Groundbreaking Blood Test for Lung Cancer Prediction

In a recent publication in the journal Cell, over 80 researchers from four continents unveiled a promising 14-protein blood test that could potentially foresee and avert lung cancer diagnoses more than five years before they typically occur. The current focus in cancer research largely revolves around treatment and early detection, with prevention being somewhat overlooked.

This impressive study leveraged machine learning combined with high-throughput proteomics, validating the 14-protein signature across eight different cohorts, including individuals with lung cancer who had never smoked. Researchers extensively investigated factors such as air pollution, the biology of lung organoids, and related healthy tissue adjacent to tumors. Coverage of this research even made it to the front page of the New York Times, albeit with a focus on its broader implications rather than its unique details.

In this discussion, I’ll summarize the findings and their potential impact on lung cancer prevention. It’s a vast area to cover, so I’ll stick to the key highlights.

Let’s take a brief look back at the groundwork laid by the CANTOS trial conducted in 2017. This large-scale randomized clinical trial tested an anti-inflammatory drug (canakinumab) among more than 10,000 participants with a history of heart attacks. Although the results were not particularly impressive for cardiovascular issues, what emerged unexpectedly was a significant reduction in lung cancer diagnoses over a five-year follow-up period. At the highest dosage, there was nearly an 80% reduction in fatal lung cancer cases. However, the challenge remained in identifying who might actually benefit from the treatment, as the number needed to treat (NNT) for one case of lung cancer prevention exceeded 1,000.

The study involved high-throughput plasma proteomics among over 48,000 participants from the UK Biobank and sought to understand lung cancer-related proteins. By analyzing nearly 3,000 proteins, researchers identified 14 that stood out, primarily linked to signs of lung cell distress, including inflammation and surfactant production. On average, these proteins appeared 5.6 years before a lung cancer diagnosis among UK Biobank participants. This 14-protein signature was confirmed in a Taiwanese cohort, most of whom never smoked, and it outperformed previous models based only on demographics and smoking history.

Interestingly, analysis from the TRACERx clinical study revealed that these proteins weren’t directly produced by tumor cells, with no correlation seen in advanced lung cancer cases. Instead, they seemed to emanate from healthy cells detecting stress from precancerous conditions, given the right mutations and environmental triggers.

Particular attention was given to the role of air pollution. Exposure to particular particulate matter activated macrophages that released interleukin-1β, which in turn spurred the production of the key 14 proteins. This interplay between environmental factors and genetic mutations led to instances of heightened protein release. Notably, in the CANTOS trial, subjects showing the 14-protein signature had over double the risk of developing lung cancer, but they also displayed a significant benefit from the interleukin-1β antibody treatment, reducing their lung cancer risk considerably.

Now, despite this pivotal study being retrospective, it does lay the groundwork for a prospective clinical trial to definitively assess the potential of reducing lung cancer cases by half through the 14-protein biomarkers and antibody treatment. Looking ahead, if this course of prevention proves successful, it could represent a seminal moment in cancer prevention strategies.

The findings highlight both the importance of established trials like CANTOS and the critical clinical resources that enabled this groundbreaking work. This combination of proteomic analysis and patient data sets a new pathway for potential health interventions and could aid in significant cancer prevention efforts moving forward.