Fax: (734) 615-2109
Use of positron emission tomography in localized extremity soft tissue sarcoma treated with neoadjuvant chemotherapy
Article first published online: 1 DEC 2004
Copyright © 2004 American Cancer Society
Volume 103, Issue 2, pages 339–348, 15 January 2005
How to Cite
Schuetze, S. M., Rubin, B. P., Vernon, C., Hawkins, D. S., Bruckner, J. D., Conrad, E. U. and Eary, J. F. (2005), Use of positron emission tomography in localized extremity soft tissue sarcoma treated with neoadjuvant chemotherapy. Cancer, 103: 339–348. doi: 10.1002/cncr.20769
- Issue published online: 5 JAN 2005
- Article first published online: 1 DEC 2004
- Manuscript Accepted: 28 SEP 2004
- Manuscript Revised: 20 SEP 2004
- Manuscript Received: 2 JUL 2004
- National Institutes of Health/National Cancer Institute. Grant Numbers: CA65537, CA76930, CA87721
- neoadjuvant therapy;
- fluorodeoxyglucose F18;
- emission-computed scan
Patients with high-grade soft tissue sarcomas are at high risk of developing local disease recurrence and metastatic disease. [F-18]-fluorodeoxy-D-glucose (FDG) positron emission tomography (PET) scans are hypothesized to detect histopathologic response to therapy and to predict risk of tumor progression in patients with various malignancies. Serial FDG-PET scans were taken to determine the correlation between FDG uptake and patient outcomes in patients receiving multimodality treatment of extremity sarcomas.
Forty-six patients with high-grade localized sarcomas were studied. The maximum standardized uptake values (SUVmax) of tumors were measured before receipt of neoadjuvant chemotherapy and again before surgery. Resected specimens were examined for residual viable tumor. Patients were followed up at least annually for evidence of local and distant recurrence of disease and survival.
Patients with a baseline tumor SUVmax ≥ 6 and < 40% decrease in FDG uptake were at high risk of systemic disease recurrence estimated to be 90% at 4 years from the time of initial diagnosis. Patients whose tumors had a ≥ 40% decline in the SUVmax in response to chemotherapy were at a significantly lower risk of recurrent disease and death after complete resection and adjuvant radiotherapy.
The FDG-PET scan was found to be a useful method with which to predict the outcomes of patients with high-grade extremity soft tissue sarcomas treated with chemotherapy. The pretreatment tumor SUVmax and change in SUVmax after neoadjuvant chemotherapy independently identified patients at high risk of tumor recurrence. The FDG-PET scan showed promise as a tool to identify the patients with sarcoma who are most likely to benefit from chemotherapy. Cancer 2005. © 2004 American Cancer Society.
Soft tissue sarcomas are a heterogeneous mix of uncommon malignancies that exhibit mesenchymal differentiation. In 2004, an estimated 8680 new cases were diagnosed and 3660 people died of the disease.1 In the majority of patients, soft tissue sarcomas arise in an extremity.2 Curative treatment of extremity sarcomas requires surgical resection, but complete removal of macroscopic tumor is sometimes inadequate because sarcomas are invasive and often have a high propensity to recur locally and spread hematogenously. The addition of adjuvant radiotherapy has been shown to improve local control without affecting the risk of distant disease recurrence or survival.3 Patient and tumor characteristics at the time of diagnosis have been studied to identify prognostic factors for disease recurrence. Tumor grade, size, and depth in relation to the muscle fascia are important determinants of disease recurrence and development of metastasis. Metastases develop in 25–50% of patients with high-grade tumors that are ≥ 5 cm in dimension.4 This increased risk of distant disease recurrence suggests that undetectable metastatic disease is present in a significant percentage of patients with large, high-grade soft tissue sarcomas at the time of surgery. Adjuvant chemotherapy after complete resection of high-grade large extremity sarcomas may improve the chance for a cure but clearly does not benefit every patient.5, 6
The assessment of soft tissue sarcoma response to chemotherapy is challenging. Only 10–30% of soft tissue sarcomas respond to neoadjuvant doxorubicin-based regimens as assessed by conventional radiographic imaging,7–11 and to our knowledge it has not been well established whether a significant change in the size of the tumor mass is a meaningful surrogate of patient outcome. Standard radiographic response determined by a change in the size of the tumor mass to preoperative chemotherapy has not correlated consistently with histologic response or with disease-free or overall survival.7–9, 11 Therefore, other methods to identify patients with limited, potentially curable disease who are likely to benefit from chemotherapy would be useful.
Intermediate-grade and high-grade sarcomas accumulate and metabolize the glucose analog, [F-18]-fluorodeoxy-D-glucose (FDG), to a greater extent than normal tissues, many benign tumors, and low-grade sarcomas presumably because higher grade sarcomas have increased energy requirements to support more rapid proliferation.12, 13 The calculated standardized uptake value (SUV) correlates with the metabolic rate of accumulation of FDG.14 Therefore, the SUV of higher grade sarcomas may be a readily measurable feature that could be a surrogate measure of tumor viability during therapy. Changes in FDG accumulation after neoadjuvant chemotherapy in small studies have correlated with histologic response in pediatric bone sarcomas,15 breast carcinoma,16, 17 head and neck carcinoma,18 adenocarcinoma of the esophagus,19 and gastric carcinoma.20 More importantly, tumor response determined by a reduction in the FDG SUV of gastrointestinal stromal tumors,21 adenocarcinoma of the esophagus,19 and gastric carincoma20 treated with chemotherapy shows promise as an early indicator of improved disease progression-free and overall survival. The pretreatment tumor FDG maximum standardized uptake values (SUVmax) and changes in FDG SUVmax after neoadjuvant doxorubicin-based chemotherapy in adolescents and adults with extremity, high-grade soft tissue sarcomas imaged at the University of Washington Medical Center (Seattle, WA) were reviewed to determine whether tumor FDG SUVmax correlates with histopathologic findings at the time of surgery, risk of local and metastatic tumor recurrence, and patient survival.
MATERIALS AND METHODS
The current study is an analysis of a subset of patients presenting to the University of Washington Medical Center or Children's Hospital and Regional Medical Center with a suspected or known soft tissue sarcoma who were enrolled prospectively in a study of FDG-positron emission tomography scans (FDG-PET). Written informed consent was obtained from all patients before their participation in the PET scan study. The study was approved by the University of Washington Medical Center Institutional Review Boards of Human Subjects and Radiation Safety in accordance with institutional and federal guidelines. All of the patients referred to the Sarcoma Service at the University of Washington Medical Center with localized, intermediate-grade soft tissue sarcomas ≥ 5 cm in greatest dimension or with high-grade soft tissue sarcoma irrespective of size were offered doxorubicin-containing chemotherapy if it was not contraindicated by concurrent medical conditions. The data for patients with localized, intermediate-grade, or high-grade soft tissue sarcoma involving the extremity who underwent chemotherapy and consented to the FDG-PET scan study were analyzed. Patients presenting with metastatic disease were excluded. The histologic features of the tumor were reviewed by one pathologist (B.P.R.) and graded according to the Federation Nationale des Centres de Lutte Contre le Cancer (FNCLCC) system.22 The extent of sarcoma was staged according to the American Joint Committee on Cancer.23 Patients were treated with a combination of doxorubicin and cisplatin24; doxorubicin and ifosfamide25; or cyclophosphamide, doxorubicin, and vincristine alternating with ifosfamide and etoposide.26 Briefly, 60–90 mg/m2 (median dose of 75 mg/m2) of doxorubicin, 120 mg/m2 of cisplatin, and 8–9 g/m2 of ifosfamide were administered per cycle. All patients underwent complete resection of macroscopic disease. Tumor was present at the microscopic margin in 8 of 46 patients. Thirty-seven of 46 patients received adjuvant radiotherapy after the completion of chemotherapy and 9 patients declined radiotherapy for personal reasons. Preoperative radiotherapy was not used.
PET Scan Imaging
Methods for PET scan imaging of sarcomas have been published previously.12 In all cases, the pretherapy FDG-PET scan was done within 34 days of the date of diagnosis. The median interval from diagnosis to the baseline FDG-PET scan was 7 days. Patients fasted for ≥ 6 hours before an intravenous injection of 7–10 mCi of FDG over 2 minutes. Blood glucose level was measured before administration of FDG, was recorded for 41 of the 46 patients, and was < 120 mg/dL in all but 1 patient. Intrapatient blood glucose levels before injection of FDG varied < 20% in 33 patients. After a 45-minute equilibration period during which the patient rested, emission and transmission scans were obtained over the known site of tumor using a GE Advance Positron Tomograph (General Electric, Waukesha, WI). The same dedicated PET scanner was used throughout the study period. Typically, the tumor image was captured in 2 adjacent 15-cm fields of view. Emission scans were attenuation corrected. Reconstructed data were rendered into 3-dimensional images using a Hanning filter at a resolution of 4.2 mm. Three-dimensional image sets were available for review in slice thicknesses of 4.2–12.0 mm. Regions of interest for determination of tumor SUV were drawn around the area of tumor uptake on the workstation using magnetic resononace imaging and/or computed tomography scan images for reference. The tumor SUV was calculated automatically by the tomography software. After careful assessment of the SUV values throughout the tumor volume, the maximum tumor SUV was recorded for analysis.
Excised tumor specimens were processed in the standard manner for soft tissue sarcoma at our institution. One section per cm of tumor tissue was submitted for histologic examination. The pathologist was unaware of the change in SUVmax or patient outcome. The amount of viable tumor specimen (expressed as a percentage) was determined semiquantitatively by subtracting the percent of necrotic and fibrotic areas from the tumor mass. For each section, necrosis and fibrosis were estimated based on the aggregate area of necrosis and fibrosis divided by the area of tumor on the slide. The total amount of necrosis and fibrosis was obtained by dividing the sum of the percentages by the number of sections examined. Areas of necrosis were characterized by ghosts of tumor cells and debris without crisp hematoxylin staining. Areas of fibrosis were characterized as containing bland-appearing fibroblasts and/or hyalinized, collagenous tissue. These areas frequently contained hemosiderin and/or foamy macrophages.
Analyses were performed using Stat View (Version 5.0.1) for Macintosh (SAS Institute, Cary, NC). Comparisons between factors were achieved using the chi-square test, the Wilcoxon signed rank test, and the Kruskal–Wallis test. The date of diagnosis was the starting point for survival analysis. The date of last medical evaluation was used as the censure time for disease-free and metastasis-free survival, and the date of last contact was used as the censure time for overall survival. The most recent date of last follow-up was June 1, 2004. Survival was analyzed using the Kaplan–Meier method with differences compared by the log-rank test.27 Multivariate analysis of relative risks with 95% confidence intervals was estimated with the Cox proportional hazards model.28 Significance was set at P ≤ 0.05. Survival analyses were performed using a baseline SUVmax of < 6 versus ≥ 6 based on our previous study of bone and soft tissue sarcomas.29 A reduction in the SUVmax of < 40% versus ≥ 40% was chosen for survival analyses because this cutoff value was higher than the median and mean changes but did not exclude more than two-thirds of the patients from the group considered to have a metabolic response to chemotherapy. Assistance with the statistical analyses was provided by the University of Washington Department of Statistics.
Forty-seven patients with localized, intermediate-grade, or high-grade extremity soft tissue sarcomas entered the study between August 1994 and August 2001 and received serial tumor FDG-PET scan imaging. One patient was excluded from the analysis because chemotherapy was not administered. The clinical characteristics of the remaining 46 patients (Table 1) are similar to previous reports of preoperative chemotherapy for extremity soft tissue sarcoma.7, 11 The 3 most common histologic subtypes of sarcoma in this series (i.e., pleomorphic undifferentiated sarcoma/malignant fibrous histiocytoma, synovial sarcoma, and leiomyosarcoma) accounted for 76% of the cases. The rhabdomyosarcomas were pleomorphic and spindle cell variants (one of each). Three of the 4 liposarcomas were myxoid/round cell variants with a > 40% round cell component and the fourth was a pleomorphic liposarcoma. Twenty-eight tumors were FNCLCC Grade 2 and 18 tumors were FNCLCC Grade 3. In all 5 patients with Stage II disease (Grade 3 T1bN0M0, (i.e., high grade [Grade 3], deep tumor ≤ 5 cm in greatest dimension [T1b], no regional lymph node metastasis [N0], no distant metastasis [M0])), the tumor was ≥ 4cm in greatest dimension. Nine patients received 2 cycles of a doxorubicin-based chemotherapy regimen between the baseline FDG-PET scan at the time of diagnosis of soft tissue sarcoma and the presurgical FDG-PET scan, 29 patients received 3 cycles, and 8 patients received 4 cycles of chemotherapy. Patients treated early in the study period received the doxorubicin/cisplatin combination, whereas patients treated in later years received the doxorubicin/ifosfamide combination. Both patients with rhabdomyosarcoma and 4 of the 12 patients with synovial sarcoma received cyclophosphamide, doxorubicin, and vincristine, with ifosfamide and etoposide given in alternating cycles. Patients received a median of seven cycles of chemotherapy including preoperative and postoperative treatments.
|Mean age (yrs) (range)||47 (10–73)|
|Pleomorphic undifferentiated sarcoma||12|
|Malignant peripheral nerve sheath tumor||3|
|AJCC tumor stage at diagnosis|
|Mean tumor size (cm) (range)||10.8 (4.0–21)|
|Median no. of therapy courses between PET scans (range)||3 (1–4)|
|Median no. of therapy courses from 2nd PET scan to surgery (range)||0 (0–1)|
Tumor FDG Standard Uptake Values
The majority of patients had sarcomas with high levels of FDG accumulation. The baseline median and mean SUVmax of the sarcomas were 6.8 and 8.0, respectively (range, 2.0–40.3) and the baseline SUVmax was ≥ 6 in 25 patients (Fig. 1). The median and mean SUVmax of the chemotherapy-treated sarcomas were 4.2 and 5.4, respectively (range, 1.4–17.2). The changes in SUVmax ranged from a reduction of 88% to an increase of 177% (the median change was a reduction of 35%; Fig. 1). Three patients had an increase in SUVmax of > 20% (26%, 116%, and 177%, respectively) after 2, 2, and 3 cycles of chemotherapy, respectively, all of whom had a baseline SUVmax < 6. There was no significant difference noted in the change in SUVmax among the groups receiving the three different chemotherapy regimens. The change in SUVmax of tumors did not correlate with the total dose of neoadjuvant doxorubicin received or the number of cycles of chemotherapy delivered between the baseline and preoperative FDG-PET scans.
The histologic findings of sarcoma after treatment with chemotherapy correlated with the observed change in SUVmax of the tumor. Resection of tumor occurred a median of 3 days after the posttreatment PET scan. In 37 (80%) patients, the tumor was excised within 2 weeks of the postchemotherapy PET scan. Macroscopic total surgical resection of the sarcoma was obtained in each patient. Surgical margins were microscopically free of tumor (R0 resection) in 38 patients and contained tumor (R1 resection) in 8 patients. The percent residual viable tumor in the resected mass was determined visually by the pathologist. The mean percent residual tumor viable was 47% (range, 0–100%). In 13 patients, there was < 10% residual viable tumor after preoperative chemotherapy. A correlation was detected between the change in SUVmax and the amount of residual viable tumor in the excised mass (P = 0.001). A > 40% reduction in tumor SUVmax also correlated with < 10% residual viable neoplasm in the resected mass (P = 0.05).
The sarcoma SUVmax was a strong predictor of disease recurrence and the development of metastasis. With a median follow-up of 46 months (range, 24–108 months) from the time of diagnosis of patients without disease recurrence, 24 (52%) patients remain free of sarcoma, 18 (39%) had distant disease recurrence, and 4 (9%) had isolated local disease recurrence as the first event. Two patients had simultaneous local and distant disease recurrences. The risk of developing metastasis correlated (P = 0.03) with a baseline tumor SUVmax ≥ 6 (Fig. 2), suggesting that sarcomas with a high SUVmax are biologically more aggressive. The baseline SUVmax was not significantly different between FNCLCC Grade 2 and Grade 3 tumors, and tumor grade did not correlate with disease-free survival in the current series that excluded patients with low-grade sarcomas. Other pretreatment characteristics including patient age ≥ 50 years, tumor size ≥ 10 cm, and tumor located in the lower extremity did not correlate with disease-free (P = 0.57, 0.10, and 0.12, respectively) or distant metastasis-free (P = 0.85, 0.14, and 0.22, respectively) survival.
A change in the sarcoma SUVmax was the only factor dependent on chemotherapy that correlated with disease recurrence. Patients with a ≥ 40% decrease in tumor SUVmax (defined as a metabolic response of tumor in the current article) had a significantly lower risk of disease recurrence (P = 0.01; Fig. 3) and of metastasis (P = 0.02), whereas < 10% residual viable tumor did not correlate (P = 0.79) with disease-free survival. Significant differences in disease recurrence and development of metastasis were also observed when tumors with a baseline SUVmax ≥ 6 were analyzed separately (Fig. 4). Metastasis developed in 11 of 13 patients with a baseline tumor SUVmax ≥ 6 and a < 40% reduction in SUVmax to preoperative chemotherapy compared with 4 of 12 patients with a baseline tumor SUVmax ≥ 6 and a ≥ 40% decrease in tumor SUVmax after chemotherapy (P = 0.006; Table 2). If the 2 patients with Stage II disease were excluded from the analysis, the risk of disease recurrence and development of metastasis remain significantly greater for patients with a reduction in tumor SUVmax of < 40% (P = 0.009 and P = 0.01, respectively). By multivariate analysis using Cox regression, both an initial SUVmax ≥ 6 and a < 40% decrease in tumor SUVmax were associated independently with an increased relative risk of disease recurrence and metastasis (Table 3). Metastasis developed in the lungs, regional lymph nodes, bone, distant soft tissue, and brain in nine, four, two, two, and one patient, respectively, as the initial site of disease recurrence.
|FDG SUVmax values||No. of disease recurrences by site|
|Local||Distant||Local and distant||Total (% of patients)|
|Baseline SUVmax < 6|
|≥ 40% decrease||0||0||0||0/5 (0)|
|< 40% decrease||2||4||1||7/16 (44)|
|Baseline SUVmax ≥ 6|
|≥ 40% decrease||0||4||0||4/12 (33)|
|< 40% decrease||2||8||1||11/13 (85)|
|Factor||Disease recurrence-free survival||Metastasis-free survival||Overall survival|
|Relative risk||95% CI||P value||Relative risk||95% CI||P value||Relative risk||95% CI||P value|
|Initial SUVmax ≥ 6||3.2||1.3–8.2||0.015||3.7||1.4–9.8||0.008||1.5||0.7–4.0||0.41|
|No metabolic response (reduction of SUVmax < 40%)||5.6||1.8–17.8||0.004||5.3||1.6–17.0||0.005||4.4||1.2–15.9||0.026|
A significant increase in tumor SUVmax did not portend systemic disease progression in all patients. Three patients had an increase in the tumor SUVmax of > 20% from baseline and were considered to have progressive disease by FDG-PET scans. All 3 patients had a baseline tumor SUVmax < 6. At the time of last follow-up, 2 of the patients were alive without disease 40 months and 100 months, respectively, after diagnosis, and the third patient developed pulmonary metastasis 10 months after diagnosis and died of disease. It is not known why two of the patients with metabolically progressive disease had good outcomes.
Local disease recurrence was infrequent but presaged systemic disease recurrence. Local disease recurrence occurred 13 months after tumor resection in a patient without tumor present at the margin (R0 resection) who received adjuvant radiotherpy; 36 months after resection in a patient with tumor present at the margin (R1 resection) who received adjuvant radiotherapy; and 14 months and 19 months, respectively, after resection in 2 patients without tumor present at the margin (R0 resection) who did not receive adjuvant radiotherapy. All 4 patients with local disease recurrence as the initial event had a < 40% reduction in tumor SUVmax after chemotherapy. Accounting for all patients with local disease recurrence (including the two patients with concurrent local and systemic disease recurrence), disease recurrence was found to be correlated with lack of adjuvant radiotherapy (P = 0.04) but not with the presence of tumor at the microscopic margin (R1 resection; P = 0.27). Metastasis developed in all 4 patients with local disease recurrence at 7, 8, 12, and 16 months after initial disease recurrence.
A change in tumor SUVmax, but not the baseline tumor SUVmax or histologic response, predicted survival in patients with high-grade extremity soft tissue sarcoma treated with preoperative chemotherapy. With a median follow-up of 47 months (range, 29–108 months) from the time of diagnosis of surviving patients, 29 (63%) patients were alive, 16 had died of disease recurrence of sarcoma, and 1 patient without disease recurrence died of an adenocarcinoma of the esophagus. A reduction in the presurgery sarcoma SUVmax of ≥ 40% relative to the baseline SUVmax correlated with overall survival (P = 0.02; Fig. 3). Of 17 patients with a ≥ 40% reduction in tumor SUVmax, 1 died of leiomyosarcoma metastatic to the lungs, 1 died of leiomyosarcoma metastatic to the brain, and 1 died of a cause unrelated to sarcoma. Conversely, 14 of 29 patients with a < 40% reduction in tumor SUVmax had died of disease at the time of last follow-up. The median overall survival for patients with a < 40% reduction in tumor SUVmax was 42 months compared with not yet reached for patients with a reduction in tumor SUVmax of > 40%. The correlation was not significant (P = 0.06) when the five patients with Stage II disease were removed from the analysis. However, 2 of the 3 patients with Stage II disease and a < 40% reduction in tumor SUVmax had died of disease at the time of last follow-up, whereas both Stage II patients with a ≥ 40% reduction in tumor SUVmax were alive without disease. A baseline tumor SUVmax ≥ 6 did not correlate with an increased risk of death (Table 3). Overall survival did not correlate with residual viable neoplasia (> 10% vs. ≤ 10%), FNCLCC tumor grade (Grade 2 vs. Grade 3), size of tumor at diagnosis (< 10 cm vs. ≥ 10 cm), tumor location (upper vs. lower extremity), or chemotherapy regimen used (P = 0.62, 0.16, 0.90, 0.08, and 0.52, respectively). When patients with a baseline tumor SUVmax ≥ 6 were analyzed separately, the risk of death was significantly (P = 0.05) greater for patients with a < 40% reduction in tumor SUVmax (Fig. 4). The median overall survival for patients with a < 40% reduction in tumor SUVmax was 39 months versus not yet reached for patients with a ≥ 40% reduction in tumor SUVmax after chemotherapy.
The current study shows that serial FDG-PET scans identify patients with intermediate-grade or high-grade soft tissue sarcomas of the extremity who are at high risk for distant disease recurrence after doxorubicin-based chemotherapy and complete surgical resection of the tumor mass. A metabolic response as determined by FDG-PET scans correlated with a pathologic response to preoperative doxorubicin-based chemotherapy but was a stronger determinant of distant recurrence of disease than the pathologic response. More importantly, a metabolic response correlated with an improved overall survival in all patients and in the subset of patients at a much higher risk of developing metastasis.
Differences in the clinical characteristics of patients and the treatments administered may limit the application of these findings to the general clinical practice of sarcoma management. However, a strength of the current analysis of 48 patients is that it is to our knowledge the largest series published to date to examine changes in FDG-PET scan after neoadjuvant chemotherapy in extremity soft tissue sarcoma. Previous studies of FDG-PET scans in monitoring soft tissue sarcoma response to neoadjuvant therapy were conducted on a limited number of patients and conclusive results could not be generated.30 Patients included in the current analysis were entered prospectively into a study of FDG-PET scans in sarcomas. However, standardized treatment was not required and information regarding chemotherapy and radiotherapy was gathered retrospectively. The lack of consistent treatment may have introduced bias, but a correlation between the chemotherapy regimen used and risk of disease-free or overall survival was not found. The majority of patients received 3 cycles of chemotherapy between the baseline PET and the presurgical PET scans, whereas 35% of patients received either 2 or 4 cycles of chemotherapy. It is unlikely that a difference in the number chemotherapy cycles administered between PET scans biased the results significantly. In six patients, three PET scans were performed: a scan at diagnosis, a scan after two cycles of chemotherapy, and a scan after three or four cycles of chemotherapy. In all six patients, there was no difference in the tumor SUV response (< 40% vs. ≥ 40%) after 3 or 4 cycles of chemotherapy compared with the tumor SUV response after 2 cycles of chemotherapy. The total number of chemotherapy cycles administered to patients also was not consistent among all patients, but most of the patients received five or more cycles. Only two patients were known to receive fewer than five cycles of chemotherapy.
One important caveat is that we chose to limit the current study to patients diagnosed with high-grade sarcoma in the extremity because data supporting the use of chemotherapy in patients with localized disease are most convincing in this group of patients.5 Data regarding the risk of disease progression and patient survival in relation to FDG SUV changes in visceral and retroperitoneal sarcomas have not been reported previously and were not included in our analysis. Therefore, caution should be used in extrapolating our results to soft tissue sarcomas involving sites other than the extremity and proximal limb girdle.
Computation of the tumor SUVmax is a relatively simple task using tomography software. Intrapatient variability in the tumor SUVmax has been shown to be < 20% when serial scans are done using the same machine and imaging protocol.31 By defining a sarcoma FDG metabolic response as a decline in the tumor SUVmax of ≥ 40%, inaccuracies in the measurement technique and spontaneous variability in the accumulation of FDG should not have influenced the assessment of response substantially. The FDG SUV is hypothesized to be proportional to the proliferative rate of the neoplastic cells.32 The SUVmax of tumors was analyzed because we believed that the most active portion of tumor would dictate the biologic aggressiveness of the lesion. We reasoned that if a portion of the mass contained viable sarcoma resistant to chemotherapy, the corresponding area would remain metabolically active, could be identified pathologically, and would influence the risk of tumor metastasis.33 Other metabolic characteristics, such as heterogeneity in FDG accumulation, may prove to be predictive of tumor behavior.
The risk of distant recurrence of disease is higher in patients with a baseline tumor SUVmax ≥ 6. We reported previously that a baseline SUVmax ≥ 6 was associated with a lower likelihood of disease-free survival and overall survival for patients with low-grade to high-grade soft tissue sarcomas involving any site.29 One possible influence on the earlier results is a lower SUVmax of low-grade sarcomas compared with high-grade sarcomas and the better prognosis of low-grade tumors. In the current study, low-grade sarcomas were not included. Moreover, FNCLCC Grade 2 tumors did not differ from FNCLCC Grade 3 tumors in the baseline SUVmax. The 5-year estimated distant metastasis-free survival rate among patients with a baseline tumor SUVmax ≥ 6 is 32%. This is lower than the 5-year distant metastasis-free survival rate of 75% reported among patients with a baseline tumor SUVmax < 6 and of 52% in a larger series of patients with Stage III (Grade 3 T2bN0M0 i.e., high-grade [Grade 3], deep tumor > 5 cm in greatest dimension [T2b], no regional lymph node metastasis [N0], and no distant metastasis [M0]) extremity soft tissue sarcomas who are treated with chemotherapy.7 To our knowledge the biologic connection between increased uptake and/or accumulation of FDG and distant spread of sarcoma is not known, but FDG uptake has been shown to correlate with tumor proliferation.32
In our selected group of patients with intermediate-grade to high-grade large extremity soft tissue sarcomas, the change in the maximum accumulation of FDG in sarcomas closely correlated with survival. The projected 5-year overall survival rate for patients with a ≥ 40% response in tumor SUVmax after chemotherapy is 76% compared with 40% for patients with a < 40% reduction in SUVmax. The pathologic response to neoadjuvant chemotherapy did not correlate with patient outcome. Although a correlation was detected between the tumor SUV response and the pathologic response, the tumor SUV response was more significant in stratifying the risk of disease recurrence and death from progression of disease. The findings in the current study suggest that the sarcoma FDG SUV at the time of diagnosis is predictive of the risk of disease recurrence after complete resection and that changes in tumor SUV in response to chemotherapy can distinguish those patients who are likely to benefit from systemic therapy from those who will not.
The noninvasive nature of the PET scan and its ability to detect early responses to therapy make it an ideal modality with which to evaluate the potential effectiveness of antitumor treatments if changes in SUV are predictive of patient outcomes. Additional prospective studies of PET scans in intermediate-grade and high-grade sarcomas treated with a current standard chemotherapy regimen are needed to confirm these results and to assess the validity of this approach in the management of soft tissue sarcomas. Information regarding the response to chemotherapy of a patient with a sarcoma could be used to select those patients for whom continued treatment is likely to be of benefit from those whose tumors are unlikely to be affected substantially by standard chemotherapy. The latter group could be spared the toxicities of chemotherapy and be considered for observation or novel treatments.
- 23AJCC cancer staging manual, 5th edition. Philadelphia: Lippincott-Raven, 1997., , , et al.
- 24Neoadjuvant chemotherapy of soft tissue sarcoma: clinical and pathologic correlates of response [abstract]. Proc Am Soc Clin Oncol. 1993; 12: 469a., , , , , .
- 28The statistical analysis of failure time data. New York: Wiley, 1980., .