AI Predicts Prostate Cancer Spread from Medical Scans

This patent describes how artificial intelligence can analyze medical scans of a prostate tumor to automatically identify suspect regions and predict if the cancer will spread to other parts of the body.

What it covers

This patent details systems and methods that use artificial intelligence, specifically a convolutional neural network (CNN), to analyze medical images of the prostate. The system automatically identifies 'suspect regions' or 'hotspots' within images like PET, CT, or PET/CT scans. It then uses these identified regions to predict whether localized prostate cancer has spread or will spread, a process known as metastasis. For example, a doctor could upload a patient's prostate PET/CT scan, and the AI system would highlight areas of concern and provide a prediction on the likelihood of the cancer spreading.

What it doesn't cover

• —Does not cover predicting cancer spread from images of organs other than the prostate.
• —Does not cover using different types of AI models beyond convolutional neural networks for this specific prediction.
• —Does not cover predicting other cancer outcomes, such as tumor growth rate or response to treatment, only metastasis.
• —Does not cover using other imaging modalities like MRI or ultrasound for this specific prediction of prostate cancer metastasis.
• —Does not cover manual identification of suspect regions without AI assistance.

The clever bit

The novel aspect is the automated identification of specific 'suspect regions' within prostate images by a convolutional neural network, which is then directly used to predict the likelihood of cancer spreading, rather than just detecting the presence of a tumor.

Why it matters

Accurate and early prediction of cancer metastasis is critical for effective treatment planning and improving patient outcomes. This technology aims to provide clinicians with a powerful tool to assess the risk of prostate cancer spreading, potentially enabling more personalized and timely interventions. It represents a significant step towards integrating advanced AI capabilities into routine oncology diagnostics, enhancing the precision of medical prognoses.

Real-world examples

• 1.AI-powered diagnostic tools for prostate cancer staging
• 2.Clinical decision support systems for oncology
• 3.Medical imaging analysis software for predictive analytics