| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Magnetic Resonance Science Center, 1 Irving St., Rm.
AC-109, Box 1290, University of California, San Francisco, San Francisco, CA
94143.
2 Department of Radiation Oncology, University of California, San Francisco, San
Francisco, CA.
3 Comprehensive Cancer Center, University of California, San Francisco, San
Francisco, CA.
Received May 8, 2007;
accepted after revision December 12, 2007.
Address correspondence to N. Hylton
(nola.hylton{at}radiology.ucsf.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Signal enhancement ratio analysis of nontumor breast stroma was performed on dynamic contrast-enhanced MRI scans of 42 patients who received neoadjuvant chemotherapy for invasive breast cancer performed before chemotherapy (scan 1) and after one cycle of chemotherapy (scan 2). Stromal signal enhancement ratio values were then correlated to several clinical parameters and to clinical outcome using univariate and multivariate analyses. Median follow-up for the group was 52.1 months.
RESULTS. On univariate analysis, factors that were significantly associated (p < 0.05) with disease-free survival included the mean stromal signal enhancement ratio at scan 2 (hazard ratio [HR] = 0.11, 95% CI = 0.013-0.88, p = 0.03), pretreatment tumor size (HR = 1.33, 95% CI = 1.07-1.66, p = 0.012), pretreatment tumor volume (HR = 1.04, 95% CI = 1.01-1.07, p = 0.006), and number of involved axillary lymph nodes (HR = 1.18, 95% CI = 1.05-1.32, p = 0.005). These factors were then analyzed in a multivariate Cox proportional hazards model. The only factor that was associated with disease-free survival was mean stromal signal enhancement ratio at scan 2 (HR = 0.11, 95% CI = 0.012-0.95, p < 0.045).
CONCLUSION. These findings indicate that breast stroma tissue outside the incident tumor can be quantified using signal enhancement ratio analysis on dynamic contrast-enhanced MRI. Stromal signal enhancement ratio is a potential indicator for response to treatment and for overall outcome in patients with breast cancer; however, these results should be validated in a prospective study.
Keywords: breast cancer • breast stroma • microvascular density • MRI • neoadjuvant chemotherapy • oncologic imaging
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI has had an evolving role in the management of breast cancer, and studies have confirmed the value of dynamic contrast-enhanced MRI in cancer detection and diagnosis [6, 7]. There is also substantial support in the literature that dynamic contrast-enhanced MRI has the ability to accurately predict the extent of residual disease in the breast after neoadjuvant chemotherapy [8, 9]. The signal enhancement ratio is a calculated parameter that quantifies the kinetics of contrast enhancement on high-spatial-resolution dynamic contrast-enhanced MRI. Our group used dynamic contrast-enhanced MRI with signal enhancement ratio analysis previously to correlate signal enhancement ratio patterns with histopathology and microvessel density in breast tumors [10]. We found that areas with high signal enhancement ratio values (early enhancement with rapid washout) were significantly correlated with high tumor vascularity.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patient Population
Between 1995 and 2002, 62 patients were enrolled in a neoadjuvant
chemotherapy breast cancer protocol
[11]. All the patients had
stage II or III locally advanced invasive breast cancer, defined as tumors
that have not spread beyond the breast and regional lymph nodes but that may
involve the skin of the breast or the chest wall. All patients had invasive
breast cancer confirmed by pathology of core biopsy or fine-needle aspiration
before treatment, and the histology results were verified on postsurgical
pathology specimens. Four patients were excluded from our analysis for the
following reasons: no pretreatment MRI was performed, inability to complete
therapy, deviation from the therapeutic protocol, and loss to follow-up. An
additional 16 patients were found to have insufficient fibroglandular stroma
for evaluation, which was defined as having less than 2 cm of radiographically
normal (nonenhancing) breast stroma extending from the edge of visibly
enhancing tumor, and were excluded from the analysis.
The remaining 42 patients were included in this study. All but one patient received four cycles of doxorubicin and cyclophosphamide chemotherapy administered every 3 weeks, and 11 received additional treatment consisting of weekly taxane. All patients underwent dynamic contrast-enhanced MRI before chemotherapy (scan 1), and 33 patients were scanned after one cycle of chemotherapy (scan 2). Patient characteristics, clinical variables, and recurrence and outcome data were obtained from the preexisting data set.
Clinical Variables
Clinical variables were retrieved from the original cohort database.
Patient age was recorded at the beginning of treatment. Pathologic data were
determined from pathology reports. Recurrence and disease-free survival were
assessed for each patient based on clinical examination and mammographic
imaging at 6-month or 1-year intervals after surgery. Length of disease-free
survival was defined as the time between the primary surgery (after
neoadjuvant chemotherapy) and local or distant recurrence or as the time to
the last follow-up in patients with no evidence of recurrence.
MRI Acquisition
As described in the original prospective study
[11], MRI was performed on the
involved (ipsilateral) breast only. Images were acquired on a 1.5-T scanner
(Signa, GE Healthcare) using a dedicated bilateral phased-array breast coil. A
fatsuppressed 3D fast gradient-recalled echo sequence was used (TR/TE, 8/4.2;
flip angle, 20°; 2 repetitions)
[12]. The entire breast was
covered with 60 slices, each 2 mm thick and acquired in the sagittal
orientation. The contrast agent used was gadopentetate dimeglumine (Magnevist,
Bayer HealthCare), injected at a dose of 0.1 mmol/kg of body weight, followed
by a 10-mL saline flush. Three time points (t0, t1,
t2) were acquired during each MRI examination: a baseline scan
before contrast agent injection (t0), followed by two sequential
scans after contrast injection (t1, t2), yielding a
temporal sampling of 0, 2.5, and 7.5 minutes, respectively.
Signal Enhancement Ratio Analysis
Three different patterns of signal increase and washout after contrast
injection are shown in Figure
1. The bottom curve shows a slow gradual increase in enhancement,
more characteristic of normal tissue; the middle curve shows early enhancement
with little washout, essentially a plateau in signal intensity; the top curve
shows a pattern of early enhancement with quick washout, which is more
characteristic of highly vascularized tissue and neoangiogenic vessels.
These patterns are quantified with the signal enhancement ratio, which
compares enhancement in the first contrast-enhanced image with enhancement in
the second contrast-enhanced image
[13].
Figure 1 displays signal
intensity (S) on the y-axis and time (t) on the x-axis.
Images are acquired at three time points: t0, or before contrast
administration; t1, at 2.5 minutes after contrast injection; and
t2, at 7.5 minutes after contrast injection. S0,
S1, and S2 represent the signal intensity at
t0, t1, and t2, respectively. The equation
for signal enhancement ratio (SER), a unitless value, is normalized
and comparable from patient to patient and from scan to scan:
 |
MRI Postprocessing and Signal Enhancement Ratio
To characterize normal-appearing breast fibroglandular tissue in the
ipsilateral breast, regions of interest (ROIs) were created on the first
contrast-enhanced image (image acquired at t1, 2.5 minutes after
contrast injection), as shown in Figures
2A and
2B. On a single representative
sagittal slice containing the largest dimension of the visible enhancing
tumor, five circular ROIs, each 5 mm in diameter, were placed extending
radially from the tumor edge. The first ROI was placed within the visible
tumor and the next four in normal-appearing breast fibroglandular stroma. This
is shown schematically in Figures
2A and
2B: "T" inside the
circle represents the tumor ROI, and "S" represents the stromal
ROIs. A second set of similarly placed ROIs was obtained along a different
radius of the same image if enough normal stroma was present. This process was
repeated for scan 2 after identifying the representative sagittal slice that
matched the tumor characteristics at scan 1; ROIs were placed in approximately
the same locations at scan 1 (before treatment) and scan 2 (after one cycle of
chemotherapy). This method was adequate for all patients because the tumor did
not significantly change between those two time points.
|
|
Signal enhancement ratio maps, which assign signal enhancement ratio values on a voxel-byvoxel basis, were generated from contrast-enhanced fat-suppressed T1-weighted breast MR images using a customized software program developed at our institution [13]. The signal enhancement ratio algorithm was initially designed to characterize breast tumors and used a minimum threshold of at least 70% enhancement from baseline to 2.5 minutes after contrast injection to define malignancy. In this study, the aim was to assess breast fibroglandular stroma, a tissue compartment that enhances less avidly, so the initial enhancement threshold was reduced to 20%. The ROIs created on the original t1 images were superimposed on the signal enhancement ratio maps to extract the mean signal enhancement ratio within each ROI excluding zero values. The ROIs were generated with the users blinded to the volumetric signal enhancement ratio map so they could avoid selecting areas of interest with higher signal enhancement ratio values. Users were also blinded to the clinical outcome of each patient while conducting MRI postprocessing and signal enhancement ratio analysis.
Pretest for Interuser Variability
To assess the variability in ROI selection, two independent users collected
data as described on MRI scans of 15 patients. The values obtained for stromal
signal enhancement ratio were analyzed for any differences with respect to the
user. The same two users conducted the MRI postprocessing and signal
enhancement ratio acquisition for the entire cohort.
Statistical Analysis
Stromal signal enhancement ratio values were evaluated as a function of the
distance from visible tumor and for the effect of chemotherapy. In addition,
the associations between stromal signal enhancement ratio and pathologic
characteristics, clinical response, and disease recurrence were analyzed using
Wilcoxon's and Student's t tests. Stromal signal enhancement ratio
values were dichotomized for the analyses, as described in the Results
section.
When analyzing disease-free survival, data for patients without disease recurrence were censored. Univariate analyses using Kaplan-Meier log-rank tests and the Cox proportional hazards model were performed to identify variables associated with disease-free survival. A hazard ratio (HR) and p value were reported for each variable in comparison with length of disease-free survival. Variables that were found to be significant in the univariate analysis were entered into multivariate Cox regression models to identify significant predictors of disease-free survival. Disease-free survival curves were produced using the Kaplan-Meier method. Statistical significance was established at a p value of < 0.05.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patient Characteristics
Table 1 shows the
characteristics of the study cohort. The median age was 48.3 years (range,
29.7-71.5 years). The median follow-up was 52.1 months (range, 17.6-86.7
months) in the patients who remained disease-free. In patients with
recurrence, the median time to recurrence was 25.2 months (range, 3.6-73.9
months). The mean size of the tumor before treatment, measured as the longest
diameter of the tumor on MRI, was 4.69 cm. Most patients had invasive ductal
carcinoma (83.3%), 15% had lobular histology, and 1.7% had medullary
carcinoma. More than half of the patients (52.4%) had grade 3 disease
according to the Scharff-Bloom-Richardson grading system, and 57.1% had
positive axillary lymph nodes.
|
All but one patient received doxorubicin and cyclophosphamide chemotherapy, and 11 of these patients received additional treatment with taxane. All patients underwent surgical resection after chemotherapy, with a fairly equal split between mastectomy and lumpectomy. The mean pathologic size, measured as the longest diameter of the lesion after surgical resection, was 2.93 cm. One patient had a complete pathologic response with neoadjuvant chemotherapy and showed no residual disease on resection.
Disease-Free Survival and Recurrence
Disease-free survival of the entire study population, calculated using the
Kaplan-Meier method, is displayed in Figure
3. The 2-year disease-free survival rate was 80% for the group of
patients studied (n = 42). There were 15 recurrences (35.7% of
patients): 11 distant metastases (26.2%) and four local recurrences (9.5%). Of
the 15 patients with recurrence, seven had received taxane. There was no
statistical difference in recurrence rates based on taxane use.
|
We attempted to assess the relationship between mean stromal signal
enhancement ratio values and recurrence. A mean stromal signal enhancement
ratio of less than 0.7 for either scan 1 (before treatment) or scan 2 (after
one cycle of chemotherapy) was found to be significantly associated with
recurrence (p 
0.03) (Figs.
4A and
4B). Thus, 0.7 was used to
dichotomize mean stromal signal enhancement ratio as a variable for subsequent
analyses. A third scan after completion of four cycles of neoadjuvant
chemotherapy and before surgery was part of the original study protocol
[11] and was available in 40
of the patients enrolled in the study. However, the analysis of signal
enhancement ratios from the third scan in our cohort did not reveal
significant findings.
|
|
0.7.
The variables that were significantly associated with the length of disease-free survival in the univariate analysis are shown in boldface in Table 1. Both pretreatment tumor size (p = 0.012) and pretreatment tumor volume (p = 0.006) were significantly associated with disease-free survival. The number of positive lymph nodes was also significantly correlated with disease-free survival in this analysis. The mean stromal signal enhancement ratio at scan 2, analyzed as a dichotomous variable, was significantly associated with disease-free survival (p = 0.03). With a hazard ratio of 0.11, a higher value of mean stromal signal enhancement ratio at scan 2 is associated with a decreased risk of recurrence.
The four factors found to be significant in the univariate analysis were incorporated into the multivariate Cox proportional hazards model. Only mean stromal signal enhancement ratio at scan 2 (HR = 0.11, p < 0.045) remained independently associated with disease-free survival.
Subset Analysis of Stromal Signal Enhancement Ratio at Scan 2 and Disease-Free Survival
Figure 5 displays
graphically the Kaplan-Meier curves for length of disease-free survival with
patients divided on the basis of a mean stromal signal enhancement ratio value
at scan 2 of < 0.7 or 0.7. There was a significant difference in
disease-free survival in the two groups (p = 0.012, log-rank test),
with a 3-year disease-free survival rate of 82.4% in the group with a mean
stromal signal enhancement ratio at scan 2 of 0.7 and 47.9% in the group
with values < 0.7.
|
0.7 (p = 0.03). Changes in tumor size and volume after
chemotherapy, based on dichotomized mean stromal signal enhancement ratio
values at scan 1, are shown in Table
2. There was no significant difference between the two groups.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The univariate analysis showed that pretreatment tumor size, pretreatment tumor volume, and mean stromal signal enhancement ratio at scan 2 were significant predictors of disease-free survival (Table 1). These strong correlations validate the use of dynamic contrast-enhanced MRI in the neoadjuvant setting to help predict clinical outcome. The number of positive axillary lymph nodes was the only other prognostic factor associated with disease-free survival in the univariate analysis. These measurements did not identify any patients who did not respond to therapy. Other than the patient who had a complete pathologic response, all of the remaining patients had partial responses. The degree of response did not correlate with stromal signal enhancement ratio measurements, and we chose not to analyze this parameter further in the present study.
In the multivariate model for disease-free survival, the mean stromal
signal enhancement ratio at scan 2 was found to be a significant independent
predictor of disease-free survival (p < 0.045), associated with a
dramatically reduced risk of recurrence in patients with values of 0.7.
This finding indicates that changes associated with one cycle of chemotherapy
may further predict outcome in patients receiving neoadjuvant treatment.
Interestingly, patients with recurrence had significantly lower values (< 0.7) of mean stromal signal enhancement ratios at both scan 1 and scan 2. Mean stromal signal enhancement ratio values at scan 1 were also associated with a number of positive lymph nodes. Patients with lower values were more likely to be node-positive, and those with values < 0.7 had more positive nodes and a smaller decrease in tumor size after four cycles of chemotherapy. These findings may be related: Patients with lower mean stromal signal enhancement ratio values at scan 1 had more lymph node involvement, showed less clinical response to treatment, and were more likely to experience recurrence. When we stratified patients according to the number of axillary lymph nodes involved, the subsets were too small to reach statistical significance.
Our data reveal an inverse correlation between stromal signal enhancement ratio and recurrence in this neoadjuvant population. A high stromal signal enhancement ratio may reflect greater microvessel density and thus better delivery of the chemotherapeutic agent to the tumor, which would result in a better clinical response and decreased likelihood of recurrence after surgery. Investigators have previously reported that the effectiveness of molecular medicines to treat cancer may be jeopardized if they cannot efficiently penetrate tumor tissue [16] because of compromised vasculature. Investigators who used 15O-water PET before and after 2 months of chemotherapy reported a decline in tumor blood flow among patients who responded compared with those who did not [17]; this finding indicates that a decline in malignant angiogenesis correlated with response [17]. Indeed, cell culture and animal models have shown that normalization in the physiology of cancerous blood vessels was indicative of response to antiangiogenic therapy [18]. Our data indicate that beyond the tumor vasculature, the stromal vasculature may also play a critical role in determining response to chemotherapy.
There are a few limitations to this study. Potential confounders such as time of menses, the presence of fibrocystic disease, and change in body habitus between time points were not recorded in the present study. Part of the reason is that at the time this study was initiated, collection of MRI data that studied these factors was ongoing and existing data to support accounting for all the factors did not exist. Certainly, these parameters will be studied and accounted for in future studies. Also, we could not correlate the stromal signal enhancement ratio with the final histopathology of the resection specimens because the patients in this study received neoadjuvant chemotherapy and, thus, the tumors were still intact at the time of scan 1 and scan 2. We plan to investigate the histopathologic correlation of stromal signal enhancement ratio in future studies, however.
The study population is relatively young with large high-grade tumors—most of which were invasive ductal carcinoma. Because of the high number of recurrences in this cohort, we were able to study the association of imaging findings with clinical outcome. However, the distinct nature of this cohort limits our ability to generalize these findings to populations of patients with better prognoses—that is, those patients with tumors that were detected earlier. Also, the automated algorithm that measures signal enhancement ratio was optimized for use in tumor where there is avid enhancement. Although we lowered the peak enhancement threshold in our methodology, there might be other adjustments that could be made to best analyze the contrast kinetics of stromal tissue. We believe that these data indicate that stromal signal enhancement ratio potentially may be an important biologic indicator of tumor response and warrants further investigation in a larger prospective study.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A.-M. Tokes, A. M. Szasz, A. Farkas, A. I. Toth, M. Dank, L. Harsanyi, B. A. Molnar, I. A. Molnar, Z. Laszlo, Z. Rusz, et al. Stromal Matrix Protein Expression Following Preoperative Systemic Therapy in Breast Cancer Clin. Cancer Res., January 15, 2009; 15(2): 731 - 739. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), 10
had both scan 1 and scan 2 values. Of 27 patients without recurrence
(
), 23 had both scan 1 and scan 2 values.