Pinpointing poor outcomes
In the current study, Dr. Schabath and colleagues used publicly available data and LDCT images from the NLST to generate radiomic features from screen detected, incidentally-diagnosed lung cancers. Radiomic features describing size, shape, volume, and textural characteristics were then calculated from both the intratumoral and peritumoral regions.
Patients were divided into training and test cohorts, and an external cohort of non-screen-detected lung cancer patients was used for further validation. There were no statistically significant differences between training and test cohorts for most demographics, including age, sex, smoking status, number of pack-years smoked, treatment, stage, and baseline screening result. However, self-reported chronic obstructive pulmonary disease (COPD) was significantly higher in the test cohort compared with the training group (16% vs. 7%; P = .02).
A total of 91 stable and reproducible radiomics features (peritumoral and intratumoral) were identified and 40 (26 peritumoral and 14 intratumoral) were significantly associated with overall survival in the training cohort. The features were subsequently narrowed to four, and backward elimination analyses identified a single model. Patients were then stratified into three risk-groups: low risk, intermediate risk, and high risk.
According to their model, the high-risk group had worse overall survival (hazard ratio, 9.91; 25% 2.5-year and 0% 5-year OS) as compared with the low-risk group (HR, 1.00; 93% 2.5-year and 78% 5-year OS).
The final model was validated in the test group and then replicated in the non–screen-detected patients with adenocarcinoma patients. Since the disease stage differed significantly across the risk groups, the model was stratified by stage and the authors found “compelling” results among early-stage patients, who generally have good outcomes. In this subset, the high-risk group was associated with a worse overall survival (HR, 2.63; 56% 2.5-year and 42% 5-year OS) vs. the low-risk group (HR, 1.00; 75% 2.5-year and 75% 5-year OS).
“We have ongoing studies to determine if these results are consistent in the real-world setting of lung cancer screening across multiple centers,” said Dr. Schabath. “If the NELSON, LUSI, or MILD trial data become publicly available, we will certainly pursue validating our results in those clinical trials.”
The study was funded by the National Cancer Institute. Dr. Schabath and Dr. Madabhushi have disclosed no relevant financial relationships.
This article first appeared on Medscape.com.