ETD system

Electronic theses and dissertations repository


Tesi etd-06212017-101205

Thesis type
Tesi di specializzazione (5 anni)
Molecular breast imaging: correlation with mammographic, sonographic and histopathological features of suspicious breast lesions.
Corso di studi
relatore Prof. Caramella, Davide
correlatore Dott.ssa Marini, Carolina
Parole chiave
  • ultrasound
  • sensitivity
  • molecular breast imaging
  • digital mammography
  • breast cancer
Data inizio appello
Riassunto analitico
The current guidelines suggest combined imaging modalities including digital mammography (DM), ultrasound (US) and magnetic resonance imaging (MRI) for cancer detection and staging. Magnetic resonance imaging (MRI) represents the most sensistive technique to evaluate suspicious breast lesions and to identify multifocal, multicentric and contralateral cancer, however many conditions (obesity, allergies, pacemakers, metallic devices and claustrophoby) may hinder its execution. In this scenario, radionuclide imaging with 99mTc-Sestamibi Molecular Breast Imaging (MBI) may represent an adjunct modality to radiologic traditional imaging, in particular when MRI cannot be performed or in selected cases. The purpose of this study is to compare the diagnostic performance of MBI compared to DM and DM+US in a retrospective work.
We retrospectively enrolled 50 women (mean age 63.5 years ±11.7) between May 2015 and May 2017, with 70 suspicious breast lesions identified by clinical examination, DM, US and biopsy (CNB). They underwent MBI followed by pathologic analysis from surgical specimens or from CNB. Histopathological analyses of tumoral lesions included nuclear grading and molecular subtype evaluation. DM/US findings were categorized according to BI-RADS® and considered as positive with a score ≥R4b/U4b. Mammographic breast density was visually estimated. MBI was performed 10 minutes after iv injection of 296 MBq of 99mTc-MIBI by using cameras in semiconductor cadmium zinc telluride (CZT) for single photon planar imaging of the immobilized breast. MBI was interpreted in conjunction with DM+US and qualitatively evaluated as positive (focal moderate or marked increased uptake) or negative (completely negative or mild uptake). Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy were calculated in order to assess the diagnostic capacity of MBI, DM and DM+US compared to histological examination, by using SPSS 24®.

Pathological correlation of the 70 suspicious breast lesions demonstrated 59 malignant foci of tumor:
38 no special type , 14 ductal carcinoma in situ, 5 invasive lobular carcinoma, 1 invasive tubulo-lobular carcinoma and 1 mucinous carcinoma. Molecular subtypes of invasive lesions were: 18 luminal A, 18 luminal B, 3 HER 2 and 1 basal-like (hormonal data not available for 5 tumors). Eleven benign lesions were detected: 1 lobular carcinoma in situ, 2 papillomas, 2 fibroadenomas, 6 adenosis and/or fibrosis.
MBI demonstrates the highest sensitivity (86.4% vs 83.1% by DM+US and 67.8% by DM) but the lowest specificity (36.4% vs 54.6% by DM+US and 72.7% by DM). MBI sensitivity is superior or equal to DM and DM+US in all categories of in situ/invasive, molecular subtypes, grading, breast density and lesion size (dichotomized as < or >1 cm). On the contrary, MBI sensitivity is inferior in premenopausal women and superior in postmenopausal women compared both to DM and DM+US. Furthermore, our results show that MBI is influenced by menopausal status (p=0.029) while it is not correlated to breast density and lesion size (p=0.859 and p=0.807, respectively).

Our initial experience with MBI supports the use of this modality in addition to DM/US, especially when MRI may not be performed. Currently, MBI remains a second choice compared to MRI due to its limitations of exposing women to a certain dose of radiation and lacking detailed spatial definition. However, our results about MBI independence from breast density and lesion size may represent a valid starting point to promote further larger studies.