• btc = $99 711.00 220.18 (0.22 %)

  • eth = $3 980.01 -6.49 (-0.16 %)

  • ton = $6.78 -0.10 (-1.42 %)

  • btc = $99 711.00 220.18 (0.22 %)

  • eth = $3 980.01 -6.49 (-0.16 %)

  • ton = $6.78 -0.10 (-1.42 %)

14 Jan, 2024
1 min time to read

PRISM neural network surpasses current screening standards, identifying pancreatic cancer cases 35% more effectively.

MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) has introduced the "PRISM" neural network, featuring two machine learning models designed to enhance the early detection of pancreatic cancer.

Utilizing diverse electronic health records from over 5 million patients, PRISM surpasses current screening criteria, identifying pancreatic ductal adenocarcinoma (PDAC) cases 35% more frequently.

Traditional PDAC screening catches only about 10% of cases, making PRISM a significant advancement. The neural network analyzes patient data, including demographics, diagnoses, medications, and lab results, predicting cancer probability. The six-year development of PRISM addresses the urgent need for early detection of pancreatic cancer, as 80% of patients are diagnosed too late.

While currently limited to MIT labs and select U.S. patients, scaling efforts involve diversifying datasets and exploring global health profiles for broader accessibility.

MIT is not new to artificial intelligence-based cancer risk prediction: they have previously developed models to predict breast cancer risk based on mammogram recordings. The success of these AI models relies on diverse datasets, demonstrating the potential for improved cancer diagnosis across various races and populations.