Identification of Plasma LGALS3BP via Proteomics and Development of a Non-invasive Diagnostic Model for Early-Stage Lung Adenocarcinoma

Lung cancer remains the leading cause of cancer-related death worldwide, with a specific type called lung adenocarcinoma being the most common form. A major hurdle in treating this disease is the difficulty of telling the difference between a harmless lump in the lung and a dangerous early-stage tumor. Doctors often rely on low-dose CT scans to find these lumps, known as pulmonary nodules, but the scans cannot always say for certain if a nodule is benign or malignant. This uncertainty creates a difficult situation for patients: many people with harmless nodules undergo unnecessary invasive procedures like biopsies or surgery because the scan looks suspicious, while some early cancers are missed or diagnosed too late, allowing the disease to spread. Current blood tests for cancer markers are not sensitive enough to solve this problem on their own, leaving doctors without a reliable, non-invasive way to make the right call before cutting into a patient.

Researchers at the First Affiliated Hospital of Kunming Medical University set out to find a better way to distinguish these nodules by looking directly at the proteins floating in the blood. They began by examining the actual tissue of tumors and comparing it to healthy lung tissue, using a technique that measures thousands of proteins at once without needing to label them with chemicals. This initial screen revealed that a specific protein, known as LGALS3BP, was significantly more abundant in the cancerous tissue than in healthy tissue. The team then looked for this same protein in the blood plasma of patients and found it was also elevated in those with lung adenocarcinoma compared to those with benign nodules. To ensure this finding was robust, they validated the results in separate groups of patients, confirming that the levels of this protein in the blood matched the levels of its genetic instructions, or mRNA, found in the same samples. This consistency suggested that measuring the genetic instructions in the blood could serve as a reliable and cost-effective proxy for measuring the protein itself.

Building on this discovery, the researchers developed a new diagnostic tool that combines the level of this specific blood marker with other information doctors already collect, such as a patient's family history of cancer, the size and type of the nodule seen on a scan, and the levels of a traditional marker called CEA. They tested this combined model on a group of 138 patients to train the system and then checked its accuracy on a separate group of 47 patients. The results showed that this new model was far better at identifying cancer than looking at any single factor alone. In the training group, the model correctly identified the disease with an accuracy score of 0.893, a significant improvement over the standard clinical approach. When tested on the second group, it maintained strong performance with a score of 0.856. Most importantly, the model proved highly effective at identifying patients who truly had cancer, achieving a success rate of over 91 percent in the group where the risk was considered intermediate to high.

The study highlights that this approach offers a practical way to navigate the "gray zone" of lung nodule diagnosis, where the evidence is often unclear. By integrating a molecular signal from the blood with standard clinical observations, the model helps doctors stratify risk more precisely. This means that patients with a high probability of cancer can be directed toward necessary treatment more quickly, while those with a low probability might be spared from invasive procedures. However, the researchers noted that the model is not perfect for every type of nodule; it was less sensitive for a specific kind of early tumor that appears as a faint, hazy spot on a scan, known as a pure ground-glass nodule. In these cases, the model sometimes missed the cancer, suggesting that such patients still require careful monitoring rather than relying solely on this test.

Ultimately, this work identifies a new, promising biomarker in the blood that reflects the biological activity of early-stage lung cancer. The diagnostic model created from this discovery provides a concrete, data-driven method to support clinical decisions, potentially reducing the number of unnecessary surgeries while ensuring that real cancers are caught early. While the study was conducted at a single center and requires further testing in larger, diverse populations to confirm its broad applicability, it represents a significant step toward solving the persistent challenge of distinguishing harmless lung nodules from life-threatening ones without invasive surgery.