Revolutionary blood test improves lung cancer screening

In a recently published study, Cancer DetectionResearchers developed and validated a blood-based cell-free deoxyribonucleic acid (cfDNA) fragmentome test for lung cancer detection, with positive results followed by low-dose computed tomography (LDCT).

study: Clinical Validation of a Cell-Free DNA Fragment Assay to Facilitate Early Detection of Lung CancerImage credit: MMD Creative/Shutterstock.com

introduction

Lung cancer is a leading cause of death, making annual screening essential, yet chest LDCT is not widely available due to barriers such as lack of patient awareness, radiation concerns, and limited access.

Other challenges include inadequate recording of smoking histories, lack of clear practices, and insufficient specialist follow-up.

Blood-based lung malignancy screening tests, such as fragmentome technology, can increase screening rates by analyzing specific chromatin configurations in peripheral blood.

About the Research

In the DELFI-L101 study, researchers developed a hematological test that uses machine learning to analyze DNA fragments to identify individuals at risk for lung cancer. Individuals who test positive will receive LDCT.

Researchers enrolled 958 individuals aged 50 to 80 years with a smoking history of 20 or more pack years at 47 US sites beginning in March 2021. Eligibility requirements were similar to the 2015 National Health Interview Survey (NHIS) LDCT screening criteria.

Patients were excluded if they had received treatment for cancer within the past year, had a history of hematologic malignancy or myelodysplasia, had undergone organ tissue transplantation, had received a transfusion of blood products within 120 days of enrollment, were pregnant, or were participating in other trials.

The research team divided the study participants into three groups: A (lung cancer), B (non-cancer control group), and C (cancers other than lung cancer).

Disease stage was determined according to the criteria of the American Joint Committee on Cancer Cancer Staging Manual (AJCC).Alterations in cfDNA fragmentation patterns (fragmentome) in blood revealed genomic and chromatin signatures of lung cancer.

The researchers trained their classifier on 576 cases and controls and then validated it on an additional 382 cases and controls.

They used the whole genome sequence from the training dataset to evaluate the fragmentation of 504 highly mappable non-overlapping 5.0 MB sections, each containing 80,000 pieces and covering a genome size of 2.50 GB.

The team investigated genome-wide changes in Hi-C open (A compartment) and closed (B compartment) chromatin.

They created a classifier using principal component analysis (PCA) and logistic regression incorporating chromosome arm-level alterations, cfDNA fraction derived from the mitochondrial genome, and cfDNA fragment length distribution.

The researchers ran Monte Carlo simulations involving 15 million people under three scenarios.

• Baseline scenario: current practice of not performing hematological screening.
• Low scenario: Among those eligible for lung cancer screening but who do not receive LDCT, blood test uptake is 10% in the first year, increasing to 25% after 5 years.
• High Scenario: Blood test uptake in the same group is 20% in the first year and increases to 50% after five years.

result

The researchers observed a test specificity of 58%, sensitivity of 84%, and negative predictive value (NPV) of 99.8%. Applying the remainder to the screening group with a lung cancer prevalence of 0.7%, the number needed to screen (NNS) was 143 people.

Study validation detected 414 and 76 NNS-related negative and positive results from LDCT imaging, respectively, with a relative risk value of 5.5. The positive predictive value (PPV) was nearly twice that of the LDCT eligibility requirement alone.

cfDNA fragments from patients with lung squamous cell carcinoma (LUSC) consisted of elements resembling cfDNA profiles from non-cancer patients and elements resembling the A/B type compartments seen in LUSC tissue.

Individuals without cancer displayed cfDNA patterns that approximated lymphoblastic Hi-C findings. Within common locations, fragmentation between samples provided by individuals with and without cancer was similar.

Lung cancer patients There was increased representation of cell-free DNA for 1q, 3q, 5p, 8q, and 12p, and decreased levels of 1p, 3p, 4q, 5q, 10q, and 17p. cfDNA fragmentation differed from controls, revealing more compact chromatin in cfDNA from closed LUSC spaces, whereas the opposite effect was seen in lymphoblast reference regions.

At a cutoff of 0.2, 10-fold cross-validation with 10 repeats within the training population yielded an overall specificity of 50% and sensitivity of 75%, 90%, 96%, and 97% for stages I, II, III, and IV, respectively. Sensitivity was consistent across age groups, with younger individuals showing higher specificity. Using 2015 NHIS data, we found a sensitivity of 80% and a specificity of 58%.

From the “base” scenario (24,489 cases), lung cancer cases identified by screening increased to 63,523 (“low” scenario) and 100,346 (“high” scenario). In contrast, stage I cases increased by 4.80% and 9.70%, respectively, and stage IV diagnoses decreased by 4.20% and 8.70%, respectively.

In total, 4,720 deaths from lung malignancies could be avoided in the “base” scenario, 7,652 in the “low” scenario, and 14,264 in the “high” scenario. Using LDCT for screening could reduce the number of tests needed to identify lung cancer from 202 (“base” scenario) to 150 (“low” scenario) and 139 (“high” scenario).

Conclusion

Based on our findings, the DNA fragmentome assay provides a new, accurate, and affordable blood-based tool for early assessment of lung cancer with LDCT follow-up.

The test has the potential to contribute to preventing deaths from lung cancer, and with modest uptake it may reduce late-stage diagnoses and mortality.