Are you interested in combining AI with digital pathology to better understand cancer-related cell changes? A new model might soon be available that could help.

NCI-funded researchers are testing a model to detect extreme “nuclear wrinkling”—a unique cellular feature that they discovered is a hallmark of cancer.

According to the authors, some nuclear wrinkling and folding in human tissues is normal. However, by using AI to assess the degree of nuclear wrinkling, the researchers found that extreme wrinkling is a measurable and potentially useful biomarker for diagnosing cancer. And their model worked for nuclei in a variety of tissue types.

As noted by corresponding author, Dr. Tanmay Lele, of Texas A&M University, “Our model helps to assess the degree of nuclear wrinkling, offering a quantifiable biomarker for cancer. Physical characteristics like these could offer further insight into the diagnosis and prognosis of cancer and apply to multiple cancer types, including head and neck, skin, breast, and thyroid cancers.”

He added, “Although we used immunofluorescence staining for the nuclear lamina in our study, immunohistochemistry, which is a more common approach for diagnostic pathology, may be a possible alternative.”

NCI’s Division of Cancer Biology funded this work through the Physical Sciences – Oncology Network.