Potential findings on whole genome sequencing for cancer treatment was made

SHIRLEY, NY, UNITED STATES - Nov 25, 2019 - In recent years, individualized treatments will benefit more cancer patients as the time and cost of whole genome sequencing decreases. A recent study published in Nature Medicine showed that the development of new machine learning algorithms to analyze tumor genome-wide sequence information can help predict patient prognosis and help patients choose the best treatment.

Whole-genome sequencing (WGS) technology can obtain almost all genetic information of patients, including whole-genome information of cancer cells and healthy cells, which make abnormal cancer cells can be found at the level of single nucleotide, copy number, epigenetic modification, etc.

To explore the clinical value of WGS, researchers at the University of Cambridge collaborated with researchers at Lund University in Sweden to develop a population-wide project called SCAN-B (Sweden Cancerome Analysis Network-Breast) (ClinicalTrials.gov ID: NCT02306096). The project has recruited breast cancer patients in Sweden since 2010 and has collected a large amount of clinical data so far.

Project process

In the SCAN-B project, the researchers performed a genome-wide sequencing of 254 triple-negative breast cancer (TNBC) in SCAN-B between 2010 and 2015. Triple-negative breast cancer refers to breast cancer patients with negative estrogen receptor, progesterone receptor and human epidermal growth factor receptor 2, accounting for 9% of breast cancer, and the prognosis is poor.

The researchers then used a machine learning algorithm called "HRDetect" to classify the tumors. This algorithm was originally developed to detect tumors with a BRCA1/BRCA2 gene mutation signature, and mutation of either BRCA1/BRCA2 gene would greatly increase the risk of breast cancer. Currently, a new targeted drug PARP inhibitor can be used for breast cancer patients with BRCA1/BRCA2 gene mutations.

According to the HRDetect algorithm, the patients were classified into three categories: high (HRDetect-high), medium (HRDetect-intermediate), and low (HRDetect-low). 59% of patients had homologous recombination repair loss (HRDetect-high): 67% of them were germline/somatic mutations of BRCA1/BRCA2, BRCA1 promoter hypermethylation, RAD51C hypermethylation or PALB2 biallelic loss.

The HRDetect algorithm provides a unique diagnostic message. The researchers found that patients with high-risk (HRDetect-high) triple-negative breast cancer had the best treatment and were more sensitive to PARP inhibitors than those with low scores (HRDetect-low).

Surprisingly, patients with moderate (HRDetect-intermediate) treatments had the worst treatment, and even though some patients had some better drug targets, the prognosis was still poor. Therefore, for patients with low scores, it is necessary to adopt a new treatment plan. Patients with low scores (HRDetect-low) also have poor prognosis. However, the WGS sequence information of these patients' tumors is also abnormal, which may be caused by the use of fixed drugs in clinical trials, such as immunological checkpoint inhibitors (PD-1) or AKT inhibitors.

Dr. Johan Staaf, the lead author of the study, said: "Three-negative breast cancer is difficult to treat, but by genome-wide sequencing, we can identify which triple-negative breast cancer patients are more sensitive to the drugs currently used clinically. Importantly, this method allows us to provide clues for studying the mechanisms of poor prognosis and develop new drugs for such patients."

With the rapid development of sequencing technology, whole genome sequencing can be completed in 24 hours, and the analysis of sequencing data can be completed in 24-48 hours. Therefore, in theory, genome-wide sequencing and analysis can be provided for each patient, and the optimal treatment plan can be selected based on the patient's own tumor genome information.

“Genome-wide sequencing opens up new avenues for individualized treatment of cancer,” said co-first author Dr. Nik Zainal. “In the past, the management and analysis of large amounts of data was a major obstacle to its widespread use. But now we are reducing the time for data analysis. Let each patient get individualized treatment, which will greatly change the treatment of cancer, even some refractory cancer treatment."

CD Genomics has been providing the accurate and affordable whole genome sequencing service for couple of years. The company combines both Illumina (short reads) and PacBio (long reads) platforms to achieve whole genome de novo assemblies and re-sequencing for virus, microbes, plants, animals and humans.