Accurate measurement of tumour size is essential for the preoperative staging of breast carcinoma,
in order to establish correct surgical treatment planning,
especially when breast-conserving surgery is being contemplated5.
Although the final histological examination is considered as the gold standard,
surgical and therapeutic decisions must be made on the basis of imaging,
mainly with conventional modalities (DM and US),
and supported by MRI in selected cases.
The US tecnique often underestimates tumour size,
even though it is especially useful in patients with dense breasts,
in which mammographic lesions are difficult to assess due to overlying tissues or the summation of different structures 2,5.
However,
while US is capable of measuring in multiple planes,
it should be noted that often the largest dimension of the lesion cannot be measured,
especially if we consider that many cancers are vertical lesions with prevalence of height on the width.
These lesions are characterized by the presence of a rear shadow cone that obscures the deep margin,
thus making the measurement on the major axis more difficult2.
According to the literature data1,9,10,11,
MRI is the most accurate and sensitive preoperative imaging modality for the evaluation of tumour extent (although it has a substantial risk of overestimation) and for the detection of multifocal,
multicentric breast cance3,6.
In addition,
MRI is currently considered the most reliable method for preoperative staging of invasive lobular carcinomas; in fact,
as reported in a recent literature review12,
MRI has a sensitivity of 93% in detecting this special type of tumour and a high level of agreement with histology; MRI also allows the detection of contralateral occult lesions in 7% of cases,
changing the surgical approach in about 25% of patients4.
The MRI,
according to the literature,
is the most accurate and sensitive exam among all the pre-operative imaging modalities for the evaluation of the real tumor extension (although it tends to overestimate it) and for the diagnosis of multifocality and multicentricity.
As outlined above,
DM lesion measurements are often difficult due to the effect of masking and to the tissues superposition.
Moreover,
DM tumour measurement can be influenced by variations in the distance between the tumour and the detector,
by poorly delineated tumour outlines and by compression of the breast during examination.
In addition,
standard mammographic projections do not always capture the maximum tumour size.
The use of additional mammographic views (magnification or compression) can improve visualisation of lesion margins2.
Actually,
there is only one published study5 that analysed the accuracy of DBT in the evaluation of tumour size,
and no studies comparing DBT and MRI in the preoperative assessment of breast cancer size.
In our study,
MRI and DBT measurements had the highest correlation with pathological tumour size and thus provided better results than DM.
DBT and DM (the latter had the lowest agreement with pathology compared to the other imaging modalities) had an equal number of over- and underestimated cases.
Most literature reports,
however,
state that DM underestimates the true size of lesions,
especially spiculated lesions,
due to difficulties in assessing their margins5.
MRI,
consistent with other studies1,4,6 tends to overestimate true tumour size (in 16.9% of cases in our series),
but it proved to be the most accurate modality,
as it had the least standard deviation from the size at pathology and therefore showed the least dispersion in the sample9.
MRI and DBT can provide different types of information about suspicious breast lesions,
thanks to the use of contrast medium; MRI is able to provide information about vascularized lesion which DBT cannot provide.
The first clinical experiences,
however,
seem to indicate a favourable impact of DBT in the detection of breast lesions.
Tomographic reconstruction allows to isolate breast lesions which are more affected by tissue overlap and thus by overlying structures on conventional mammography.
Therefore,
DBT seems to be able to play a major role in breast imaging and,
according to our data,
even in the preoperative staging of breast cancer 13,14.
Despite the good results in terms of dimensional accuracy and breast lesion detection demonstrated by DBT,
compared to DM and US,
we have to consider some limitations of our study.
The retrospective review of the DM,
DBT,
US and MRI images and the lesion measurements were performed by the same two radiologists who,
although unaware of pathological size,
were not blinded to the individual methods,
so it cannot be excluded that one type of imaging may have partly influenced the other ones.
In conclusion,
although DBT is superior to DM in the evaluation of tumour size and our results confirm that MRI still remains the most accurate imaging technique in preoperative staging of breast cancer.