AJR: Kinetic variable for identifying MRI-detected breast lesions discovered

Computer-aided kinetic information can help significantly in distinguishing benign from malignant suspicious breast lesions on MRI, according to a study in the September issue of the American Journal of Roentgenology. These findings support the American College of Radiology (ACR) BI-RADS Breast MRI Lexicon recommendation to report the “worst looking” kinetic curve.

Computer-aided evaluation programs for breast MRI provide automated lesion kinetic information. According to the authors, the computer-aided evaluation kinetic parameters that best predict malignancy have not been established. The investigators compared three computer-aided evaluation kinetic features of suspicious breast MRI lesions to determine associations with benign or malignant outcomes.

In the study, performed at the University of Washington Medical Center in Seattle, researchers analyzed and compared the computer-aided evaluation variables of 125 suspicious breast lesions. Three different kinetic curves (washout, plateau and persistent), were compared along with lesion morphology (size and shape).

"We wanted to clarify which of the many variables that reflect kinetics were most predictive of malignancy,” said lead author Constance Lehman, MD. “We found overlap in kinetic patterns across benign and malignant lesions, but we did determine that the ‘most suspicious’ curve type, washout, was useful in separating benign from malignant lesions," Lehman said.

The investigators identified 125 lesions (42 malignant, 83 benign) in the analysis set. There were no significant differences in initial peak enhancement (p = 0.28) or delayed kinetics categorized by largest percentage enhancement types (p = 0.39) between benign and malignant lesions. There was a significant difference in delayed kinetics categorized by the most suspicious enhancement types (p = 0.005).

"Of lesions with the most suspicious curve type (any washout), 45.7 percent were malignant compared with 20 percent with plateau and 13.3 percent with entirely persistent enhancement," Lehman said.

"We continue to study the specific features on MRI that are most predictive of breast cancer. We know that the morphology of the lesion is extremely important, but our study also supports the use of kinetic features in lesion assessment. The "most suspicious" curve, washout, does seem to help distinguish benign from malignant lesions," she said.

“In breast MRI, it is important to know which variables are most important for predicting malignancy because they help us in determining whether or not a lesion needs to be biopsied or not," she said.

Despite its single-site limitations, the authors wrote that their study showed that this particular kinetic measure provided predictive information when considered in conjunction with lesion morphology.

“Because computer-assisted evaluation for breast MRI is increasingly used in clinical practice and provides automated and standardized measurements of kinetics, further studies would be useful to clarify the key computer-aided evaluation-enhancement variables most appropriate to guide clinical decision making,” the researchers wrote.

Around the web

As debate simmers over how best to regulate AI, experts continue to offer guidance on where to start, how to proceed and what to emphasize. A new resource models its recommendations on what its authors call the “SETO Loop.”

FDA Commissioner Robert Califf, MD, said the clinical community needs to combat health misinformation at a grassroots level. He warned that patients are immersed in a "sea of misinformation without a compass."

With generative AI coming into its own, AI regulators must avoid relying too much on principles of risk management—and not enough on those of uncertainty management.

Trimed Popup
Trimed Popup