor diameter on cross-sectional imaging (computed tomography or magnetic resonance imaging) before treatment. Tumors were grouped into 2 size categories: ≤10 cm and >10 cm.

Postoperative follow-up included physical examination, chest radiograph, and computed tomography of the abdomen and pelvis at 6-month intervals for 5 years and yearly thereafter. Endpoints for evaluation included date of death from disease and date of last follow-up. Survival was measured from the date of surgery at UCLA to the date of last follow-up or death from disease. All patient data, including percentage of pathologic necrosis, were collected prospectively, and follow-up in this group of 55 patients was 100%.

The variables used for nomogram analysis included tumor size (≤5 cm, 5-10 cm, and >10 cm), depth (deep), site (retroperitoneal), histology (liposarcoma, leiomyosarcoma, malignant peripheral nerve sheath tumor [MPNST], malignant fibrous histiocytoma, or other), age at diagnosis, and grade (high).29, 30

DSS was estimated using the Kaplan-Meier method and compared using the log-rank test. Comparisons between patients who responded and those who did not respond were performed using the chi-square test for dichotomous variables, the Student t test for continuous variables, and the Mann-Whitney U test for ranked categorical variables. Predicted DSS probabilities were calculated using the corrected group-prognosis method31 applied to the regression model underlying the Memorial Sloan-Kettering Cancer Center nomogram for STSs.29 By using this method, the probabilities at each failure time were then averaged across all patients, and the predicted survival curve was graphed as mean survival versus time.

RESULTS

Clinical, Pathologic, and Treatment Variables

From 1987 to 2007, 55 patients with primary high-grade retroperitoneal sarcomas received protocol neoadjuvant therapy (chemotherapy ± radiation). All patients underwent complete surgical resection, and response was assessed histopathologically. The clinical, pathologic, and treatment variables of the 55 patients are listed in Table 1. The median age of the patients was 56 years (range, 17-77). Twenty-two (40%) patients were female, and 33 (60%) were male. Fourteen (25%) patients had ≥95% histopathologic necrosis and were classified as responders, and 41 (75%) were classified as nonresponders. The median tumor necrosis was 50% (range, 5%-100%). Median tumor size was 15 cm (range, 5-40 cm). The majority of the tumors were >10 cm (n = 46, 84%), and there was no significant difference between the median size of the responders (14 cm; range, 5-30 cm) and nonresponders (15 cm; range, 6-40 cm) (P = .7).

Table 1. Clinical and Pathologic Characteristics
CharacteristicAll PatientsNonrespondersResponders
  • Leiomyo indicates leiomyosarcoma; Lipo, liposarcoma; MPNST, malignant peripheral nerve sheath tumor; MFH, malignant fibrous histiocytoma.

  • a

    P<.01.

No. of patients5541 (75%)14 (25%)
Median age, y (range)56 (17-77)57 (17-77)42 (20-75)
Sex   
 Male33 (60%)23 (44%)10 (71%)
 Female22 (40%)18 (56%)4 (29%)
Median size (range) cm15 (5-40)15 (6-40)14 (5-30)
Microscopic margin   
 Positive10 (18%)9 (22%)1 (7%)
 Negative45 (82%)32 (78%)13 (93%)
Histologic diagnosis   
 Leiomyo17 (31%)16 (40%)a1 (7%)
 Lipo11 (20%)8 (20%)3 (21%)
 MPNST5 (10%)5 (12%)0 (0%)
 MFH4 (7%)4 (10%)0 (0%)
 Other18 (32%)8 (18%)10 (71%)a
Neoadjuvant chemotherapy   
 Doxorubicin13 (24%)10 (24%)3 (21%)
 Ifosfamide33 (60%)22 (54%)11 (79%)
 Dacarbazine5 (9%)5 (12%)0 (0%)
 Gemcitabine/docetaxel4 (7%)4 (10%)0 (0%)
Neoadjuvant radiotherapy   
 Yes31 (56%)21 (51%)10 (71%)
 No24 (44%)20 (49%)4 (29%)

Patients received either doxorubicin (n = 13, 24%), ifosfamide (n = 33, 60%), dacarbazine (n = 5, 9%), or gemcitabine/docetaxel (n = 4, 7%)-based preoperative therapy. Thirty-one (56%) patients received neoadjuvant external beam radiation therapy. There was not a significant difference in the percentage of patients treated with neoadjuvant radiation therapy among the responders and nonresponders.

The 3 most common histologies were leiomyosarcoma (n = 17, 31%), liposarcoma (n = 11, 20%), and MPNST (n = 5, 10%). Only 1 of the 17 patients with leiomyosarcomas responded to treatment. This was significantly lower (P = .02) as compared with nonleiomyosarcoma histologies. “Other” histologies were significantly more likely to respond to treatment (P = .0004). Of the “other” group, all 6 of primitive neuroectodermal tumors (PNETs) responded to treatment.

DSS

The median follow-up for surviving patients was 68 months (Table 1). Thirty-one patients died of disease, and 4 patients died of another cause. Of the surviving patients, at the time of last follow-up, 1 (2%) patient was alive with disease, and 19 (35%) had no evidence of disease. The 5-year DSS for all patients was 47%. The 5-year DSS for responders was 83%, which was significantly better than the 34% for nonresponders (P = .002) (Fig. 1).

Figure 1.

Disease-specific survival is shown in observed responders (≥95% histopathologic necrosis) versus nonresponders (≤ 95% histopathologic necrosis).

On univariate analysis (Table 2), younger age (P = .003) and histopathologic response (P = .002) were significantly associated with an improved DSS. Size scored as either a continuous variable or a nominal variable, separated into 2 groups (<10 cm or ≥10 cm) was not significant. On multivariate analysis, both histopathologic response (hazard ratio [HR], 2.36; P = .007) and age (HR, 1.04; P = .003) were associated with DSS.

Table 2. Survival: Regression Analysis
OverallUnivariate PMultivariate PHazard Ratio
  1. Leiomyo indicates leiomyosarcoma; Lipo, liposarcoma; MPNST, malignant peripheral nerve sheath tumor; MFH, malignant fibrous histiocytoma.

Age, y (continuous).0032.0031.04
Male sex.88  
Size, cm (continuous).63  
Positive micromargin.12  
Histologic diagnosis   
 Leiomyo.27  
 Lipo.57  
 MPNST.33  
 MFH.56  
 Other.14  
Neoadjuvant chemotherapy   
 Doxorubicin.46  
 Ifosfamide.22  
 Dacarbazine.16  
 Gemcitabine/docetaxel.49  
Neoadjuvant radiotherapy.15  
≥95%Necrosis.0007.0072.36

Nomogram Analysis

Patient and tumor variables for all 55 patients in the study were entered into the Memorial Sloan-Kettering Cancer Center nomogram, and a predicted DSS was generated. The predicted DSS was then compared with the observed DSS of each of our patients (Fig. 2). There was not a significant difference between the observed and predicted DSS for all patients (P = .44). When the nonresponders were entered into the nomogram, there was not a significant difference between predicted and observed DSS (P = .51) (Fig. 3). However, the observed DSS for responders was significantly better than the nomogram prediction (P = .018) (Fig. 4).

Figure 2.

Disease-specific survival is shown in all patients (n = 55), observed versus nomogram predicted.

Figure 3.

Disease-specific survival is shown in nonresponders (n = 31), observed versus nomogram predicted.

Figure 4.

Disease-specific survival is shown in responders (n = 14), observed versus nomogram predicted.

DISCUSSION

Retroperitoneal sarcomas have a worse prognosis than STSs of other sites, including the extremity and trunk.2, 32-35 Despite increased experience in managing these tumors, the overall disease-specific mortality has not significantly changed, with an overall 5-year DSS of 50%.1, 2 Factors that have been consistently shown to significantly predict worse survival for patients with retroperitoneal sarcomas include high grade, advanced stage at presentation, histologic subtype, macroscopic positive margins, and local recurrence.3, 4, 6, 7, 35-39 Up to 60% of patients with high-grade retroperitoneal sarcomas die of locally recurrent disease. On the basis of an analysis of 500 patients with retroperitoneal sarcomas at Memorial Sloan-Kettering Cancer Center, Lewis et al3 emphasize that the initial resection of these tumors critically affects the outcome, and once a recurrence develops, the course of disease is primarily determined by the biology of the tumor. Re-resections have limited impact on preventing additional recurrences, and the prospect for cure in this setting is negligible. As has been shown in multiple malignancies, including sarcomas, neoadjuvant therapy often provides patients with the best opportunity for complete resection of the primary tumor and importantly gives an insight into the biology of the particular tumor by measuring the response to the chosen therapy.

To improve DSS in patients with retroperitoneal sarcomas, better adjuvant treatments are needed. Different combinations of therapies have been tried, with mixed results.3, 4, 7, 8, 11, 36-44 These approaches include adjuvant chemotherapy or radiation therapy, intraoperative radiation therapy, and neoadjuvant radiation therapy.11, 44-48 The most significant obstacle to radiation therapy includes toxicity to adjacent organs.34, 47, 48 Trials examining the role of adjuvant chemotherapy have suggested inferior overall survival for patients with poor prognostic factors such as metastatic disease, leiomyosarcomas, and high-grade tumors.11, 35, 36, 39 Because of the low prevalence and histologic heterogeneity of retroperitoneal sarcomas, most of these single-institution trials are quite small and of low power. As a result, the Radiation Therapy Oncology Group and American College of Surgeons Oncology Group (Z9031) began multi-institutional attempts at multimodality therapy for retroperitoneal sarcomas. Unfortunately, both studies were closed because of poor patient accrual.

Histopathologic response to neoadjuvant treatment is correlated with improved DSS for many solid malignancies and can be used as a marker of response to treatment.12-18 In this study, responders were defined as those patients having >95% histopathologic necrosis. This level of necrosis was chosen for 2 principal reasons. Although there were only 2 pathologists, both specializing in sarcoma, who evaluated the specimen, there is still a potential for interobserver variability. Choosing a high level of necrosis minimizes such variability, especially for tumors that are borderline responders. Also, a previous study on neoadjuvant treatment for high-grade extremity sarcomas defined response as at least 95% histopathologic necrosis. The responders had a significantly lower rate of local recurrence and improved overall survival.12

The postoperative sarcoma nomogram was developed as a more accurate and updated tool to predict 12-year sarcoma-specific death.29 This nomogram was generated from a Cox analysis of the prospectively collected database of 2135 patients with primary disease at Memorial Sloan-Kettering Cancer Center to identify significant predictor variables for DSS. The nomogram has since been internally validated by this same Memorial Sloan-Kettering Cancer Center patient cohort and externally validated by 2 large independent cohorts of patients with primary STSs. These external cohorts consisted of 929 patients from all sites, including the retroperitoneum, from UCLA30 and 642 patients with extremity sarcomas from the Milan Cancer Institute.49 Each of these validations yielded high concordance indices and precise calibration curves. In addition, the sarcoma nomogram has been previously used as a control to evaluate the impact of neoadjuvant ifosfamide-based chemotherapy for extremity sarcomas. Rather than using a historical control, the authors found that the observed survival of patients treated with ifosfamide was significantly better than that of untreated patients and the nomogram prediction for these same treated patients.19

On the basis of the perceived success of neoadjuvant treatment for extremity sarcomas and the poor outcomes with surgery alone, a select group of patients with primary, high-grade retroperitoneal sarcomas were given a similar trial of neoadjuvant treatment. Although this is a select population, the patient and tumor characteristics (age, size, etc) are similar to those of other reported large series of high-grade primary retroperitoneal sarcomas.3-5 The neoadjuvant therapy protocols were administered in a nonrandomized fashion and reflect the evolution of treatment at UCLA. All 55 patients were able to tolerate the neoadjuvant therapy and undergo complete resection. The finding that none of the patients progressed to unresectability and that all patients completed the therapy is likely a result of selecting patients with appropriate performance statuses and/or the administration of neoadjuvant chemotherapy over a relatively short course.

The overall 5-year survival was 47% for all 55 patients, 83% for the 14 responders, and 34% for the 41 nonresponders (Table 2). The significant predictors of poor survival included <95% histopathologic necrosis and advanced age. Leiomyosarcomas were statistically significantly less likely to respond to neoadjuvant treatment, whereas other more favorable histologies, including all 6 patients with PNET, were significantly more likely to respond. In the nomogram analysis, there was no significant difference in the observed DSS, and nomogram predicted DSS for all 55 patients, confirming the calibration of the nomogram to our cohort of patients, as was seen in the previous validation by the UCLA data set. Responders had a significantly better DSS than the nomogram prediction for such patients; nonresponders did not do significantly worse than their nomogram prediction.

Although the pathologic response rate was higher, and there were more survivors from 1990 on (1 [11%] of 9 responders and 1 [11%] of 9 survivors before 1990 versus 13 [28%] of 46 responders and 15 [33%] of 46 survivors from 1990 on), there are really too few patients in the time period before 1990 to make any meaningful comparisons as we have done with extremity disease.50, 51 After 1990, we do not see any differences in either response rates or survival as both are distributed through this time period.

It is important to note that despite treating a select patient population, we do not show that patients who received neoadjuvant therapy for primary high-grade retroperitoneal sarcomas have a better survival. Although this could be because of several contributing factors, the central issue is likely that the response rates to the chemotherapy regimens used were simply too low (25%). If the 6 patients with PNETs, a member of the Ewing sarcoma family that is known to be chemosensitive as measured by histopathologic response and DSS,52 are excluded, the response rate is only 16% (8 of 49). Although the response rate is lower with the exclusion of the PNETs, the histopathologically responding patients still demonstrated a significantly improved survival compared with the nomogram prediction for such patients (P = .032) (Fig. 5).

Figure 5.

Disease-specific survival is shown in responders, excluding patients with primitive neuroectodermal tumors (n = 8), observed versus nomogram predicted.

Our analysis shows that responders do significantly better than nonresponders, and that responders do better than the nomogram prediction. Thus, by giving these cytotoxic drugs preoperatively and assessing histopathologic response, we were able to identify individual patients who will and who will not benefit from the particular therapy administered. Furthermore, because all 55 patients, including the nonresponders, did not appear to do worse than the nomogram prediction, it does not appear that this approach adversely affects the survival of these patients.

Future directions for the treatment of retroperitoneal sarcomas include identification of better/targeted drugs for which there is currently no long-term outcome data. Preliminary evidence to justify a large-scale multi-institutional randomized clinical trial can be generated using similar methods of analysis: assessing histopathologic response or pathway inhibition to the specific agent and comparing the observed versus predicted outcome using nomogram analysis (www.nomograms.com). In addition, noninvasive measures of response to therapy, such as 18F-fluorodeoxyglucose positron emission tomography, will allow for quicker treatment modifications.53-55

In conclusion, treatment strategies for high-grade retroperitoneal sarcomas that rely on resection alone have poor DSS. Therefore, we investigated a neoadjuvant therapy approach to treating patients with this high-risk malignancy. In this study, we found that neoadjuvant therapy was not associated with an overall improvement in DSS in patients with primary high-grade retroperitoneal sarcomas compared with the sarcoma nomogram prediction; however, histopathologic response to neoadjuvant therapy was associated with a significantly improved DSS. In addition, the nomogram appears to be an effective model to evaluate new therapies for retroperitoneal sarcomas and other diseases when a randomized clinical trial is not available.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary