Photo Credit: Yacobchuk

Using preoperative chest CT scans, researchers identified novel preoperative biomarkers for lung cancer recurrence post-surgical resection.

Using information derived from preoperative CT chest scans, computer models successfully identified patients who did and did not, experience lung cancer recurrence after undergoing surgical resection, researchers reported in Cancers.

“A distinctive feature of this study is its reliance on chest CT scans, routinely obtained as part of standard care for [patients with lung cancer],” Xin Meng, PhD, and colleagues wrote. “This eliminates the need for additional radiation exposure or costly procedures, offering a practice and accessible method for recurrence prediction.”

Researchers studied lung cancer recurrence in patients after surgical resection, in an attempt to identify novel image biomarkers from preoperative chest CT scans.

“The goal is to improve the ability to predict the risk and time of recurrence in seemingly ‘cured’ patients,” they explained, “enabling personalized surveillance strategies to minimize lung cancer recurrence.”

Factors Associated With Survival

The study focused on data from 309 patients with non-small cell lung cancer who underwent lung resection.

Cox proportional hazards regression analysis, alongside machine learning methods for predicting recurrence, identified several factors that were significant determinants of recurrence risk: surgical procedure, TNM staging, lymph node involvement, body composition, and tumor characteristics. The factors were associated with local/regional and distant recurrence, recurrence-free survival, and overall survival.

“The interesting findings are that detailed body composition and pulmonary vasculature are associated with recurrence-free survival and overall survival,” the authors reported.

The Cox and machine learning models showed comparable predictive performance.

“The [area under the receiver operative characteristic curve] values for predicting recurrence within 2 years ranged from 0.676 to 0.737, while those for predicting recurrence within 5 years ranged from 0.738 to 0.770,” Dr. Meng and colleagues wrote.

Going forward, the team plans to validate the predictive models in more diverse and larger cohorts to enhance their generalizability and applicability.

“Furthermore, postoperative recurrence is a challenge in lung cancer and in other cancers, such as liver, colon, and esophageal cancer, which typically require CT imaging,” the authors wrote. “Therefore, the methodology presented in this study holds the potential for broader applicability, providing a valuable framework for studying recurrence across different cancer types.”