Monitoring the size and response of locally advanced ...

Breast Cancer Research and Treatment 78: 51–58, 2003. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

Report

Monitoring the size and response of locally advanced breast cancers to neoadjuvant chemotherapy (weekly paclitaxel and epirubicin) with serial enhanced MRI Yun-Chung Cheung1,6 , Shih-Cheh Chen2 , Min-Ying Su8 , Lai-Chu See7 , Swei Hsueh3 , Hsien-Kun Chang4 , Yung-Chang Lin4 , and Chien-Sheng Tsai5 1 Department

of Diagnostic Radiology, 2 Department of Surgery, 3 Department of Pathology, 4 Division of Hematology–Oncology, Department of Internal Medicine, 5 Department of Radiation Oncology, Chang Gung Memorial Hospital, 6 College of Medicine and School of Medical Technology, 7 Department of Public Health, Chang Gung University, Kwei Shan, Tao Yuan Hsien, Taiwan; 8 Health Sciences Research Imaging Center, College of Medicine, University of California, CA, USA

Key words: advanced breast cancer, chemotherapeutic response, chemotherapy, magnetic resonance imaging, preoperative planning, size monitoring, treatment

Summary Purpose. To determine if early cancer size reduction seen on enhanced magnetic resonance imaging (MRI) can serve as a response predictor and to correlate final tumor sizes on MRI and excised gross tumor size to microscopic findings in patients with locally advanced breast cancers treated with preoperative neoadjuvant chemotherapy. Methods and materials. Thirty-three patients with advanced breast cancer entered this prospective chemotherapeutic study. Serial, dynamic, enhanced MRI was performed before chemotherapy induction, after the first course of chemotherapy and after the third course of chemotherapy prior to surgery. Responses were measured by image subtraction of tumor size on subsequent axial MRIs using the response evaluation criteria in solid tumors (RECIST). Early tumor size reduction, percentage of relative early tumor size reduction and final tumor size response were calculated and analyzed statistically. Sizes of residual tumors measured on MRI and gross tumors in excised breasts were correlated with microscopic findings. Results. Based on tumor sizes measured with enhanced MRI, four complete responders (CR), 19 partial responders (PR) and 10 non-responder were documented. Twelve (52%) of the 23 responders (CR and PR) had reached the criteria for PR (≥30% size reduction) after the first course of chemotherapy. All CR had a marked early size reduction (ESR) of more than 45%. Using the receiver operating characteristic (ROC) curve, a good cutoff point for early tumor size reduction was 7.4 cm, with a false positive rate of 0.1 and a false negative rate of 0.13. The percentage of ESR was 8.8%, with a false positive rate of 0.1 and a false negative rate of 0.09. Residual tumor size on MRI correlated well with microscopic findings (r = 0.982, p < 0.001) and gross tumor size in excised breasts correlated moderately with microscopic findings (r = 0.640, p < 0.001). Conclusion. Serial, dynamic, enhanced MRI monitoring of chemotherapeutic response in patients with locally advanced breast cancer can be used to assess early response to chemotherapy and post-chemotherapy tumor size change. Although the residual tumor size on MRI correlated well with the microscopic findings, surgical determination of residual cancer load is still recommended to avoid underestimation. Introduction Neoadjuvant chemotherapy was introduced in the 1970s; it has now become standard for locally

advanced breast cancer and has been shown to improve both relapse-free and overall survival [1–8]. The multimodal treatment typically comprises chemotherapy induction before surgery, surgery, post-operative

52

Y-C Cheung et al.

chemotherapy and external radiation therapy. The presence or absence of residual cancer has been proposed as a critical prognostic factor for prolonged disease-free and overall survival [6, 7, 9]. Postchemotherapeutic assessment of cancer size as well as the detection of residual cancer is difficult by means of physical examination, mammography or sonography due to the development of chemotherapyinduced fibrosis [5, 6, 10, 11]. Recently, a series of reports documented how magnetic resonance imaging (MRI) can accurately evaluate chemotherapeutic response, offering high resolution, enhancement techniques, rapid image subtraction processing and simple assessment using subtracted images [12–15]. This report describes the use of dynamic, enhanced MRI to serially monitor the chemotherapeutic responses in locally advanced breast cancers treated with weekly paclitaxel and epirubicin. MRI examinations were performed before chemotherapy induction, after the first course of chemotherapy and after the third course of chemotherapy prior to surgery. Sizes of residual tumors as measured using subtracted MRI imaging, palpable residual tumors in the excised breasts and microscopic evaluation were recorded. These data were analyzed statistically. The two aims of this prospective study were to determine if early size reduction (ESR) of tumor after the first course of chemotherapy enabled predictions regarding the final chemotherapeutic response to be made and to establish how well MRI detected residual cancer after chemotherapy. Both of these aims are important issues in the chemotherapeutic management of locally advanced breast cancer.

Materials and methods This project was sponsored by a grant from the National Science Council of Republic of China and by a clinical research project from Chang Gung Memorial Hospital. Patient selection Between December 1999 and November 2001, 33 women with ages ranging from 29 to 63 years (mean, 44.94 years) were recruited. All patients had histologically proven locally advanced breast cancer based according to minimal excisional biopsies or core needle biopsies and were to receive preoperative neoadjuvant chemotherapy. The pathological diagnoses included 31 cases of invasive ductal carcinoma, one case of mucinous carcinoma and one of invasive

lobular carcinoma. Based on clinical examination, mammography and ultrasonography, candidates with large breast cancers of more than 4 cm in their greatest dimension (29 cases) or of any size with skin invasion (four cases) were included in the study. Furthermore, they had normal renal, hepatic and cardiac function, no prior history of chemotherapy or hormone therapy and were without chest X-ray, abdominal ultrasound and bone scan evidence of distant metastasis. Patients with inflammatory cancer or with other systemic conditions were excluded from the study. Chemotherapeutic protocol The therapeutic schedule and treatment plan were explained to every patient. After obtaining informed consent, all patients enrolled in the study had a Port-A subclavian vein catheter inserted for chemotherapeutic drug administration. The preoperative chemotherapy regimen consisted of the three courses as described in Figure 1. A course of chemotherapy comprised weekly administration of paclitaxel and epirubicin for 3 weeks. After completing the third course, surgery was performed, and then followed by six cycles of postoperative chemotherapy using cyclophosphamide, fluorouracil and epirubicin, as well as radiotherapy. MRI protocol MRI examinations were performed with a 1.5 T MR scanner (Siemens, Magnetom Vision, Erlangen, Germany). The patient was placed in a prone position with their breasts placed in dedicated bilateral breast coils, 15 cm in diameter (Siemens, Magnetom Vision, Erlangen, Germany). A three-dimensional (3D) slab covering the entire breast was obtained for shimming. The study protocol consisted of an axial, T1-weighted, spin-echo pulse sequence (TR = 928 ms, TE = 14 ms, matrix size = 256 × 256, 32 slices of 4 mm thickness with no gaps, scan time = 3 min and 7 s); an axial, fat suppressed, T2-weighted pulse sequence (TR = 5930 ms, TE = 90 ms, matrix size = 256 × 256, 32 slices of 4 mm thickness with no gaps, scan time = 3 min and 39 s) and an axial, 3D, Turbo FLASH pulse sequence for dynamic imaging (TR = 8.1 ms, TE = 4.0 ms, flip angle = 20◦ , 32 partitions of 4 mm thickness). Both breasts were examined at the same time. The scan time for the dynamic study was 39 s per acquisition. The sequence was continuously repeated 14 times with four precontrast and 10 post-contrast images for dynamic acquisition. The dose of the contrast agent was based on patient body mass (0.1 mmol/kg Magnevist agent) with injection

Serial MRI monitoring of breast cancer size and response during chemotherapy

53

Figure 1. Epirubicin and taxol chemotherapeutic protocol for advanced breast cancers.

began after four precontrast acquisitions were completed. The total injection time was fixed at 15 s, so that in all subjects, the contrast agent was administered at a constant rate and finished in exactly 15 s. After finishing the dynamic study, fat suppressed, T1-weighted pulse sequences in the sagittal plane were immediately acquired of the whole breast. For easy comparison of the images acquired using the pulse sequences, the number of slices, slice thickness and locations for all three pulse sequences were kept the same. Three serial MRI examinations were performed for each patient. Each patient was examined before chemotherapy induction (day 0, baseline), within the week preceding the second course of chemotherapy (days 15–20), and within the week preceding surgery (days 58–63). All the examinations, including the pulse sequences and enhancement techniques, were kept the same. Response analysis The chief project investigator, a radiologist, performed all the breast MRI examinations in the hospital MRI department and used the subtracted images obtained using dynamic enhanced MRI at the ninth acquisition (scanning time between 157 and 195 s after contrast medium administration) to measure tumor size. The ninth acquisition was chosen for measurement of tumor size because of better visualization of the tumor margins, just before washout of the gadolinium that predominantly occurs after the third minute following injection. The radiologist was blinded to the patients’ pathological findings. As per the response evaluation criteria in solid tumors (RECIST) [16], tumor size was calculated by summing the maximum diameters

of the individual enhanced tumors on the slices of axial MRIs. The absence of obvious enhancement was taken to indicate an absence of residual cancer. The response documented in this study was based on the changes of tumor size demonstrated by serial dynamic enhanced MRI. The ESR was equal to the size difference between the first MRI (the baseline tumor size before chemotherapy) and the second MRI (the tumor size after one cycle of chemotherapy). The final response was defined as the change in size between the first (the baseline tumor size before chemotherapy) and the third MRI (the tumor size after three cycles of chemotherapy). Percentages of relative ESR (ESR/original size × 100) were then calculated and recorded. In accordance with the RECIST response categories, responses based on enhanced MRI examinations were classified as complete response (CR: no residual cancer), partial response (PR: ≥30% size reduction) and no response (NR: 50 y/o

26 7

78.79 21.21

21 12

63.63 36.37

10 23

30.3 69.7

4 19 10

12.12 57.58 30.3

Relative size reduction After first course C/T