Related Experiment Videos
Breast lesions detected on MR imaging: features and positive predictive value
Laura Liberman1, Elizabeth A Morris, Melissa Joo-Young Lee
1Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10021, USA.
AJR. American Journal of Roentgenology
|June 22, 2002
Summary
This study analyzed breast lesions found on MRI, determining that 25% were cancerous. Key features like spiculated margins and irregular shapes in mass lesions, and segmental enhancement in non-mass lesions, showed the highest positive predictive value for malignancy.
Area of Science:
- Radiology
- Oncology
- Pathology
Background:
- Magnetic Resonance (MR) imaging is crucial for detecting breast lesions.
- Accurate characterization of these lesions is vital for timely diagnosis and treatment.
- Understanding the predictive value of specific MR imaging features can improve diagnostic accuracy.
Purpose of the Study:
- To analyze the features of breast lesions detected by MR imaging that were subsequently biopsied.
- To determine the positive predictive value (PPV) of these MR imaging features for malignancy.
Main Methods:
- Retrospective review of 100 solitary MR imaging-detected breast lesions requiring biopsy and surgical excision.
- Lesions were described using a standardized breast MR imaging lexicon.
- Histopathological findings were correlated with MR imaging features.
Main Results:
- Carcinoma was identified in 25% of lesions, with half being ductal carcinoma in situ (DCIS) and half infiltrating carcinoma.
- For mass lesions, spiculated margins (80% PPV), rim enhancement (40% PPV), and irregular shapes (32% PPV) had the highest PPV for carcinoma.
- For non-mass lesions, segmental (67% PPV) or clumped linear and ductal enhancement (31% PPV) were most predictive of carcinoma.
- Washout on kinetic assessment was more common in infiltrating carcinomas (70%) than DCIS (9%).
- Lesions categorized as highly suggestive of malignancy had a significantly higher rate of carcinoma (67%) compared to suspicious lesions (19%).
Conclusions:
- MR imaging detected breast lesions referred for biopsy have a 25% malignancy rate, with half being DCIS.
- Specific MR imaging features, including spiculated margins, rim enhancement, irregular shapes (for masses), and segmental or clumped enhancement (for non-masses), are strong predictors of malignancy.
- Final assessment categories based on MR imaging features significantly predict the likelihood of carcinoma, aiding in biopsy decisions.