Early cancer detection is performed from cell-free DNA in blood samples

Early detection remains key to the successful treatment of many cancers, and early detection through circulating cellular free DNA (cfDNA) circulating in the blood — the so-called "liquid biopsy" — has become a research focus
.
However, due to the low concentration of tumors in blood DNA fragments, as well as the genetic diversity of cancers, using this method to detect cancers at an early stage is challenging
.

Now, researchers and collaborating organizations at UCLA's Johnson Center for Comprehensive Cancer have reported the successful results of an experimental cancer detection system that appears to have overcome these challenges
in a novel and economical way.

Their study, published in the journal Nature Communications, highlights a method that is more than 12 times more cost-effective than traditional cfDNA methyl group sequencing methods, as well as a computational model for extracting information from DNA sequencing to aid in early detection and diagnosis
.

Cell free DNA methylation has proven to be one of
the most promising biomarkers for early cancer detection.
However, cfDNA aberration features from different cancer types, subtypes, stages, and etiologies are heterogeneous, which leads to challenges
in identifying methylation markers suitable for early detection.
This is particularly worrying
about the small sample sizes currently available compared to the diversity of diseases and patient populations (age, sex, ethnicity, and comorbidities).
Analyzing the cfDNA methyl genome can solve this challenge because it preserves a genome-wide epigenetic profile of cancer abnormalities, thus allowing taxonomic models to learn and develop new important features as the training cohort grows, and extend its scope to more cancer types
.
However, traditional methods of analyzing the cell's free DNA methyl group (genome-wide bisulfite sequencing) are too costly
in clinical applications.

"Our method, cfMethyl-seq, makes cfDNA methyl group sequencing a viable option for clinical use," said
Xianghong "Jasmine" Zhou, professor of pathology and laboratory medicine at UCLA and corresponding author of the study.
"Despite the inherent challenges, our research shows great potential
for accurate early diagnosis of certain cancers with a single blood test.
"

Colleagues in the UCLA lab focused on precision medicine — using patient genomic information to develop more personalized and targeted treatments — and large-scale biodata analytics that integrate complex data from different platforms and models into practical methods that can be used clinically
.

In the study, Zhou and collaborators tested their new method to see if it could accurately detect four common cancers — colon, liver, lung and stomach — and do so
at an early stage.

The researchers collected blood samples from 408 study participants and applied a blood test based on the methyl group that can identify a wide range of markers for different cancer types and possible causes
.
Of these, 217 were cancer patients and 191 were cancer-free control subjects
.
Samples are collected at UCLA hospitals or purchased from commercial laboratories for cross-source validation
.
The researchers also performed cross-batch validation, age-match validation, and independent validation to prevent bias
in the study.

After collecting and validating the measures, the researchers fed the data into their sophisticated computer model to measure its accuracy in detecting cancer, as well as the specific location of the tumor, known as the "tissue of origin
.
"

Their model has an accuracy rate of 80.
7% in detecting all stages of cancer and about 74.
5% accuracy in detecting early-stage cancer (stage I or II) with a specificity of just under 98%.

There is only one normal sample with a misclassification (false positive).

For the accuracy of the source tissue, the model correctly identified tumor locations, with an average accuracy of 89.
1% for all cancer stages and about 85%
for early stage patients.

"The key to early cancer detection is to identify true cancer biomarkers, which require a large cohort of trained samples to cover cancer and population heterogeneity, especially pan-cancer detection
.
" As the training cohort grows, our cfDNA methyl group approach allows for better weights
to include new labels and existing labels.
In fact, our data show that the detection capabilities of our method continue to enhance as the training sample size increases," Zhou said
.
"cfMethyl-seq's cost-effective methyl group sequencing can truly facilitate big data methods
for cancer detection.
"

The team is currently raising funds for large clinical trials to validate the technology, hoping to use it for the benefit of patients
.

Mary L.
Stackpole, Weihua Zeng, Shuo Li, Chun-Chi Liu, Yonggang Zhou, Shanshan He, Angela Yeh, Ziye Wang, Fengzhu Sun, Qingjiao Li, Zuyang Yuan, Asli Yildirim, Pin-Jung Chen, Paul Winograd, Benjamin Tran, Yi-Te Lee, Paul Shize Li, Zorawar Noor, Megumi Yokomizo, Preeti Ahuja, Yazhen Zhu, Hsian-Rong Tseng, James S.
Tomlinson, Edward Garon, Samuel French, Clara E.
Magyar, Sarah Dry, Clara Lajonchere, Daniel Geschwind, Gina Choi, Sammy Saab, Frank Alber, Wing Hung Wong, Steven M.
Dubinett, Denise R.
Aberle, Vatche Agopian, Steven-Huy B.
Han, Xiaohui Ni, Wenyuan Li, Xianghong Jasmine Zhou.
Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer.
Nature Communications, 2022; 13 (1) DOI: 10.
1038/s41467-022-32995-6