Breast cancer is a major health problem. Unlike many other
forms of cancers, awareness among women of the risks associated
with the breast cancer is high and derives from many
sources including health education programs, extensive media
coverage and firsthand knowledge from friends and relatives
Despite this public awareness, the best screening tool is mammography,
which has a false negative rate of 10’25%.
Furthermore, mammography has limitations in its ability to
accurately establish the extent of the disease in the breast cancer
for some subsets of women undergoing treatment. It may
underestimate the extent of lobular carcinoma up to 25% of
cases. That’s why interest has focused on MRI as an adjunct
to mammography (1).
Dynamic contrast enhancement MRI (DCE-MRI) of the
breast has a high sensitivity for breast cancer detection and
has recently been shown to be the most sensitive breast screening
technique for women at high risk. DCE-MRI is also more
accurate than mammography or U/S. for delineation of the
extent of the disease in patient with recent diagnosis of cancer.
The high sensitivity of clinical breast DCE-MRI is due to its
differential enhancement between normal and malignant tissue
on TIWI (2).
Researches on new MRI technique are being conducted to
further increase the specificity of breast MRI. Diffusion
weighted imaging (DWI) was recently integrated into the standard
breast MRI examination for this purpose. It is a noninvasive
technique that measures the random motion of free
water protons (Brownian motion) and characterizes the tissue
with a mechanism that is different from T1 and T2 relaxation.
The motion of water protons in the tissue is affected by fluid
viscosity, membrane permeability, blood flow and cellularity
of the tissue, for quantification of this motion, Apparent
Diffusion Coefficient (ADC) values are used (3).
Diffusion weighted MR imaging detects early changes in
the morphology and physiology of tissues, such as changes
in the permeability of membrane, cell swelling and/or cell lysis
(4,5). Since 2002 many studies revealed the usefulness of DWI
in differentiation of malignant from benign tumor of the
breast. In these studies, sensitivity in the range of 80’96%
and specificity in the range of 46’91% were reported (6,7).
Moreover, DWI has a potential role for characterization of
breast masses and treatment monitoring after chemotherapy
Promising findings from the preliminary DWI studies of the
breast have shown significantly lower ADC measures for
breast carcinoma than for benign breast lesion or normal tissue.
The lower ADC in malignancy is primarily attributed to
higher cell density causing increased restriction of extracellular
matrix and increased fraction of signal coming from intracellular
water. A recent study reported high accuracy for characterizing
enhancing breast masses through multivariate
combination of DWI and DCE-MRI (10).
In addition to morphologic and kinetic analyses, molecular
information has been expected to be useful for diagnosis of the
breast disease. In vivo proton (H) Magnetic Resonance
Spectroscopy (MRS) of the breast which provides molecular
information obtained in non-invasive manner, has shown that
the choline is generally not detectable in normal breast tissue.
In several studies performed on 1.5 T MR imagers, investigators
have reported sensitivities of 70’100% and of 67’100%
specificities for breast MRS (11).
4. Discussion
The use of MRI for screening high risk patients is now recommended
by almost all major medical societies. Breast cancers
in the high risk populations generally present at younger age
and screening with both mammography and MRI is recommended
beginning at age 30 ys. Breast MRI is clearly the most
sensitive method for breast cancer detection and specificities
are comparable if not superior to other breast imaging
methods (12).
Dynamic contrast enhanced MRI (DCE-MRI) of the
breast has a high sensitivity for the breast cancer detection
and has a recently been shown to be the most sensitive breast
screening technique for women at high risk. It is also more
accurate than mammography or ultrasound for the delineation
of the extent of the disease in patients with recent diagnosis of
cancer (2).
Up to now breast MRI is analysed according to morphological
criteria, enhancement kinetics and T2 characteristics
of the breast lesion. However an overlap between benign and
malignant lesions which leads to a reported specificity of about
40’80%. There is an increasing number of congress abstracts
and published studies which proves that the specificity of
breast MR could be increased using DWI and MRS studies
(13).
MRI lesions characteristics in the our study included size in
ml, type (mass, non-mass like enhancement, cystic lesion) and
BI-RADS category, this agreed with the study of Savanah and
Patridage (14). We visually analysed the enhancement characteristics
of the lesion from the post-contrast subtracted images
and this agreed with Kvistad (15) who made detection of
enhancing lesion easier by subtracted image.
For generation of time intensity curve (TIC) we set ROIs
based on visual inspection. TIC was obtained with the use of
a small ROI inside the mass and avoidance of central hemorrhagic
necrosis or fibrosis (16).
In our study type III curve was the most common type in
the pathologically proved malignant cases (6/10) (60%). This
correlated with Jack’s (17) study in which the type III (washout)
curve was detected in 32 patients out of 37 ones
(86.5%), on other hand type I a curve was seen in 5 patients
out of 10 patients with malignant lesions (50%) of our study,
compared to 10.8% in Jack, s study, Also type 1a curve was
detected in 13 benign lesions out of 21 ones (13/21) (61.9%)
of our study and this was in agree with the results of Jack, s
study in which type 1 curve was detected in 65% of benign
breast lesions. We agreed also with Jack, s results (17) in that
there was significant difference between malignant and benign
lesions at the distribution of curve type and in that the TIC is
useful in differentiating malignant lesions from benign ones.
The results of our study were in agree with that of Yi et al.
(18) who stated that we could acquire general information
about tumor vascular physiology, interstitial space volume
and prognostic factor by analyzing TIC without a complicated
acquisition process.
In our study we found that DCE-MRI had sensitivity
(92.3%), specificity (81%) and accuracy (85.3%).These results
were compared to that of Kul et al. (19) who reported 75.7%
sensitivity, 97.5% specificity and 88.1% accuracy. Overlap in
morphologic characteristics and kinetic features of malignant
and benign lesions caused improper classifications. In an
attempt to increase diagnostic efficacy, mainly the specificity
of breast MR, we evaluated the additional role of DWI and
MRS.
We ensured that DWI was performed prior to contrast
enhancement to avoid the effect of contrast material. In some
studies however, DWI was performed after injection of contrast
like that of the study done by S.C. Partridge et al. (20)
but actually they considered this one of the limitation of their
study and they stated that it might be preferable to acquire
DWI sequence before contrast injection.
Contrarily Janka et al. (21) study have been done to compare
the DWI image and ADC results before and after administration
of contrast, showing that DWI after contrast
administration gives a slightly better lesion discrimination.
In our study we assessed the DWI for all cases with breast
lesions in conjunction with DCE-MRI. This was in agree with
that of Kuroki et al. (9) who stated that DWI is not a complete
method of diagnosis. In most applications the diffusion gradients
are integrated in echo planar imaging (EPI) sequences
which exhibit high signal intensity in areas with restricted diffusion
as well as in fatty tissue. This make fat saturation techniques
necessary to identify the lesions in the diffusion
weighted images (13).
In our study we selected b values of 50, 400 and 800 which
were the same values chosen by Wenkel et al. (13). Liberman
et al. (22) which concluded that for good image quality and
valid differentiation between malignant and benign tumor,
the optimized b value of DWI is in the range of
600’1200 s/mm2 at 1.5 T. Our 34 studied cases were classified
according to diffusion pattern in the detected lesion into two
groups; G.I. included 25 patients without restricted diffusion
(74%) and G.II. Included 9 patients with restricted diffusion
(26%). The mean ADC value was significantly lower
(0.5’0.9 ‘ 10_3 mm2/s) for malignant lesion in comparison
with that of benign lesions (1.7’2.7 ‘ 10_3 mm2/s).
Out of 10 patients with malignant lesion of our study 6
patients showed restricted diffusion and out of 21 patients with
benign breast lesions 19 patients showed non-restricted diffusion.
In our study we found significant difference between
ADC values of malignant and benign breast lesions, assuming
a threshold of 1.2 ‘ 10_3 mm2/s. Similarly an 2014 by Nogeria
et al. (23) study, proved DWI with complementary ADC values
to be useful for the detection and characterization of breast
lesion where mean ADCs of 1.99 ” 0.27 ‘ 10_3 mm2/s,
1.08” 0.25 ‘ 10_3 mm2/s, and 1.74” 0.35 ‘ 10_3 mm2/s,
were obtained for normal tissue, malignant, and benign lesion
respectively.
Our study reported a raised specificity from 81% to 90.5%
and a slightly improved PPV from 75% to 77.8% after
combining DWI to DCE-MRI without any improvement in
sensitivity (53.8%). NPV was 76% and accuracy was 76.5%.
These results agreed with Kul et al. (19) who reported an
improved specificity (86.5%), sensitivity (91.5%), PPV
(89.6%), NPV (88.9%) and accuracy (89.3%). An older study
by Sonmez (24) shared nearly the same results.
The correlation between the findings of DWI and pathological
results of different breast lesions showed the value of this
sequence as an additive tool that augment the results of
dynamic MRI and increase the overall specificity of the study.
This fact gains a wide agreement with a large number of studies
(25,26). DWI has some important advantages for use in combined
MR protocols. It is available on most commercial MR
scanners and does not need secondary gadolinium use. It has
a very short imaging time with the use of EPI. The evaluation
of the image obtained is quantitative and rather easy (24).
In our study, we have demonstrated the clinical utility of
breast 1H MRS to distinguish between malignant and benign
breast lesions by use of the composite Choline signal. All published
results suggest that there is a relationship between the
choline metabolic activity and angiogenic activity. As choline
is involved in cellular proliferation, it is logical that angiogenesis
increases to support tumor metabolic requirements
(27,28).
In our study we found that the single-voxel proton MRS of
the breast is clinically feasible. It can be performed after standard
unenhanced and contrast breast study in an examination
time of approximately 40 min, with relatively failure rate (6%),
similarly 3% failure rate reported in meta-analysis by Tse et al.
(29) of more than 280 patients. In our study, breast 1H-MR
spectroscopy was predominantly performed with qualitative
analysis of choline peak integral.
According to spectroscopic analysis of the breast lesions,
the 34 patients of our study were classified into two groups;
G.I. included 17 patients with MRS suspicious results (probably
malignant where choline trace was detected) and G.II.
included 17 patients with non-suspicious MRS (probably
benign where no detectable choline trace). We found that
out of 17 patients with choline trace 9 patient showed malignant
lesion (9/17) (52.9%), 3 patients showed high risk lesions
(3/17)while the remained 5 patient were having benign lesions
(29.4%), while out of 17 patients without detectable choline
trace 16 patients were pathologically proven to be benign
(16/17) (94.1%). In our study we found that MRS sensitivity
was (92.3%), specificity (76.2%) and accuracy (82.4%) and
our these results were in agree with that of Rachel et al. (30)
study where sensitivity and specificity of SV-MRS were 71%
and 85% respectively.
The combination of choline presence and ADC values
achieved higher level of accuracy and specificity in discriminating
malignant from benign lesions over choline presence or
ADC results alone (27). Huang et al. (31) also reported the
increase in sensitivity and specificity of the breast cancer detection
when DCE-MRI, SV 1H -MRS and T2* weighted perfusion
MR imaging results were combined within the
examination. Specificity improved from 62.5% to 87% when
MRS finding were integrated to DCE-MRI and increased further
to 100% once perfusion MR results were considered. This
clearly highlights the benefit of incorporating secondary MR
modalities into the routine breast MR examination.
Using the optimal threshold for absolute tCho, we reported
one false ‘ve (recurrent mammary ductal carcinoma) and 5
false +ve cases (two lactating mothers, two fibroadenomas
and a case of fat necrosis). The histologic type of the false ‘
ve lesion is not surprising because of the possibility of a relatively
low level of tCho in ductal carcinoma which is different
from invasive ductal cancer (29). Yeung et al. (32) reported
nine false ‘ve four ductal carcinoma in situ and three invasive
ductal carcinoma with an extensive in situ component.
Conversely, it is already known that some fibroadenoma
may present high levels of tCho at both in vitro and in vivo
MRS (33).
Our MRS study for 34 patients with breast lesion has several
limitations. In addition to small sample size already mentioned,
we should consider the variable filling factor caused by
the impossibility of reducing the VOI below 1 ml, thus almost
always including surrounding fat or healthy gland parenchyma.
More advanced hardware (e.g., field strength higher than
1.5 tesla, multichannel coils) and dedicated post-processing
software could provide MR spectra of better quality than
those was obtained. Moreover approach of our study, based
on the use of arbitrary units, may not allow the application
of our cutoff value for tCho peak integral to different technical
and clinical settings. Finally the long acquisition time of our
MRS sequence (nearly 13 min) could have reduced the spectral
resolution because of the probability of artifacts from respiratory
and other patient, s motion.
In conclusion, our experience first showed that in vivo 1.5 T
single voxel water and fat suppressed proton MRS of the
breast can be added as a last phase after unenhanced and
DCE-MRI, with an entire examination time not longer than
40 min, Moreover we showed that breast MRS using tCho
peak integral allows high sensitivity and specificity. Studies
of large clinical series are warranted.
We agreed with 2014 multi-parametric MRI (MP MRI)
study done by Pinker et al. (34) who stated that the breast
MRI with 3 parameters (DCE-MRI, DWI, and MRS)
increased the diagnostic accuracy of breast cancer in comparison
with the DCE-MRI alone and MP MRI with 2 parameters,
yielding significantly higher AUC (>0.90 for small
tissue sample) in comparison with DCE-MRI alone resulting
in elimination of false ‘ve lesions and significantly reducing
the false +ve ones.
5. Conclusion
DCE-MRI of the breast has a high sensitivity for breast cancer
detection and has been recently shown to be the most sensitive
breast screening technique for women at high risk and more
accurate than sonomammography in delineation the extent
of the disease in patients with recent diagnosis of cancer.
DWI MR detects early changes in the morphology and
physiology of the tissue. It has a potential role for characterization
of the breast masses and treatment monitoring after
chemotherapy.
Combining DWI to DCE-MRI improves the discrimination
power of malignant from benign breast lesion and
increases the overall accuracy of MRI and reduces the unnecessary
invasive procedures.
MRS of the breast provides molecular information in noninvasive
manner.
The combination of choline presence and ADC values
achieved higher level of accuracy and specificity in
discriminating malignant from benign lesion over choline presence
or ADC alone.
The specificity improved also when MRS findings were integrated
to DCE-MRI. This clearly highlights the benefit of
incorporating secondary modalities into routine breast MR
examination for elimination of the false ‘ve lesions and reducing
the false +ve lesions.
Conflict of interest
No conflict of interest.