Page 89 - Read Online

Page 4 of 16 D’Amico et al. J Cancer Metastasis Treat 2021;7:3 I http://dx.doi.org/10.20517/2394-4722.2020.93
[12]
extraction and analyses through high-throughput sequencing . Indeed, the sensitivity of conventional
next generation sequencing (NGS) in detecting DNA alterations is limited to a relatively high fraction
[13]
of mutant to wild-type DNA (> 1%) . Nonetheless, the conceptual value of ctDNA-based driver gene
mutations identification stands in tumor cells clonality, which might in fact determine a high degree
of specificity compared to other blood-based biomarkers, endorsing its clinical utility in the diagnostic
process.

[14]
Beaver et al. prospectively analyzed plasma ctDNA of 29 eBC patients before and after surgery through
digital polymerase chain reaction (ddPCR) and compared it with tissue-based Sanger sequencing and
ddPCR analyses. PIK3CA mutations were identified in 15 tumor tissues and 13 of them had a matching
pre-operatory plasma sample mutation. Of note, in 5 cases PIK3CA mutations were still detectable in
post-surgical plasma samples. ddPCR high sensitivity allowed for ctDNA mutation detection with even
low levels of tumor burden, providing an initial proof-of-concept to support the feasibility of ctDNA-
[15]
based analyses in the detection of somatic DNA mutations. Phallen et al. developed the targeted error
correction sequencing (TEC-Seq) approach to retrospectively detect cfDNA sequence changes through
ultra-sensitive massive genome sequencing. Overall, 58 cancer-related driver genes were analyzed in 200
patients with early-stage colorectal, breast, lung, and ovarian cancer and somatic mutations were detected
in 71%, 59%, 59%, and 68% of cases, respectively. Of note, BC ctDNA samples had the lowest mutant
allele fraction. High concordance between plasma and tumor-detected alterations was observed and the
apparent lack of concordance between some specific alterations, potentially determined by intratumoral
heterogeneity, might in fact be overcome by plasma-based analyses.

[16]
Recently, a single, multi-analyte blood-test, named CancerSEEK, was developed by Cohen et al. to assess
cfDNA mutations and it was applied to 1.005 clinically diagnosed stage I-III cancer patients to screen for
eight common cancer types, including eBC. The circulating levels of eight serum protein biomarkers were
simultaneously evaluated to overcome the issue of low detectable levels of ctDNA, therefore improving
sensitivity. This approach also included machine-learning tools to accurately narrow down the location of
a tumor to a small number of anatomic sites. A relatively small yet robust and highly selective 61-amplicon
panel was designed through multiplex-polymerase chain reaction to detect ctDNA mutations with an
acceptable degree of screening sensitivity and specificity. Median sensitivity was 70% overall and 33% in
eBC patients, increasing alongside tumor stage, with 43% for stage I, 73% for stage II, and 78% for stage
III cancers. Specificity was higher than 99%, with only 7 out of 812 healthy controls scoring positive and
localization of tissue of origin (TOO) was possible in a median of 83% patients. Intriguingly, a significant
concordance between ctDNA- and tissue-detected mutations was observed in 90% of cases among all
tumors.

Other circulating biomarkers, including exosomes, miRNA, and methylated DNA sequences could
be integrated in multi-analyte blood tests to provide complementary early information through gene
expression profiling. Proof-of-concept studies have demonstrated that miRNA, packed into tumor-released
exosomes, can be detected at higher concentrations compared to ctDNA in early-stage cancer patients.
Moreover, their qualitative analyses might be informative both of DNA somatic mutations and epigenetic
alterations [17,18] . Indeed, large-scale epigenetic alterations might improve sensitivity in early cancer detection.
For instance, the enrichment of methylated cfDNA fragments, which are cancer-specific, could potentially
overcome the current technical constraints of mutation-based ctDNA detection methods, mostly
determined by the limited amount of recurrent mutations that discriminate tumor and normal cfDNA.
[19]
Aiming to profile the methylome of small amounts of cfDNA, Shen et al. developed a non-invasive
sequencing approach that included cell-free methylated DNA immunoprecipitation and high throughput
sequencing (cfMeDIP-seq) for genome-wide bisulfite-free plasma DNA. An optimized Me-DIP-sequencing
protocol was first analyzed in a cohort of early-stage pancreatic cancer tissues and healthy controls and