| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University of California, San Francisco, Magnetic
Resonance Science Center, 1 Irving St., Rm. AC-109, San Francisco, CA
94143-1290.
2 Department of Surgery, University of California, San Francisco, Mount Zion
Cancer Center, 2356 Sutter St., San Francisco, CA 94115.
3 Department of Pathology, Marin General Hospital, 250 Bon Air Rd., Greenbrae,
CA 94904.
Received March 4, 2002;
accepted after revision May 10, 2002.
Presented in part at the annual meeting of the International Society for
Magnetic Resonance in Medicine, Honolulu, May 2002.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Before undergoing surgery, 52 patients were imaged before and after receiving neoadjuvant chemotherapy. For each patient, specific malignancy criteria were applied to MR images before chemotherapy to identify the location of tumor, and residual disease was then identified as any remaining enhancement in the same area on the MR images after chemotherapy. Residual tumor size was measured using both the MR technique and the clinical examination findings, and the degree of measurement error for each method was assessed in comparison with the pathologic findings.
RESULTS. The correlation with pathology was an r value of 0.89 for MR measurements compared with an r value of 0.60 for clinical measurements. In addition, MR imaging revealed all cases of residual disease, whereas clinical assessment resulted in five false-negative interpretations in the 52 treated lesions.
CONCLUSION. The high correlation between measurements of residual disease obtained on MR images and those obtained at pathology validates the sensitivity of MR imaging of the breast after chemotherapy.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Four recent studies investigated the accuracy of using MR imaging after chemotherapy and reported mixed results. Generally, MR imaging was found to enable a high sensitivity for detecting the presence of residual disease in the breast [3,4,5]. However, other studies have reported high rates of false-negative interpretations. One group of researchers found that MR imaging failed to reveal four of 13 residual tumors after preoperative chemotherapy [6], and another group reported that only 16 of 33 lesions were accurately depicted after treatment (Gligorov J et al., presented at the American Society of Clinical Oncology meeting, June 2001). Poor sensitivity was attributed to absent or low enhancement of the lesions after treatment.
The few studies that directly compared size measurements of residual tumor obtained on MR images with those obtained at pathology also gave mixed results. One study described a high correlation between both types of measurements (r = 0.93) [5], whereas other studies found that MR imaging notably underrepresented lesion size [6] or was generally inaccurate (Gligorov J et al., presented at the ASCO meeting, June 2001) in more than half the cases. In addition to conflicting results, the studies we reviewed had relatively small cohorts (13-33 patients). Clearly, more investigations are needed to develop MR imaging techniques after treatment and to establish the accuracy of using MR imaging to detect residual disease in the breast.
Our previous research illustrated the importance of measuring both contrast enhancement uptake and washout in the tissue to improve diagnostic sensitivity and specificity using MR imaging. We have used these enhancement parameters to successfully segment and measure the extent of malignancies in the breast revealed on MR images (Partridge SC et al., presented at the International Society for Magnetic Resonance in Medicine meeting, May 1999). However, this criterion was refined for untreated lesions and must be further adapted to the neoadjuvant setting on the basis of the changes that occur in the vascular and enhancement dynamics of the tumor during treatment.
We are currently developing new methods for characterizing change in tumor volume using MR imaging because volumetric changes in tumors should better depict treatment response than changes in tumor diameter. However, accurate measurement of residual disease is a prerequisite to the accurate assessment of volumetric changes. Therefore, this study was undertaken to optimize and evaluate our MR imaging and analysis techniques for measuring residual tumor after neoadjuvant chemotherapy in comparison with the results of the clinical examination and pathology.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MR Imaging Acquisition
MR imaging was performed on a 1.5-T Signa scanner (General Electric Medical
Systems, Milwaukee, WI) using a bilateral phased array breast coil (Open
Breast Coil; MRI Devices, Waukesha, WI). Specifications of our
contrast-enhanced imaging sequence included full coverage of the symptomatic
breast with no gap between slices, fat suppression, and good spatial
resolution to maximize sensitivity for detecting residual breast disease. We
used a three-dimensional fast gradient-recalled echo imaging sequence that
produces high-resolution contiguous thin sections with a greater
signal-to-noise ratio than does a two-dimensional acquisition with the same
number of excitations. Imaging parameters included TR/TE, 8/4.2; flip angle,
20°; and repetitions, 2 (oversampling to remove phase wrap). The field of
view was typically 18-20 cm with a slice thickness of 2 mm and an acquisition
matrix of 256 x 192. The resulting in-plane resolution was approximately
0.7 x 0.94 mm, and 60 slices were acquired in the sagittal orientation,
covering the entire symptomatic breast. The contrast agent used was
gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) at a dose of
0.1 mmol/kg of body weight.
Because enhancing lesions may become isointense to the bright fat signal on T1-weighted images, fat suppression was important for improving the conspicuity of abnormal tissue. Fat suppression was achieved using a frequency-selective inversion recovery preparatory pulse to eliminate the signal from fat before image acquisition [7] based on a fat—water chemical shift. The total scanning time with this fat-suppression technique was 5 min, significantly less than scanning times achievable with conventional chemical saturation techniques using the same scanning parameters. The low-order phase-encoding data for this sequence were acquired at the center of the scan, resulting in temporal sampling of 2.5 min from the start of the scan. Three time points (t0, t1, and t2) were acquired during each MR examination; a baseline scan was obtained before contrast injection (t0 = 0), followed by two sequential scans (t1 = 2 min 30 sec and t2 = 7 min 30 sec) obtained after contrast injection.
Measurements of Lesion Enhancement
Twenty-five focal and 27 diffuse tumors with varying enhancement
morphologies were depicted in this study (Figs.
1,2,3).
We initially characterized the change in signal enhancement observed in the
treated tumors compared with the pretreatment levels. Tissue contrast
enhancement shown on the MR images was assessed quantitatively by sampling the
contrast uptake curve at three time points to characterize both initial
enhancement and washout in the lesions. The percentage of enhancement (PE)
after the injection of contrast material represents uptake and is given by
 | (1) |
 | (2) |
|
|
|
Measurements of Lesion Size
Initial tumor extent was identified on the MR images obtained before
treatment. Lesion location and extent were determined on the basis of
malignancy criteria that were previously optimized by receiver operating
characteristic analysis for identifying breast lesions (Hylton NM et al.,
presented at the RSNA meeting, November 1998). Lesions meeting the criteria
for percentage of enhancement greater than 80% and a signal enhancement ratio
greater than 1.0 were defined as malignant tissue for calculation of tumor
extent in pretreated lesions.
Two other studies have indicated that after treatment, measurements of residual disease in the breast may necessitate reduced thresholds to avoid underestimation of lesion size [5, 6]. To account for the dampening of contrast uptake response after chemotherapy, we relaxed the criteria for determining malignancy on posttreatment MR images as compared with pretreatment levels to include all tissue with enhancement above normal in the previous region of tumor. Tumor changes in response to chemotherapy included both change in size and extent, as well as change in levels and washout of enhancement (Figs. 4A,4B and 5A,5B). Maximum intensity projections were created to identify the full lesion extent in three dimensions, and tumor diameter was taken in its longest dimension (Fig. 6A,6B). Four research associates measured the tumor sizes obtained on the MR images. In 20 cases, two independent operators performed the same measurements to assess the reproducibility of the MR technique for identification of residual disease.
|
|
|
|
|
|
Clinical tumor size was assessed before and after treatment by lesion palpation. Oncologists, surgeons, or nurse practitioners who were monitoring the patients' treatment performed the physical examinations. Clinical size measurements were not available for one patient included in the study. Pathologic size was taken from pathology reports after surgery. In 15 cases in which the extent of disease was not provided in the pathology report, the cases were independently rereviewed by a pathologist and a surgeon to determine pathologic size.
Statistical Analysis
The interobserver reproducibility of our method for measuring the tumor's
longest diameters using MR imaging was determined by comparing the results
obtained by two independent observers. The difference in measurements was
quantified by the root-mean-square error, and the coefficient of variation was
calculated as the root-mean-square error divided by the overall mean size of
the measurements.
Pathologic determination of tumor size was used as the gold standard for measurements obtained after treatment. The Bland—Altman technique [9] of plotting the differences versus the means was used to assess the measurement error for both the MR imaging and the clinical examinations and to identify any systematic differences or biases. Correlation coefficients were calculated to determine the association of lesion sizes measured by the three methods, and 95% confidence intervals (CI) were calculated for each correlation.
To assess changes in tissue enhancement with treatment, we compared percentage of enhancement and signal enhancement ratio values before and after treatment using a paired t test analysis.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MR Imaging Measurements
The interobserver reproducibility for measuring tumor size on MR images was
good. In the 20 cases in which reproducibility was tested, the
root-mean-square error was 0.54 cm with a coefficient of variation of 11%. We
next compared the results that were obtained from the MR images and pathology
(Fig. 7). In the analysis of
all patients, the correlation between posttreatment MR imaging size and
pathology was high (r = 0.89, p < 0.001), with a 95% CI
of 0.82-0.93. The Bland—Altman analysis suggested that MR assessment had
a negligible bias toward overestimation of lesion size compared with pathology
(mean error, 0.09 cm) and a tight 95% CI
(Fig. 8). The SD of
measurement error was 1.57 cm. The mean size of residual tumor found at
pathology was 3.66 cm. All three complete responses that were determined on MR
imaging also showed no residual disease at pathology.
|
|
Five Problematic Cases
Five patients had pathologic measurements that were problematic for
comparison with MR images. In two of the patients, residual disease was
present across the whole breast, and the pathologic size was larger than the
entire breast diameter shown on MR images. The discrepancy in size was thought
to be due to distortion of the mastectomy specimen when it was detached from
the chest wall and axilla. Two of the five patients had separate measurements
obtained at pathology for ductal carcinoma in situ and invasive components,
but the abnormalities could not be distinguished on the MR images, and a
single, inclusive measurement was taken across the entire area of lesion
enhancement. One patient had measurable enhancement on MR imaging, although no
residual disease was found at lumpectomy. This patient presented with a large
local recurrence 39 weeks after surgery, suggesting that the area of
enhancement was not removed at lumpectomy.
When these five problematic cases were excluded from the analysis, the correlation between MR imaging and the pathologic size was strengthened (r = 0.94, p < 0.001), with a 95% CI of 0.89-0.97. The mean size of residual disease for the remaining patients was 3.40 cm at pathology. The mean error for estimation of residual tumor size on MR imaging was 0.12 cm, and the SD of the measurement error was reduced to 1.00 cm.
Clinical Examination Measurements
Clinical examination measurements showed a moderate correlation with
pathologic size (r = 0.60, p < 0.001), with a 95% CI of
0.39-0.75 (Fig. 9). Removal of
the five problematic cases from the statistical analysis did not improve the
correlation between the clinical measurements and the pathologic findings
(r = 0.55, p 0.001). The Bland—Altman analysis showed
that the clinical examination systematically underestimated lesion size
compared with the pathologic findings (mean error, -0.51 cm) and had a wide
95% CI (Fig. 10). The SD of
the measurement error was 2.81 cm. At pathology, five of the eight complete
clinical responders were found to have residual disease (mean size, 4.7
cm).
|
|
Influence of Lesion Morphology on MR Imaging Measurements
In our study, no significant difference was found between the accuracy of
measuring focal or diffuse lesions on MR images. However, four of the five
problematic cases that were described earlier were diffuse lesions, which may
indicate that interpretation of these lesions also proves difficult for the
pathologic assessment.
Characterization of Changes in Signal Enhancement
Lesion enhancement values were measured before and after treatment to
characterize changes in tumor vascularity. In general, decreases were observed
in both the peak percentage of enhancement and the peak signal enhancement
ratio. The median tumor peak percentage of enhancement was 210% before
treatment (range, 148-357%) and decreased to 166% after treatment (range,
68-316%). Statistical analysis showed that values for peak percentage of
enhancement decreased significantly after treatment (p < 0.001,
paired t test), with a median change of -24%. Of the 52 lesions in
the study, 42 showed a decrease in peak enhancement after treatment and 10
showed an increase in enhancement.
Similarly, the median value for the peak signal enhancement ratio dropped from 1.96 before treatment (range, 1.42-2.54) to 1.52 after treatment, (range, 0-2.52), indicating lower rates of contrast washout in the treated tumors. The decrease observed in the signal enhancement ratio values was also significant (p < 0.001, paired t test), and the median change was -21% after neoadjuvant chemotherapy. Of the 52 lesions in the study, 42 decreased in peak signal enhancement ratio and 10 increased (three of which also showed increases in peak percentage of enhancement). These findings are consistent with the visual observation of dampened enhancement after treatment.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our finding of reduced enhancement in lesions of most patients after receiving neoadjuvant treatment correlates with reports from previous studies [5, 6, 10]. Changes in tumor vascularity in response to chemotherapy may explain the decrease or delay in enhancement observed in the tumors after treatment. In one prior study, this decrease in tumor contrast uptake contributed to a high rate (4/13) of false-negative interpretations and caused a significant underestimation of the extent of residual tumor in two other patients [6]. The most likely factor contributing to the poor sensitivity was the relatively low imaging resolution used in the study (slice partitioning, 4 mm; pixel size, 1.4 x 2.2 mm), resulting in voxel sizes significantly larger than those used in the other studies. For this reason, we used a higher imaging resolution (0.703 x 0.94 x 2 mm) with fat suppression and chose to reduce our criterion for identifying carcinoma in the breast after treatment to any notable enhancement in the region of prior tumor. This approach increased our sensitivity for the detection of carcinoma after treatment, allowing us to correctly identify and measure 44 of 44 residual lesions. These results compared well with the prior findings obtained by Gilles et al. (17/18) [3], Abraham et al. (30/31) [4], and Weatherall et al. (20/20) [5], further validating the use of MR imaging to accurately measure residual disease after presurgical chemotherapy. Although our criterion for identifying residual disease allowed any amount of enhancement to be considered, we found the resulting false-positive rate and overestimation of disease to be minimal.
Clinical examination was found to be less accurate than our MR technique for assessing residual cancer in the breast. These results also agree well with findings obtained in previous studies. Five false-negative interpretations occurred through clinical measurements, whereas no residual cancers were missed on MR imaging. The mean size of cancers missed at clinical examination was 4.7 cm. In addition, clinical examinations frequently underestimated size in large tumors and overestimated size in tumors less than 4 cm. The underestimation of disease in patients in whom breast conservation is an option could result in positive margins, morbidity, cost, and the stress of further surgical procedures. In smaller tumors, an overestimation of residual size may lead to aggressive surgery when breast conservation is possible. Patients were treated at a comprehensive breast cancer center, and physical examinations were performed by various clinicians, which may have contributed to measurement inconsistencies.
Mammography was not performed after treatment in most cases; therefore, we did not include mammographic representation of disease extent in the comparison.
The breast tumor morphologies observed in our study varied from very focal lesions to those diffusely spread throughout the breast parenchyma. However, no significant difference was found in the accuracy of our technique for measuring focal or diffuse lesions in this study. In general, it is more challenging to measure the extent within the breast tissue of diffusely enhancing lesions, partially because they may have a lower signal intensity and therefore may be less visible on maximum intensity projections. These diffuse lesions may also have extensions that are difficult to measure and correlate with pathologic sampling.
The results of our study validate the accuracy of using MR imaging to identify residual breast disease in patients who receive neoadjuvant chemotherapy. Investigations are currently underway to characterize early treatment response in tumors. Because of variability in tumor distribution and shrinkage patterns, we hypothesize that measurements of tumor volume rather than longest diameter will more accurately represent changes in tumor size in response to treatment. Preliminary results show that significant changes in tumor volume are measurable after only one cycle of chemotherapy (Partridge SC et al., presented at the ISMRM meeting, April 2001). Also, changes in tumor volume show correlation with disease-free survival (Patridge SC et al., presented at the ISMRM meeting, May 2002). This finding indicates that measurements of tumor volume obtained on MR images could provide a novel means for detecting response to chemotherapy early in the course of treatment when the information can be used to tailor treatment for individual patients.
Acknowledgments
We thank Lorna Beccaria, Margarita Watkins, Niles Bruce, and Evelyn Proctor
for their technical and administrative assistance in this study.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Hattangadi, C. Park, J. Rembert, C. Klifa, J. Hwang, J. Gibbs, and N. Hylton Breast Stromal Enhancement on MRI Is Associated with Response to Neoadjuvant Chemotherapy Am. J. Roentgenol., June 1, 2008; 190(6): 1630 - 1636. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Orel Who Should Have Breast Magnetic Resonance Imaging Evaluation? J. Clin. Oncol., February 10, 2008; 26(5): 703 - 711. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. A. Buchholz, C. D. Lehman, J. R. Harris, B. A. Pockaj, N. Khouri, N. F. Hylton, M. J. Miller, T. Whelan, L. J. Pierce, L. J. Esserman, et al. Statement of the Science Concerning Locoregional Treatments After Preoperative Chemotherapy for Breast Cancer: A National Cancer Institute Conference J. Clin. Oncol., February 10, 2008; 26(5): 791 - 797. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. J. Solin, S. G. Orel, W.-T. Hwang, E. E. Harris, and M. D. Schnall Relationship of Breast Magnetic Resonance Imaging to Outcome After Breast-Conservation Treatment With Radiation for Women With Early-Stage Invasive Breast Carcinoma or Ductal Carcinoma in Situ J. Clin. Oncol., January 20, 2008; 26(3): 386 - 391. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. M. Duarte, C. Cabello, R. Z. Torresan, M. Alvarenga, G. H. Q. Telles, S. T. Bianchessi, N. Caserta, S. R. Segala, M. d. C. L. de Lima, E. C. S. de Camargo Etchebehere, et al. Fusion of Magnetic Resonance and Scintimammography Images for Breast Cancer Evaluation: A Pilot Study Ann. Surg. Oncol., October 1, 2007; 14(10): 2903 - 2910. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Hsiang, M. Yamamoto, R. S. Mehta, M.-Y. Su, C. H. Baick, K. T. Lane, and J. A. Butler Predicting Nodal Status Using Dynamic Contrast-Enhanced Magnetic Resonance Imaging in Patients With Locally Advanced Breast Cancer Undergoing Neoadjuvant Chemotherapy With and Without Sequential Trastuzumab Arch Surg, September 1, 2007; 142(9): 855 - 861. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. K. Kuhl Current Status of Breast MR Imaging * Part 2. Clinical Applications Radiology, September 1, 2007; 244(3): 672 - 691. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. D. Friedman, S. V. Swaminathan, K. Herman, and L. Kalisher Breast MRI: the importance of bilateral imaging. Am. J. Roentgenol., August 1, 2006; 187(2): 345 - 349. [Full Text] [PDF]
|
|
|
|
|
|
M. Tozaki, T. Kobayashi, S. Uno, K. Aiba, H. Takeyama, H. Shioya, I. Tabei, Y. Toriumi, M. Suzuki, and K. Fukuda Breast-Conserving Surgery After Chemotherapy: Value of MDCT for Determining Tumor Distribution and Shrinkage Pattern Am. J. Roentgenol., February 1, 2006; 186(2): 431 - 439. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Shah, S. K. Shah, and K. V. Greatrex Current Role of Magnetic Resonance Imaging in Breast Imaging: A Primer for the Primary Care Physician J Am Board Fam Med, November 1, 2005; 18(6): 478 - 490. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Partridge, J. E. Gibbs, Y. Lu, L. J. Esserman, D. Tripathy, D. S. Wolverton, H. S. Rugo, E. S. Hwang, C. A. Ewing, and N. M. Hylton MRI Measurements of Breast Tumor Volume Predict Response to Neoadjuvant Chemotherapy and Recurrence-Free Survival Am. J. Roentgenol., June 1, 2005; 184(6): 1774 - 1781. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Esserman Integration of Imaging in the Management of Breast Cancer J. Clin. Oncol., March 10, 2005; 23(8): 1601 - 1602. [Full Text] [PDF]
|
|
|
|
|
|
N. Hylton Magnetic Resonance Imaging of the Breast: Opportunities to Improve Breast Cancer Management J. Clin. Oncol., March 10, 2005; 23(8): 1678 - 1684. [Full Text] [PDF]
|
|
|
|
|
|
E. Yeh, P. Slanetz, D. B. Kopans, E. Rafferty, D. Georgian-Smith, L. Moy, E. Halpern, R. Moore, I. Kuter, and A. Taghian Prospective Comparison of Mammography, Sonography, and MRI in Patients Undergoing Neoadjuvant Chemotherapy for Palpable Breast Cancer Am. J. Roentgenol., March 1, 2005; 184(3): 868 - 877. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. P. Mamounas Tailoring Loco-Regional Therapy with Neoadjuvant Chemotherapy: Another Step in the Right Direction Ann. Surg. Oncol., October 1, 2004; 11(10): 888 - 891. [Full Text] [PDF]
|
|
|
|
|
|
F. Thibault, C. Nos, M. Meunier, C. El Khoury, L. Ollivier, B. Sigal-Zafrani, and K. Clough MRI for Surgical Planning in Patients with Breast Cancer Who Undergo Preoperative Chemotherapy Am. J. Roentgenol., October 1, 2004; 183(4): 1159 - 1168. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ring, A. Webb, S. Ashley, W.H. Allum, S. Ebbs, G. Gui, N.P. Sacks, G. Walsh, and I.E. Smith Is Surgery Necessary After Complete Clinical Remission Following Neoadjuvant Chemotherapy for Early Breast Cancer? J. Clin. Oncol., December 15, 2003; 21(24): 4540 - 4545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. L. Rosen, K. L. Blackwell, J. A. Baker, M. S. Soo, R. C. Bentley, D. Yu, T. V. Samulski, and M. W. Dewhirst Accuracy of MRI in the Detection of Residual Breast Cancer After Neoadjuvant Chemotherapy Am. J. Roentgenol., November 1, 2003; 181(5): 1275 - 1282. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
). Correlation between measurements was strong with and without these
cases (r = 0.89 and r = 0.94, respectively; p <
0.001 for both). Trend line (black line) depicts least-squares fit
for data.
