At the Advances in Decision Analysis sponsored by the INFORMS Decision Analysis Society, a conference that took place June 16-18 at Georgetown University, a new approach to examining mammograms was presented in which a woman’s health risk profile would be taken into account, therefore reducing the number of cancer instances missed and also cutting the number of false positives.
The findings were presented by Mehmet U.S. Ayvaci of the University of Texas Dallas and discussed the role of risk profiling in the interpretation of mammograms.
In order to reduce false negatives by 3.7% and reduce false positives by 3.23, the researchers concluded that the patient’s risk profile information for breast cancer should be provided to the radiologist at the most advantageous time when examining the mammogram, together with statistical weighting based on profile risk. This would help alert women whose early cancer would have gone undiagnosed at an early stage.
Family history, reproductive history, age, and ethnicity, as well as others, are included as risk factors.
Providing risk profile information about women being screened for cancer may bias radiologists and the paper examines these questions in full. It also examines whether this bias actually helps make readings more accurate. In the past, available clinical evidence has been inconclusive on the use of profile information when interpreting mammograms. Profile information helps radiologists make better decisions and should be employed when reading mammograms, according to one position. A contradictory position says that profile information may bias the radiologists, however, it is unclear if bias causes harm at all.
Profile information and potential bias in mammography interpretation was explored by the authors and used a decision science technique called linear opinion pooling, which assigns weights to better aggregate probability estimates.
Three groups were analyzed on decision performance: 1.) A mammogram-only reading, with no profile information about the patient, 2.) an unbiased reading, in which radiologists consult the risk profile after examining the mammogram and 3.) biased or “influenced” readings, in which radiologists consult a women’s risk profile as they examine the mammogram. Then the conditions in which profile information could help improve biopsy decisions was examined.
The use of profile information with an appropriate weight could reduce the false positives and the number of missed cancers were revealed by numerical analysis using a clinical dataset from the Breast Cancer Surveillance Consortium, especially in comparison to cases where profile information had not been examined at all.
Breast cancer is the second most non-skin cancer. In 2013, approximately 232,000 breast cancer diagnoses were made and about 39,000 women died from the disease.
This comment has been removed by a blog administrator.
ReplyDelete