Tumor volume as a prognostic factor for sarcomatosis

Authors


Abstract

BACKGROUND

The appropriate therapeutic interventions for sarcomatosis, or sarcoma characterized by intraabdominal dissemination, remain unclear. The authors performed a retrospective analysis of their recent experience with patients diagnosed with sarcomatosis to determine the overall survival and the effects of clinicopathologic features on survival rates at two and four years.

METHODS

A query of the authors' prospective soft tissue sarcoma database identified 51 patients with a diagnosis of sarcomatosis who were evaluated at the authors' institution between June 1996 and June 1999. Clinical and pathologic factors were evaluated, and survival was calculated using a Kaplan-Meier survival analysis. Disease was categorized as low or high volume based on findings at surgical exploration or computed tomography scan evaluation. Disease was classified as low/intermediate grade or high grade based upon histologic examination.

RESULTS

Twenty five patients were male and 26 were female. The median time from the initial diagnosis of sarcoma to the development of sarcomatosis was 0.9 years (range, 0-26 years). Thirty nine patients were treated with surgery, whereas 32 received primarily nonsurgical treatment. Histology revealed gastrointestinal stromal tumor (GIST) in 33 patients and other histologies in 18 patients. The two year overall survival rate of patients with GIST was similar to that of patients with other types of sarcoma (38% versus 42%, respectively, P = 0.77). Patients with low volume disease had an overall two year survival rate of 82%, compared with only 24% for patients with high volume disease (P = 0.008). There was no difference in the overall survival rates of patients with low grade (n = 18) versus high grade tumors (n = 33, P = 0.29). With a median followup of 2.7 years (range, 0.5–26.4 years), the median time from sarcomatosis to death was 13 months (range, 4–42 months).

CONCLUSIONS

Evaluating volume of disease at the time of diagnosis permits stratification of patients into prognosis based subsets. We found no significant difference in two or four year survival rates in patients with GIST and those with non-GIST sarcomatosis. Cancer 2002;94:2441–6. © 2002 American Cancer Society.

DOI 10.1002/cncr.10504

Disseminated intraabdominal sarcoma, or sarcomatosis, is commonly seen in patients with end-stage intraabdominal sarcoma. About 18% of all sarcomas arise within the abdominal cavity: 15% of all sarcomas arise in the retroperitoneum and 3% in the primary viscera. Of the intraabdominal sarcomas that recur, approximately 80-90% progress to sarcomatosis.2 Although the prognosis of sarcomatosis is believed to be poor, the overall five year survival rates and time to development have not been fully elucidated. More important, no clearly defined prognostic features have been shown to distinguish patients who will experience accelerated progression of disease from those who will have a more indolent course that will allow a window of opportunity for therapeutic intervention.

Sarcomatosis varies in the volume of tumor, the presence of ascites, and the distribution of tumor within the abdomen. However, disease characterized by regional intraabdominal dissemination of sarcoma carries a poor prognosis and is nearly always incurable.3, 4 The poor survival rate has led most surgeons to suggest nonoperative measures or supportive care in nearly all cases.5 Surgical intervention is usually reserved for gastrointestinal obstruction, bleeding, perforation, or, occasionally, for the control of ascites.

Recently, investigators have used intraperitoneal hyperthermic peritoneal perfusion in an attempt to control the disease and possibly extend survival times.6 Others have found that the use of tumor debulking and postoperative intraperitoneal chemotherapy lowered the rate of disease recurrence in certain patients.7 Owing to its high morbidity, intraperitoneal chemotherapy requires careful patient selection. Studying the natural course and identifying prognostic features of this disease may reveal which patients are best suited for this aggressive therapy. We performed a retrospective analysis of patients diagnosed with sarcomatosis who were admitted to The University of Texas M.D. Anderson Cancer Center between June 1996 and June 1999.

PATIENTS AND METHODS

The records of all patients presenting from June 1996 to June 1999 at The University of Texas M.D. Anderson Cancer Center with a diagnosis of disseminated intraperitoneal sarcoma were retrospectively reviewed. A soft tissue sarcoma pathologist gave histologic confirmation of sarcoma for all patients. One patient was excluded from the study because his intraabdominal tumor volume could not be accurately assessed. Fifty one patients were included in the evaluation. For each patient, pathologic characteristics of tumor grade and histology were assessed. Patient and clinical features of age, sex, tumor volume, prior treatment of sarcoma, and treatment at M.D. Anderson Cancer Center were evaluated for each patient.

Tumor volume of patients who underwent surgery at M.D. Anderson was measured at the time of surgical exploration. Stratification of patients into high volume versus low volume disease was based on previously described parameters established by both Chu et al.3 and Sugarbaker.8 In these cases, high volume disease was distinguished from low volume disease by the presence of more than 10 separate tumor nodules with a single tumor at least 5 cm in diameter or a total of more than 20 separate tumor nodules of any size. Tumor volume of patients who had undergone primary surgery elsewhere was measured using computed tomography (CT) scan analysis. In these cases, high volume disease was distinguished from low volume disease by the presence of more than 10 separate areas of tumor implantation of any size, or more than five separate areas of tumor implantation with any tumor measuring at least 5 cm in diameter.

Survival time was measured from the date of diagnosis of sarcomatosis to the date of death. All living patients had their most recent followup visit after June 1999.

Clinical and pathologic factors were evaluated by chi-square analysis. Kaplan-Meier analysis was used to evaluate survival data. A P value less than 0.05 was considered significant.

RESULTS

Patients

The 51 study patients with a diagnosis of sarcomatosis had a median age of 54 years (range, 27–70 years). Twenty five patients were men and 26 patients were women (Table 1). The median time from the initial diagnosis of sarcoma to the development of sarcomatosis was 11 months (range, 0.25–24 years). The median followup was 2.7 years (range, 0.5–26.4 years) from the initial diagnosis of sarcomatosis. Twenty two (43%) of the 51 patients died of their disease, with a median of 13 months (range, 4-42 months) from the date of diagnosis to the date of death.

Table 1. Demographics and Therapeutic Modalities in Patients with Sarcomatosis
 Patients (n = 51)Percent
  1. Chemo: chemotherapy; rad: radiation; surg: surgery.

Gender
 Male2549%
 Female2651%
Therapy
 Surgery1936%
 Surgery and chemo1224%
 Surgery and radiation816%
 Chemotherapy alone36%
 Surg and chemo and rad714%
 Other24%

Histology

All histologic diagnoses were confirmed by M.D. Anderson pathologists and are listed in Table 2. The most prevalent histologic diagnosis was leiomyosarcoma of gastrointestinal origin, also known as gastrointestinal stromal tumor (GIST), which accounted for 33 (65%) of the diagnoses. Other histologic diagnoses were uterine leiomyosarcoma, liposarcoma, malignant fibrous histiocytoma, and desmoplastic small cell tumor.

Table 2. Pathologic Features in Patients with Sarcomatosis
 Patients (n = 51)Percent
  1. GIST: gastrointestinal stromal tumor.

Histology
 GIST3365%
 Uterine leiomyosarcoma1020%
 Liposarcoma48%
 Malignant fibrous histiocytoma24%
 Small cell desmoplastic24%
Grade
 Low or intermediate1835%
 High3365%

To determine whether histology is a significant prognostic factor, the rates of survival for patients with GIST were compared with those of patients with other diagnoses. The two year survival for patients with GIST was 38%, compared with 42% for patients with other histologies (P = 0.77). There was no statistically significant difference between the two groups (Fig. 1).

Figure 1.

Kaplan-Meier survival curves for patients with gastrointestinal stromal tumors (GIST) versus other histologies.

The patients were divided into two groups based on tumor grade: the high grade group comprised patients with Grade 3 histology, and the low and intermediate grade group comprised patients with Grade 1 or 2 histology. Thirty three patients (65%) were in the high grade group, and 18 patients (35%) had intermediate or low grade tumors (Table 1). The two year overall survival rate for patients with high grade tumors was 35%, compared with 52% for patients with low or intermediate grade tumors (P = 0.29). The overall survival curves diverge at two years, showing a higher survival rate through four years for patients with low or intermediate grade disease than for patients with high grade disease (Fig. 2); however, the difference is not statistically significant.

Figure 2.

Kaplan-Meier survival analysis for patients with low or intermediate grade (histologic Grade 1 or 2) versus high grade (histologic Grade 3) sarcomatosis.

Tumor Volume

Tumor volume at the time of diagnosis of sarcomatosis was determined by surgical exploration or by CT scan. High volume disease was observed in 33 patients (65%), whereas low volume disease was noted in 18 patients (35%). The overall two year survival rate for patients with high volume disease was 24%, compared with 82% for patients with low volume disease (P = 0.008). Figure 3 shows patient survival curves based on tumor volume.

Figure 3.

Kaplan-Meier survival analysis for patients with low versus high volume sarcomatosis.

Figure 4.

Computed tomography scan showing peritoneal sarcomatosis in a patient with low volume disease. The white arrows denote tumor implants.

Figure 5.

Computed tomography showing peritoneal sarcomatosis in a patient with high volume disease. The tumor is filling the pelvis and displacing the rectum.

Adjuvant Therapy

Of the 51 patients, 32 (63%) received chemotherapy and/or radiation therapy in addition to surgery for sarcomatosis. Therapy was most often given postoperatively; only 12 patients (24%) received a component of their therapy preoperatively. Chemotherapeutic regimens varied widely and were a component of therapy in 22 patients (43%). Radiation therapy was used in 17 patients (33%). Table 1 shows the various treatment regimens and their frequency. There was no difference in survival rates between patients who received adjuvant therapies and those treated with surgery alone (P = 0.16).

DISCUSSION

We found that tumor volume was a prognostic factor for survival in patients with sarcomatosis. No other clinical or pathologic feature examined was found to have a significant effect on survival. The general prognosis for patients with peritoneal dissemination of sarcoma remains poor, as evidenced by a median survival of only 13 months in the current series. However, the patient survival times varied widely, ranging from just 3 months to over 24 years. Identification of prognostic features at the time of diagnosis would be useful for identifying patients who might benefit the most from aggressive therapies or who would be good candidates for clinical trials.

The results of the current study suggest that tumor volume has prognostic value for patients with sarcomatosis. An obvious concern is that the difference in survival between low and high volume groups is secondary to a lead-time bias rather than a true marker of tumor biology. It is logical to infer that the reason patients with low volume disease appear to have higher survival rates is that they presented earlier and were therefore followed longer. However, one argument against lead-time bias is that the four year survival rate of patients with low volume disease was better than that of patients with high volume disease, suggesting a true difference in tumor biology between the two groups.

Other studies support our observation that low volume disease is predictive of improved survival in disseminated intraperitoneal malignancy. In a study of patients with intraperitoneal dissemination of colon carcinoma, Sugarbaker8 reported that the survival rate of patients with lesions less than 5 cm in diameter was significantly higher than that of patients with lesions greater than 5 cm. Similarly, Eisenhauer et al.9 found that tumor burden (as measured by the diameter of the largest tumor and by the total number of tumors) in ovarian carcinoma patients with intraabdominal dissemination was an important prognostic factor for predicting both response to therapy and survival time.

Leiomyosarcomas, particularly those of primary gastrointestinal origin, have long been believed to carry a worse prognosis than other histologic types of intraabdominal sarcomas. This belief is based largely on the findings of Ng et al.,10 who documented that 78% of patients with GIST eventually developed peritoneal disease and that 72% developed liver metastases. However, in the current study of sarcomatosis patients, we found no association between histologic type and overall survival. This finding is not surprising because the histologic type has little effect on a tumor that has already shown biologic aggressiveness.

We evaluated tumor grade to determine whether high grade sarcomas were associated with a shorter overall survival time in the presence of sarcomatosis. Previously both Conlon et al.11 and Dougherty et al.12 showed that, compared with low grade sarcomas, high-grade GISTs are associated with decreased patient survival time. However, in these studies, survival was determined for patients presenting for the first time with the diagnosis of sarcoma. In contrast, we measured survival from the time of development of sarcomatosis. Therefore, the current results reflect survival of patients with the most aggressive forms of intraabdominal sarcoma, regardless of tumor grade.

Although the survival curve (Fig. 2) shows a trend toward improved survival for patients with intermediate or low grade tumors, the difference in survival times of these patients versus those with high grade tumors is not significant. Tumor grade is probably important in determining the prognosis of patients presenting with intraabdominal sarcoma but is likely to be of less prognostic value for patients who go on to develop sarcomatosis.

We also evaluated nonoperative therapies for impact on overall survival. The current results confirm the findings of other retrospective studies that reported that nonoperative therapies offered no survival benefit for patients with sarcomatosis. Using various regimens, investigators have studied the primary treatment of intraabdominal sarcomas with chemotherapy. A regimen of doxorubicin with dacarbazine in patients with soft tissue sarcomas showed a response rate of only 7%.13 Ifosfamide used in leiomyosarcoma patients in which previous doxorubicin based therapy failed had a correspondingly low response rate.14 Although one study showed response rates to chemotherapy as high as 12-43% in patients with gastrointestinal sarcomas,15 retrospective reviews showed that chemotherapy did not improve survival rates in patients with primary gastrointestinal sarcomas and retroperitoneal sarcomas.1, 11, 16 External beam radiation therapy is a standard component in the treatment of extremity sarcoma. However, this therapy is toxic when used within the abdominal cavity and therefore is of limited potential benefit in this group of patients.17

The current study shows that patients with a diagnosis of sarcomatosis have a poor prognosis, with a median survival of 13 months. Traditional predictors of survival in sarcoma patients, such as tumor grade and tumor histologic type, appear to have little prognostic significance for patients with sarcomatosis. However, tumor volume appears to be a valuable prognostic factor. Patients with low volume disease had a significantly higher overall survival rate (82% at two years) compared with patients with high volume disease (24% at two years). Tumor volume may prove to be an important prognostic factor for selecting patients for novel therapies, particularly because of the poor outcome with standard chemotherapy or radiation therapy regimens. The identification of prognostic factors is essential for selecting suitable candidates for clinical trials.

Acknowledgements

The authors thank Jacquilyn Valentine and Abigail Corona for article preparation and Kate Ó. Súilleabháin for editorial assistance.

Ancillary