A newly published peer-reviewed study of the CIZ1B-FibA lung cancer biomarker has reported 94% sensitivity in a cohort containing stage I and stage II cancers, adding new evidence behind a blood test Cizzle Bio is positioning for the evaluation of patients with suspicious pulmonary nodules. In the relevant study set, the assay correctly identified 17 of 18 patients with early-stage lung cancer and produced a negative predictive value of 97%. The result is encouraging for early-detection research, but the paper simultaneously identified variability between control samples collected in different countries and concluded that additional validation of analytical and biological variables is required.
That caution makes the study more interesting than a conventional promotional accuracy claim. The researchers evaluated 143 plasma samples distributed across three independent sample sets and found area-under-the-curve values of 0.82, 0.81 and 0.86. Although these results demonstrate discrimination between lung cancer and control samples, differences in baseline measurements associated with site-specific plasma processing showed that sample handling could materially influence the assay, an important issue for any blood test intended eventually to function consistently across many laboratories.
What exactly does the CIZ1B blood test measure?
CIZ1B is an alternatively spliced form associated with the CIZ1 protein, and the assay described in The Journal of Applied Laboratory Medicine measures a complex of CIZ1B with fibrinogen alpha chain known as CIZ1B-FibA. Researchers developed a sandwich ELISA using a monoclonal antibody designed to recognize the CIZ1B-specific exon junction and another antibody targeting fibrinogen. Only small quantities of patient plasma are required for the test.
The biological proposition is that CIZ1B-FibA concentrations can provide information associated with lung cancer, including early disease. Cizzle Bio has licensed the biomarker for commercial development in the United States, Canada and the Caribbean and has already begun limited commercial availability for healthcare providers in the Dallas-Fort Worth and San Antonio areas. The newly published research examines the same underlying ELISA-based biomarker platform supporting that commercial program.
How strong was the result in early-stage lung cancer?
The third sample set contained 57 specimens and produced the strongest overall discrimination, with an ROC area under the curve of 0.86. All 18 lung-cancer patients within that particular cohort had stage I or stage II disease, and the assay achieved 94% sensitivity using the set-specific threshold. The negative predictive value reached 97%, meaning that within that study population, 97% of participants with a negative result did not have lung cancer.
These numbers are potentially valuable because detecting lung cancer while it remains localized can dramatically alter prognosis. However, sensitivity and negative predictive value depend on the population being tested and the disease prevalence within that population. A 97% NPV obtained in a controlled study cohort therefore cannot automatically be assumed when the assay is used across a broader screening population or among patients with very different pretest risks.
Could CIZ1B help doctors decide what to do with an indeterminate lung nodule?
This may be one of the most plausible applications being explored. Low-dose computed tomography can identify small pulmonary nodules in people undergoing lung-cancer screening, but many nodules are not cancerous. Physicians then need to determine whether to monitor the nodule with repeat CT, obtain more advanced imaging, perform a biopsy or proceed to another diagnostic intervention.
A sufficiently validated blood biomarker could add another piece of evidence to that decision. A high-risk molecular signal might strengthen the case for faster investigation, while a reassuring result could potentially help avoid unnecessary invasive procedures in selected patients. Cizzle Bio specifically identifies indeterminate pulmonary nodules as a potential use case, but the current publication does not establish that CIZ1B should independently determine whether a patient undergoes biopsy or surveillance.
What was the biggest limitation identified by the researchers?
Pre-analytical variability may be the most consequential finding for future commercialization. Control samples used across the three independent datasets were collected in three countries and processed using site-specific plasma-isolation methods. Researchers found baseline differences between those controls, indicating that sample collection or processing variables could influence measured CIZ1B-FibA concentrations.
This is a classic laboratory-medicine challenge. A biomarker can discriminate patients effectively in one carefully handled dataset but become less reliable if temperature, anticoagulant choice, storage time, centrifugation or other pre-analytical steps change across real-world laboratories. The paper explicitly calls for additional investigation of these variables and further analytical validation, making standardization an important part of the CIZ1B development story rather than a minor technical detail.
Why does the sample size still matter despite the 94% sensitivity result?
Because the headline percentage is based on a relatively small number of cancer cases. With 18 early-stage patients, missing one patient produces roughly 94% sensitivity; missing only a few more would change the percentage substantially. Large prospective studies involving hundreds or thousands of appropriately selected patients are therefore necessary before clinicians can understand performance with greater precision.
The same principle applies to specificity and negative predictive value. Diagnostic tests ultimately need to work across smokers and nonsmokers, different ages, benign inflammatory lung disease, multiple nodule sizes and the broad biological diversity seen in actual clinical practice. Cizzle Bio says validation work is continuing through Moffitt Cancer Center and other academic and National Cancer Institute-designated centers.
Could this test compete with CT-based lung cancer screening?
It is better viewed as potentially complementary than as a replacement based on the current evidence.
Low-dose CT has randomized evidence demonstrating a reduction in lung-cancer mortality among appropriately selected high-risk patients, and it can directly visualize nodules and other thoracic abnormalities. A biomarker blood test cannot provide anatomical information or establish where a suspicious lesion is located. The potential advantage of CIZ1B is that molecular information could help interpret what imaging discovers or potentially identify patients who deserve more intensive investigation.
That distinction also protects against an easy but misleading headline narrative. A blood test achieving high sensitivity in one early-stage study cohort does not mean people can skip recommended CT screening. The more realistic near-term role would involve adding biological information to existing diagnostic pathways.
What should the diagnostics industry watch next?
The most important milestones are larger clinical-validation datasets and evidence that assay performance remains stable when samples are collected and processed across different healthcare systems. The Moffitt Cancer Center study is examining patients with identified lung nodules, while Cizzle Bio lists additional research activity with institutions including UT Health San Antonio MD Anderson Cancer Center, City of Hope, UCLA Health and Duke Health. Those studies could provide substantially more clinically representative evidence than the 143-sample laboratory dataset reported here.
CIZ1B therefore deserves attention without skipping steps. Detecting 17 of 18 stage I or II cancers in one dataset is exactly the type of signal that can justify expanded validation. The same publication also provides the reason not to declare victory too early: sample-processing differences altered baseline values, and the researchers themselves conclude that analytical and biological variables still need further study before the biomarker’s full clinical role can be defined.
