New Blood Test Enhances Lung Cancer Screening

Researchers have developed and validated a blood-based, cell-free deoxyribonucleic acid (cfDNA) fragmentome assay for lung cancer detection. If the assay yielded positive results, low-dose computed tomography (LDCT) would be the next step. This research was published in Cancer Discovery.

Overview

Lung cancer is a major death cause, and regular screening is necessary. Nevertheless, patient hurdles such as restricted availability, radiation fears, and insufficient awareness have resulted in a poor acceptance rate for chest LDCT.

Poor smoking history documentation, a lack of established procedures, and specialized follow-up are further difficulties.

Similar to the fragmentome technique, a blood-based lung cancer screening test could raise screening rates by examining particular chromatin patterns in peripheral blood.

Concerning the study

Researchers used machine learning to create a hematological test in the current DELFI-L101 project, which analyzes DNA fragmentomes to determine who is at risk of lung cancer. Those who tested positive will be sent to LDCT.

Starting in March 2021, 958 people between the ages of 50 and 80 who had smoked for at least 20 pack years were registered by the researchers in 47 US facilities. The 2015 National Health Interview Survey’s (NHIS) LDCT screening criteria were similar to the eligibility elements.

They excluded anyone who had received cancer treatment within a year, had a history of myelodysplasia or hematologic malignancy, had received organ tissue transplants, had received blood products within 120 days of enrolling, were pregnant, or had taken part in previous trials.

The researchers split the study participants into three groups: C (cancer other than lung cancer), B (non-cancer controls), and A (lung cancer).

The disease stage was determined using the Cancer Staging Manual (AJCC) guidelines established by the American Joint Committee on Cancer. Lung cancer’s genomic and chromatin characteristics were uncovered by alterations in cfDNA fragmentation patterns, or fragmentomes, in blood.

Prior to verifying the classifier on an additional 382 cases and controls, the researchers trained it on 576 cases and controls.

They evaluated fragmentations in 504 non-overlapping-type 5.0 MB sections with strong mappability using whole genome sequences from the training dataset. Every region comprised 80,000 pieces, encompassing a genome with a capacity of 2.50 GB.

The group looked at changes to Hi-C closed-type (B compartment) and open-type (A compartment) chromatin over the whole genome.

Using principle component analysis (PCA) and logistic regressions, they developed a classifier that used cfDNA fractions extracted from the mitochondrial genome, chromosomal arm-level alterations, and cfDNA fragment length distributions.

The scientists ran Monte Carlo simulations on 15 million people in three different scenarios:

Base Scenario: Present procedures without a blood test.
Low Scenario: 10% hematological screening uptake in the first year for those eligible for lung cancer screening but not exposed to low-dose CT, rising to 25% in five years.
High Scenario: in the first year, 20% of the same group will have hematological screening; in five years, this percentage will rise to 50%.

Outcomes
The test’s 99.8% negative predictive value (NPV), 58% specificity, and 84% sensitivity were noted by the researchers. The number needed to screen (NNS) was 143 when applying the remaining portion to the screening-eligible group with a lung cancer prevalence of 0.7%.

The validation of the study yielded a 5.5 relative risk value, with negative and positive results associated to NNS using LDCT imaging to detect 414 and 76 instances, respectively. The PPV, or positive predictive value, was nearly twice as high as the LDCT qualifying conditions by itself.

Patients with lung squamous cell carcinoma (LUSC) had cfDNA fragmentomes that included a part that resembled cfDNA profiles from people without cancer and another that resembled A/B-type compartments found in LUSC tissues.

People without cancer displayed cfDNA patterns that resembled lymphoblastoid Hi-C results. Fragments from samples given by people with and without cancer were identical within frequent areas.

Patients with lung cancer showed lower levels of 1p, 3p, 4q, 5q, 10q, and 17p and higher representations of cell-free DNA fpr 1q, 3q, 5p, 8q, and 12p. Compared to controls, their cfDNA fragmentations revealed more densely packed chromatin in the cfDNA of closed LUSC spaces, whereas lymphoblastoid reference sections displayed the opposite effect.

10% overall specificity and sensitivities of 75%, 90%, 96%, and 97% for stages I, II, III, and IV were obtained at the cut-off of 0.2 using ten-fold cross-validation with ten repeats within the training sample. All ages showed the same level of sensitivity, with younger people showing better specificity. The 2015 NHIS data produced results with 58% specificity and 80% sensitivity.

Lung cancer cases found by screening rose from the “base” scenario (24,489 instances) to 63,523 (the “low” scenario) and 100,346 (the “high” scenario). In comparison, diagnoses for stage IV fell by 4.20% and 8.70%, whereas cases for stage I climbed by 4.80% and 9.70%, respectively.

The “base” scenario might prevent 4,720 deaths from lung cancer, the “low” scenario 7,652 deaths, and the “high” scenario 14,264 deaths. The number of tests needed to detect lung tumors could drop from 202 (the “base”) to 150 (the “low”) and 139 (the “high”) if LDCT is used in screening.

In summary

The DNA fragmentome assay offers a new, precise, cost-effective, blood-based method for early lung cancer assessment with LDCT follow-ups, according to the study’s findings.

With reasonable acceptance rates potentially reducing late-stage diagnoses and mortality, the assay may help avoid deaths from lung cancer.

For more information: Clinical validation of a cell-free DNA fragmentome assay for augmentation of lung cancer early detection, Cancer Discov (2024), https://doi.org/10.1158/2159-8290.CD-24-0519