Published Ahead of Print on December 15, 2008 as /JCO J Clin Oncol by American Society of Clinical Oncology

Transcription

1 ublished Ahead of rint on December 15, 2008 as /JCO The latest version is at JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E O R T Monitoring rimary Systemic Therapy of Large and Locally Advanced Breast Cancer by Using Sequential ositron Emission Tomography Imaging With [ 18 F]Fluorodeoxyglucose Jörg Schwarz-Dose, Michael Untch, Reinhold Tiling, Stefanie Sassen, Sven Mahner, Steffen Kahlert, Nadia Harbeck, Annette Lebeau, Winfried Brenner, Markus Schwaiger, Fritz Jaenicke, and Norbert Avril From the Departments of Gynecology, Nuclear Medicine, and athology, Universitätsklinikum Hamburg- Eppendorf, Hamburg; Departments of Gynecology, Nuclear Medicine, and athology, Ludwig-Maximilians Universität; and Departments of Gynecology, Nuclear Medicine, and athology, Technische Universität München, Munich, Germany. Submitted March 20, 2008; accepted August 25, 2008; published online ahead of print at on December 15, Supported by Deutsche Krebshilfe and Bristol-Myers Squibb GmbH, Munich, Germany. Authors disclosures of potential conflicts of interest and author contributions are found at the end of this article. Corresponding author: Norbert Avril, MD, Department of Nuclear Medicine, Barts and The London, Queen Mary, University of London, West Smithfield (QE II), London EC1A 7BE, United Kingdom; n.e.avril@qmul.ac.uk by American Society of Clinical Oncology X/08/2799-1/$20.00 DOI: /JCO A B S T R A C T urpose To evaluate positron emission tomography (ET) using [ 18 F]fluorodeoxyglucose (FDG) for prediction of histopathologic response early during primary systemic therapy of large or locally advanced breast cancer. atients and Methods In a prospective multicenter trial, 272 FDG-ET scans were performed in 104 patients at baseline (n 104) and after the first (n 87) and second cycle (n 81) of chemotherapy. The level and relative changes in standardized uptake value (SUV) of FDG uptake were assessed regarding their ability to predict histopathologic response. All patients underwent surgery after chemotherapy, and histopathologic response defined as minimal residual disease or gross residual disease served as the reference standard. s Seventeen (16%) of 104 patients were histopathologic responders and 87 were (84%) nonresponders. All patients for whom baseline SUV was less than 3.0 (n 24) did not achieve a histopathologic response. SUV decreased by 51% 18% after the first cycle of chemotherapy in histopathologic responders (n 15), compared with 37% 21% in nonresponders (n 54;.01). A threshold of 45% decrease in SUV correctly identified 11 of 15 responders, and histopathologic nonresponders were identified with a negative predictive value of 90%. Similar results were found after the second cycle when using a threshold of 55% relative decrease in SUV. Conclusion FDG-ET allows for prediction of treatment response by the level of FDG uptake in terms of SUV at baseline and after each cycle of chemotherapy. Moreover, relative changes in SUV after the first and second cycle are a strong predictor of response. Thus, FDG-ET may be helpful for individual treatment stratification in breast cancer patients. J Clin Oncol by American Society of Clinical Oncology INTRODUCTION Approximately 5% to 10% of newly diagnosed breast cancer patients present with locally advanced disease. 1,2 An increasing number of patients undergo preoperative chemotherapy. Disease-free and overall survival as well as locoregional tumor control is similar compared with primary surgery followed by conventional adjuvant chemotherapy. However, primary chemotherapy often reduces the tumor volume, which enables more patients to be treated with breast-conserving surgery. 3 Histopathology obtained after definitive breast surgery serves as the reference standard for evaluation of response to neoadjuvant chemotherapy. Histopathologic response criteria are based on the extent of residual tumor and regressive changes within tumor tissue. atients with complete pathologic response or minimal residual disease have longer disease-free and overall survival rates compared with nonresponders. 4,5 Approximately 70% of patients demonstrate clinical response to neoadjuvant chemotherapy, but only approximately 20% achieve complete pathologic response. 3,5,6 Therefore, methods that allow prediction of therapeutic effectiveness at an early time point could help to individualize treatment and to avoid potentially ineffective chemotherapies by American Society of Clinical Oncology 1 Copyright 2008 by American Society of Clinical Oncology

2 Schwarz-Dose et al ositron emission tomography (ET) using [ 18 F]fluorodeoxyglucose (FDG) enables visualization of increased glucose metabolism of cancer tissue. FDG-ET has been shown to identify primary tumors, regional lymph nodes, and distant metastases with high diagnostic accuracy for various tumor types, including primary and recurrent breast cancer. 7 Several studies have shown that changes in tumor metabolism occur early in the course of systemic therapy and precede the reduction of tumor size These studies suggested that quantification of tumor glucose metabolism with sequential FDG- ET imaging provided a sensitive means of early detection of response to therapy in breast cancer. 12,15 The aim of this prospective multicenter study was to evaluate sequential FDG-ET for monitoring response to neoadjuvant chemotherapy in breast cancer. We tested the hypothesis that treatment response can be predicted by the level of tumor FDG uptake at baseline before treatment as well as after the first and second cycle of chemotherapy. In addition, we assessed relative changes in FDG uptake early in the course of chemotherapy. FDG-ET was compared with histopathologic response after completion of chemotherapy, which served as the reference standard. ATIENTS AND METHODS atients atients with newly diagnosed large ( 3 cm) or locally advanced noninflammatory breast cancer who participated in a prospective, randomized, multicenter trial comparing two regimens of preoperative chemotherapy (epirubicin and paclitaxel [ET]/plus adjuvant cyclophosphamide, methotrexate, and fluorouracil [CMF]) were eligible for a prospective FDG-ET treatment monitoring study. Study centers at Hamburg-Eppendorf, Ludwig- Maximilians-Universität-München, and Technische-Universität-München (Germany) participated in the ET study. From these centers, 173 patients were recruited into the ET/CMF trial, of which 104 patients participated in the FDG-ET study. atient characteristics are shown in Table 1. Details of the ET/CMF trial and the option to participate in the FDG-ET study were explained, and written informed consent for each study part was obtained from all patients. The study protocol was approved by the local institutional review boards. rimary Chemotherapy The ET/CMF trial compared two regimens: ET standard dose versus ET dose-dense sequential. ET standard dose consisted of four cycles of epirubicin 90 mg/m 2 and paclitaxel 175 mg/m 2 administered every 3 weeks. ET dosedense sequential consisted of three cycles of chemotherapy with epirubicin 150 mg/m 2 every 2 weeks followed by three cycles of paclitaxel 250 mg/m 2 every 2 weeks in combination with filgrastim. After completion of chemotherapy, all patients underwent breast-conserving surgery or mastectomy. Table 1. atient Characteristics Characteristic (n 104) % articipant age, years Median 50 Range Menopausal status remenopausal ostmenopausal ct before chemotherapy T2 (3-5 cm) T T TX 3 3 Tumor size before chemotherapy, cm (n 102) Mean 4.9 SD 1.7 Median 4.5 Range cn before chemotherapy N N N NX Estrogen receptor status ositive Negative Not available 6 6 Histology Invasive ductal carcinoma Invasive lobular carcinoma Invasive medullary carcinoma 1 1 Grade G1 0 0 G G Chemotherapy ET standard-dose regimen ET dose-dense sequential regimen Abbreviations: SD, standard deviation; ET, epirubicin and paclitaxel. ET standard-dose regimen, four cycles of combined epirubicin plus paclitaxel at intervals of 3 weeks. ET dose-dense sequential regimen, three cycles epirubicin followed by three cycles paclitaxel at intervals of 2 weeks. FDG-ET FDG-ET of the breast was obtained before chemotherapy and after the first and second cycle of chemotherapy. Only patients with clearly visible lesions at baseline underwent sequential FDG-ET. atients fasted for at least 6 hours before injection of 280 to 420 MBq [ 18 F]FDG. The mean blood glucose level was mg/dl at baseline ET, mg/dl at ET after the first cycle, and mg/dl at ET after the second cycle of chemotherapy. After an uptake period of 45 minutes, patients were positioned prone with both arms at their sides on the scanner couch. A gap in the scanner support ensured no compression of the breast. Whole-body ET scanners (ECAT951R/31 [Technische-Universität-München], ECAT- Exact47[Hamburg-Eppendorf],andECATExactHR [Ludwig-Maximilians- Universität-München]; Siemens, Knoxville, TN) were used. Emission scans of the breast (two-dimensional mode), acquired in one bed position, were obtained from 45 to 60 minutes after tracer injection followed by a transmission scan with germanium-68 rod sources (Appendix, online-only). No ad-hoc interpretation of sequential FDG-ET was provided to avoid a bias resulting from potential treatment changes on the basis of FDG-ET. After completion of chemotherapy, FDG-ET was analyzed by blinded examiners who were unaware of the results from anatomic imaging or histopathology. There were no prospectively established cutoff values for relative changes in standardized uptake value (SUV) available to define metabolic response. Therefore, it was necessary to derive SUV cutoffs from retrospective receiver operating characteristic (ROC) analysis of this prospective trial. Assessment of Histopathologic Response Histopathologic response was determined as previously described by Honkoop et al. 16 Surgical specimens were cut in 0.5-cm thick slices and evaluated for the presence of macroscopic tumor. Representative samples were taken from all areas of macroscopically visible tumor and resection margins, as well as areas with marked fibrosis or scarring. All sections were microscopically by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

3 FDG-ET Breast Cancer Monitoring analyzed for the presence of residual tumor. Immunohistochemical staining with antibodies against cytokeratins was performed on selected sections to identify or verify tumor residues. athologic complete response (pcr) required additional sampling from macroscopically suggestive as well as uninvolved areas of the surgical specimens. In addition, the tumor bed was identified by signs of tumor regression such as necrosis, presence of macrophages, or marked fibrosis. Specimens with no residual invasive tumor were classified as pcr. Specimens with only few scattered foci of microscopic residual invasive tumor ( 2 mm) were classified as minimal residual disease (MRD). Gross residual disease (GRD) comprised tumors showing macroscopic residual tumor or extensive residual tumor infiltration on microscopic examination. For the purpose of this analysis, pcr and MRD were defined as histopathologic response and GRD as no histopathologic response. Assessment of Metabolic Response First, metabolic response was assessed by the level of FDG uptake in primary breast cancer, expressed as SUV. Metabolic response was assessed independently for baseline FDG-ET and after the first and second cycle of chemotherapy, and compared with histopathologic response. Second, relative changes in tumor FDG uptake expressed as SUV were analyzed regarding their ability to predict histopathologic treatment response by comparing FDG uptake at baseline with ET after the first and second cycle of chemotherapy. Table 2 lists the relative changes. Statistical Analysis The Mann-Whitney test was used to compare SUVs between responders and GRD at a 5% level of significance. To identify an optimal threshold for prediction of histopathologic response, ROC analysis was performed by incrementally increasing cutoff values and recalculating Youden s index. Fisher s exact test was used to test for a correlation between metabolic and histopathologic response. Spearman s rank correlation coefficient ( ) was used to describe correlations between quantitative parameters. All quantitative parameters are expressed as median and range or mean 1 standard deviation. RESULTS Table 2. Relative Changes in FDG Uptake (SUV) After the First and Second Cycle of Chemotherapy Compared With Histopathologic Response Decrease in SUV (%) SD ET after first cycle Responder Nonresponder ET after second cycle Responder Nonresponder NOTE: Only patients with SUV 3.0 at baseline FDG-ET were included in this analysis. Responder and nonresponder were defined by histopathology. values were calculated for the difference in decrease in SUV between responder and nonresponder. Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; ET, positron emission tomography. Histopathologic Response Breast-conserving surgery was performed in 50 patients and mastectomy was performed in 54 patients. Histopathology revealed pcr in 12 patients (12%) and MRD in five patients (5%). Consequently, 17 (16%) of 104 patients were classified as histopathologic responders and 87 patients (84%) were classified as nonresponders. The mean time interval between initiation of chemotherapy and surgery was days, and the mean time interval between the last cycle of chemotherapy and surgery was days. FDG-ET A total of 272 FDG-ET scans were performed on 104 patients. All underwent FDG-ET days before chemotherapy; 87 patients days after the first cycle, and 81 patients days after the second cycle of chemotherapy. SUVs calculated for maximum and average activity values within a tumor region of interest showed a close correlation at baseline ( 0.94) and after the first ( 0.91) and second cycle of chemotherapy ( 0.91). The SUV based on maximum activity values was therefore used for subsequent analyses. rediction of Treatment Response by the Level of FDG Uptake The mean baseline SUV in histopathologic responders (n 17) was , compared with in nonresponders (n 87;.02; Appendix Table A6, online only). Of note, 24 breast carcinomas presented with a baseline SUV less than 3.0, and none achieved histopathologic response. For tumors with a baseline SUV more than 3.0, the SUVs of histopathologic responders and nonresponders were not significantly different (.5). There was a significant correlation between histopathologic response and the level of FDG uptake (SUV) at ET after the first and second cycle of chemotherapy, as listed in Appendix Table A2 (online only). For additional analysis, only patients with a baseline SUV greater than 3.0 were included. In histopathologic responders (n 15), the mean SUV after the first cycle was , compared with in nonresponders (n 54;.03). After the second cycle, the mean SUV in histopathologic responders (n 13) was , compared with in nonresponders (n 51;.02). Various SUV thresholds were evaluated with regard to their ability to predict histopathologic response. After the first cycle, a threshold SUV of 3.5 correctly identified 11 of 15 histopathologic responders with a sensitivity of 73% and a specificity of 59% (.03; Appendix Table A3, online only). The negative predictive value for histopathologic nonresponders was 89%. After the second cycle, a threshold SUV of 2.5 predicted histopathologic response with a sensitivity of 77% and specificity of 59% (.02; Appendix Table A3). The corresponding negative predictive value for histopathologic nonresponders was 91%. Relative Changes of FDG Uptake After the first cycle, SUV decreased by 50.5% 18.4% in histopathologic responders (n 15), compared with 28.4% 26.0% in nonresponders (n 72;.001; Appendix Table A7, online only). We subsequently excluded patients with an SUV less than 3.0 at baseline ET (n 24) from this analysis because these tumors were unlikely to achieve a substantial relative decrease in SUV, given the level of background activity in surrounding normal breast tissue. After the first cycle, SUV decreased by 51% 18% from baseline in patients who achieved histopathologic response (n 15), compared with 37% 21% in nonresponders (n 54;.01). After the second cycle, SUV decreased by 63% 19% from baseline in histopathologic responders (n 13), compared with 48% 19% in nonresponders (n 51;.01; Fig 1) by American Society of Clinical Oncology 3

4 Schwarz-Dose et al A Standardized Uptake Value (%) B Standardized Uptake Value (%) ET Baseline ET Baseline After 1st CTx Cycle After 2nd CTx Cycle rediction of Treatment Response by Relative Changes of FDG Uptake Eighty-seven patients underwent FDG-ET after the first cycle of chemotherapy; of these, 69 with an SUV greater than 3.0 at baseline ET received additional analysis. This group included 15 histopathologic responders and 54 nonresponders. atients who had a decrease in SUV (relative change to baseline) equal or above a defined threshold were classified as responders. By using a threshold of 45% decrease in SUV, 11 of 15 histopathologic responders were correctly identified after the first cycle (sensitivity 73%; specificity 63%). The negative predictive value for histopathologic nonresponders was 90% (Table 4). Sixty-four of 81 patients who underwent FDG-ET after the second cycle had an SUV 3.0 at baseline ET. Among these patients were 13 histopathologic responders and 51 histopathologic nonresponders. After the second cycle, a 55% relative decrease in SUV predicted histopathologic response with a sensitivity of 69% and specificity of 63% (Appendix Table A1, online only). The corresponding negative predictive value for histopathologic nonresponders was 89%. The ROC curve analysis is shown in Figure 2. DISCUSSION Histopathologic nonresponder Histopathologic responder Histopathologic nonresponder Histopathologic responder After 1st CTx Cycle After 2nd CTx Cycle Fig 1. Relative changes in [ 18 F]fluorodeoxyglucose uptake expressed as standardized uptake value from baseline after the first and second cycle of chemotherapy treatment (CTx) in histopathologic responders and nonresponders. Representation of mean values (dots, Fig 1A) and SEs (bars, Fig 1A and box plots (B). ET, positron emission tomography. FDG-ET identified patients with low tumor metabolic activity before treatment who did not achieve histopathologic response. Twenty-four of 104 breast carcinomas (23%) had a baseline SUV less than 3.0, and none responded to chemotherapy. Our findings are supported by recent reports in smaller series Interestingly, tumors with an SUV less than 3.0 were better differentiated and more often steroid-receptor positive as shown in Table 3. However, regarding therapeutic response, there was no correlation between responders and nonresponders and clinicopathologic parameters, including TNM stage, tumor size, histologic subtype, nuclear grade, and estrogen- and progesterone-receptor status. In addition, we found a correlation between the level of tumor FDG uptake after the first and second cycle of chemotherapy and subsequent histopathologic response (Appendix Table A2). To use FDG-ET for treatment stratification, a high negative predictive value for nonresponders is desirable to avoid discontinuing treatment in patients who might still respond to the full course of therapy. After the first and second cycle, the negative predictive values for nonresponders were 89% (threshold SUV of 3.5) and 91% (threshold SUV of 2.5), respectively (Appendix Table A3). However, no defined SUV threshold derived from a single FDG-ET during chemotherapy was clearly superior for response prediction. Table 3. Clinicopathologic Characteristics of Tumors With Baseline FDG Uptake (SUV) 3.0 Compared With Those With Baseline SUV 3.0 Baseline Characteristic SUV 3.0 (n 24) SUV 3.0 (n 80) % % Histology Invasive ductal carcinoma Invasive lobular carcinoma Invasive medullary carcinoma Grade G G G rogesterone receptor status ositive Negative Not available Estrogen receptor status ositive Negative Not available cn before chemotherapy N N N NX ct before chemotherapy T2, 3-5 cm NS T T TX Tumor size before chemotherapy, cm (n 102) Mean NS SD NS Median Range Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; NS, not significant; SD, standard deviation by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

5 FDG-ET Breast Cancer Monitoring Table 4. Various Thresholds for Relative Decrease in FDG Uptake (SUV) to redict Histopathologic Response After the First Cycle of Chemotherapy Threshold for Decrease in SUV After First Cycle (n 69) 20% 25% 30% 35% 40% 45% 50% ositive Negative True positive False positive True negative False negative % Sensitivity Specificity V NV Accuracy NOTE: Only patients with SUV 3.0 at baseline FDG-ET were included in this analysis. Histopathology served as the reference standard. Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; V, positive predictive value; NV, negative predictive value. ositive is defined as relative decrease in SUV equal to or above the threshold. Negative is defined as relative decrease in SUV below the threshold. In contrast, we found relative changes in FDG uptake to be potentially more clinically applicable. After the first cycle, SUV decreased by 51% 18% in histopathologic responders, compared with 37% 21% in nonresponders (.01; Table 2). An additional decrease of 63% 19% from baseline was observed after the second cycle in responders, versus 48% 19% in nonresponders (.01). These findings are in line with previous reports that metabolic FDG-ET imaging provides important information about the response to treatment in breast cancer and other tumors. 8,9,12,14,15,20,21 Tumors responding to chemotherapy exhibit a rapid decrease in glucose metabolism early in the course of therapy that generally precedes morphological changes. 7,10,22 FDG-ET after the first cycle of chemotherapy identified histopathologic nonresponders with a negative predictive value of 90% when using a relative decrease in SUV of less Survival robability Youden-Index Distance Decrease in Standardized Uptake Value (%) Fig 2. rediction of treatment response by changes of [ 18 F]fluorodeoxyglucose uptake after first cycle of chemotherapy. Analysis of receiver operating characteristic performed by determining maximum Youden s index and minimum distance. Lower curve peak indicates optimal threshold for prediction of response between 40% and 50% relative change (best threshold 47.6% decrease in standardized uptake value). than 45% as cutoff. Similar results were recently reported with a negative predictive value of 85% using a cutoff of 40% decrease in SUV. 19 Interestingly, the relative change in FDG uptake in breast cancer is not the same magnitude of change that is observed, for example, in lymphoma or GI stromal tumors. Our study extends previous experience and, to our knowledge, is the first prospective multicenter trial. Wahl et al 12 found in 11 locally advanced breast cancers that FDG uptake decreased promptly with treatment in eight responding patients. In 30 breast cancer patients who received eight cycles of neoadjuvant chemotherapy, the reduction in FDG uptake after the first cycle was significantly higher in tumors with complete response. 15 After a single cycle of chemotherapy, FDG-ET predicted pcr with a sensitivity of 90% and specificity of 74%. A pilot series of 22 breast cancer patients found similar results with an accuracy of 88% and 91% to predict histopathologic response after the first and second cycle of neoadjuvant chemotherapy, respectively. 14 The larger number of patients in our study, however, provides a more robust assessment of temporal changes in FDG uptake. When comparing previous reports, the definition of treatment response as reference for metabolic response is critical. In our study, pathologic response was defined as minimal residual disease or pcr. 16 However, the response criteria vary among studies; histopathologic response rates ranging between 16.3% and 55.6% limit the direct comparison between FDG-ET results. 14,15,18,19 Histopathologic response criteria have distinct limitations. Between 13% and 25% of patients experience systemic recurrence during 5-year follow-up despite pcr In addition, the rate of pcr is relatively low, ranging between 3% and 27%. 3,5,6 A subgroup of patients classified as histopathologic nonresponders might have responded but are not recognized by current criteria. Furthermore, response to chemotherapy is a continuous variable rather than a dichotomized decision. It might therefore be important to establish criteria that relate tumor response to patient outcome to guide subsequent treatment decisions. Therapy-induced changes in tumor glucose metabolism could potentially play such a role by American Society of Clinical Oncology 5

6 Schwarz-Dose et al We found that the greater the reduction in tumor metabolic activity (represented as SUV) early in the course of chemotherapy, the more likely patients will achieve a histopathologic response. rospectively validated metabolic response criteria could potentially serve as new surrogate end points in future trials. Metabolic response might potentially provide complementary information and may be better suited to quantitatively assess continuous grades of response. An important advantage of FDG-ET is that it provides an in vivo test of chemosensitivity early during therapy, when a change in treatment is still possible. The metabolic information correlates well with patient survival and is distinctly different from the histopathologic information at a later stage. In ovarian cancer, metabolic responders identified after the first cycle of chemotherapy had a median overall survival of 38.3 months, compared with 23.1 months in metabolic nonresponders. In addition, metabolic changes were superior to histopathologic criteria in defining treatment response. 27 Similar results were found in lung, esophageal, and gastric cancer. 9,20,28 In metastatic breast cancer, metabolic responders also had a better overall survival rate (19.2 v 8.8 months). 29 Similarly, the decrease in FDG uptake correlated with time to progression in bone-dominant metastatic breast cancer. 30 Although our study confirms the close correlation between an early decrease in tumor glucose metabolism and response to treatment, certain limitations have to be taken into account. Even though this study represents the largest patient group reported so far, the sample size is limited. The purpose of our study was not to compare two chemotherapy regimes, but to investigate early metabolic changes to chemotherapy regarding their ability to predict histopathologic response. Therefore, the specific dose regimen (ET standard dose v ET dose-dense sequential) was not expected to influence the ability of FDG-ET to predict outcome. However, the effect of the different time interval between initiation of chemotherapy and FDG-ET after the first and second cycle in the two treatment arms remains uncertain. Questions also remain regarding the optimal thresholds to define metabolic response. Of crucial importance is how FDG-ET treatment monitoring could potentially be integrated into the various clinical scenarios. Recently, results became available from the first phase II trial which used early FDG-ET for treatment stratification of patients with adenocarcinoma of the esophagogastric junction. 31 All patients in which tumor FDG uptake decreased by less then 35% from baseline were classified as metabolic nonresponders and went directly to surgery after the first cycle of chemotherapy, whereas metabolic responders completed 12 weeks of primary systemic therapy followed by surgery. Early termination of chemotherapy did not negatively influence clinical outcome. 31 Future studies need to address whether breast cancer patients would benefit from a FDG-ET guided treatment approach as well In addition, other ET tracers targeting tumor cell proliferation, steroid receptor expression, or blood flow might also provide relevant information regarding treatment response. 35,36 In conclusion, our results suggest a potential clinical application of FDG-ET in guiding systemic therapy of patients with locally advanced breast cancer on the basis of pretherapy tumor metabolic activity. FDG-ET could be particularly helpful in identifying metabolically active tumors that may gain the most benefit from neoadjuvant chemotherapy. This hypothesis needs additional evaluation in prospective trials. Relative changes in metabolic activity after the first and second cycle of chemotherapy seem to be a strong predictor of response. The greater the SUV reduction early in the course of chemotherapy, the more likely patients will achieve histopathologic response. FDG-ET may therefore be helpful for individual treatment stratification in breast cancer patients. AUTHORS DISCLOSURES OF OTENTIAL CONFLICTS OF INTEREST The authors indicated no potential conflicts of interest. AUTHOR CONTRIBUTIONS Conception and design: Jörg Schwarz-Dose, Michael Untch, Markus Schwaiger, Norbert Avril Administrative support: Reinhold Tiling, Markus Schwaiger, Fritz Jaenicke rovision of study materials or patients: Jörg Schwarz-Dose, Michael Untch, Reinhold Tiling, Sven Mahner, Steffen Kahlert, Nadia Harbeck, Annette Lebeau, Winfried Brenner, Markus Schwaiger, Fritz Jaenicke, Norbert Avril Collection and assembly of data: Jörg Schwarz-Dose, Reinhold Tiling, Stefanie Sassen, Sven Mahner, Steffen Kahlert, Nadia Harbeck, Annette Lebeau, Winfried Brenner, Norbert Avril Data analysis and interpretation: Jörg Schwarz-Dose, Stefanie Sassen, Winfried Brenner, Norbert Avril Manuscript writing: Jörg Schwarz-Dose, Stefanie Sassen, Nadia Harbeck, Annette Lebeau, Norbert Avril Final approval of manuscript: Jörg Schwarz-Dose, Michael Untch, Reinhold Tiling, Stefanie Sassen, Sven Mahner, Steffen Kahlert, Nadia Harbeck, Annette Lebeau, Winfried Brenner, Markus Schwaiger, Fritz Jaenicke, Norbert Avril REFERENCES 1. Giordano SH: Update on locally advanced breast cancer. Oncologist 8: , Cristofanilli M, Buzdar AU, Hortobagyi GN: Update on the management of inflammatory breast cancer. Oncologist 8: , van der Hage JA, van de Velde CJ, Julien J, et al: reoperative chemotherapy in primary operable breast cancer: s from the European Organization for Research and Treatment of Cancer Trial J Clin Oncol 19: , Feldman LD, Hortobagyi GN, Buzdar AU, et al: athological assessment of response to induction chemotherapy in breast cancer. Cancer Res 46: , Fisher ER, Wang J, Bryant J, et al: athobiology of preoperative chemotherapy: Findings from the National Surgical Adjuvant Breast and Bowel (NSAB) rotocol B-18. Cancer 95: , Bonadonna G, Valagussa, Brambilla C, et al: rimary chemotherapy in operable breast cancer: Eight-year experience at the Milan Cancer Institute. J Clin Oncol 16:93-100, Rohren EM, Turkington TG, Coleman RE: Clinical applications of ET in oncology. Radiology 231: , Flamen, Van Cutsem E, Lerut A, et al: ositron emission tomography for assessment of the response to induction chemotherapy in locally advanced esophageal cancer. Ann Oncol 13: , Mac Manus M, Hicks RJ, Matthews J, et al: ositron emission tomography is superior to computed tomography scanning for responseassessment after radical radiotherapy or chemoradiotherapy in patients with non small-cell lung cancer. J Clin Oncol 21: , Römer W, Hanauske AR, Ziegler S, et al: ositron emission tomography in non-hodgkin s lymphoma: Assessment of chemotherapy with fluorodeoxyglucose. Blood 91: , Spaepen K, Stroobants S, Dupont, et al: rognostic value of positron emission tomography (ET) with fluorine-18 fluorodeoxyglucose ([18F]FDG) after first-line chemotherapy in non-hodgkin s lymphoma: Is [18F]FDG- ET a valid alternative to conventional diagnostic methods? J Clin Oncol 19: , by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

7 FDG-ET Breast Cancer Monitoring 12. Wahl RL, Zasadny K, Helvie M, et al: Metabolic monitoring of breast cancer chemohormonotherapy using positron emission tomography: Initial evaluation. J Clin Oncol 11: , Weber WA, Ott K, Becker K, et al: rediction of response to preoperative chemotherapy in adenocarcinomas of the esophagogastric junction by metabolic imaging. J Clin Oncol 19: , Schelling M, Avril N, Nahrig J, et al: ositron emission tomography using [(18)F]Fluorodeoxyglucose for monitoring primary chemotherapy in breast cancer. J Clin Oncol 18: , Smith IC, Welch AE, Hutcheon AW, et al: ositron emission tomography using [(18)F]- fluorodeoxy-d-glucose to predict the pathologic response of breast cancer to primary chemotherapy. J Clin Oncol 18: , Honkoop AH, van Diest J, de Jong JS, et al: rognostic role of clinical, pathological and biological characteristics in patients with locally advanced breast cancer. Br J Cancer 77: , Doot RK, Dunnwald LK, Schubert EK, et al: Dynamic and static approaches to quantifying 18F- FDG uptake for measuring cancer response to therapy, including the effect of granulocyte CSF. J Nucl Med 48: , McDermott GM, Welch A, Staff RT, et al: Monitoring primary breast cancer throughout chemotherapy using FDG-ET. Breast Cancer Res Treat 102:75-84, Rousseau C, Devillers A, Sagan C, et al: Monitoring of early response to neoadjuvant chemotherapy in stage II and III breast cancer by [18F]fluorodeoxyglucose positron emission tomography. J Clin Oncol 24: , Downey RJ, Akhurst T, Ilson D, et al: Whole body 18FDG-ET and the response of esophageal cancer to induction therapy: s of a prospective trial. J Clin Oncol 21: , Weber WA, etersen V, Schmidt B, et al: ositron emission tomography in non small-cell lung cancer: rediction of response to chemotherapy by quantitative assessment of glucose use. J Clin Oncol 21: , Wieder H, Brucher BL, Zimmerman F, et al: Time course of tumor metabolic activity during chemoradiotherapy of esophageal squamous cell carcinoma and response to treatment. J Clin Oncol 22: , Kuerer HM, Newman LA, Smith TL, et al: Clinical course of breast cancer patients with complete pathologic primary tumor and axillary lymph node response to doxorubicin-based neoadjuvant chemotherapy. J Clin Oncol 17: , Fisher B, Bryant J, Wolmark N, et al: Effect of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 16: , Chollet, Amat S, Cure H, et al: rognostic significance of a complete pathological response after induction chemotherapy in operable breast cancer. Br J Cancer 86: , Gonzalez-Angulo AM, McGuire SE, Buchholz TA, et al: Factors predictive of distant metastases in patients with breast cancer who have a pathologic complete response after neoadjuvant chemotherapy. J Clin Oncol 23: , Avril N, Sassen S, Schmalfeldt B, et al: rediction of response to neoadjuvant chemotherapy by sequential F-18-fluorodeoxyglucose positron emission tomography in patients with advancedstage ovarian cancer. J Clin Oncol 23: , Ott K, Fink U, Becker K, et al: rediction of response to preoperative chemotherapy in gastric carcinoma by metabolic imaging: s of a prospective trial. J Clin Oncol 21: , Dose Schwarz J, Bader M, Jenicke L, et al: Early prediction of response to chemotherapy in metastatic breast cancer using sequential 18F-FDG ET. J Nucl Med 46: , Specht JM, Tam SL, Kurland BF, et al: Serial 2-[(18)F] fluoro-2-deoxy-d-glucose positron emission tomography (FDG-ET) to monitor treatment of bone-dominant metastatic breast cancer predicts time to progression (TT). Breast Cancer Res Treat 105:87-94, Lordick F, Ott K, Krause BJ, et al: ET to assess early metabolic response and to guide treatment of adenocarcinoma of the oesophagogastric junction: The MUNICON phase II trial. Lancet Oncol 8: , Bear HD, Anderson S, Smith RE, et al: Sequential preoperative or postoperative docetaxel added to preoperative doxorubicin plus cyclophosphamide for operable breast cancer: National Surgical Adjuvant Breast and Bowel roject rotocol B-27. J Clin Oncol 24: , von Minckwitz G, Blohmer JU, Raab G, et al: In vivo chemosensitivity-adapted preoperative chemotherapy in patients with early-stage breast cancer: The GEARTRIO pilot study. Ann Oncol 16:56-63, von Minckwitz G, Raab G, Caputo A, et al: Doxorubicin with cyclophosphamide followed by docetaxel every 21 days compared with doxorubicin and docetaxel every 14 days as preoperative treatment in operable breast cancer: The GEARDUO study of the German Breast Group. J Clin Oncol 23: , Linden HM, Stekhova SA, Link JM, et al: Quantitative fluoroestradiol positron emission tomography imaging predicts response to endocrine treatment in breast cancer. J Clin Oncol 24: , Tseng J, Dunnwald LK, Schubert EK, et al: 18F-FDG kinetics in locally advanced breast cancer: Correlation with tumor blood flow and changes in response to neoadjuvant chemotherapy. J Nucl Med 45: , 2004 Acknowledgment We deeply appreciate the help from all staff involved in carrying out this study, as well as the excellent statistical support and advice from Mrs Raymonde Busch. We are grateful to all patients for their cooperation by American Society of Clinical Oncology 7

8 Schwarz-Dose et al Appendix Emission data corrected for random events, dead time, and attenuation were reconstructed with filtered back projection or iterative reconstruction. The same reconstruction method was used for sequential imaging in individual patients. The image pixel counts were calibrated to activity-concentration (Bq/mL) and decay corrected using the time of tracer injection as reference. Regions of interest (ROIs) were placed semi-automatically over all breast tumors. The slice with the highest radioactivity concentration within the tumor was identified and a circular ROI with a diameter of 1.0 cm was placed in this area and in the directly adjacent slices. Standardized uptake values were calculated using the average and maximum activity values within the ROIs normalized to injected dose and patient s body weight. Table A1. Various Thresholds for Relative Decrease in FDG Uptake (SUV) to redict Histopathologic Response After the Second Cycle of Chemotherapy Threshold for Decrease in SUV After Second Cycle (n 64) 30% 35% 40% 45% 50% 55% 60% ositive Negative True positive False positive True negative False negative % Sensitivity Specificity V NV Accuracy NOTE: Only patients with a SUV 3.0 at baseline FDG-ET were included in this analysis. Histopathology served as the reference standard. Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; V, positive predictive values; NV, negative predictive values. ositive is defined as relative decrease in SUV equal to or above the threshold. Negative is defined as relative decrease in SUV below the threshold. Table A2. Level of FDG Uptake (SUV) in Breast Cancer at ET Baseline, After First Cycle and After Second Cycle of Chemotherapy Compared With Histopathologic Response Mean SD ET baseline Responder Nonresponder ET after first cycle Responder Nonresponder ET after second cycle Responder Nonresponder NOTE: Only patients with SUV 3.0 at baseline FDG-ET were included in this analysis. Responder and nonresponder were defined by histopathology. values were calculated for the difference in decrease in SUV between responder and nonresponder. Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; ET, positron emission tomography; SD, standard deviation. SUV by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

9 FDG-ET Breast Cancer Monitoring Table A3. Various Thresholds for the Level of FDG Uptake (SUV) in Breast Cancer to redict Histopathologic Response After the First and Second Cycles of Neoadjuvant Chemotherapy SUV Threshold SUV after first cycle CTx (n 69) ositive Negative True positive False positive True negative False negative % Sensitivity Specificity V NV Accuracy Threshold SUV after second cycle CTx (n 64) ositive Negative True positive False positive True negative False negative % Sensitivity Specificity V NV Accuracy NOTE: Only patients with SUV 3.0 at baseline FDG-ET were included in this analysis. Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; CTx, chemotherapy treatment; V, positive predictive values; NV, negative predictive values. ositive is defined as relative decrease in SUV equal to or above the threshold. Negative is defined as relative decrease in SUV below the threshold. Table A4. Level of FDG Uptake (SUV) in Breast Cancer at ET Baseline, After First Cycle and After Second Cycle of Chemotherapy With pcr As the Reference Standard Mean SD ET baseline pcr Non-pCR ET after first cycle pcr Non-pCR ET after second cycle pcr Non-pCR Abbreviations: pcr, pathologic complete response; FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; SD, standard deviation; ET, positron emission tomography. SUV by American Society of Clinical Oncology 9

10 Schwarz-Dose et al Table A5. Relative Changes in FDG Uptake (SUV) After the First and Second Cycle of Chemotherapy With pcr As the Reference Standard Decrease in SUV % SD ET after first cycle pcr Non-pCR ET after second cycle pcr Non-pCR NOTE: Only patients with a SUV 3.0 at baseline FDG-ET were included in this analysis. values were calculated for the difference in SUV between the groups of pcr and non-pcr. Abbreviations: pcr, pathologic complete response; FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; SD, standard deviation; ET, positron emission tomography. Table A6. Level of FDG Uptake (SUV) in Breast Cancer at ET Baseline, After First Cycle and After Second Cycle of Chemotherapy Compared With Histopathologic Response, Including All atients Independent of Baseline SUV Mean SD ET baseline Responder Nonresponder ET after first cycle Responder Nonresponder ET after second cycle Responder Nonresponder Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; ET, positron emission tomography; SD, standard deviation. SUV Table A7. Relative Changes in FDG Uptake (SUV) After the First and Second Cycle of Chemotherapy Compared With Histopathologic Response, Including All atients Independent of Their Baseline SUV Decrease in SUV % SD ET after first cycle Responder Nonresponder ET after second cycle Responder Nonresponder NOTE: Responder and nonresponder were defined by histopathology. values were calculated for the difference in SUV between responder and nonresponder. Abbreviations: FDG, 18 F fluorodeoxyglucose; SUV, standardized uptake value; SD, standard deviation; ET, positron emission tomography by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

11 FDG-ET Breast Cancer Monitoring Table A8. atient Characteristics, Clinicopathologic Tumor arameters, and Histopathologic Response for Both Treatment Arms Characteristic Standard CTx (n 59) Sequential CTx (n 45) % % articipant age, years.4 Median Range Menopausal status remenopausal ostmenopausal ct before chemotherapy T2 (3-5 cm) T T TX Tumor size before chemotherapy (cm; n 102) Mean SD Median Range cn before chemotherapy N N N NX Estrogen receptor status ositive Negative Not available Histology Invasive ductal carcinoma Invasive lobular carcinoma Invasive medullary carcinoma Grade G G G Histopathologic response Responder Nonresponder athologic complete response pcr Non-pCR NOTE: Standard CTx was ET standard-dose regimen, consisting of four cycles of combined epirubicin plus paclitaxel at intervals of 3 weeks. Sequential CTx was ET dose-dense sequential regimen, consisting of three cycles epirubicin followed by three cycles paclitaxel at intervals of 2 weeks. values were calculated for differences between the two treatment arms (standard v sequential CTx). Abbreviations: CTx, chemotherapy treatment; SD, standard deviation; pcr, pathologic complete response; ET, epirubicin and paclitaxel by American Society of Clinical Oncology 11