- Email: [email protected]
Contrast enhanced magnetic resonance imaging underestimates residual disease following neoadjuvant docetaxel based chemotherapy for breast cancer
EJSO (2004) 30, 1069–1076
www.ejso.com
Contrast enhanced magnetic resonance imaging underestimates residual disease following neoadjuvant docetaxel based chemotherapy for breast cancer F. Denisa,*, A.V. Desbiez-Bourciera, C. Chapironb, F. Arbionc, G. Bodyd, L. Brunereaub a
´rapie, Tours, France Clinique d’Oncologie-Radiothe Groupement d’Imagerie INSERM U619, Tours, France c Laboratoire d’Anatomo-Pathologie, CHU Bretonneau, Tours, France d ´partement de Gyne ´cologie-Obste ´trique, CHU Bretonneau, Tours, France De b
Accepted for publication 29 July 2004 Available online 15 September 2004
KEYWORDS Preoperative chemotherapy; Docetaxel; Magnetic resonance imaging; Breast cancer
Abstract Aims. We prospectively compared the ability of magnetic resonance imaging (MRI) to measure residual breast cancer in patients treated with different neoadjuvant chemotherapy regimen. Methods. Forty patients with locally advanced breast carcinoma underwent neoadjuvant chemotherapy. Twelve patients received 5-fluoro-uracyl-epirubicincyclophosphamide (FEC-group, six cycles), 28 (DXL-group) received docetaxel-based chemotherapy (six cycles DXL-epirubicin: 13 patients, eight cycles DXL alone: 15 patients). All patients had baseline and preoperative MRI. The spread of pathologic residual disease (PRd) was compared to preoperative MRI measures according to chemotherapy regimen. Results. MRI over/underestimation of the spread of residual tumour was never superior to 15 mm in FEC group, whereas it appeared in 11/28 (39%, 30–48%—95% CI) patients in DXL group (pZ0.017). Tumour shrinkage led to single nodular residual lesions in FEC group, whereas vast numerous microscopic nests were observed in docetaxel group in pathology. Conclusion. Among tumours treated with a taxane-containing regimen, residual disease was frequently underestimated by MRI because of PRd features. q 2004 Elsevier Ltd. All rights reserved.
Introduction * Corresponding author. Address: CORAD, Centre Hospitalier Universitaire Bretonneau, 37044 Tours, France. Tel.: C33-11332-47-47-82-61; fax: C33-11-332-47-47-60-12. E-mail address: [email protected] (F. Denis).
Neoadjuvant chemotherapy is used to treat women with locally advanced breast carcinoma. It allows conservative treatment in half of patients, and may
0748-7983/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejso.2004.07.024
1070 be associated with better outcome for responders.1–3 New drugs such as taxanes are often used alone or in association in this situation and seem to enhance response rate and conservative surgical treatment.4 Recent studies suggest that docetaxel may have an anti-angiogenic activity which would explain, in part, its anti-tumoural effect by reducting the synthesis of pro-angiogenic molecules (such as bFGF and VEGF) by tumour cells.5 Contrast-enhanced magnetic resonance imaging (MRI) is used to measure tumour shrinkage following neoadjuvant chemotherapy. 6 Visualization of tumour by contrast-enhanced MRI is allowed by tumour vascularization. However, some studies reported that anti-angiogenic treatment may lead to vascular changes assessed by Doppler sonography7 or MRI.8,9 Contrast enhancement may be affected by both cytotoxic and potential anti-angiogenic effects of drugs such as docetaxel, leading to over or underestimate residual tumour size, and to mislead preoperative tumour response assessment. No comparison of the accuracy of MRI for measuring residual tumour according to chemotherapy regimen has been performed yet. This prospective study was undertaken to compare the ability of MRI to measure residual breast cancer between patients treated with or without docetaxel-based.
Patients and methods Patients and treatment Between September 2001 and March 2004, 40 patients with histologically confirmed locally advanced breast carcinoma were treated with intravenous neoadjuvant chemotherapy. Patients with inflammatory breast cancer or synchronous metastasis were excluded. No previous treatment had been performed before chemotherapy. Twelve patients (mean age 47 years old, range 32–58) received 5 fluoro uracyl (500 mg/m2)-epirubicin (Farmorubicinw 100 mg/m2, Pharmacia and Upjohn Co)-cyclophosphamide (FEC group) and 28 (mean age 49 years old, range 29–64) received docetaxelbased chemotherapy (Taxoterew, Aventis Pharmaceuticals (DXL group)). In DXL group, patients were randomly assigned to receive DXL alone (100 mg/m2) or an association of DXL (75 mg/m2) and epirubicin (75 mg/m 2 ). Thirteen patients received six cycles DXL-epirubicin and 15 received eight cycles DXL alone. Patients who refused to be randomized (four patients) or for which inclusion in
F. Denis et al. DXL regimen protocol was closed (eight patients) received standard chemotherapy FEC and were assigned to control arm. All cycles were repeated every 3 weeks. Informed consent was obtained from all patients before randomization and the study was approved by ethical committee. Surgery was performed in all patients after chemotherapy was completed. Conservative or radical surgical treatment decision was decided according to clinical and MRI findings by a committee consisted of one radiologist, two medical oncologists, two radiation oncologists, two surgeons and one pathologist.
MRI technique Each patient underwent contrast-enhanced MRI of the breast immediately before the first cycle of chemotherapy and 3 weeks after the last cycle. Surgery occurred within 10 days following MRI. All MRI examinations were performed on a commercially available 1.5-T echo-speed General Electric system using a dedicated breast coil. The imaging sequences were similar for all patients and for the pre- and post-chemotherapy examinations. After 3 plans localizer image was obtained, a sagitally oriented fat-suppressed fast spin-echo T2-weighted sequence was performed (TR/TE 6500;102; 256!320 matrix; 4.0 mm slice thickness; no gap). This sequence was followed by a 3D FSPGR (fast spoiled gradient-recalled) axial-T1-weighted fat-suppressed gradient-recalled echo sequence (224!256 matrix; 3.8 mm slice thickness; no gap) in steady state. This sequence was followed by IV contrast administration of gadopentetate meglumine (Dotarem 0.2 ml/kg of body weight) using a rapid injector through an indwelling IV catheter. The same sequence was performed after contrast administration and contrast-enhanced imaging was initiated immediately after the contrast material was completely injected. It consisted of three dynamic volume acquisitions through the breast. The first was performed 1 min after the initial injection, the second 1 min later and the last 3 min later. All contrast-enhanced imaging were performed with the same T1-weighted sequence optimized before contrast injection. Approximate total imaging time was 20 min per examination. The acquired images were transferred to a work-station (Advantage Window 4.1). Post-processing consisted in a subtraction of pre and post-injection acquisition. Image subtraction of basal acquisition from all series of contrast-enhanced slices was performed to improve visualization of areas of contrast
Breast MRI and neoadjuvant chemotherapy uptake and to evaluate its morphology and to calculate its size. In every subtracted slice, we thus defined the area of the tumour by a selection with electronic paintbrush. No evaluation of the enhancement kinetic was performed because of hardware limitation. Tumour size determination was performed two minutes after contrast administration which corresponded to the time point with maximum signal intensity. The best image depicting the abnormality was selected. Lesion location and extend were determined and measured using electronic caliper according to previously described interpretation criteria.10,11 The maximum dimension of the abnormally enhancing lesion or lesions corresponding to the known malignancy was measured before treatment and after neoadjuvant chemotherapy. The RECIST criteria were used to assess clinical and radiological response rate. Clinical and radiological objective responses were defined by the disappearance or at least a 30% decrease in the sum of the longest diameter of breast lesion. 12 All enhanced regions were measured to assess the response to chemotherapy according to RECIST criteria in patients having multifocal lesions. No dynamic analysis of specific region of interest was performed. All MRIs were prospectively and retrospectively evaluated by two radiologists with extensive breast MRI experience. The radiologists were not provided with pathological outcome or treatment regimen. Sonography (power-Doppler sonography or contrast enhanced sonography) and/or mammography were also performed to assess tumour shrinkage.
Histologic assessment of lumpectomy and mastectomy specimens Breast specimens were evaluated by pathologic examination. The pathologist was not provided with radiological results or chemotherapy regimen. All tumoural masses or nodules were measured, and representative tumour samples were embedded in paraffin. Any other suspect area detected was sampled. Lumpectomies without residual tumour on macroscopic examination were sampled entirely. Residual invasive carcinoma was confirmed histologically and measured. Intraductal carcinoma and vessel tumour emboli were noted. The margins were evaluated and all axillary lymph nodes were submitted for microscopic evaluation. Only infiltrative component was measured (not ductal carcinoma in situ).13 Pathological complete response was claimed when no residual invasive
1071 carcinoma were found in microscopic examination of lumpectomy or mastectomy specimens
Statistical analysis Histologic measurement of tumour size was used as the gold standard and was compared with tumour measurements from MRI after treatment. Parametric distribution of data allowed the use of Pearson’s correlation coefficients to determine the association between the MRI measurements and histologic size. Linear regression and Bland-Altman analysis were performed to further characterize the nature of the relationship between MRI and histology. Fisher exact test was performed to compare under/over estimation of tumour size between MRI and histology according to treatment arm. Results were considered statistically significant at pZ0.05.
Results Tumour response rates according to treatment arm Clinical objective response rate was 82% in DXL group and 58% in FEC group (pZ0.13). Radiological and histological response were also not statistically significant (pZ0.45 and pZ0.15, respectively) Table 1.
Tumour measurements There were one patient with multifocal lesions in FEC group and two in DXL. Tumour MRI measurements after neoadjuvant chemotherapy were compared with histologic size for each of the 40 patients (Table 2). The average tumour size predicted by MRI was 20.3 mm (SDZ17.0 mm, range 0–65 mm) compared with 22.8 cm (SDZ 19.5 mm, range 0–70 mm) at histology. There were no significant difference between tumour sizes assessed by MRI or pathology between FEC and DXL groups (p Mann–Whitney U-testZ0.68 and pZ0.38 for pathology and MRI, respectively). The Pearson’s correlation between MRI after chemotherapy and histology was 0.89, (95% confidence interval (CI)Z0.65–0.97) in FEC group and 0.5 (95% (CI)Z0.15–0.73) in DXL group. Linear regression analysis and Bland-Altman analysis showed that MRI over or underestimation was never superior to 15 mm in FEC group whereas it was in 39% of patients in DXL group (p Fisher exact testZ0.017) compared with the pathologic findings. It concerned both DXL alone (7/15) and
1072 Table 1
F. Denis et al. Clinical, radiological and pathological response rates according to treatment arms (CI 95%)
Clinical objective response rate (%) Radiological objective response rate (%) Complete pathological response rate (%)
FEC 100
DOCETAXEL
DOCETAXEL-EPI
p
58 (44–72)
87 (78–96)
77 (66–88)
NS
58 (44–72)
87 (78–96)
62 (49–75)
NS
25 (13–38)
13 (4–22)
0
NS
DXL-epirubicin (4/13) subgroups (Fig. 1). Underestimation of tumour residual disease by MRI tended to be greater than overestimation in DXL group (mean of the size of overestimated lesion: 13.8 mm, mean of the size of underestimated lesion: 20.2 mm, pZ0.37) (Fig. 2).
Pathologic characteristics of residual tumours There were three pathological complete responses in FEC group and two in DXL group. In DXL group, among 26 lesions which contained 5 mm or more residual disease 12 had multiple microscopic area of infiltrative carcinoma. This feature was observed in 2/9 tumours in FEC group. In DXL group, 5/6 lesions with at least 15 mm underestimation by MRI were also consisted of numerous tumour cell nests. Tumour shrinkage led to confined nodular residual lesions in 8/12 patients from FEC group for which size were similar to MRI measurements (Fig. 3).
Surgical treatment Fifteen patients were surgically treated with lumpectomy (13/28 in DXL group and 2/12 in FEC group, pZNS) and 25 underwent modified radical mastectomy. Radical mastectomy has been performed after lumpectomy in three patients in DXL group because of positive surgical margins subsequent to major underestimations of infiltrative Table 2 Correlation between modalities of tumour size measurements of residual lesions according to chemotherapy regimens
FEC DXL DXL-EPI
Type of evaluation MRI mm Pathology mm (standard (standard deviation) deviation)
r
18.6 (14.7) 15.1 (17.9) 27.7 (16.4)
0.89 0.64 0.16
18.5 (17.7) 17.9 (18.3) 32.5 (20.3)
FEC, control arm; DXL docetaxel-based chemotherapy; r, correlation coefficients.
carcinoma in residual tumour by MRI (superior to 2.5 cm compared to pathology). In this group, a fourth patient had positive surgical margins that required further radical mastectomy. Residual tumour size of this patient was 15 mm at MRI and 28 mm at histology. Overall, one third of patients of DXL group who were candidate for conservative surgery required further radical mastectomy. One patient in FEC group had to undergo modified radical mastectomy following lumpectomy because of positive surgical margin containing high grade in situ carcinoma.
Discussion MRI accuracy and neoadjuvant chemotherapy We prospectively compared accuracy of MRI to pathology to measure residual breast cancer after FEC or docetaxel preoperative chemotherapy in a homogeneous population of patient. We found a significant over/underestimation rate of tumour size by preoperative MRI and specific pathologic features were observed after docetaxel-based chemotherapy. Neoadjuvant chemotherapy is administered to patients especially to allow breast conservation surgery in locally advanced breast cancer.1 Although, the impact on survival is unknown, patients have to undergo 5–7 months chemotherapy before surgical option assessment. Accurate tools for assessing tumour size and response to chemotherapy are needed for optimizing the clinical management of those patients. Since MRI appears to be more effective than clinical evaluation to assess residual tumour following neoadjuvant chemotherapy,14 we need to appreciate its accuracy according to new treatment regimen. Rieber et al. also found numerous discordances between MRI and pathology and observed 33% falsenegative rate for preoperative MRI and attributed this result to the dramatic reduction or lack of contrast enhancement in the breast after
Breast MRI and neoadjuvant chemotherapy
1073
Figure 1 Relationship between MRI and pathology measurements of residual tumour size according to neoadjuvant chemotherapy. Graph shows relation ship between data points and best-fit line between MRI and histology. Solid lineZ identity, dashed lineZDXL group, dotted lineZFEC group.
treatment.15 In a more recent study, the same author concluded that MRI was unreliable for determining the size of residual tumour.16 Moreover, in his study, the number of patients receiving paclitaxel, a cytotoxic having similar anti-tumour proprieties than docetaxel and anti-angiogenic activity, is unknown and no comparison was performed between MRI and pathology according to treatments.
Reduction of MRI accuracy by docetaxel Docetaxel-based neoadjuvant chemotherapies are increasingly used to allow conservation of breast but interaction between drug and tumour vascularization is unknown. Moreover, tumour neo-vessels may in part contribute to contrast enhancement of breast carcinoma and tumour visualization may be altered by cytotoxic. We found in this prospective,
Figure 2 Relationship between MRI and pathology measurements of residual tumour size according to neoadjuvant chemotherapy. Bland-Altman technique. Graph shows difference versus mean for each measurement of residual disease obtained at MRI compared with pathologic findings according to treatment.
1074
Figure 3 False negative MRIs of 56 years old woman with T2 carcinoma of right breast treated with docetaxel alone. (A) Image acquired before neoadjuvant chemotherapy. (B) Image acquired after neoadjuvant chemotherapy: complete response. (C) Large area of residual disease on pathology.
comparative study that docetaxel-based regimen was associated with an enhancement of discordance between MRI-assessed and pathologic tumour size. Underestimations of tumour size by MRI were
F. Denis et al. more often observed than overestimations suggesting that direct or indirect anti-vascular effects of docetaxel may be involved. Hence, tumour pathologic characteristics following docetaxel-based chemotherapy were often associated with the persistence of numerous tumour cells nests that had no sufficient size and vascularization to allow MRI detection. This feature of residual tumours may explain in part the reduction of MRI accuracy after docetaxel-based chemotherapy. In addition, adjunction of epirubicin to docetaxel did not reduce discordance and residual tumours had similar pathologic aspects than in docetaxel alone arm, suggesting a specific role of docetaxel in tumour shrinkage. Numerous hypotheses may be plausible to explain underestimation of tumour by MRI after chemotherapy in our study. Tumour excision and paraffin fixation may alter tumour size and enhance the mismatch between tumour sizes determined by MRI and by pathological evaluation after docetaxel treatment. However, tumours in FEC-group were similarly excised and fixed as in DXL group and no such underestimation of residual disease was observed. Moreover, there might be misinterpretation of tumour tissue using MRI in both groups because we could not perform dynamic study of contrast enhancement. Hence, post-processing assessment allowed us to subjectively select images with maximal signal intensity which was mainly observed 2 min after contrast injection. However, dynamic contrast enhancement MRI (DC-MRI) of breast cancer seems to be more effective, robust and is a non invasive method to monitor the pharmacodynamic effects of cytotoxic and anti-angiogenic agents in patients.17,18 Administration of a MRI contrast agent such as Gadolinium-DTPA, allows the visualization of the vasculature into which that solute distributes, and the monitoring of the rate of influx of blood into tumour and other organs, as well as the retention and the rate of efflux.19 Dynamic contrast enhancement MRI may be a promising tool to depict functional parameter like perfusion and permeability of the tumoural capillary network and may have provided high-resolution images that may facilitates image selection reproducibility of tumour size assessment.20 Nevertheless, in our study, the interpretations of both radiologists were concordant although they were not provided with pathological outcome or treatment regimen. Another confounding factor has also been reported: tumour response may be accompanied by an inflammatory reaction which may make the tumour appear bigger than it is.21 However, no specific enhancement of inflammatory reaction
Breast MRI and neoadjuvant chemotherapy subsequent to docetaxel has been reported and no patient with inflammatory breast cancer was included in our study. A direct or indirect anti-vascular effect of docetaxel may also be involved in the lack of contrast enhancement. We previously found in a rat mammary tumour system that anticancer treatment may induce tumour vascular changes.7,22,23 In animal, epirubicin-sensitive tumours showed marked reduction of vascularization and epirubicin-resistant had stable or enhanced vascularization suggesting that the unspecific inhibition of mitotic tissue may be involved in the reduce release of pro-angiogenic factors by the tumour cells due to their degradation.
Pathologic features in FEC and docetaxel treated tumours Moreover we did not observe similar pathological characteristics in FEC and in DXL groups, it is also plausible than residual tumour cell nests were not detectable using MRI in docetaxel-treated tumours. We reported in animal study TNP-470, an antiangiogenic drug, to have a major anti-vascular properties leading to a reduction of vascular powerDoppler index superior to those observed in epirubicin-sensitive tumours.7 In Human, antiangiogenic drugs also seemed to have anti-vascular effect as assessed by dynamic contrast enhancement MRI.8 Anti-tumour effects of anti-angiogenic treatment seems to be associated with cellular scarcity and necrosis,9 and pathologic aspects of residual tumours in our study (in docetaxel group) may be explain, in part, by a specific anti-vascular effect of docetaxel as suggested by the reduction of synthesis of pro-angiogenic molecules by tumour cells.5 However, we did not perform pathological microvessel density assessment because microvessel density may not be an indicator of antiangiogenic and anti-vascular treatment efficacies. Thus, if vascular properties of docetaxel may be involved in tumour shrinkage, then microvessel density which is neither a measure of tumour angiogenic activity nor a measure of tumour angiogenic dependence may not be useful to assess this modality of action. Moreover, contrary to MRI, measures of microvessel density are not sufficient to reveal the functional status of tumour neovasculature and thus may not allow evaluation tumour perfusion changes subsequent to treatment.24
Limitations of the study The main limitation of this study concerned the lack
1075 of randomization between docetaxel-based chemotherapy and FEC groups. Patients who received docetaxel were already included (and randomized) in a prospective clinical trial comparing docetaxel to docetaxel and epirubicin. Hence, patients in FEC group who were treated during the same period and who had the same inclusion criteria as in docetaxel clinical trial were matched in a case controlled fashion. Our study only concerned patients from a larger multicenter clinical trial who where treated in our center. So, no relevant demonstration in superiority of any regimen can be performed until pulled data of all involved centers are known and overall findings of pathological response may probably change after study completion. Mammography and/or sonography were also performed to assess tumour changes subsequent to chemotherapy. However, the differences in sonography examination (such as the use of contrast injection, power-Doppler, different hardware) and the important number of operators did not allow any relevant comparisons between imaging modalities. Moreover, although tumour measurement have always been performed using MRI, mammography was also used in some patients, sonography in some others (almost the youngest) or both. This lack of homogeneity in imaging modality assessment reduced the power of any statistical comparison between MRI and others imaging modalities. Further research is needed using (dynamic) contrast enhancement MRI to confirm the underestimation enhancement of residual tumour size by MRI after docetaxel-based chemotherapy, but caution should be used in basing surgical treatment options on the results of MRI in this population.
References 1. Bonadonna G, Veronesi U, Brambilla C, et al. Primary chemotherapy to avoid mastectomy in tumors with diamaters of three centimetres or more. J Natl Cancer Inst 1990; 82:1539–45. 2. Wolmark N, Wang J, Mamounas E, Bryant J, Fisher B. Preoperative chemotherapy in patients with operable breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst 2001;30: 96–102. 3. van der Hage JA, van der Velde CJ, Julien JP, TubianaHulin M, Vandervelden C, Duchateau L. Preoperative chemotherapy in primary operable breast cancer: results from European Organization for Research and Treatment of Cancer trial 10902. J Clin Oncol 2001;19:4224–37. 4. Hutcheon AW, Heys SD, Sarkar TK. Aberdeen Breast Group. Neoadjuvant docetaxel in locally advanced breast cancer. Breast Cancer Res Treat 2003;79(Suppl. 1):S19–S24.
1076 5. Guo XL, Lin GJ, Zhao H, et al. Inhibitory effects of docetaxel on expression of VEGF, bFGF and MMPs of LS174T cell. World J Gastroenterol 2003;9:1995–8. 6. Rosen EL, Blackwell KL, Baker JA, et al. Accuracy of MRI in the detection of residual breast cancer after neo-adjuvant chemotherapy. Am J Roentgenol 2003;181:1275–82. 7. Denis F, Colas S, Chami L, et al. Changes in tumor vascularization following irradiation, anthracyclin or antiangiogenic treatment in autochthonous rat mammary tumors. Clinical Cancer Res 2003;9:4546–52. 8. Morgan B, Thomas AL, et al, Drevs J. Dynamic contrastenhanced magnetic resonance imaging as a biomarker for the pharmacological response of PTK787, an inhibitor of the VEGF receptor tyrosine kinases, in patients with colorectal cancer and liver metastases: results from two phase 1 studies. J Clin Oncol 2003;21:3955–64. 9. Bhujwalla ZM, Artemov D, Natarajan K, Solaiyappan M, Kollars P, Kristjansen PE. Reduction of vascular and permeable regions in solid tumors detected by macromolecular contrast magnetic resonance imaging after treatment with anti-angiogenic agent TNP-470. Clin Cancer Res 2003;9: 355–62. 10. Harms SE, Schnall MD, Ikeda DM, Orel SG, Hylton N, Mumtaz H. Technical reports of the international working group on breast MRI. J Magn Reson Imaging 1999;10: 980–1015. 11. Ikeda DM, Hylton NM, Kinkel K, et al. Development standardization, and testing of a lexicon for reporting contrast-enhancement breast magnetic resonance imaging studies. J Magn Reson Imaging 2001;13:889–95. 12. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 2000;92:205–16. 13. Matsuo K, Fukutomi T, Watanabe T, Hasegawa T, Tsuda H, Akashi-Tanaka S. Concordance in pathological response to neoadjuvant chemotherapy between invasive and noninvasive components of primary breast carcinoma. Breast Cancer 2002;9:75–81. 14. Partridge SC, Gibbs JE, Lu Y, Esserman LJ, Sudilovsky D, Hylton NM. Accuracy of magnetic resonance imaging for
F. Denis et al.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
revealing residual breast cancer in patients who have undergone neoadjuvant chemotherapy. Am J Roentgenol 2002;179:1193–9. Rieber A, Zeitler H, Rosenthal H, et al. Magnetic resonance imaging of breast cancer: influence of chemotherapy on sensitivity. Br J Radiol 1997;70:452–8. Rieber A, Brambs HJ, Gabelmann A, Heilmann V, Kreienberg R, Kuhn T. Breast magnetic resonance imaging for monitoring response of primary breast cancer to neoadjuvant chemotherapy. Eur Radiol 2002;12:1711–9. Port RE, Knopp MV, Hoffmann U, Milker-Zabel S, Brix G. Multicompartment analysis of gadolinium chelate kinetics: blood-tissue exchange in mammary tumors as monitored by dynamic MR imaging. J Magn Reson Imaging 1999;10:233–41. Daldrup-Link HE, Brasch RC. Macromolecular contrast agents for MR mammography: current status. Eur Radiol 2003;13: 354–65. Brix G, Bellemann ME, Gerlach L, Haberkorn U. Intra- and extracellular fluorouracil uptake: assessment with contrastenhanced metabolic F-19 MR imaging. Radiology 1998;209: 259–67. Hawighorst H, Weikel W, Knapstein PG, et al. Angiogenic activity of cervical carcinoma: assessment by functional magnetic resonance imaging-based parameters and a histomorphological approach in correlation with disease outcome. Clin Cancer Res 1998;4:2305–12. Wasser K, Sinn HP, Fink C, et al. Accuracy of tumor size measurement in breast cancer using MRI is influenced by histological regression induced by neoadjuvant chemotherapy. Eur Radiol 2003;13:1213–23. Denis F, Bougnoux P, de Poncheville L, Prat M, Catroux R, Tranquart F. In vivo quantitation of tumour vascularization assessed by doppler sonography in rat mammary tumours. Ultrasound in Med Biol 2002;28:431–7. Denis F, Bougnoux P, Paon L, Aget H, le Floch O, Tranquart F. Radiosensitivity of rat mammary tumors correlates with early vessel changes assessed by power-doppler sonography. J Ultras Med 2003;22:921–9. Hlatky L, Hahnfeldt P, Folkman J. Clinical application of anti-angiogenic therapy: microvessel density, what it does and doesn’t tell us. J Natl Cancer Inst 2002;94:883–93.
www.ejso.com
Contrast enhanced magnetic resonance imaging underestimates residual disease following neoadjuvant docetaxel based chemotherapy for breast cancer F. Denisa,*, A.V. Desbiez-Bourciera, C. Chapironb, F. Arbionc, G. Bodyd, L. Brunereaub a
´rapie, Tours, France Clinique d’Oncologie-Radiothe Groupement d’Imagerie INSERM U619, Tours, France c Laboratoire d’Anatomo-Pathologie, CHU Bretonneau, Tours, France d ´partement de Gyne ´cologie-Obste ´trique, CHU Bretonneau, Tours, France De b
Accepted for publication 29 July 2004 Available online 15 September 2004
KEYWORDS Preoperative chemotherapy; Docetaxel; Magnetic resonance imaging; Breast cancer
Abstract Aims. We prospectively compared the ability of magnetic resonance imaging (MRI) to measure residual breast cancer in patients treated with different neoadjuvant chemotherapy regimen. Methods. Forty patients with locally advanced breast carcinoma underwent neoadjuvant chemotherapy. Twelve patients received 5-fluoro-uracyl-epirubicincyclophosphamide (FEC-group, six cycles), 28 (DXL-group) received docetaxel-based chemotherapy (six cycles DXL-epirubicin: 13 patients, eight cycles DXL alone: 15 patients). All patients had baseline and preoperative MRI. The spread of pathologic residual disease (PRd) was compared to preoperative MRI measures according to chemotherapy regimen. Results. MRI over/underestimation of the spread of residual tumour was never superior to 15 mm in FEC group, whereas it appeared in 11/28 (39%, 30–48%—95% CI) patients in DXL group (pZ0.017). Tumour shrinkage led to single nodular residual lesions in FEC group, whereas vast numerous microscopic nests were observed in docetaxel group in pathology. Conclusion. Among tumours treated with a taxane-containing regimen, residual disease was frequently underestimated by MRI because of PRd features. q 2004 Elsevier Ltd. All rights reserved.
Introduction * Corresponding author. Address: CORAD, Centre Hospitalier Universitaire Bretonneau, 37044 Tours, France. Tel.: C33-11332-47-47-82-61; fax: C33-11-332-47-47-60-12. E-mail address: [email protected] (F. Denis).
Neoadjuvant chemotherapy is used to treat women with locally advanced breast carcinoma. It allows conservative treatment in half of patients, and may
0748-7983/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejso.2004.07.024
1070 be associated with better outcome for responders.1–3 New drugs such as taxanes are often used alone or in association in this situation and seem to enhance response rate and conservative surgical treatment.4 Recent studies suggest that docetaxel may have an anti-angiogenic activity which would explain, in part, its anti-tumoural effect by reducting the synthesis of pro-angiogenic molecules (such as bFGF and VEGF) by tumour cells.5 Contrast-enhanced magnetic resonance imaging (MRI) is used to measure tumour shrinkage following neoadjuvant chemotherapy. 6 Visualization of tumour by contrast-enhanced MRI is allowed by tumour vascularization. However, some studies reported that anti-angiogenic treatment may lead to vascular changes assessed by Doppler sonography7 or MRI.8,9 Contrast enhancement may be affected by both cytotoxic and potential anti-angiogenic effects of drugs such as docetaxel, leading to over or underestimate residual tumour size, and to mislead preoperative tumour response assessment. No comparison of the accuracy of MRI for measuring residual tumour according to chemotherapy regimen has been performed yet. This prospective study was undertaken to compare the ability of MRI to measure residual breast cancer between patients treated with or without docetaxel-based.
Patients and methods Patients and treatment Between September 2001 and March 2004, 40 patients with histologically confirmed locally advanced breast carcinoma were treated with intravenous neoadjuvant chemotherapy. Patients with inflammatory breast cancer or synchronous metastasis were excluded. No previous treatment had been performed before chemotherapy. Twelve patients (mean age 47 years old, range 32–58) received 5 fluoro uracyl (500 mg/m2)-epirubicin (Farmorubicinw 100 mg/m2, Pharmacia and Upjohn Co)-cyclophosphamide (FEC group) and 28 (mean age 49 years old, range 29–64) received docetaxelbased chemotherapy (Taxoterew, Aventis Pharmaceuticals (DXL group)). In DXL group, patients were randomly assigned to receive DXL alone (100 mg/m2) or an association of DXL (75 mg/m2) and epirubicin (75 mg/m 2 ). Thirteen patients received six cycles DXL-epirubicin and 15 received eight cycles DXL alone. Patients who refused to be randomized (four patients) or for which inclusion in
F. Denis et al. DXL regimen protocol was closed (eight patients) received standard chemotherapy FEC and were assigned to control arm. All cycles were repeated every 3 weeks. Informed consent was obtained from all patients before randomization and the study was approved by ethical committee. Surgery was performed in all patients after chemotherapy was completed. Conservative or radical surgical treatment decision was decided according to clinical and MRI findings by a committee consisted of one radiologist, two medical oncologists, two radiation oncologists, two surgeons and one pathologist.
MRI technique Each patient underwent contrast-enhanced MRI of the breast immediately before the first cycle of chemotherapy and 3 weeks after the last cycle. Surgery occurred within 10 days following MRI. All MRI examinations were performed on a commercially available 1.5-T echo-speed General Electric system using a dedicated breast coil. The imaging sequences were similar for all patients and for the pre- and post-chemotherapy examinations. After 3 plans localizer image was obtained, a sagitally oriented fat-suppressed fast spin-echo T2-weighted sequence was performed (TR/TE 6500;102; 256!320 matrix; 4.0 mm slice thickness; no gap). This sequence was followed by a 3D FSPGR (fast spoiled gradient-recalled) axial-T1-weighted fat-suppressed gradient-recalled echo sequence (224!256 matrix; 3.8 mm slice thickness; no gap) in steady state. This sequence was followed by IV contrast administration of gadopentetate meglumine (Dotarem 0.2 ml/kg of body weight) using a rapid injector through an indwelling IV catheter. The same sequence was performed after contrast administration and contrast-enhanced imaging was initiated immediately after the contrast material was completely injected. It consisted of three dynamic volume acquisitions through the breast. The first was performed 1 min after the initial injection, the second 1 min later and the last 3 min later. All contrast-enhanced imaging were performed with the same T1-weighted sequence optimized before contrast injection. Approximate total imaging time was 20 min per examination. The acquired images were transferred to a work-station (Advantage Window 4.1). Post-processing consisted in a subtraction of pre and post-injection acquisition. Image subtraction of basal acquisition from all series of contrast-enhanced slices was performed to improve visualization of areas of contrast
Breast MRI and neoadjuvant chemotherapy uptake and to evaluate its morphology and to calculate its size. In every subtracted slice, we thus defined the area of the tumour by a selection with electronic paintbrush. No evaluation of the enhancement kinetic was performed because of hardware limitation. Tumour size determination was performed two minutes after contrast administration which corresponded to the time point with maximum signal intensity. The best image depicting the abnormality was selected. Lesion location and extend were determined and measured using electronic caliper according to previously described interpretation criteria.10,11 The maximum dimension of the abnormally enhancing lesion or lesions corresponding to the known malignancy was measured before treatment and after neoadjuvant chemotherapy. The RECIST criteria were used to assess clinical and radiological response rate. Clinical and radiological objective responses were defined by the disappearance or at least a 30% decrease in the sum of the longest diameter of breast lesion. 12 All enhanced regions were measured to assess the response to chemotherapy according to RECIST criteria in patients having multifocal lesions. No dynamic analysis of specific region of interest was performed. All MRIs were prospectively and retrospectively evaluated by two radiologists with extensive breast MRI experience. The radiologists were not provided with pathological outcome or treatment regimen. Sonography (power-Doppler sonography or contrast enhanced sonography) and/or mammography were also performed to assess tumour shrinkage.
Histologic assessment of lumpectomy and mastectomy specimens Breast specimens were evaluated by pathologic examination. The pathologist was not provided with radiological results or chemotherapy regimen. All tumoural masses or nodules were measured, and representative tumour samples were embedded in paraffin. Any other suspect area detected was sampled. Lumpectomies without residual tumour on macroscopic examination were sampled entirely. Residual invasive carcinoma was confirmed histologically and measured. Intraductal carcinoma and vessel tumour emboli were noted. The margins were evaluated and all axillary lymph nodes were submitted for microscopic evaluation. Only infiltrative component was measured (not ductal carcinoma in situ).13 Pathological complete response was claimed when no residual invasive
1071 carcinoma were found in microscopic examination of lumpectomy or mastectomy specimens
Statistical analysis Histologic measurement of tumour size was used as the gold standard and was compared with tumour measurements from MRI after treatment. Parametric distribution of data allowed the use of Pearson’s correlation coefficients to determine the association between the MRI measurements and histologic size. Linear regression and Bland-Altman analysis were performed to further characterize the nature of the relationship between MRI and histology. Fisher exact test was performed to compare under/over estimation of tumour size between MRI and histology according to treatment arm. Results were considered statistically significant at pZ0.05.
Results Tumour response rates according to treatment arm Clinical objective response rate was 82% in DXL group and 58% in FEC group (pZ0.13). Radiological and histological response were also not statistically significant (pZ0.45 and pZ0.15, respectively) Table 1.
Tumour measurements There were one patient with multifocal lesions in FEC group and two in DXL. Tumour MRI measurements after neoadjuvant chemotherapy were compared with histologic size for each of the 40 patients (Table 2). The average tumour size predicted by MRI was 20.3 mm (SDZ17.0 mm, range 0–65 mm) compared with 22.8 cm (SDZ 19.5 mm, range 0–70 mm) at histology. There were no significant difference between tumour sizes assessed by MRI or pathology between FEC and DXL groups (p Mann–Whitney U-testZ0.68 and pZ0.38 for pathology and MRI, respectively). The Pearson’s correlation between MRI after chemotherapy and histology was 0.89, (95% confidence interval (CI)Z0.65–0.97) in FEC group and 0.5 (95% (CI)Z0.15–0.73) in DXL group. Linear regression analysis and Bland-Altman analysis showed that MRI over or underestimation was never superior to 15 mm in FEC group whereas it was in 39% of patients in DXL group (p Fisher exact testZ0.017) compared with the pathologic findings. It concerned both DXL alone (7/15) and
1072 Table 1
F. Denis et al. Clinical, radiological and pathological response rates according to treatment arms (CI 95%)
Clinical objective response rate (%) Radiological objective response rate (%) Complete pathological response rate (%)
FEC 100
DOCETAXEL
DOCETAXEL-EPI
p
58 (44–72)
87 (78–96)
77 (66–88)
NS
58 (44–72)
87 (78–96)
62 (49–75)
NS
25 (13–38)
13 (4–22)
0
NS
DXL-epirubicin (4/13) subgroups (Fig. 1). Underestimation of tumour residual disease by MRI tended to be greater than overestimation in DXL group (mean of the size of overestimated lesion: 13.8 mm, mean of the size of underestimated lesion: 20.2 mm, pZ0.37) (Fig. 2).
Pathologic characteristics of residual tumours There were three pathological complete responses in FEC group and two in DXL group. In DXL group, among 26 lesions which contained 5 mm or more residual disease 12 had multiple microscopic area of infiltrative carcinoma. This feature was observed in 2/9 tumours in FEC group. In DXL group, 5/6 lesions with at least 15 mm underestimation by MRI were also consisted of numerous tumour cell nests. Tumour shrinkage led to confined nodular residual lesions in 8/12 patients from FEC group for which size were similar to MRI measurements (Fig. 3).
Surgical treatment Fifteen patients were surgically treated with lumpectomy (13/28 in DXL group and 2/12 in FEC group, pZNS) and 25 underwent modified radical mastectomy. Radical mastectomy has been performed after lumpectomy in three patients in DXL group because of positive surgical margins subsequent to major underestimations of infiltrative Table 2 Correlation between modalities of tumour size measurements of residual lesions according to chemotherapy regimens
FEC DXL DXL-EPI
Type of evaluation MRI mm Pathology mm (standard (standard deviation) deviation)
r
18.6 (14.7) 15.1 (17.9) 27.7 (16.4)
0.89 0.64 0.16
18.5 (17.7) 17.9 (18.3) 32.5 (20.3)
FEC, control arm; DXL docetaxel-based chemotherapy; r, correlation coefficients.
carcinoma in residual tumour by MRI (superior to 2.5 cm compared to pathology). In this group, a fourth patient had positive surgical margins that required further radical mastectomy. Residual tumour size of this patient was 15 mm at MRI and 28 mm at histology. Overall, one third of patients of DXL group who were candidate for conservative surgery required further radical mastectomy. One patient in FEC group had to undergo modified radical mastectomy following lumpectomy because of positive surgical margin containing high grade in situ carcinoma.
Discussion MRI accuracy and neoadjuvant chemotherapy We prospectively compared accuracy of MRI to pathology to measure residual breast cancer after FEC or docetaxel preoperative chemotherapy in a homogeneous population of patient. We found a significant over/underestimation rate of tumour size by preoperative MRI and specific pathologic features were observed after docetaxel-based chemotherapy. Neoadjuvant chemotherapy is administered to patients especially to allow breast conservation surgery in locally advanced breast cancer.1 Although, the impact on survival is unknown, patients have to undergo 5–7 months chemotherapy before surgical option assessment. Accurate tools for assessing tumour size and response to chemotherapy are needed for optimizing the clinical management of those patients. Since MRI appears to be more effective than clinical evaluation to assess residual tumour following neoadjuvant chemotherapy,14 we need to appreciate its accuracy according to new treatment regimen. Rieber et al. also found numerous discordances between MRI and pathology and observed 33% falsenegative rate for preoperative MRI and attributed this result to the dramatic reduction or lack of contrast enhancement in the breast after
Breast MRI and neoadjuvant chemotherapy
1073
Figure 1 Relationship between MRI and pathology measurements of residual tumour size according to neoadjuvant chemotherapy. Graph shows relation ship between data points and best-fit line between MRI and histology. Solid lineZ identity, dashed lineZDXL group, dotted lineZFEC group.
treatment.15 In a more recent study, the same author concluded that MRI was unreliable for determining the size of residual tumour.16 Moreover, in his study, the number of patients receiving paclitaxel, a cytotoxic having similar anti-tumour proprieties than docetaxel and anti-angiogenic activity, is unknown and no comparison was performed between MRI and pathology according to treatments.
Reduction of MRI accuracy by docetaxel Docetaxel-based neoadjuvant chemotherapies are increasingly used to allow conservation of breast but interaction between drug and tumour vascularization is unknown. Moreover, tumour neo-vessels may in part contribute to contrast enhancement of breast carcinoma and tumour visualization may be altered by cytotoxic. We found in this prospective,
Figure 2 Relationship between MRI and pathology measurements of residual tumour size according to neoadjuvant chemotherapy. Bland-Altman technique. Graph shows difference versus mean for each measurement of residual disease obtained at MRI compared with pathologic findings according to treatment.
1074
Figure 3 False negative MRIs of 56 years old woman with T2 carcinoma of right breast treated with docetaxel alone. (A) Image acquired before neoadjuvant chemotherapy. (B) Image acquired after neoadjuvant chemotherapy: complete response. (C) Large area of residual disease on pathology.
comparative study that docetaxel-based regimen was associated with an enhancement of discordance between MRI-assessed and pathologic tumour size. Underestimations of tumour size by MRI were
F. Denis et al. more often observed than overestimations suggesting that direct or indirect anti-vascular effects of docetaxel may be involved. Hence, tumour pathologic characteristics following docetaxel-based chemotherapy were often associated with the persistence of numerous tumour cells nests that had no sufficient size and vascularization to allow MRI detection. This feature of residual tumours may explain in part the reduction of MRI accuracy after docetaxel-based chemotherapy. In addition, adjunction of epirubicin to docetaxel did not reduce discordance and residual tumours had similar pathologic aspects than in docetaxel alone arm, suggesting a specific role of docetaxel in tumour shrinkage. Numerous hypotheses may be plausible to explain underestimation of tumour by MRI after chemotherapy in our study. Tumour excision and paraffin fixation may alter tumour size and enhance the mismatch between tumour sizes determined by MRI and by pathological evaluation after docetaxel treatment. However, tumours in FEC-group were similarly excised and fixed as in DXL group and no such underestimation of residual disease was observed. Moreover, there might be misinterpretation of tumour tissue using MRI in both groups because we could not perform dynamic study of contrast enhancement. Hence, post-processing assessment allowed us to subjectively select images with maximal signal intensity which was mainly observed 2 min after contrast injection. However, dynamic contrast enhancement MRI (DC-MRI) of breast cancer seems to be more effective, robust and is a non invasive method to monitor the pharmacodynamic effects of cytotoxic and anti-angiogenic agents in patients.17,18 Administration of a MRI contrast agent such as Gadolinium-DTPA, allows the visualization of the vasculature into which that solute distributes, and the monitoring of the rate of influx of blood into tumour and other organs, as well as the retention and the rate of efflux.19 Dynamic contrast enhancement MRI may be a promising tool to depict functional parameter like perfusion and permeability of the tumoural capillary network and may have provided high-resolution images that may facilitates image selection reproducibility of tumour size assessment.20 Nevertheless, in our study, the interpretations of both radiologists were concordant although they were not provided with pathological outcome or treatment regimen. Another confounding factor has also been reported: tumour response may be accompanied by an inflammatory reaction which may make the tumour appear bigger than it is.21 However, no specific enhancement of inflammatory reaction
Breast MRI and neoadjuvant chemotherapy subsequent to docetaxel has been reported and no patient with inflammatory breast cancer was included in our study. A direct or indirect anti-vascular effect of docetaxel may also be involved in the lack of contrast enhancement. We previously found in a rat mammary tumour system that anticancer treatment may induce tumour vascular changes.7,22,23 In animal, epirubicin-sensitive tumours showed marked reduction of vascularization and epirubicin-resistant had stable or enhanced vascularization suggesting that the unspecific inhibition of mitotic tissue may be involved in the reduce release of pro-angiogenic factors by the tumour cells due to their degradation.
Pathologic features in FEC and docetaxel treated tumours Moreover we did not observe similar pathological characteristics in FEC and in DXL groups, it is also plausible than residual tumour cell nests were not detectable using MRI in docetaxel-treated tumours. We reported in animal study TNP-470, an antiangiogenic drug, to have a major anti-vascular properties leading to a reduction of vascular powerDoppler index superior to those observed in epirubicin-sensitive tumours.7 In Human, antiangiogenic drugs also seemed to have anti-vascular effect as assessed by dynamic contrast enhancement MRI.8 Anti-tumour effects of anti-angiogenic treatment seems to be associated with cellular scarcity and necrosis,9 and pathologic aspects of residual tumours in our study (in docetaxel group) may be explain, in part, by a specific anti-vascular effect of docetaxel as suggested by the reduction of synthesis of pro-angiogenic molecules by tumour cells.5 However, we did not perform pathological microvessel density assessment because microvessel density may not be an indicator of antiangiogenic and anti-vascular treatment efficacies. Thus, if vascular properties of docetaxel may be involved in tumour shrinkage, then microvessel density which is neither a measure of tumour angiogenic activity nor a measure of tumour angiogenic dependence may not be useful to assess this modality of action. Moreover, contrary to MRI, measures of microvessel density are not sufficient to reveal the functional status of tumour neovasculature and thus may not allow evaluation tumour perfusion changes subsequent to treatment.24
Limitations of the study The main limitation of this study concerned the lack
1075 of randomization between docetaxel-based chemotherapy and FEC groups. Patients who received docetaxel were already included (and randomized) in a prospective clinical trial comparing docetaxel to docetaxel and epirubicin. Hence, patients in FEC group who were treated during the same period and who had the same inclusion criteria as in docetaxel clinical trial were matched in a case controlled fashion. Our study only concerned patients from a larger multicenter clinical trial who where treated in our center. So, no relevant demonstration in superiority of any regimen can be performed until pulled data of all involved centers are known and overall findings of pathological response may probably change after study completion. Mammography and/or sonography were also performed to assess tumour changes subsequent to chemotherapy. However, the differences in sonography examination (such as the use of contrast injection, power-Doppler, different hardware) and the important number of operators did not allow any relevant comparisons between imaging modalities. Moreover, although tumour measurement have always been performed using MRI, mammography was also used in some patients, sonography in some others (almost the youngest) or both. This lack of homogeneity in imaging modality assessment reduced the power of any statistical comparison between MRI and others imaging modalities. Further research is needed using (dynamic) contrast enhancement MRI to confirm the underestimation enhancement of residual tumour size by MRI after docetaxel-based chemotherapy, but caution should be used in basing surgical treatment options on the results of MRI in this population.
References 1. Bonadonna G, Veronesi U, Brambilla C, et al. Primary chemotherapy to avoid mastectomy in tumors with diamaters of three centimetres or more. J Natl Cancer Inst 1990; 82:1539–45. 2. Wolmark N, Wang J, Mamounas E, Bryant J, Fisher B. Preoperative chemotherapy in patients with operable breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst 2001;30: 96–102. 3. van der Hage JA, van der Velde CJ, Julien JP, TubianaHulin M, Vandervelden C, Duchateau L. Preoperative chemotherapy in primary operable breast cancer: results from European Organization for Research and Treatment of Cancer trial 10902. J Clin Oncol 2001;19:4224–37. 4. Hutcheon AW, Heys SD, Sarkar TK. Aberdeen Breast Group. Neoadjuvant docetaxel in locally advanced breast cancer. Breast Cancer Res Treat 2003;79(Suppl. 1):S19–S24.
1076 5. Guo XL, Lin GJ, Zhao H, et al. Inhibitory effects of docetaxel on expression of VEGF, bFGF and MMPs of LS174T cell. World J Gastroenterol 2003;9:1995–8. 6. Rosen EL, Blackwell KL, Baker JA, et al. Accuracy of MRI in the detection of residual breast cancer after neo-adjuvant chemotherapy. Am J Roentgenol 2003;181:1275–82. 7. Denis F, Colas S, Chami L, et al. Changes in tumor vascularization following irradiation, anthracyclin or antiangiogenic treatment in autochthonous rat mammary tumors. Clinical Cancer Res 2003;9:4546–52. 8. Morgan B, Thomas AL, et al, Drevs J. Dynamic contrastenhanced magnetic resonance imaging as a biomarker for the pharmacological response of PTK787, an inhibitor of the VEGF receptor tyrosine kinases, in patients with colorectal cancer and liver metastases: results from two phase 1 studies. J Clin Oncol 2003;21:3955–64. 9. Bhujwalla ZM, Artemov D, Natarajan K, Solaiyappan M, Kollars P, Kristjansen PE. Reduction of vascular and permeable regions in solid tumors detected by macromolecular contrast magnetic resonance imaging after treatment with anti-angiogenic agent TNP-470. Clin Cancer Res 2003;9: 355–62. 10. Harms SE, Schnall MD, Ikeda DM, Orel SG, Hylton N, Mumtaz H. Technical reports of the international working group on breast MRI. J Magn Reson Imaging 1999;10: 980–1015. 11. Ikeda DM, Hylton NM, Kinkel K, et al. Development standardization, and testing of a lexicon for reporting contrast-enhancement breast magnetic resonance imaging studies. J Magn Reson Imaging 2001;13:889–95. 12. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 2000;92:205–16. 13. Matsuo K, Fukutomi T, Watanabe T, Hasegawa T, Tsuda H, Akashi-Tanaka S. Concordance in pathological response to neoadjuvant chemotherapy between invasive and noninvasive components of primary breast carcinoma. Breast Cancer 2002;9:75–81. 14. Partridge SC, Gibbs JE, Lu Y, Esserman LJ, Sudilovsky D, Hylton NM. Accuracy of magnetic resonance imaging for
F. Denis et al.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
revealing residual breast cancer in patients who have undergone neoadjuvant chemotherapy. Am J Roentgenol 2002;179:1193–9. Rieber A, Zeitler H, Rosenthal H, et al. Magnetic resonance imaging of breast cancer: influence of chemotherapy on sensitivity. Br J Radiol 1997;70:452–8. Rieber A, Brambs HJ, Gabelmann A, Heilmann V, Kreienberg R, Kuhn T. Breast magnetic resonance imaging for monitoring response of primary breast cancer to neoadjuvant chemotherapy. Eur Radiol 2002;12:1711–9. Port RE, Knopp MV, Hoffmann U, Milker-Zabel S, Brix G. Multicompartment analysis of gadolinium chelate kinetics: blood-tissue exchange in mammary tumors as monitored by dynamic MR imaging. J Magn Reson Imaging 1999;10:233–41. Daldrup-Link HE, Brasch RC. Macromolecular contrast agents for MR mammography: current status. Eur Radiol 2003;13: 354–65. Brix G, Bellemann ME, Gerlach L, Haberkorn U. Intra- and extracellular fluorouracil uptake: assessment with contrastenhanced metabolic F-19 MR imaging. Radiology 1998;209: 259–67. Hawighorst H, Weikel W, Knapstein PG, et al. Angiogenic activity of cervical carcinoma: assessment by functional magnetic resonance imaging-based parameters and a histomorphological approach in correlation with disease outcome. Clin Cancer Res 1998;4:2305–12. Wasser K, Sinn HP, Fink C, et al. Accuracy of tumor size measurement in breast cancer using MRI is influenced by histological regression induced by neoadjuvant chemotherapy. Eur Radiol 2003;13:1213–23. Denis F, Bougnoux P, de Poncheville L, Prat M, Catroux R, Tranquart F. In vivo quantitation of tumour vascularization assessed by doppler sonography in rat mammary tumours. Ultrasound in Med Biol 2002;28:431–7. Denis F, Bougnoux P, Paon L, Aget H, le Floch O, Tranquart F. Radiosensitivity of rat mammary tumors correlates with early vessel changes assessed by power-doppler sonography. J Ultras Med 2003;22:921–9. Hlatky L, Hahnfeldt P, Folkman J. Clinical application of anti-angiogenic therapy: microvessel density, what it does and doesn’t tell us. J Natl Cancer Inst 2002;94:883–93.