الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 169 |  |  |
| | | from | 169 | | |
Abstract
This study was performed to propose a diagnostic imaging approach for the evaluation of mammographic non mass findings.
It was conducted on 180 female patients proved to have non mass findings on FFDM, including asymmetry (n=104), microcalcifications (n=70), architectural distortion (n=35) and associated features (n=63). Associated features were noted in conjunction with the other non-mass findings but not the predominating features.
All patients underwent FFDM, digital breast tomosynthesis and ultrasound taking histopathological correlation as gold standard diagnosis. MRI and CESM were done in selected cases.
Their ages ranged from twenty-six to sixty-nine years, mean age of 51 years. Young patients were referred for mammography due to suspicious sonographic features or strong positive family history, as suspicious calcifications should be excluded.
Cases with breast masses evident in 2D were excluded from our study as we focused on non-mass findings and breast masses in 2D would affect the initial BIRADS of the breast, pregnant females were excluded for radiation hazards as well as breast implant as proper visualization of breast tissue will need additional implant displacement (ID) views, also cases with contraindications to contrast material and MRI were excluded.
Among the 180 cases, FFDM suggested 209 mammographic non mass findings. Asymmetry showed the highest prevalence of 104 lesions, with cancer yield of 44 malignant lesions (42.3%), While architectural distortion was the least incident finding in 30 lesions (16.6 %) with cancer yield in 17 cases. Microcalcifications were detected in 70 cases, 43/70 (61.4%) yielded malignant diagnosis while the remaining lesions were benign.
Based on final diagnosis, FFDM correctly diagnosed 85/ 96 benign lesions as true negatives, 54/84 malignant lesions considered as true positives. The overall “sensitivity and specificity” for non-mass findings were 64% and 88.5%, while “PPV, NPV and accuracy” were 83.1%, 73.9% and 77.2%.
Digital breast tomosynthesis was done for all cases and this increased the count of the lesions into 234; Seventeen asymmetries were overestimated by FFDM due to tissue overlap and downgraded their BIRADS into 2. While five distortions were successfully detected, that were not seen in FFDM and were counted as false negatives for mammography.
Two radiologists evaluated non mass findings and assigned a BI-RADS category to each lesion according to BI-RADS lexicon 2013 classification using FFDM, DBT alone and combined FFDM and DBT. The overall results showed improved diagnostic performance of DBT in the evaluation of mammographic non mass findings with estimated “sensitivity and specificity” into 96% and 93%, also the “PPV, NPV and accuracy” to 92%, 97% and 94 % respectively, due to better depiction of distortions, masses, excluding the false lesions imposed by tissue overlap and asymmetries, compared to FFDM with “sensitivity, sensitivity, PPV, NPV and accuracy” of 64%, 88.5%, 83.1%, 73.9% and 77.2%.
Combined FFDM and DBT further improved the diagnostic performance for cancer breast detection, with “sensitivity and specificity” reaching up to 96%, 95% with “PPV, NPV and accuracy” up to 94%, 97% and 95.6%.
The current study revealed DBT showed significant changes of BIRADS category and better performance compared to FFDM.
Ultrasound was done for all cases and was able to detect 221 lesions, It successfully diagnosed 90/96 benign lesions and 81/84 malignant lesions. The “sensitivity”of ultrasound for the non-mass findings was 96%, “specificity” was 94%, “PPV” was 93%, “NPV” was 97% and “accuracy” was 95%. US couldn’t visualize microcalcifications especially benign types and subtle distortion detected in DBT.
It was also suggested that cases of questionable subtle architectural distortion with no sonographic or MR imaging correlate, In low-risk patients without relevant surgical histories, it is reasonable to consider these lesions probably benign, with (BI-RADS) final assessment of 3 and recommended for short-interval follow-up.
MRI and contrast enhanced mammography were done in 39 cases which showed non conclusive or discordant imaging data as problem solving.
MRI/CESM provides both anatomic and functional information of breast lesions and should be reserved for cases with non-conclusive sono mammographic data.