European Journal of Nuclear Medicine and Molecular Imaging

, Volume 34, Issue 12, pp 1925–1932

Prediction of tumor response by FDG-PET: comparison of the accuracy of single and sequential studies in patients with adenocarcinomas of the esophagogastric junction

Authors

    • Department of Nuclear MedicineKlinikum rechts der Isar, Technische Universitaet Muenchen
    • Department of RadiologyKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Katja Ott
    • Department of SurgeryKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Florian Lordick
    • Department of SurgeryKlinikum rechts der Isar, Technische Universitaet Muenchen
    • Department of Medicine IIIKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Karen Becker
    • Department of Medicine IIIKlinikum rechts der Isar, Technische Universitaet Muenchen
    • Department of PathologyKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Alexander Stahl
    • Department of Nuclear MedicineKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Ken Herrmann
    • Department of Nuclear MedicineKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Ulrich Fink
    • Department of SurgeryKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Jörg Rüdiger Siewert
    • Department of SurgeryKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Markus Schwaiger
    • Department of Nuclear MedicineKlinikum rechts der Isar, Technische Universitaet Muenchen
  • Wolfgang A. Weber
    • Department of Nuclear MedicineKlinikum rechts der Isar, Technische Universitaet Muenchen
    • Department of Nuclear MedicineUniversitaetsklinikum Freiburg, Albrecht-Ludwigs-Universitaet Freiburg
Original article

DOI: 10.1007/s00259-007-0521-3

Cite this article as:
Wieder, H.A., Ott, K., Lordick, F. et al. Eur J Nucl Med Mol Imaging (2007) 34: 1925. doi:10.1007/s00259-007-0521-3

Abstract

Purpose

Positron-emission-tomography with the glucose analog fluorodeoxyglucose (FDG-PET) has shown encouraging results for prediction of tumor response to chemotherapy. However, there is no consensus as to what time after initiation of therapy FDG-PET should be performed. To address this question we studied the time course of changes in tumor FDG-uptake in patients with locally advanced adenocarcinomas of the esophagogastric junction (AEG) treated with preoperative chemotherapy.

Methods

Twenty-four patients with AEG were included and underwent FDG-PET prior to therapy (PET1), 2 weeks after initiation of therapy (PET2), and preoperatively (PET3). Tumor metabolic activity was assessed by standardized uptake values (SUV) and correlated with histopathologic response and patient survival.

Results

Baseline tumor SUV was 8.3 ± 3.5 and decreased to 5.0 ± 1.8 at PET2 (p  <  0.0001). At PET3 there was further decrease to 3.5 ± 1.9 (p < 0.0001). The relative decrease of tumor FDG-uptake from PET1 to PET2 and from PET1 to PET3 were both significantly correlated with histopathologic response. Reduction of tumor SUV from PET1 to PET2 was significantly correlated with survival (p = 0.03) and there was a similar trend for changes from PET1 to PET3 (p = 0.09). In contrast, absolute SUVs did not demonstrate a significant correlation with histopathological response or patient survival at any of the studied time points.

Conclusion

In patients with AEG, relative changes in tumor FDG uptake are better predictors for treatment outcome than absolute SUVs. Metabolic changes within the first 2 weeks of therapy are at least as efficient for prediction of histopathologic response and patient survival as later changes.

Keywords

18F-FDG-PETOncologyNeoadjuvant chemotherapyEsophageal cancerTherapy response

Introduction

Patients with locally advanced adenocarcinomas of the esophagus or esophagogastric junction (AEG) are frequently offered preoperative chemotherapy in order to increase the rate of complete surgical resection (R0). However, the role of preoperative chemotherapy in patient management and its influence on overall patient outcome and survival still remain controversial because randomized studies have not consistently shown an improvement of patient survival by preoperative chemotherapy [17]. Nevertheless, there is considerable evidence that patients with a significant histopathologic response to preoperative therapy are characterized by an excellent prognosis that is markedly better than the prognosis of patients treated with surgery alone [4, 811]. The need for a noninvasive test to predict histopathologic response to preoperative therapy and survival is therefore widely acknowledged.

Positron emission tomography with the glucose analogue fluorodeoxyglucose (FDG) has recently been introduced as a new technique for the staging of esophageal cancer [1214]. Furthermore, changes in tumor metabolic activity during and after chemotherapy have been shown to correlate with histopathologic response in esophageal and gastric cancer patients [1521]. However, there is currently no consensus on the most appropriate time point for monitoring tumor response by FDG-PET. Furthermore, it is currently unknown whether it is preferable to measure absolute tumor FDG-uptake or relative changes in tumor FDG-uptake. To address these questions we studied patients with FDG-PET before chemotherapy, two-weeks after start of chemotherapy, and 3–4 weeks after completion of chemotherapy. Absolute tumor FDG-uptake at these time points and relative changes during therapy were correlated with histopathologic tumor response and patient survival.

Patients and methods

Patient population

Twenty-four patients with locally advanced adenocarcinoma of the distal esophagus scheduled to undergo neoadjuvant chemotherapy at the Klinikum rechts der Isar (Munich, Germany) were included in this study.

Eligibility requirements included the presence of biopsy proven, locally advanced (clinical stage T3/T4, N0/+, M0/1lymph, M0organ) adenocarcinoma of the distal esophagus (AEG type I) or cardia (AEG type II) [22]. Individual patient data are shown in Table 1. The study protocol was approved by the Ethics Committee of the Technische Universitaet Muenchen.
Table 1

Patients’ characteristics, changes in tumor FDG-uptake, and histopathologic tumor response

Patient number

Gender

Age (y)

AEG type

Histopathologic evaluation

Tumor FDG uptake (SUV)

Stage

Response

Baseline

After 2 weeks

Preoperation

1

M

71

1

T2N1

Nonresponder

9.9

7.8

6.3

2

M

61

1

T2N0

Nonresponder

6.2

4.5

2.4

3

M

65

1

T2N0

Responder

11.7

4.4

2.1

4

F

58

1

T2N1

Nonresponder

4.7

3.5

3.0

5

M

63

1

T3N1

Nonresponder

8.8

6.7

4.1

6

M

52

1

T2N1

Responder

9.2

4.8

3.4

7

F

67

2

T2N1

Nonresponder

4.5

2.0

2.3

8

M

58

1

T1N0

Nonresponder

9.2

5.0

2.3

9

M

48

2

T2N1

Nonresponder

6.7

5.4

3.6

10

M

52

2

T2N2

Nonresponder

4.8

5.1

4.0

11

M

70

2

T2N1

Responder

14.5

7.3

2.5

12

M

59

2

T2N2

Nonresponder

5.2

4.2

4.2

13

M

70

2

T2N0

Nonresponder

5.4

3.7

2.6

14

M

65

1

T1N1

Responder

11.0

6.6

6.9

15

M

57

2

T1N0

Responder

3.7

1.8

1.7

16

M

60

1

T3N1

Nonresponder

10.9

4.8

3.3

17

M

52

1

T3N1

Nonresponder

10.0

8.2

9.8

18

M

71

2

T0N0

Responder

16.6

6.9

3.0

19

F

38

1

T0N0

Responder

5.2

3.5

1.8

20

M

69

2

T1N0

Responder

13.3

6.1

2.5

21

F

62

2

T4N1

Nonresponder

3.7

1.9

1.7

22

M

71

1

T3N1

Nonresponder

7.8

7.2

4.1

23

M

33

2

T2N1

Nonresponder

8.5

5.7

4.0

24

M

67

1

T3N1

Nonresponder

7.8

3.7

3.1

Neoadjuvant chemotherapy

Neoadjuvant therapy consisted of two cycles of combination chemotherapy, each of 36 days’ duration [23]. On day 1, cisplatin, at a dose of 50 mg/m2 body-surface area (BSA), was given as intravenous infusion over a period of 1 h. Thereafter, patients received leucovorin (500 mg/m2 BSA) over a period of 2 h, followed by 5-fluorouracil (2 g/m2 BSA) over a period of 24 hours. Treatment with cisplatin was repeated on days 15 and 29. Infusion of leucovorin and 5-fluorouracil was repeated on days 8, 15, 22, 29, and 36. AEG I tumors were additionally treated by paclitaxel (80 mg/m2 BSA) over a period of 3 h, one day prior to infusion of cisplatin. The tumor resection was scheduled 3–4 weeks after completion of chemotherapy.

PET-imaging

A total of 72 FDG-PET scans were performed in the 24 patients. The patients underwent PET scans prior to initiation of therapy, on day 14 of the first chemotherapy cycle and 3–4 weeks after the end of therapy (preoperatively, approximately 3 months after initiation of therapy). The changes in FDG uptake from the baseline to the first follow-up study have been reported previously [24]. However, this study did not address the findings of the third PET scan.

Patients fasted at least 6 h before PET scanning to minimize the blood insulin level and ensure standardized metabolism in all patients. Blood glucose levels were measured before each PET examination. All measured values were less than 150 mg/dl and showed no significant changes during chemotherapy. Static emission images of 20 minutes’ duration covering the area of the tumor were acquired 40 minutes after injection of 300–400 MBq FDG using an ECAT EXACT PET scanner (Siemens CTI, Knoxville, TN). After the emission scan, transmission measurements were performed for attenuation correction. Images were reconstructed iteratively using an attenuation weighted ordered subset expectation maximization algorithm (8 iterations, 4 subsets) and then smoothed in 3D using a 4 mm Gauss filter.

PET data analysis

Attenuation-corrected PET images were normalized for the injected dose of FDG and the patients’ body weight, resulting in parametric images representing regional standardized uptake values (SUVs). For the quantitative evaluation of regional FDG uptake, regions of interest (ROIs) were manually placed over each primary tumor as previously described [15, 16]. Briefly, a circular ROI of 1.5 cm (corresponding to 10 pixels) was placed in the slice with maximum FDG uptake. This approach for ROI definition avoids the rather subjective delineation of the “whole” tumor activity since the maximum tumour activity can easily be identified. On the other hand, average counts from this small ROI are less dependent on the reconstruction and postprocessing filters than SUVmax [25].

In the follow-up examinations, the ROI was placed in the same location as the previously identified lesion using the landmarks of the transmission images (apex of the lung, bifurcation of trachea) as a reference. SUVs were calculated using the average activity values in the ROI.

Surgical therapy and histopathologic response evaluation

The surgical procedure was a transhiatal esophagectomy for patients with AEG I. In patients with AEG II tumors, a transhiatal extended gastrectomy and an extended D2-lymphadenectomy, including a left retroperitoneal lymphadenectomy, were performed.

Histopathologic analysis of the resected specimens was performed by one pathologist who was unaware of the results of PET imaging and patient outcome data. Tumor regression was assessed semiquantitatively according to a published scoring system [26]. Briefly, the grading of tumor regression in response to chemotherapy is based on an estimation of the percentage of vital tumor tissue in relation to the macroscopically identifiable, completely histologically examined tumor bed. The scoring system uses the following grades: 1a complete tumor regression; 1b less than 10% residual tumor; 2 10–50% residual tumor; and 3 more than 50% residual tumor.

For the purposes of this study, all patients with less than 10% viable residual tumor cells (regression score, grade 1) were classified as responding as previously described [15, 16, 19]. All other patients were classified as nonresponding.

Statistical analysis

Statistical analyses were performed using the StatView program (SAS Institute Inc., Cary, NC). Quantitative values were expressed as mean ±1 standard deviation. Intra- and interindividual comparisons of tumor FDG uptake were performed by the Wilcoxon signed-rank and the Mann-Whitney U test, respectively.

Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic accuracy of FDG PET for assessment of histopathologic response. For calculation of these curves the threshold value for definition of a tumor response in PET-imaging was systematically varied over the whole range of the observed changes in tumor FDG-uptake. For each of these threshold values the percentage of correctly predicted histopathologic responses (true positive rate on the y-axis) was plotted against the rate of incorrectly predicted histopathologic responses (false positive rate on the x-axis). The optimum threshold value for differentiation of responding and non-responding tumors was defined by the point of the ROC-curve with minimum distance from the 0% false-positive rate and the 100% true-positive rate. For this threshold value sensitivity, specificity and positive and negative predictive value were calculated using standard formulas.

Overall survival was calculated from the first day of chemoradiotherapy. Survival rates were calculated according to Kaplan-Meier and differences between groups of patients tested with a log-rank test. All statistical tests were performed at a two-sided 5% level of statistical significance.

Results

Histopathological response evaluation and patient survival

Thirteen of 24 patients (54%) had adenocarcinomas of the distal esophagus (AEG I), and 11 patients (46%) had adenocarcinomas of the cardia (AEG II). All patients underwent two cycles of preoperative chemotherapy. The histopathologic response evaluation following preoperative chemotherapy and surgery revealed complete tumor regression (regression grade 1a) in two of 24 patients (8%), less than 10% tumor cells (regression grade 1b) in six patients (25%), regression grade 2 in five patients (21%), and regression grade 3 in 11 patients (46%). Thus, the group of histopathological responders consisted of eight patients (33%), and the group of nonresponders was 16 patients (67%). Median follow-up time was 25 months. Histopathologic response correlation was highly significant with overall survival. Two year survival in responders was 87% as compared to 44% in nonresponders (p = 0.02).

Correlation of tumor FDG uptake with histopathologic response

Mean tumor FDG uptake was 8.3 ± 3.5 SUV at baseline (PET 1) decreasing to 5.0 ± 1.8 (−37% ± 18%, p < 0.0001) 2 weeks after the onset of chemoradiotherapy (PET 2) and showing a further decrease to 3.5 ± 1.9 (−28% ± 23%, p < 0.0001) prior to surgery (PET 3, Fig. 1a). The time course of relative changes in tumor FDG-uptake over time is presented in Fig. 1b. In none of the patients was there a relevant increase of the FDG-signal from the 2-weeks scan to the preoperative scan (Table 1). Figure 2 shows examples of FDG PET scans in histopathologically responding and nonresponding tumors.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-007-0521-3/MediaObjects/259_2007_521_Fig1_HTML.gif
Fig. 1

a Time course of tumor FDG uptake during chemotherapy in histopathologic responding (solid line) and nonresponding tumors (dashed line). Error bars denote one standard deviation. b Time course of relative changes in tumor FDG uptake during chemotherapy in histopathologic responding and nonresponding tumors. Error bars denote one standard deviation

https://static-content.springer.com/image/art%3A10.1007%2Fs00259-007-0521-3/MediaObjects/259_2007_521_Fig2_HTML.gif
Fig. 2

Examples of FDG-PET studies (axial slices) in histopathologically responding and nonresponding tumors. In the responding tumor, FDG uptake decreases to background level as soon as 14 days after the beginning of therapy. At this time point FDG uptake is almost unchanged for the nonresponding tumor

At baseline, there was no significant difference in tumor SUV between patients subsequently defined as histopathological responders compared to nonresponders (p = 0.08, Fig. 1a). However, there was a tendency toward higher tumor SUVs in responders. After 2 weeks of preoperative chemotherapy, subsequent responders and nonresponders showed virtually equal mean tumor SUVs (p = 0.8, Fig. 1a). Prior to surgery, tumor SUV was also not significantly lower in histopathologic responders compared to nonresponders (p = 0.5, Table 2).
Table 2

Findings in FDG-PET, histopathologic response, and patient survival

 

Absolute SUV

Relative decrease of SUV

PET 1

PET 2

PET 3

PET 1 to PET 2

PET 2 to PET 3

PET 1 to PET 3

Histopathologic responders (n = 8)

10.7 ± 4.5

5.2 ± 1.9

3.0 ± 1.7

50% ± 10%

39% ± 26%

68% ± 17%

Histopathologic nonresponder (n = 16)

7.1 ± 2.3

5.0 ± 1.8

3.8 ± 1.9

30% ± 18%

22% ± 20%

46% ± 20%

p-value

0.08

0.9

0.5

0.03

0.2

0.05

Cut-off for a “metabolic response”

<−35%

≤−63%

Accuracy

75%

83%

Sensitivity

88%

75%

Specificity

69%

87%

AUC

0.81

0.85

95% CI

0.638 to 0.957

0.648 to 0.961

2 year survival rate “PET-responder”

83% (n = 12)

73% (n = 11)

2 year survival rate “PET-non-responder”

33% (n = 12)

46% (n = 13)

p-value

0.03

0.09

In contrast, relative changes in tumor SUV were related with histopathological response as follows. Two weeks after the onset of preoperative chemotherapy, there was a significantly more pronounced SUV decrease in histopathologic responders than in nonresponders. The mean difference in SUV changes for histopathologic responders (−50%) and nonresponders (−30%) was 20% (Table 2). The optimum threshold value for differentiation of responding and nonresponding tumors was achieved by applying a cut-off value of −33% (Fig. 3). Using this cut-off value, all responding and ten of the 16 nonresponding tumors were correctly identified. Corresponding accuracy, sensitivity, and specificity to distinguish between responders and nonresponders were 75%, 100% and 63%, respectively. When applying a cut-off value of −35% prospectively, as recently published [15, 16], one more nonresponder was detected at the expense of one responder less yielding an accuracy, sensitivity, and specificity of 75%, 88% and 69%, respectively.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-007-0521-3/MediaObjects/259_2007_521_Fig3_HTML.gif
Fig. 3

ROC curves for assessment of histopathologic response by FDG-PET. Changes in FDG-uptake from the baseline scan to the scan after 14 days of chemotherapy (solid line); from the baseline to the preoperative scan (dotted line)

Prior to surgery, the relative SUV decrease in responders was also more pronounced than in nonresponders. The difference in relative changes for subsequent responders (−68%) and nonresponders (−46%) was similar to the PET scan after 2 weeks of therapy (Table 2). The ROC curve demonstrated that the highest accuracy for differentiation of histopathologically responding and nonresponding tumors was achieved by applying a cut-off value of 63% decrease of baseline FDG uptake (Fig. 3). Using this cut-off value, six of the eight responding and 14 of the 16 nonresponding tumors were correctly identified providing a sensitivity of 75%, a specificity of 87%, and an accuracy of 83%, respectively. The ROC curve was almost identical to the ROC curve for changes from baseline to day 14 (area under the ROC curve 0.82 vs. 0.85, p = 0.9).

Correlation of tumor FDG uptake with survival

Relative changes from PET1 to PET2 were significantly correlated with overall survival (p = 0.03), whereas the predictive value of relative changes from PET1 to PET3 did not reach statistical significance (p = 0.09). Two-year survival was 83% (n = 12) and 33% (n = 12) for patients with a decrease in tumor FDG-uptake by more/less than 35% (PET1 to PET2, Fig. 4b). When analysing changes from the baseline to the preoperative scan (PET1 to PET3), two-year survival rates for patients with a decrease in tumor FDG-uptake by more/less than 63% were 73% (n  = 11) and 46% (n = 13), respectively (Fig. 4c). Absolute SUVs did not demonstrate a significant correlation with patient survival at any of the studied time points (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-007-0521-3/MediaObjects/259_2007_521_Fig4_HTML.gif
Fig. 4

a Reduction of tumor FDG-uptake 14 days after initiation of chemotherapy and patient survival. b Reduction of tumor FDG-uptake after completion of chemotherapy and patient survival

Discussion

This study provides two new findings regarding the use of FDG-PET for monitoring chemotherapy of locally advanced esophageal cancer. First, it indicates that early metabolic changes (14 days after start of therapy) provide at least the same accuracy for prediction of treatment outcome than late changes (3 months after start of therapy). Second, it shows that the predictive value of relative changes in tumor FDG-uptake is superior to measurements of absolute tumor FDG-uptake.

One might have expected that PET imaging after completion of chemotherapy would result in a higher accuracy for assessment of histopathologic response than imaging during the first chemotherapy cycle (14 days after start of chemotherapy). Tumors were resected within 1 week after the preoperative PET scan and their metabolic status in this scan should therefore more closely reflect histopathologic tumor regression than a PET scan performed 9–10 weeks earlier. For example, a tumor might initially respond well to treatment and demonstrate a marked decrease in metabolic activity. Later on, a resistant cell population may emerge leading to a subsequent rise in FDG-uptake. However, our data indicate that this is a rare event in patients with AEG, since none of the studied tumors demonstrated a relevant rise in FDG-uptake from the 2-week scan to the preoperative scan.

Furthermore, mean tumor SUV in the preoperative PET scan was only 3.5 ± 1.9, which is close to the SUV of mediastinal blood pool (≈2.5). Thus FDG-uptake of all tumors had markedly decreased at this point in time and differences between responders (mean SUV 3.0 ± 1.7) and nonresponders (mean SUV 3.8 ± 1.9) were small. This made it difficult to discriminate responding and nonresponding tumors by measuring their FDG-uptake in the preoperative PET scan. When tumor response was defined by a relative decrease in tumor FDG-uptake from the baseline to the preoperative scan, the large variability of baseline tumor FDG-uptake resulted in an additional methodological problem. Since all histopathologically responding tumors demonstrated no or only minimal focal FDG-uptake in the preoperative scan, relative changes in FDG-uptake from the baseline to the preoperative scan were necessarily larger for tumors with high baseline uptake than for tumors with low baseline FDG-uptake. As a consequence, criteria for assessment of tumor response in PET became too strict for tumors with low baseline FDG-uptake. In the present study, histopathologically responding tumors were only detected with a sensitivity of 75% when using relative changes from the baseline to the preoperative scan as a criterion. The preoperative SUV of tumors with a low baseline FDG-uptake would have to be less than the mediastinal blood pool in order to fulfill the criterion for a metabolic response derived from the whole patient population (−63% change, see Table 2). In this situation, visual analysis of FDG-PET scans [21] or a combination of visual and quantitative analysis may thus be preferable to measurement of relative changes. On visual analysis tumors are generally considered as responding when their FDG-uptake after therapy is not higher than in surrounding normal tissue [21]. Therefore, the problem of classifying tumors with relatively low baseline metabolic activity as nonresponding is avoided.

We have recently made a similar observation in patients with squamous cell carcinomas of the esophagus treated by preoperative chemoradiotherapy [19]. In this study, tumor FDG-uptake after a 4 week course of chemoradiotherapy was low, and there were only small differences between histopathological responders and nonresponders. It is currently not clear whether the low FDG-uptake of histopathologically nonresponding tumors is due to a significant loss of viable tumor cells that does not fulfill the criterion for a histopathologic response (<10% viable tumor cells) or reflects a lower metabolic activity of viable tumor cells (“metabolic stunning”). Nevertheless, the results of these two studies indicate that imaging after completion of therapy may not represent the optimum point in time to assess histopathologic response.

If PET imaging is performed during therapy the methodological problems discussed above are less severe, since residual FDG-uptake is higher and the relative change in FDG-uptake is smaller at this point in time. The major clinical advantages of using early changes in tumor FDG-uptake to predict tumor response is, however, the ability to adjust treatments to the chemosensitivity of the tumor tissue. Patients classified as metabolic responders would continue to receive standard treatment, whereas an alternative treatment would be used in metabolic nonresponders. Examples of alternative treatments could be immediate surgical resection instead of continued neoadjuvant therapy, the use of different, second-line, chemotherapy regimens, or the use of “targeted” drugs. This may avoid unnecessary toxicity, costs, and time loss caused by ineffective therapy.

Therefore, it is very encouraging that the present study confirmed that FDG-PET can predict histopathologic tumor response and patient survival within 14 days after the start of chemotherapy.

For the clinical application of FDG-PET for prediction of treatment response it is also encouraging that relative changes provided the same or higher accuracy for predicting or assessing tumor response than absolute SUVs (Table 2). Absolute SUVs are much more sensitive to differences in data acquisition, image reconstruction, and data analysis than relative changes. For example, Boellaard et al. have shown in a simulation study that differences in defining regions of interest can result in more than 50% differences of measured absolute SUVs. In contrast, SUV ratios measured by different methods varied only minimally [25]. In a clinical study, Stahl et al. [27] found that the SUVs of gastric carcinomas increased by 60% between 40 and 90 min post injection indicating that absolute tumor SUVs are highly dependent on the timing of the data acquisition. In contrast, therapy induced changes in tumor SUVs were only mildly affected by the timing of data acquisition. The mean decrease of tumor SUVs after chemotherapy was 31% for patients imaged 40 min post injection and 29% for patients imaged 90 min post injection [27]. Thus, relative changes represent more robust parameters than absolute SUVs and are therefore preferable for establishing response criteria that can be used at multiple institutions, e.g. in multicenter trials.

Copyright information

© Springer-Verlag 2007