A total of 76 lesions in 37 women were observed on breast MRI examinations. Of the 76 lesions, 20 were malignant and 56 were benign based on MRI findings. Twenty-two of the benign lesions also had histopathologic diagnosis (3 granulomatous mastitis, 11 fibroadenoma, 3 adenosis, 2 papilloma, 1 intramammarian lymph node, 1 post-operative scar tissue and radiation necrosis (Tables 1 and 2).
Due to the analysis with Chi-square test, 55,4% of the benign lesions were displaying Type I, 33,9% Type II, and 10,7% Type III contrast-time curve, whereas 5% of the malignant lesions were displaying Type I, 50% Type II, and 45% Type III contrast-time curve. MR imaging findings were in harmony with histopathological results (p<0.01). Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MR imaging with 1 molar contrast agent in detecting malignant lesions were 45%, 89,3%, 60%, 82%, and 77,63%, respectively.
Table 1: Distribution of contrast–time curves in benign lesions
Histopathology | Type-I contrast –time curve | Type-II contrast –time curve | Type-III contrast–time curve | Total |
Fibroadenoma | 15 | 14 | 1 | 30 |
Adenosis | 6 | 1 | - | 7 |
Fat necrosis | 5 | - | - | 5 |
Mastitis | 1 | 1 | 1 | 3 |
Intramammarian lymph node | - | - | 3 | 3 |
Papilloma | - | 2 | 1 | 3 |
Scar tissue | 2 | - | - | 2 |
Fibroadenolipoma | 2 | - | - | 2 |
Infected cyst | - | 1 | - | 1 |
Total | 31 | 19 | 6 | 56 |
Table 2: Distribution of contrast–time curves in malignant lesions
Histopathology | Type-I contrast–time curve | Type-II contrast –time curve | Type-III contrast–time curve | Total |
Invasive ductal carcinoma | 1 | 10 | 8 | 19 |
Mucinous adenocarcinoma | | | 1 | 1 |
Total | 1 | 10 | 9 | 20 |
One malignant lesion (invasive ductal carcinoma) displayed Type-I contrast–time curve, which was accompanied by enhancement on the breast skin at the level of the lesion. Six benign lesions showed Type-III contrast–time curve (1 fibroadenoma, 1 granulomatous mastitis, 3 intramammarian lymph nodes, and 1 papilloma).
Of the lesions with Type-I contrast–time curve, 31 (96,9%) were benign, and 1 (3,1%) was malign. Of the lesions with Type-II contrast–time curve 19 (65,5%) were benign, 10 (34,5%) were malign. Of the lesions with Type-III contrast–time curve, 6 (40%) were benign, while 9 (60%) were malignant. There was significant correlation between the types of contrast–time curve and histopathologic results (p<0.05) (Graphic 1).
In terms of diagnosing benign pathology, Type-I contrast–time curve had a sensitivity of 55%, specificity of 95%, positive predictive value of 96%, negative predictive value of 56%, and accuracy rate of 65%.
In terms of diagnosing malignant pathology, Type-II contrast–time curve had a sensitivity of 50%, specificity of 66%, positive predictive value of 34%, negative predictive value of 21%, and accuracy rate of 61%, whereas these values for Type-III contrast–time curve were 45%, 89%, 60%, 18%, and 77%, respectively.
In terms of diagnosing benign pathology, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy rate of homogenous contrast enhancement were 54%, 75%, 86%, 63%, and 60%, respectively (p=0.059).
Positive predictive value of peripheral ring-like contrast enhancement to determine malignancy was found to be 41%.
Table 3: Histopathologic correlation of increased vascularity in the involved breast
Increased vascularity | Histopathologically malignant | Histopathologically benign |
present | 11 (91%) | 5 (20%) |
absent | 1 (8,3%) | 20 (80%) |
When increased vascularity was evaluated in terms of diagnosing malignancy, its sensitivity was 91%, specificity 80%, positive predictive value 68%, and negative predictive value 95% (p<0.05).
The mean increase in contrast-noise ratio (CNR) was significantly higher in malignant lesions (450%) than in benign lesions (60%) (p<0.05).