FDG-PET for the diagnosis and management of soft tissue sarcoma
John P.A. Ioannidis, MD
Joseph Lau, MD
AHRQ Technology Assessment
New England Medical Center EPC
Contract No. 290-97-0019
April 5, 2002
Abstract
Purpose: To perform an assessment on the use of 18-fluorodeoxyglucose positron emission tomography (FDG-PET) in the diagnosis, grading, and management of soft tissue sarcoma (STS).
Materials and Methods: We performed a systematic review of relevant studies identified in Medline and Embase. The diagnostic and grading performance were evaluated for four diagnostic criteria: qualitative visualization; standard uptake value (SUV) with cut-offs of 2.0 and 3.0; and metabolic rate of glucose (MRG) with cut-off of 6.0 µmol/100g/min.
Results: Twenty studies were included in this systematic review. By qualitative interpretation, the sensitivity of FDG-PET ranged from 91 to 100% and the specificity varied between 26 to 88% in diagnosing primary lesions. When an SUV cut-off of ≥2.0 was used to evaluate primary lesions, the sensitivity ranged between 64 to 100% and the specificity ranged between 71 to 100%. Diagnostic performance was similar for primary and recurrent lesions. Limited data on comparisons with MRI and CT scan showed no differences between FDG-PET and these imaging modalities in diagnosing recurrent and metastatic disease.
Conclusion: FDG-PET has very good discriminating ability in the evaluation of both primary and recurrent soft tissue lesions. FDG-PET may be helpful in tumor grading, but offers inadequate discrimination between low-grade tumors and benign lesions. There is insufficient data on the impact of FDG-PET on clinical outcomes and on the usefulness of FDG-PET in assessing the response to therapy.
Background
Soft tissue sarcoma (STS) is a relatively rare group of tumors that represents about 1% of all cancers diagnosed each year in the United States. The prognosis for many of these tumors is poor and about 40% of patients with STS die from distant metastatic disease. Early diagnosis, accurate grading and staging, and early detection of recurrence and metastatic disease offer hope for improved outcomes. Tumor grade is a strong predictor of eventual outcome with a poor prognosis for high- and intermediate-grade tumors and a more indolent course for low-grade tumors that have been excised. Although no large randomized trials are available, current data suggest that good response to chemotherapy is possible even for aggressive tumors. Diagnosing and staging of STS is complicated by the fact that tissue diagnosis may sometimes be inaccurate, since these tumors can exhibit substantial histological variability and grading may not be straightforward. Therefore, accurate imaging modalities that can provide reliable information on tumor grade and staging would be desirable. Computed tomography (CT) and magnetic resonance imaging (MRI) have been used for these purposes.
Positron emission tomography (PET) is increasingly being used to diagnose, grade, and stage diverse types of tumors and to assess tumor response to therapy. PET is generally performed with 18F-fluorodeoxyglucose (FDG) administered intravenously to patients who have fasted overnight or for several hours. Scans may be performed on the whole-body or focused on specific areas. FDG-PET results can be interpreted according to various protocols and parameters. These include: a) simple visualization of a lesion based on subjective appearance; b) a qualitative assessment of visualization, where specific features, such as heterogeneous and multifocal uptake are considered to be suggestive of malignancy, while simple visualization is not; c) evaluation of the tumor to background ratio (TBR), where glucose uptake is compared between the suspect lesion and some surrounding tissue, typically muscle; d) evaluation of a standardized measure of positron emission, typically the standardized uptake value (SUV), that corrects for the administered dose and the weight of the patient as well as the plasma glucose level; and e) evaluation of dynamic parameters with several measurements performed over a period of time during which one can estimate the glucose metabolic rate of the suspect lesion.
Requests by CMS
CMS requested assistance from AHRQ to perform an assessment on the use of FDG-PET to diagnose, grade, and manage STS.
The following questions were formulated for this report:
1. What is the diagnostic test performance (sensitivity and specificity) of FDG-PET for:
- distinguishing benign lesions from malignant soft tissue sarcoma
- distinguishing low grade from high grade soft tissue sarcoma
2. How does the test performance of FDG-PET compare with conventional anatomic imaging (CT, MRI, etc) among patients with soft tissue sarcoma with respect to:
- primary diagnosis
- diagnosing locoregional recurrence
- diagnosing distant metastasis.
3. A review of studies on changes in patient management or improved outcomes for patients with soft tissue sarcoma with the use of FDG-PET
4. A review of studies on using FDG-PET to determine tumor response to therapeutic interventions for patients with soft tissue sarcoma
Method
Scope of the literature
In November 2001, we searched Medline and EMBASE for FDG-PET and STS studies. In addition, we perused the references of retrieved articles for additional studies. The search, using various search terms for FDG-PET, exploding sarcoma, and using the names of specific histologic types of STS yielded about 200 English language abstracts on human subjects (Table 1, [PDF, 993KB]). Using a low threshold for the purpose of screening abstracts, we identified 35 potentially relevant articles that might have addressed one of the research questions. We retrieved the full text of these articles and evaluated them for eligibility. Several included only 3-5 patients with STS. We rejected single case reports as well as articles that were deemed irrelevant to the research questions upon close scrutiny. Fifteen articles were rejected at this stage. Twenty articles were considered eligible for the research questions and had more than one patient with STS. These articles are Adler (1990), Dimitrakopoulou-Strauss (2001), Eary (1998), El-Zeftawy (2001), Ferner (2000), Folpe (2000), Griffeth (1992), Hain (1999), Jones (1996), Kern (1988), Kole (1997), Lodge (1999), Lucas (1998), Lucas (1999), Nieweg (1996), Schulte (1995), Schulte (1999), Schwarzbach (2000), van Ginkel (1996), and Watanabe (2000).
Of the 20 articles, there were two pairs with probably largely overlapping study populations (as inferred by overlap of authors, eligibility criteria, chronological proximity). Data were extracted from each of the articles, but only one from each pair was considered in the data synthesis. Of the two articles by Dimitrakopoulou-Strauss (2001) and Schwarzbach (2000), only the latter is included in the quantitative synthesis, since the former provided unclear data. Of the two articles by the team of Folpe (2000) and Eary (1998), data are used accordingly once in each analysis, as pertinent. This study team combined data for both bone and soft tissue lesions that could not be separated. We contacted the authors but were unable to obtain more data. We did not include mixed data in this report. There were three publications by the same Netherlands team (Nieweg 1996, van Ginkel 1996, Kole 1997). All three studies could be included in answering the research questions since the study populations were different in the Nieweg (1996) and Kole (1997) reports. Despite a partial overlap in the study population of Nieweg (1996) and van Ginkel (1996), the questions addressed in each report were different. Finally, investigators from the St. Thomas team had published five reports [Hain (1999), Ferner (2000), Lodge (1999), Lucas (1998) and Lucas (1999)], but they dealt largely with non-overlapping subjects and provided complementary information.
Data extraction
We extracted from each study information on the publication year, general characteristics (including location, specialties involved, age, race, total number of subjects and lesions evaluated, number of malignant and benign lesions, number of participating sites and location), inclusion criteria, exclusion criteria, procedures performed (including technical characteristics for FDG-PET and other imaging studies as well as information on whether evaluators were blinded), study design (prospective, retrospective, unclear), tumor types included and types of benign lesions, imaging parameters considered for diagnostic inferences and whether cut-offs had been pre-specified or not, definition and interpretation of reference standard (biopsy, operative histological diagnosis, other imaging, clinical, other, unspecified) as well as the potential for verification bias. Verification bias (also called work-up bias) refers to incomplete verification of the results of the test-under-investigation with the reference test.
For each report, we also extracted data on the diagnostic performance pertaining to the relevant research questions. In addition to data on pre-specified cut-offs for various FDG-PET parameters, we extracted information on the number of true positives, false positives, true negatives and false negatives for FDG-PET in diagnosing malignant vs. benign lesions using the following pre-specified definitions: a) visualization (simple or qualitative interpretation, if no data on simple visualization available); b) SUV ≥ 2.0; c) SUV ≥ 3.0; d) glucose metabolic rate ≥ 6 micromol/100g/min. To extract the relevant information we combined information presented in the text, tables and figures in the published reports to derive exact counts, whenever possible. We also recorded the type of lesions that were false positive or false negative diagnoses, according to the various key diagnostic criteria of FDG-PET. Furthermore, whenever information was provided on tumor grade, we recorded the number of lesions that were positive by FDG-PET (based on each of the above standardized definitions), separately for high/intermediate (G III/II) grade tumors, low-grade (G I) tumors, non-inflammatory benign lesions, and inflammatory benign lesions (including active infection, myositis ossificans, pigmented villonodular synovitis, fasciitis, non-specified inflammation).
For studies where all patients were evaluated with both FDG-PET and CT or MRI, we evaluated the diagnostic performance of each test in diagnosing primary disease, local recurrence, and distant metastasis. Data were compared on the same patients where two imaging procedures were performed. We recorded the number of patients correctly diagnosed with only one of the two imaging procedures.
For reports focusing on the impact of FDG-PET in the management of patients with STS, we extracted information regarding the diagnostic accuracy and impact of FDG-PET on the management of each patient.
Data presentation and analysis
We summarized the sensitivity and specificity of studies that addressed each of the questions as well as the prevalence of malignant disease in each study. Data were combined to provide summary information across studies, whenever feasible.
Table 2 [PDF, 993KB] (Parts I, II, and III) reports the detailed information extracted from each of the 20 eligible articles pertinent to the research questions. Eighteen articles provided data for question 1. There were zero, two, and three articles pertinent to questions 2A, 2B, and 2C, respectively. Four articles provided data to questions 3 and 4.
Results
Question 1. What is the diagnostic test performance (sensitivity and specificity) of FDG-PET for:
a. distinguishing benign lesions from malignant soft tissue sarcoma
b. distinguishing low grade from high grade soft tissue sarcoma
1a. Distinguishing benign lesions from malignant soft tissue sarcoma
Tables 3 to 6 [PDF, 993KB] summarize the available data for the diagnostic performance of FDG-PET to discriminate between malignant and benign soft tissue lesions. The studies that addressed this question are remarkably different in terms of the prevalence of malignant disease in their population (13 to 88%). Data pertain both to primary lesions as well as evaluations for recurrent lesions. The sensitivity and specificity depend on the criteria used to define positive FDG-PET result.
Visualization method
When the positive FDG-PET result is defined on the basis of visualization of the lesion without quantitative assessments (Table 3, [PDF, 993KB]), sensitivity is typically 100% (with one exception of 91% and another of 95%) in diagnosing primary lesions, while it varies between 74 to 93% in the three studies of recurrent lesions with at least 5 tumors included. Specificity varies between 26 to 88% for primary lesions, but is consistently very high (92 to 94%) in the three studies of recurrent lesions that include at least five benign lesions. This means that for primary lesions, almost all tumors are visualized, but a variable, potentially high, proportion of benign primary lesions may also be visualized. The variable specificity may depend on the case-mix of the benign lesions, but also on the definition of positive visualization that may vary across studies. The criteria for visualization seem to be subjective in most studies. Conversely, for evaluation of recurrence, most recurrent tumors are visualized, while very few benign lesions are visualized.
Semi-quantitative SUV values
When positive FDG-PET is defined on the basis of semi-quantitative SUV values, the sensitivity and specificity depend on the cut-off used. Typically, studies have not used pre-specified cut-offs and some of them presented sensitivity and specificity estimates based on post hoc cut-offs that were selected so as to maximize the joint sensitivity and specificity. This would clearly introduce bias. In a standardized analysis, Tables 4 and 5 present data using the standard cut-offs of 2.0 and 3.0 for SUV.
When a cut-off of SUV ≥ 2.0 (Table 4) was used to evaluate primary lesions, the sensitivity ranges between 64 to 100% and the specificity ranges between 71 to 100%. The largest study (n=37) on patients evaluated for recurrences reported a sensitivity of 58% and specificity of 92%. A considerable number of recurrent malignant tumors would be missed with the low sensitivity, although false positives are unlikely. A smaller study (n=10) shows 100% specificity and sensitivity.
With a cut-off of SUV ≥ 3.0 (Table 5) and for studies with at least 5 patients, there is a loss of sensitivity in most studies (36 to 100% for primary lesions, 46% to 80% for recurrent lesions). The improvement in specificity is more modest and is seen in only three studies (range 66 to 100% for primary lesions, 92%-100% for recurrences). A cut-off of SUV ≥ 2.0 appears to be more informative compared with a cut-off of SUV ≥ 3.0, although the data are not very robust due to small number of studies and total number of patients.
Glucose metabolic rate
When a positive FDG-PET result is defined on the basis of glucose metabolic rate, the four studies with available data (Table 6) suggest that diagnostic performance is only modest. Sensitivity ranges from 58 to 75% in studies with at least five patients. The specificity is 76% in the sole study that has more than four benign lesions.
1b. Distinguishing low grade from high grade soft tissue sarcoma
Some of the variability in the diagnostic performance between the various studies in diagnosing malignant vs. benign lesions may be due to a variable case-mix of benign lesions and low-grade malignancies in each study. Tables 7 to 10 present separate data for FDG-PET positivity in high/intermediate-grade (G III/II) tumors, low-grade (G I) tumors, all benign lesions, and specifically inflammatory benign lesions, according to various definitions for FDG-PET positivity.
Visualization method
Based on visualization interpretation of FDG-PET results (Table 7), all G II/III tumors are visualized. The same applies to the large majority of low-grade tumors. Variable proportions of benign tumors are also visualized and almost all inflammatory lesions are visualized, while visualization is uncommon in benign tumors. The high visualization rate of 20 of 27 primary benign non-inflammatory lesions in Watanabe (2000) is an outlier compared to other studies. However, as stated earlier, the definitions of visualization were not clear in these studies.
Semi-quantitative SUV values
As shown in Table 8, using an SUV cut-off of 2.0, high values are seen in 59/66 high- or intermediate-grade tumors, 8/24 low-grade tumors, and 19/98 benign lesions. In the latter group 11 or 12 of the 13 benign inflammatory lesions have a high value, while high SUV values are seen in only 8/85 benign non-inflammatory lesions. Using a cut-off SUV of 3.0 (Table 9), high values are seen in 51/66 high/intermediate-grade tumors, 3/24 low-grade tumors, 9/85 benign non-inflammatory lesions and 11 or 12/13 benign inflammatory lesions.
Overall, an SUV cut-off of 2.0 can identify all high/intermediate-grade tumors, but it will identify only about half of the low-grade tumors, and will give false positive results in about one of six benign non-inflammatory lesions and will misclassify almost all benign inflammatory lesions. Moving the cut-off to 3.0 offers no meaningful advantage in separating these types of lesions.
Glucose metabolic rate
Data on differentiating tumor grade based on glucose metabolic rate are limited (Table 10). In all, using a cut-off of 6 micromol/100g/min, high values are seen in 32/35 (91%) high/intermediate-grade tumors, 1/13 (8%) low-grade tumors, and 6/24 (25%) benign lesions (no inflammatory lesions were included in any of the studies). These data suggest that there is limited discrimination ability between low-grade malignant and benign lesions. While almost all high/intermediate-grade malignant lesions have high glucose metabolic rate, the same is true for one of four benign lesions.
Overall, the combined results of these studies suggest that semi-quantitative and quantitative FDG-PET measures can help differentiate tumor grades, but there is some overlap between categories. Inflammatory lesions are almost always categorized as false positives, while a significant proportion of benign non-inflammatory lesions may also give false positive results.
Question 2. How does the test performance of FDG-PET compare with conventional anatomic imaging (CT, MRI, etc) among patients with soft tissues sarcoma with respect to:
- primary diagnosis
- diagnosing locoregional recurrence
- diagnosing distant metastases
2a. Primary diagnosis
There are currently no available data to answer reliably the question of how FDG-PET compares against CT or MRI for the diagnosis of primary STS. Several studies evaluating FDG-PET for the diagnosis of primary lesions stated that part of the inclusion criteria was the CT, MRI, or US result that was suggestive of malignancy, along with the clinical appearance. Even studies that do not specify the prior use of other imaging modalities are likely to have used some of these imaging modalities as part of the pre-PET work-up in some patients. The substantial variability in the prevalence of soft tissue malignancy among the primary lesion series included in addressing question 1 suggests that these imaging modalities must have been used to various extents before FDG-PET in defining the eligible population. Since FDG-PET is selectively performed after suggestive CT or MRI results, rather than in all unselected patients, the currently available data cannot be used to reliably address the comparative accuracy of CT or MRI vs. FDG-PET for primary soft tissue lesions.
2b. Diagnosing locoregional recurrence
Similar considerations relate for the most part also to the question of how FDG-PET compares against CT or MRI for the diagnosis of recurrent disease. Again, most studies have been performed in selected patients. Nevertheless, there are two studies that address specifically question 2b. These studies are Lucas (1998) and Kole (1997).
In Lucas (1998), 72 FDG-PET scans were performed on 60 patients. FDG-PET was positive in 14/19 recurrences (sensitivity 74%), while it was positive in only 3/53 non-recurrences (specificity 94%). There were 67 MRI studies performed with an overall sensitivity and specificity of 88% (15/17) and 96% (48/50), respectively. Based on these data, MRI seems to have better sensitivity than FDG-PET, but the difference is not statistically significant, while there is no difference in specificity. Of the three false positive FDG-PET results, MRI also reported false positive in two of them. Of the five false negative FDG-PET results, MRI had been performed in four of them and it was also false negative in two. The study has several limitations. First, the definition of positive FDG-PET is apparently based on qualitative interpretation of the scans, without exact description of the criteria used to interpret that a scan is positive. Second, the MRI and FDG-PET were not done on exactly the same patients and lesions (a few patients or lesions were examined with only FDG-PET). If the comparison is limited to the 67 cases where both studies were performed, FDG-PET had a sensitivity of 76% (13/17) and specificity of 94% (47/50). There does not seem to be any case where FDG-PET was correct while MRI was wrong, while the opposite occurred in three of these cases. Third, it is unclear if biopsies were performed on all cases; apparently they were performed only when changes in imaging suggested a recurrence. Several patients were classified as negative apparently based on no clinical evolution during follow-up. The authors conclude that "all three methods (CT, MRI, PET) accurately define the extent of disease", but the presented data do not provide additional information that demonstrates the incremental benefit of FDG-PET over MRI for detecting local recurrence.
In Kole (1997), only 17 subjects were included who were being evaluated for potential local recurrence. FDG-PET had sensitivity of 93% (14/15) and specificity of 100% (2/2). MRI had sensitivity of 77% (10/13) and gave false positive reading for both benign lesions (scar and Ascaris mass), where FDG-PET had given correct negative readings. Of the three false negative MRIs, FDG-PET was correct in two and was also false negative in one. MRI was not done in two cases. The study suggests that FDG-PET was correct in three cases where MRI was wrong, while the opposite situation was not seen in any case. This study has too few patients to draw reliable conclusions. Moreover, no information was provided about how the MRI was performed and whether a contrast agent was also given or which weighting method was used.
Overall, the limited available data suggest approximately similar diagnostic performance of MRI and FDG-PET for the evaluation of local recurrence. However, modest differences could have been missed due to small numbers or design flaws. Cases are presented where FDG-PET can correct MRI, or MRI can correct FDG-PET, but each phenomenon occurs in about only 3% of the examined cases.
2c. Diagnosing distant metastases
There are three studies that address the comparative performance of FDG-PET vs. other imaging modalities for the diagnosis of distant metastasis. These studies are Lucas (1998), Lucas (1999), and El-Zeftawy (2001).
In Lucas (1998), 70 FDG-PET scans were performed on 62 patients. FDG-PET was positive in 13/15 lung metastases (sensitivity 87%), while it was not positive in any of the 55 cases without metastasis (specificity 100%). There were also 70 CT scans performed. The sensitivity was 15/15 (100%) and the specificity was 48/50 (96%). Based on these data, FDG-PET seems to have equal diagnostic performance with CT scan for diagnosing lung metastasis. FDG-PET failed to detect two lung metastases that were detected by CT, while FDG-PET gave correct negative readings in two cases where CT scan was falsely positive. FDG-PET also identified 13 metastatic sites other than lung and gave one false positive FDG-PET in a patient who had negative MRI. These data cannot be compared with other imaging modalities, as other imaging was either performed very selectively or not reported.
This study also has several limitations. First, the definition of positive FDG-PET is apparently based on qualitative interpretation of the scans, without exact description of the additional criteria used to state that a scan is positive, once the uptake is found to be higher than the liver uptake. Second, histopathology was not confirmed in all cases. This was limited to cases where there was a suggestion for metastasis, while most patients are classified as negative apparently based on no clinical evolution during follow-up. The authors conclude that "all three methods (CT, MRI, PET) accurately define the extent of disease", but the presented data do not clearly demonstrate that the additional information provided by FDG-PET provided incremental benefit over CT for detecting metastatic disease.
In Lucas (1999), 31 lesions in 30 patients were evaluated. Data on metastatic disease are very sparse. Three patients had distant metastasis at presentation and one had multiple primaries. FDG-PET correctly identified two of three metastases (alveolar rhabdomyosarcoma, high-grade leiomyosarcoma), but failed to diagnose the presence of multiple bone metastasis in the long bones and spine of a third patient with high-grade leiomyosarcoma. Metastatic lesions were detected by CT (first two patients) and MRI (third patient). FDG-PET correctly identified and graded the multiple primaries in a patient with neurofibromatosis type 1 with malignant nerve sheath tumors. No false positives were obtained. Besides limited sample size, the study has several other limitations. First, it is unclear how many patients had CT or MRI along with FDG-PET and whether they were performed approximately at the same time. Second, diagnosis was evaluated histologically only when imaging was suggestive, while most patients without imaging or clinical evidence were simply evaluated for metastatic disease based on clinical follow-up (potential for verification bias). While the authors conclude that "PET ... may have a role in staging malignant tumors", the study did not demonstrate that FDG-PET provides incremental benefit over the use of CT or MRI.
In El-Zeftawy (2001), both CT and FDG-PET scans were performed at baseline for staging purposes in seven of the eight patients with STS. There was full concordance in six cases (no metastatic disease in three, metastases in three). In one patient, FDG-PET found three metastases vs. two with CT. Both CT and FDG-PET were performed on follow-up in three patients with surgery with or without chemotherapy. New lesions were observed in one patient and were detected 2 months earlier with CT than with FDG-PET, but the FDG-PET performed 2 months later also showed adrenal involvement that was not seen on CT. Both CT and FDG-PET were performed on follow-up in four patients treated with additional radiotherapy. There was good concordance between CT and MRI on seven occasions, and in one instance FDG-PET also showed kidney metastasis in addition to lung involvement shown on CT. The study has severe limitations. It is a small retrospective case series. Imaging studies were not performed systematically on all patients. Diagnosis may be subject to verification bias, although this is not clear. The authors avoided concluding that FDG-PET offers an incremental diagnostic benefit over CT or MRI, but they claim that FDG-PET "proved to have an impact on the clinical management." The supporting evidence for this claim is limited.
Question 3. A review of studies on changes in patient management or improved outcomes for patients with soft tissue sarcoma with the use of FDG-PET.
There are no randomized trials or other controlled studies that addressed these questions. Thus we do not know the comparative outcomes of patients who had FDG-PET vs. those who did not have FDG-PET as part of their patient management. There are four small studies that addressed the role of FDG-PET in guiding the management of STS during longitudinal follow-up. These studies are Jones (1996) (n=4), Shulkin (1995) (n=3), van Ginkel (1996) (n=20), and El-Zeftawy (2001) (n=8). A few other case reports are presented in other articles, but these are isolated cases that cannot be interpreted reliably. All four studies present anecdotal experience without a control group. Occasional patients are said to have benefited from the diagnostic information received by FDG-PET (El-Zeftawy, 2001), but the claims cannot be documented against a control arm. Other findings from these studies are presented under question 4 below.
Question 4. A review of studies on using FDG-PET to determine tumor response to therapeutic interventions for patients with soft tissue sarcoma.
As above, there are four small studies that addressed the role of FDG-PET in guiding the management of STS during longitudinal follow-up. These studies are Jones (1996) (n=4), Shulkin (1995) (n=3), van Ginkel (1996) (n=20), and El-Zeftawy (2001) (n=8). A few other case reports are presented in other articles, but they are isolated cases that cannot be interpreted reliably.
In Jones (1996), response to radiotherapy/hyperthermia was evaluated in four patients with STS: scans showed an extension of the intratumoral region of absent uptake in three early therapy scans and absent central uptake with a peripheral rim uptake in a late post-therapy pre-surgery scan. Parallel changes were seen on MRI (development of high intratumoral SI on T2WI, peripheral rim enhancement, absent central enhancement).
Shulkin (1995) reported pertinent data on only one patient. Specifically, one patient with Ewing 's sarcoma had baseline FDG-PET, as well as evaluations at 3-months and 6-months post-diagnosis (after chemotherapy and after chemotherapy and radiotherapy, respectively). FDG-PET correctly showed regression of disease, while MRI showed no changes on the second imaging at 3 months.
In van Ginkel (1996), of 20 patients with locally advanced primary (n=14) or recurrent (n=6) sarcoma, seven showed complete response and 12 showed partial response on pathological examination after treatment with hyperthermic isolated limb perfusion with tumor necrosis factor alpha, interferon gamma and melphalan (one patient had no post-treatment pathology examination). Pre-perfusion glucose consumption was higher in the complete response group than in the partial response group (p<0.05). The glucose consumption in the complete response group decreased significantly at 2 and 8 weeks post-treatment, respectively, as compared with the partial response group. The authors concluded "PET indicated the pathologic tumor response to hyperthermic isolated limb perfusion, although the lack of specificity of FDG, in terms of differentiating between an inflammatory response and viable tumor tissue, hampered the discrimination between partial and complete response." This is a balanced conclusion. The study is well reported, but it suffers from small sample size and questionable generalizability since a specific therapeutic modality was used.
Finally, El-Zeftawy (2001), based on the information that we discussed under question 2C, concluded "repeated PET examinations proved to have an impact on the clinical management of patients with bone and soft tissue sarcoma." Limitations of the study have been discussed under question 2C.
Of all the studies, only van Ginkel (1996) provided substantial information on the role of FDG-PET to monitor therapeutic response, while the other studies describe very limited experience. The data are promising, but they are insufficient and inconclusive.
Conclusions
The main conclusions of this evidence report are:
- ≥ 2.0 and glucose metabolic rate ≥ 6 micromol/100g/min; many low-grade (G I) tumors are also visualized, but they usually have SUV < 2.0 and rarely have glucose metabolic rate ≥ 6 micromol/100g/min. Benign lesions are visualized almost as frequently as grade I malignant lesions and these two types of lesions are not easy to differentiate based on SUV or glucose metabolic rate. Thus although FDG-PET has very good to excellent performance in differentiating high- or intermediate-grade tumors from low-grade tumors, it is not a good diagnostic test for differentiating low-grade tumors from benign lesions. Of note, benign tumors are usually not visualized by FDG-PET and almost always have low SUV and glucose metabolic rate values, while benign inflammatory lesions, such as infection and myositis ossificans are likely to be mistaken for tumors based on their qualitative and quantitative scoring.
- There are no good quality data on the comparative diagnostic performance of FDG-PET against CT or MRI for diagnosis of primary soft tissue lesions, since comparative evidence has been obtained with tests used in series. There is limited evidence suggesting approximately equivalent diagnostic performance of FDG-PET and MRI for diagnosing local recurrence and even more limited evidence suggesting approximately equivalent performance of FDG-PET and CT scan for diagnosing distant metastatic disease.
- Evidence on the impact of FDG-PET on clinical outcomes and on the management of patients is extremely limited, and there are no controlled studies to answer these important questions.
- There is very limited data on the usefulness of FDG-PET in assessing the response to therapy. The evidence suggests that FDG-PET can be used to follow therapeutic responses, but it is difficult to separate complete from partial response, and there is insufficient evidence to compare the performance of FDG-PET against CT or MRI in this regard.
References
Adler LP, Blair HF, Williams RP, Pathria MN, et al. Grading liposarcomas with PET using [18F]FDG. J Computer Assisted Tomography 1990;14:960-2.
Dimitrakopoulou-Strauss A. Strauss LG. Schwarzbach M. Burger C. Heichel T. Willeke F. Mechtersheimer G. Lehnert T. Dynamic PET 18F-FDG studies in patients with primary and recurrent soft-tissue sarcomas: impact on diagnosis and correlation with grading. J Nucl Med 2001;42:713-20.
Eary JF. Conrad EU. Bruckner JD. Folpe A. Hunt KJ. Mankoff DA. Howlett AT. Quantitative [F-18]fluorodeoxyglucose positron emission tomography in pretreatment and grading of sarcoma. Clinical Cancer Research 1998;4:1215-20.
El-Zeftawy H, Heiba SI, Jana S, Rosen G, et al. Role of repeated F-18 fluorodeoxyglucose imaging in management of patients with bone and soft tissue sarcoma. Cancer Biotherapy & Radiopharmaceuticals 2001;16:37-46.
Ferner RE, Lucas JD, O;Doherty MJ, Hughes RA, Smith MA, Cronin BF, Bingham J. Evaluation of 18fluorodeoxyglucose positron emission tomography in the detection of malignant peripheral nerve sheath tumours arising from within plexiform neurofibromas in neurofibromatosis 1. J Neurol Neurosurg Psychiatry 2000;68:353-7.
Folpe AL. Lyles RH. Sprouse JT. Conrad EU 3rd. Eary JF. (F-18) fluorodeoxyglucose positron emission tomography as a predictor of pathologic grade and other prognostic variables in bone and soft tissue sarcoma. Clinical Cancer Research 2000;6:1279-87.
Griffeth LK, Dehdashti F, McGuire AH, McGuire DJ, Perry DJ, Moerlein SM, Siegel BA. PET evaluation of soft tissue masses with fluorine 18 fluoro-2-deoxy-D-glucose. Radiology 1992;182:185-94.
Hain SF, O'DohertyJ, Lucas JD, Smith MA. Fluorodeoxyglucose PET in the evaluation of amputation for soft tissue sarcoma. Nuclear Medicine Communications 1999;20:845-8.
Kern KA. Brunetti A. Norton JA. Chang AE. Malawer M. Lack E. Finn RD. Rosenberg SA. Larson SM. Metabolic imaging of human extremity musculoskeletal tumors by PET. J Nucl Med 1988;29:181-6.
Kole AC, Nieweg OE, van Ginkel RJ, Pruim J, et al. Detection of local recurrence of soft-tissue sarcoma with positron emission tomography using [18F]fluorodeoxyglucose. Ann Surg Oncol 1997;4:57-63.
Jones DN, McCowage GB, Sostman HD, Brizel DM, Layfield L, Charles HC, Dewhirst MW, Prescott DM, Friedman HS, Harrelson JM, Scully SP, Coleman RE. Monitoring of neoadjuvant therapy response of soft-tissue and musculoskeletal sarcoma using fluorine-18-FDG-PET. J Nucl Med 1996;37:1438-44.
Lodge MA. Lucas JD. Marsden PK. Cronin BF. O'Doherty MJ. Smith MA. A PET study of 18FDG uptake in soft tissue masses. Eur J Nucl Med 1999;26:22-30.
Lucas JD. O'Doherty MJ. Wong JC. Bingham JB. McKee PH. Fletcher CD. Smith MA. Evaluation of fluorodeoxyglucose positron emission tomography in the management of soft-tissue sarcomas. J Bone & Joint Surg - British Volume 1998;80:441-7.
Lucas JD. O'Doherty MJ. Cronin BF. Marsden PK. Lodge MA. McKee PH. Smith MA. Prospective evaluation of soft tissue masses and sarcomas using fluorodeoxyglucose positron emission tomography. Br J Surg 1999;86:550-6.
Nieweg OE. Pruim J. van Ginkel RJ. Hoekstra HJ. Paans AM. Molenaar WM. Koops HS. Vaalburg W. Fluorine-18-fluorodeoxyglucose PET imaging of soft-tissue sarcoma. J Nucl Med 1996;37:257-61.
Schulte M. Brecht-Krauss D. Heymer B. Guhlmann A. Hartwig E. Sarkar MR. Diederichs CG. Von Baer A. Kotzerke J. Reske SN. Grading of tumors and tumorlike lesions of bone: evaluation by FDG-PET. [see comments]. J Nucl Med 2000;41:1695-701.
Schwarzbach MH. Dimitrakopoulou-Strauss A. Willeke F. Hinz U. Strauss LG. Zhang YM. Mechtersheimer G. Attigah N. Lehnert T. Herfarth C. Clinical value of [18-F]] fluorodeoxyglucose positron emission tomography imaging in soft tissue sarcomas. Annals of Surgery 2000;231:380-6.
Shulkin BL, Mitchell DS, Ungar DR, Prahash D, et al. Neoplasms ion a pediatric population: 2-[F-18]-fluoro-2-deoxy-D-glucose PET studies. Radiology 1995;194:495-500.
van Ginkel RJ. Hoekstra HJ. Pruim J. Nieweg OE. Molenaar WM. Paans AM. Willemsen AT. Vaalburg W. Koops HS. FDG-PET to evaluate response to hyperthermic isolated limb perfusion for locally advanced soft-tissue sarcoma. J Nucl Med 1996;37:984-90.
Watanabe H. Shinozaki T. Yanagawa T. Aoki J. Tokunaga M. Inoue T. Endo K. Mohara S. Sano K. Takagishi K. Glucose metabolic analysis of musculoskeletal tumours using 18fluorine-FDG-PET as an aid to preoperative planning. J Bone & Joint Surg - British Volume 2000; 82:760-7.
