BREAST ULTRASOUND COMPUTER-AIDED LESION ASSESSMENT BASED ON BI-RADS

The National Cancer Institute estimates that approximately 1,700,000 women undergo breast biopsies (surgical or needle) in the U.S. each year. Approximately 80% of tumors biopsied are benign, 20% are malignant. Surgical biopsies--the most common--cost between $2,500 and $5,000 while needle biopsies cost from $750 to $1,200. Patients experience both physical and emotional effects when undergoing biopsy procedures and internal scarring may be problematic since it complicates interpretation of future mammograms. Until fairly recently, ultrasound in the U.S. has been used only to distinguish cystic from solid breast masses and to guide needle biopsies. A number of positive studies in Europe, Asia and the U.S. indicate that high-quality ultrasound can provide radiologists with a high degree of confidence in differentiating many benign from malignant or suspicious lesions detected by mammography. Results suggest that ultrasound could help reduce the number of biopsies of benign masses by 40% with a cost savings of as much as $1 billion per year in the U.S.

Work to improve the accuracy of diagnostic breast ultrasound has led to the development by American College of Radiologists of a well-defined system for scoring the Level of Suspicion (LOS) based on parameters describing the ultrasound appearance of breast lesions (known as BI-RADS^®). BI-RADS outlines six different possible findings, ranging from a Category 0 to a Category 5. The American Institute of Ultrasound in Medicine launched in 1998 a program to accredit the breast ultrasound practices of radiologists and sonographers. These steps by professional organizations are designed to improve quality and uniformity.

We developed a sophisticated software tool for imaging applications that targets breast ultrasound and uses BI-RADS lexicon and categories to quantify the computerized BI-RADS Assessment also known as LOS score. Radiologists report results with the BI-RADS classification. The developed software system can provide practitioners with a high degree of confidence to differentiate benign lesions from multitude of benign or/and suspicious lesions and also generate computerized score for Level of Suspicion to cancer.

The specific guidelines for differentiation of breast lesions are shown in Table 1 while the LOS score is assigned based on biopsy proven pathology of the most similar cases in the reference database of lesions with known findings.

Table 1.

Criteria Associated with Benign Lesions

Criteria Associated with Malignant Lesions

Spherical/ovoid/lobulated

Irregular shape

Linear margin

Poorly defined margin

Homogeneous texture

Central shadowing

Isoechoic/anechoic

Distorted architecture

Edge shadow

Calcifications

Parallel to the skin

Skin thickening

Distal enhancement

Dilated duct/mobile

Breast Imaging Reporting and Data System (BIRADS) Categories

(American College of Radiology Copyright © 2000-04)

Abbreviated version:

(click here to download full version directly from ACR web site)

Category 0 No findings:

Lesion is absent

Category 1 Negative:

There is nothing to comment on. The breasts are symmetrical, no masses or suspicious calcifications are present.

Category 2 Benign Finding:

Involuting, calcified fibroadenomas, multiple secretory calcifications, fat containing lesions such as oil cysts, lipomas, galactoceles, and mixed density hemartomas all have characteristic appearances, and may be labeled with confidence. Describe intramammary lymph nodes, implants, etc., while still concluding that there is no evidence of malignancy.

Category 3 Probably Benign Finding :

Should have a very high probability of being benign. Not expected to change over the follow-up interval, but the radiologist would prefer to establish its stability. Data are becoming available that shed light on the efficacy of short interval follow-up. These will likely undergo future modification as more data accrue as to the validity of an approach, the interval required, and the type of findings that should be followed.

Category 4 Suspicious Abnormality:

These are lesions that do not have the characteristic morphologies of breast cancer but have a definite probability of being malignant. The radiologist has sufficient concern to urge a biopsy. If possible, the relevant probabilities should be cited so that the patient and her physician can make the decision on the ultimate course of action.

Category 5 Highly Suggestive of Malignancy:

These lesions have a high probability of being cancer.