However, as of 2019, nearly half of the half million variants in ClinVar remain in the
VUS category [6]. Therefore, deep bioinformatic analyses are urgently needed to assess the
increasingly large number of variants and attempt to prioritize those that may be important
for disease risk and progression.
Another exciting application of high-throughput sequencing to cancer diagnosis is the
analysis of liquid biopsies. Liquid biopsies refer to the non-invasive sampling of body
fluids such as blood, urine, saliva, and cerebrospinal fluids, among others. The early
detection of cancer in this way is theoretically possible because tumors shed cells and
DNA (referred to as “circulating tumor cells,” CTCs, and “circulating tumor DNA,”
ctDNA, respectively) into the bloodstream by apoptosis or necrosis. Several specialized
techniques, such as antibody capture, depletion of white and red blood cells, and size
exclusion have been applied to the detection of CTCs; cell-free DNA (cfDNA) in the blood,
of which ctDNA represents between <0.1 and 10%, can be analyzed for common driver
mutations [7]. This technology offers great potential, as for example, several studies have
shown that RAS- or TP53-cancer associated mutations can be discovered in sputum or
plasma several months before lung adenocarcinoma or bladder cancer diagnosis, respectively [5], and the size of cfDNA fragments may indicate a tumoral origin. Still, great
challenges need to be overcome for this technique to be deployed for the screening of an
asymptomatic population: First, the amount of ctDNA is so low that even though
specialized techniques have been developed for its analysis, such as digital PCR, droplet
digital PCR, and BEAM, these still suffer from low multiplexing capacity where only a few
mutations can be assessed. Second, high-throughput sequencing technologies, while able
to assess a large number of loci, do not have the required sensitivity to confidently detect
these mutations [7]. These complications would increase the number of false-positive
diagnoses, and though the technique is developing quickly, it is still at an early stage.
Liquid biopsies have presently better value for prognostic assessments and disease monitoring, which we will discuss in Sect. 2.6.
2.4
Genome Sequences Can Reveal Cancer Origins
As mentioned in the introduction, mutations in the genome of a cancer cell can be caused
by exogenous factors, such as exposure to carcinogenic agents like ultraviolet radiation or
cigarette smoke, or endogenous processes such as defects in the DNA damage repair
machinery. As these mutagenic agents have very diverse modes of action (e.g., bulky
DNA adducts bind covalently to DNA bases, ionizing radiation is able to induce DNA
breaks by disrupting chemical bonds, alkylating agents can add alkyl groups to guanine
bases, etc.) we can expect that the patterns of mutations these leave in the genome are also
quite different. For example, UV radiation preferentially causes C>T transitions at
dipyrimidine sites, and exposure to benzo-[α]-pyrene results mainly in C>A mutations
[8]. Therefore, the set of all mutations in a genome can be considered as an archeological
22
C. Molina-Aguilar et al.
VUS category [6]. Therefore, deep bioinformatic analyses are urgently needed to assess the
increasingly large number of variants and attempt to prioritize those that may be important
for disease risk and progression.
Another exciting application of high-throughput sequencing to cancer diagnosis is the
analysis of liquid biopsies. Liquid biopsies refer to the non-invasive sampling of body
fluids such as blood, urine, saliva, and cerebrospinal fluids, among others. The early
detection of cancer in this way is theoretically possible because tumors shed cells and
DNA (referred to as “circulating tumor cells,” CTCs, and “circulating tumor DNA,”
ctDNA, respectively) into the bloodstream by apoptosis or necrosis. Several specialized
techniques, such as antibody capture, depletion of white and red blood cells, and size
exclusion have been applied to the detection of CTCs; cell-free DNA (cfDNA) in the blood,
of which ctDNA represents between <0.1 and 10%, can be analyzed for common driver
mutations [7]. This technology offers great potential, as for example, several studies have
shown that RAS- or TP53-cancer associated mutations can be discovered in sputum or
plasma several months before lung adenocarcinoma or bladder cancer diagnosis, respectively [5], and the size of cfDNA fragments may indicate a tumoral origin. Still, great
challenges need to be overcome for this technique to be deployed for the screening of an
asymptomatic population: First, the amount of ctDNA is so low that even though
specialized techniques have been developed for its analysis, such as digital PCR, droplet
digital PCR, and BEAM, these still suffer from low multiplexing capacity where only a few
mutations can be assessed. Second, high-throughput sequencing technologies, while able
to assess a large number of loci, do not have the required sensitivity to confidently detect
these mutations [7]. These complications would increase the number of false-positive
diagnoses, and though the technique is developing quickly, it is still at an early stage.
Liquid biopsies have presently better value for prognostic assessments and disease monitoring, which we will discuss in Sect. 2.6.
2.4
Genome Sequences Can Reveal Cancer Origins
As mentioned in the introduction, mutations in the genome of a cancer cell can be caused
by exogenous factors, such as exposure to carcinogenic agents like ultraviolet radiation or
cigarette smoke, or endogenous processes such as defects in the DNA damage repair
machinery. As these mutagenic agents have very diverse modes of action (e.g., bulky
DNA adducts bind covalently to DNA bases, ionizing radiation is able to induce DNA
breaks by disrupting chemical bonds, alkylating agents can add alkyl groups to guanine
bases, etc.) we can expect that the patterns of mutations these leave in the genome are also
quite different. For example, UV radiation preferentially causes C>T transitions at
dipyrimidine sites, and exposure to benzo-[α]-pyrene results mainly in C>A mutations
[8]. Therefore, the set of all mutations in a genome can be considered as an archeological
22
C. Molina-Aguilar et al.