AI in Mesothelioma Diagnosis
Artificial intelligence is helping doctors diagnose mesothelioma more accurately. This rare cancer is hard to detect because it mimics other diseases and shows complex patterns under the microscope. AI tools now analyse imaging scans and pathology slides to spot subtle signs that human eyes can miss.
Radiologists use AI on CT and PET/CT scans. These systems measure tumour volume automatically. They also distinguish malignant pleural mesothelioma from benign pleural disease. In some studies, deep learning models reached accuracy levels above 80 percent when combining imaging with basic clinical data such as age and sex. Therefore, AI supports earlier identification and better staging.
Pathologists benefit as well. Researchers have trained neural networks on thousands of whole-slide images. One self-supervised study created a histomorphological atlas from more than 3,400 slides. The resulting patterns predicted outcomes with a concordance index of 0.65 and classified subtypes with 88 percent AUC. Another model separated epithelioid from non-epithelioid mesothelioma with high accuracy, sometimes matching or exceeding expert pathologists.
AI also aids prognosis. Models analyse tissue patterns to estimate survival and predict response to treatment. This information helps multidisciplinary teams choose therapy more confidently. In addition, automated volume measurement on CT scans tracks how tumours change during chemotherapy, reducing the variability of manual readings.
Despite these advances, challenges remain. Mesothelioma is rare, so high-quality training data is limited. Models need careful validation across different hospitals and patient groups. Regulatory approval and integration into clinical workflows also take time. However, ongoing research continues to expand the available datasets and refine the algorithms.
Overall, AI does not replace doctors. It assists them by handling repetitive analysis and highlighting features that improve diagnostic confidence. As tools mature, they may shorten the path from first symptoms to confirmed diagnosis and more personalised care for mesothelioma patients.