Image Source: Gerenme | iStock
ELSA-seq, a circulating free DNA (cfDNA) methylation detection technology is able to sensitively screen 6 common cancers such as lung cancer and colorectal cancer, and significantly reduce the occurrence of advanced cancers leading to an increase of 40% in 5-year survival rate.
As we all know, Early detection, Early diagnosis, and Early treatment are the top priorities in the fight against cancer, especially as the population ages and the impact of cancer risk factors such as diet, tobacco, and alcohol becomes more and more obvious. The incidence and mortality of cancer worldwide are increasing. On the rise[1], it is really difficult to fight against this deadly disease if you can’t do the three “Earlys”.
The principle is easy to understand, but easier said than done. In order to truly implement the “Three Earlys”, first of all, there must be effective early screening methods for cancer. However, only a few types of cancer screening such as those for lung cancer and colorectal cancer meet this requirement; secondly, the screening must be convenient and fast. It is best to be able to screen for multiple cancers at one time, otherwise, if you take time off and go to the hospital just to get screened for one cancer, chances are you really don’t have enough motivation to do so[2].
It sounds like neither of these problems is easy to solve. But switching to cancer liquid biopsy, which is booming, it may not be difficult to break the situation, and with the upgrading of liquid biopsy technology, the detection becomes more and more accurate and reliable, and the comprehensive promotion of liquid biopsy for cancer screening may benefit more than imagined!
“Blood test for cancer” is becoming more and more feasible.
Recently, in the authoritative journal “Annals of Oncology”, the team led by Fan Jia of the Zhongshan Hospital affiliated with Fudan University in Shanghai, China published the latest research results: ELSA, a methylation detection technology for circulating free DNA (cfDNA) can more sensitively screen 6 common cancers such as lung cancer and colorectal cancer, and significantly reduce the occurrence of advanced cancer, thereby improving the overall 5-year survival rate of cancer patients.
Based on ELSA-seq analysis of cfDNA methylation signatures in cancer patients and non-patient populations, the researchers developed two screening models (MCDBT-1/2), in which the screening specificity of MCDBT-1 in an independent validation cohort It is as high as 98.9%, and the sensitivity is also 69.1%. According to the calculation of cancer incidence data in my country, the comprehensive promotion of this screening can reduce the number of advanced patients with 6 common cancers by 38.7-46.4%, and increase the 5-year survival rate by 40% [3].
Readers who are concerned about the field of cancer liquid biopsy should not be unfamiliar with ELSA-seq technology. It is a “star” in the industry that has repeatedly appeared in important conferences and journals in recent years. It can cover a wide range of CpG sites related to cancer. In this way, it can accurately screen cancer and trace its origin. In this study, it is responsible for screening six notorious common cancers: lung cancer, colorectal cancer, esophageal cancer, liver cancer, ovarian cancer, and pancreatic cancer.
The overall idea of this study is also similar to previous studies. First, a cfDNA methylation custom detection panel was constructed based on open-source data and initially verified, and then the panel was used to analyze 1,693 retrospective cfDNA samples (735 cancer patients and 735 non-cancer patients). 958 cases), respectively trained and validated the MCDBT-1/2 model, and finally validated the model in an independent validation prospective cohort (n=946, cancer/non-cancer 473 cases, feature matching).
In the panel construction stage, according to the co-methylation level of CpG sites and the genomic distance of adjacent CpG sites, the researchers determined the methylated regions specific to each of the six cancers, effectively distinguishing cancer patients from non-cancer population’s Methylation patterns, and the accuracy and reliability are verified by machine learning models.
The MCDBT-1/2 model based on the above-mentioned Panel is mainly different in that the former is a very high specificity model (Cut-off value is 99.5%), and the latter is a high specificity model (Cut-off value is 99.0%). Therefore, the sensitivity of the MCDBT-2 model is also different, and it is suitable for people with different risks of cancer. However, from the evaluation of the study, the overall performance of the MCDBT-2 model is similar to that of the MCDBT-1.
The training set data in the model building phase showed that the overall specificity of the MCDBT-1 model for screening 6 cancers was 99.7% (95% CI: 98.9%-100.0%), and the sensitivity was 75.2% (70.6%-79.4%), Comparable to previous studies; while in the validation set, the specificity of MCDBT-1 model screening was 100.0% (95% CI: 97.0%-100.0%) and the sensitivity was 69.4% (70.6%-79.4%).
In the independent validation cohort, the MCDBT-1 model had a specificity of 98.9% (97.6%-99.7%) and a sensitivity of 69.1% (64.8%-73.3%). In terms of refinement, the sensitivity of MCDBT-1 model for screening stage I-II cancer is 43.8%, and the sensitivity for screening stage I-III is 54.0%. The sensitivity was low (51.2%/71.3%), which may be related to the difference in the incidence of cancer between genders.
When liquid biopsy screens multiple cancers at one time, it is also necessary to accurately trace the location of cancer (Top predicted origin, TPO) to guide subsequent diagnosis and treatment. The MCDBT-1 model performs better in this item. The traceability accuracy (TPO1) in the validation cohort reached 83.2%.
After talking about the data for a long time, how can liquid biopsy translate into the huge benefits of “reducing 40% of advanced cancer and increasing the 5-year survival rate of cancer by 40%”? The researchers re-estimated the performance of the MCDBT-1 model in the real world based on the actual incidence and diagnosis stages of the six cancers covered by the model in my country.
Estimates show that the real-world screening sensitivity of the MCDBT-1 model is 70.5%. Using it to carry out comprehensive early cancer screening can enable 38.7-46.4% of patients to be diagnosed at a relatively early stage (I/II stage), changing the current About 60% of patients are diagnosed at a relatively advanced stage (III/IV stage); combined with radical treatment, the absolute value of the 5-year survival rate of the above six cancers can be increased from 31.4% to 41.8-44.1%, which is a relative improvement of 40%.
References:
[1] Chen S, Cao Z, Prettner K, et al. Estimates and Projections of the Global Economic Cost of 29 Cancers in 204 Countries and Territories From 2020 to 2050[J]. JAMA Oncology, 2023 .
[2] Callison K, Pesko MF, Phillips S, et al. Cancer Screening after the Adoption of Paid-Sick-Leave Mandates[J]. New England Journal of Medicine, 2023, 388(9): 824-832.
[3] Gao Q, Lin YP, Li BS, et al. Unintrusive multi-cancer detection by circulating cell-free DNA methylation sequencing (THUNDER): development and independent validation studies[J]. Annals of Oncology, 2023.