Artificial Intelligence Revolutionizes Breast Cancer Screening in Tennessee

Researchers from the University of Tennessee, Knoxville, have made significant progress in using artificial intelligence and machine learning to provide treatment plans to breast cancer patients more quickly. By analyzing pathology reports and other clinical records, the team has developed a system that determines the extent of breast cancer in the body, reducing the staging process from hours to mere minutes. This breakthrough has the potential to improve the lives of thousands of cancer patients in Tennessee, which has one of the highest rates of cancer incidence and mortality in the country.

Key Takeaways:

• The researchers used a five-step process to convert scanned pathology reports into a format that machines can interpret, including pre-processing to remove visual noise and employing pretrained deep learning models to recognize characters and words.
• The system requires about 14 years to manually diagnose cancer stages from 300,000 pages of reports, but can do so in under one hour using the developed algorithms.
• The tool is designed to support medical professionals, not replace them, and provides flexibility to review and validate computer-generated results.
• The technology has been proven to be more accurate than human staging, with one in 10 diagnoses containing pathological staging errors and one in five containing clinical staging errors.
• The researchers are exploring how artificial intelligence can help predict cancer recurrence and are considering licensing and commercializing the new technology.
• The collaboration between the University of Tennessee and UT Medical Center has shown that interdisciplinary leadership can lead to breakthroughs in solving complex problems, and the potential impact of this work for cancer patients could be considerable.

Statistics:

• 14 years: the estimated time it takes to manually diagnose cancer stages from 300,000 pages of reports
• 1 hour: the time it takes for the developed algorithms to diagnose cancer stages from 300,000 pages of reports
• 1 in 10: the ratio of pathological staging errors
• 1 in 5: the ratio of clinical staging errors
• 300,000: the number of pathology reports used in the research
• 10%: the proportion of diagnoses containing pathological staging errors
• 20%: the proportion of diagnoses containing clinical staging errors

Sources:

• University of Tennessee: Xueping Li and Bing Yao's research on artificial intelligence in breast cancer screening
• American Joint Committee on Cancer (AJCC): staging rules for cancer
• Tom Berg, assistant professor in the College of Nursing
• Brad Day, associate vice chancellor for research innovation initiatives at UT
• John Bell, former director of the UT Medical Center Cancer Institute and a professor of surgery