Clinlabomics‐Enabled Blending Ensemble Learning for Low‐Cost Pan‐Cancer Detection and Classification Using Routine Clinical Laboratory Data

The early detection and accurate diagnosis of cancer is of great significance for cancer patients. Artificial intelligences rapid advancement has boosted clinlabomics, bridging clinical lab data with intelligent systems. Inspired by clinlabomics, a liquid biopsy method for the detection and identification of ten cancers is reported. By applying the blending ensemble learning method to clinical laboratory data, separate cancer detection and identification models are developed to construct an intelligence system, called clinlabomics assisted for cancer identification (CACI). The CACI is applied to 19 199 individuals, including 9,047 noncancer patients in the control group and 10 152 patients with 10 types of multiple cancers from Lanzhou University Second Hospital. In the external test cohort, the cancer detection model, based on 21 indices, exhibits a sensitivity of 90.39%, specificity of 82.41%, with an area under the curve (AUC) exceeding 0.9373 in distinguishing cancer from noncancer individuals. The cancer identification model, constructed using 34 top‐rank indices after feature extraction, achieves an overall accuracy of 72.57%. This method achieves low‐cost cancer detection and demonstrates significant potential in cancer diagnosis. The results provide valuable guidance for physicians, enabling them to select more accurate testing indicators and assess disease risks with greater precision.