Early Prediction of Response to Chemotherapy in Metastatic Breast Cancer Using Sequential ¹⁸F-FDG PET

Patients

Patients with the diagnosis of metastatic breast cancer who were participating in a phase III chemotherapy trial were offered enrollment in this prospective evaluation of sequential ¹⁸F-FDG PET. The purpose of the chemotherapy trial was to compare standard chemotherapy using epirubicin/cyclophosphamide with epirubicin/paclitaxel. Inclusion criteria for sequential ¹⁸F-FDG PET were the diagnosis of metastatic breast cancer, at least 1 measurable metastatic lesion seen on conventional anatomic imaging, and an age of 18–70 y. Exclusion criteria were pregnancy, breast feeding, and diabetes mellitus. Between February 1997 and February 2000, 11 patients were enrolled to undergo sequential ¹⁸F-FDG PET for monitoring of treatment response. All patients were treated at the Department of Obstetrics and Gynecology of the University Hospital Hamburg-Eppendorf, Germany (patient characteristics are shown in Table 1). All patients were followed up either until death or until April 2004. The Committee for Human Research at the University Hospital Hamburg-Eppendorf approved the study protocol. A physician explained the details of the study to the patients, and all consented in writing to participate.

Chemotherapy

The patients were treated with chemotherapy according to a phase III study protocol with epirubicin/cyclophosphamide versus epirubicin/paclitaxel. Chemotherapy was repeated every 3 wk. Nine patients received epirubicin/paclitaxel (60/175 mg/m² of body surface), and 2 patients received epirubicin/cyclophosphamide (60/600 mg/m² of body surface). Chemotherapy was discontinued in patients who showed progressive disease on conventional imaging, and they subsequently were excluded from further participation in the study. Patients with no change, partial remission, or complete remission received additional cycles of chemotherapy up to a maximum of 10 cycles.

Diagnostic Procedures

Patients underwent conventional imaging, including ultrasound, plain film radiography, contrast-enhanced CT, and MRI, depending on the localization of the metastatic lesions. The imaging procedures were performed according to routine clinical practice. Ultrasound was performed by an experienced radiologist, plain film radiographs were obtained in at least 2 projections, and contrast-enhanced CT or MRI was performed if no contraindications were present. A total of 26 separate metastases had been identified before the patients were enrolled in sequential ¹⁸F-FDG PET. The imaging procedures were repeated after 3 cycles of chemotherapy (9 wk), 6 cycles of chemotherapy (18 wk), and 9 cycles of chemotherapy (27 wk).

Classification of Response Seen on Conventional Imaging

Response to treatment was classified according to the criteria of the World Health Organization (2): Complete response was determined as resolution of metastatic lesions; partial response, as a reduction in size (product of the 2 largest perpendicular dimensions) of more than 50%; no change, as a reduction of less than 50% or an increase of less than 25%; and progressive disease, as an increase (product of the 2 largest perpendicular dimensions) of more than 25%. For conventional imaging, patients with no change or with partial or complete response were classified as responders to chemotherapy and patients with progressive disease were classified as nonresponders to chemotherapy.

¹⁸F-FDG PET Imaging

¹⁸F-FDG PET was performed at baseline before chemotherapy, after the first cycle (3 wk) of chemotherapy, and after the second cycle (6 wk) of chemotherapy. Eleven patients underwent a total of 31 ¹⁸F-FDG PET examinations; 2 patients did not undergo a third ¹⁸F-FDG PET examination after the second cycle (6 wk) of chemotherapy. A whole-body PET scanner (ECAT EXACT 47/921; CTI Siemens, Inc.) was used, and patients fasted for at least 6 h before undergoing PET. The serum glucose level was measured before the intravenous administration of 240–400 MBq (approximately 10 mCi) of ¹⁸F-FDG. The blood glucose level in all patients was less than 150 mg/dL. All patients lay supine during the study, after being comfortably positioned on the scanner table with both arms at their sides. Emission scans (2-dimensional) were started 60 min after intravenous administration of ¹⁸F-FDG, with 10 min allowed per bed position. Depending on the location of metastases, emission scans of 1–4 bed positions were acquired, followed by transmission scanning for attenuation correction. Additionally, whole-body emission scans from the base of the skull to the groin were obtained without attenuation correction. Emission data corrected for random events, dead time, and attenuation were reconstructed with filtered backprojection (Hanning filter with cutoff frequency of 0.4 cycles per bin). The image pixel counts were calibrated to activity concentration (Bq/mL) and were decay corrected using the time of tracer injection as a reference. For visual analysis, PET images were printed on transparency films (Helios 810; Sterling Diagnostic Imaging) using a linear gray scale with the highest activity displayed in black. Standard documentation on film included 20 transverse slices with a slice thickness of 13.5 mm, 20 coronal slices with a slice thickness of 13.5 mm, and maximum-intensity projections in anterior, left lateral, right anterior oblique, and left anterior oblique views. In addition, a monitor was used with full control over the display.

PET Image Analysis

The region-of-interest technique was used for quantification of ¹⁸F-FDG uptake in metastatic lesions. Circular regions of interest were placed manually over each lesion by 1 observer. The maximum activity values within the regions of interest were normalized to injected activity and patient body weight and were corrected for variation in blood glucose levels at the time of tracer injection by normalizing to a level of 100 mg/100 mL (SUVs) (18).

An experienced nuclear medicine physician unaware of the patient’s history, clinical findings, and conventional imaging results interpreted the whole-body PET images visually. Response was classified according to the following criteria: Complete response was defined as resolution of abnormal ¹⁸F-FDG uptake in metastatic lesions; partial response, as a reduction in the intensity of uptake or in the number of metastatic lesions with increased uptake; no change, as no change in the number of metastatic lesions and in the intensity of uptake in metastatic lesions; and progressive disease, as an increase in the intensity of uptake or in the number of metastatic lesions. Patients with ¹⁸F-FDG PET scans showing partial or complete response were classified as responders to chemotherapy, and patients with scans showing no change or progressive disease were classified as nonresponders.

Statistical Analysis

The Mann–Whitney U test was used to compare SUVs between responding and nonresponding metastases at a 5% level of significance.

Conventional Imaging

Conventional imaging was repeated after the third, sixth, and ninth cycles of chemotherapy and compared with the pretreatment findings. Table 2 shows the treatment response for individual metastases after the third and sixth cycles of chemotherapy. Two of 11 patients had a mixed response, with metastases classified as no change and partial response. After the third cycle of chemotherapy, 9 patients were classified as responders (1, complete response; 5, partial response; and 3, no change) and 2 patients as nonresponders (progressive disease). In accord with the protocol of the chemotherapy trial, the 2 patients who did not respond were excluded from further participation in the study. The findings on conventional imaging changed after the sixth cycle of chemotherapy; only 6 patients were classified as responders and another 3 patients as nonresponders (Table 3). A patient with a lung metastasis (patient 8) initially responded after the third cycle (partial response), but the metastasis progressed in size after the sixth cycle of chemotherapy (progressive disease). There was no change between the sixth and ninth cycles of chemotherapy, and responders who continued to receive chemotherapy were still classified as responders after the ninth cycle of chemotherapy. In summary, 6 of 11 patients were classified as responders and 5 patients as nonresponders on the basis of conventional imaging results after completion of chemotherapy, which determined the gold standard for this study. Therefore, conventional imaging after the third cycle of chemotherapy misclassified 3 patients (27%) as responders (Table 3).

PET Quantitative Image Analysis (Lesion Based)

Conventional imaging before initiation of chemotherapy identified a total of 26 separate metastases of which 17 metastatic lesions responded after completion of chemotherapy. In those lesions, SUVs decreased to 72% ± 21% after the first cycle and to 54% ± 16% after the second cycle, when compared with the baseline PET scan. In contrast, ¹⁸F-FDG uptake in lesions not responding to chemotherapy (n = 9) declined only to 94% ± 19% after the first cycle and to 79% ± 9% after the second cycle (Fig. 1). The differences between responders and nonresponders were statistically significant after the first (P = 0.02) and second (P = 0.003) cycles.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Changes in SUV in responding and nonresponding metastases.

PET Visual Image Analysis (Patient Based)

Whole-body ¹⁸F-FDG PET images obtained after the first (n = 11) and second (n = 9) cycles of chemotherapy were viewed and compared with the baseline ¹⁸F-FDG PET image. After the first cycle of chemotherapy, 6 patients were classified as metabolic responders and 5 patients as nonresponders by ¹⁸F-FDG PET. These classifications did not change after the second cycle of chemotherapy. ¹⁸F-FDG PET correctly predicted the response in all patients as early as after the first cycle of chemotherapy. Compared with conventional imaging procedures obtained after the third cycle of chemotherapy, ¹⁸F-FDG PET was superior specifically in the identification of nonresponders. In the present study, ¹⁸F-FDG PET identified 2 nonresponding patients 3 wk (1 cycle of chemotherapy) and 3 nonresponding patients 12 wk (4 cycles of chemotherapy) before progression was detected on conventional imaging. Use of a PET-based strategy could have avoided 14 cycles of chemotherapy in 5 patients, and these patients could have received earlier second-line chemotherapy. ¹⁸F-FDG PET also correctly classified 3 patients who had no change on conventional imaging after the third cycle of chemotherapy; 1 turned out to be a responder and 2, nonresponders.

Patient Follow-up and Survival

Mean follow-up was 14.5 mo (range, 2–39 mo). During this time, 10 patients died and mean overall survival was 14.5 ± 3.6 mo. The overall survival in nonresponders identified by ¹⁸F-FDG PET after the first cycle of chemotherapy was 8.8 ± 6.7 mo, compared with 19.2 ± 13.6 mo in responders (not significant).