A study has emerged indicating that pancreatic cancer surgery can allow the prediction of survival periods using artificial intelligence (AI). Cancers that develop in the pancreas, which secretes digestive enzymes, are often found in advanced stages, resulting in a high recurrence rate even after surgery. If the surgical prognosis can be assessed accurately by AI, it will assist in determining treatment directions.
Professor Park Joo-kyung of the Digestive Internal Medicine Department and Professor Han In-woong of the Department of Hepatobiliary Surgery at Samsung Medical Center's Cancer Hospital, along with Professor Jang Gi-taek of the Pathology Department, noted on the 30th that "the density of tumor-infiltrating lymphocytes analyzed by AI is related to the survival rate of pancreatic cancer patients." The research was published in the international journal "JAMA Surgery" on the 25th.
Tumor-infiltrating lymphocytes are a type of immune cell gathered around cancer cells and are a kind of white blood cell. They have the characteristic of attacking tumors based on immune response, allowing for the prediction of treatment prognosis through analysis. However, measuring tumor-infiltrating lymphocytes one by one takes considerable time for medical staff. Due to differences in measurement among observers, there were limitations to their practical use in medical settings.
The research team analyzed tumor-infiltrating lymphocytes from 304 pancreatic cancer surgery patients using AI. A higher lymphocyte density was correlated with relatively longer survival periods post-surgery. The median survival period for patients in the immune-active group with abundant tumor-infiltrating lymphocytes was 35.11 months, while the median survival period for patients in the immune-deficient group was only 11.6 months.
The research team utilized the immune phenotype analysis AI called "Lunit Scope IO" from the medical AI corporation Lunit. Professor Park Joo-kyung stated, "This study enabled a more accurate prediction of survival outcomes for pancreatic cancer patients," and added, "A new era has opened where AI can provide critical information for determining cancer treatment directions."
References
JAMA Surgery (2025), DOI: https://doi.org/10.1001/jamasurg.2025.1999