Journal Thoracic Oncology

Abstracts Journal of Thoracic Oncology • Volume 12 Issue S1 January 2017
On the contrary, tumors from never smokers demonstrate a much lower
mutational burden (0.8-1 mutations/Mb) and are enriched for C>T transitions.
1,2
Single nucleotide variations (SNVs) and copy number alterations (CNAs)
While both LUAD and SQCC show frequent inactivation of the tumor
suppressors TP53 and CDKN2A, these alterations are considerably more
common in SQCCs. CDKN2A harbors the loci for two isoforms, p14ARF and
p16INK4A, and is inactivated in SQCC through homozygous deletion (29%),
methylation (21%), inactivating mutations (18%), or exon 1b skipping (4%).
1,2 These findings indicate a strong selective pressure for the loss of these
tumor suppressors in NSCLC. The pattern of oncogenic alterations varies
considerably between LUAD and SQCC. While LUADs typically showed
activating RTK/RAS/RAF pathway mutations, these mutations are highly
infrequent in SQCCs - which predominantly showed alterations in oxidative
stress response (NFE2L2, KEAP1 and CUL3) and squamous differentiation
pathways (SOX2, TP63, NOTCH1, etc.) in 44% of samples. 1,2KRAS is the most
commonly mutated oncogene in LUAD, followed by EGFR, BRAF, PIK3CA, and
MET. The majority of EGFR mutations in LUAD are targetable (L858R or exon 19
deletion) with tyrosine kinase inhibitors (TKIs). 1 In contrast, such alterations
are absent in SQCC. Two SQCC samples however demonstrated L861Q
mutations in EGFR, which are potentially targetable with TKIs. 1,2 Although
SQCC and LUAD shared several CNAs at the chromosomal arm level,
amplification of 3q was frequent in SQCC. This region harbors important
oncogenes such as SOX2, PIK3CA, and TP63. LUADs frequently showed
amplifications in genes such as NKX2-1, TERT, MDM2, KRAS, and
EGFR. 1,2 Oncogenic activation of kinases such as ALK, ROS1, and RET through
rearrangement has been well described in LUAD, and these fusions are
targetable with TKIs. These fusions were seen in 1-2% (ALK : 3/230, ROS1:
4/230, and RET: 2/230 samples) of LUADs. 1,2 Transcriptome analysis
Deregulated splicing can be a consequence of mutations that alter splice-sites
within a gene or splicing factors. Mutations in the proto-oncogene MET that
lead to exon 14 skipping, and abnormal splicing of proto-oncogenes such as
CTNNB1 as a result of U2AF1 mutation have been described in LUAD. 1
Transcriptome analyses have also enabled a reclassification of LUADs and
SQCCs into three and four distinct subtypes, respectively. LUAD samples can
be categorized as terminal respiratory unit (enriched for EGFR mutations and
fusions; favorable prognosis), proximal-inflammatory (NF1 and TP53
co-mutation), or proximal-proliferative (KRAS and STK11 alterations) subtypes.
Similarly, SQCCs can be classified as classical, basal, secretory, or primitive.
Alterations in genes that participate in the oxidative stress response
pathway, hypermethylation, and chromosomal instability are characteristic
of the classical subtype (associated with heavy smoking and poor prognosis).
1,2
Key pathogenic alterations TCGA analysis revealed alterations in well
known oncogenic drivers involving RAS signaling pathway in 62% of LUAD..
These samples with readily identifiable oncogenic driver alterations were
collectively labeled ‘oncogene-positive’. Additional analyses of the ‘oncogenenegative’
sample cohort showed enrichment for RIT1, and NF1 mutations.
Given the role of RIT1 and NF1 in RTK/RAS/RAF signaling, samples with these
mutations were reclassified as oncogene positive, increasing the overall
percentage of oncogene positive samples in LUAD to 76%. Nearly 69% of SQCC
samples showed alterations in genes regulating PI3K/AKT, or RTK/RAS
signaling. 1,2 The inability to readily identify an oncogenic driver in nearly a
third of sequenced lung cancer samples highlights the need for greater
powering of subsequent studies to identify novel low frequency genomic
alterations. For instance, previously uncharacterized alterations in the RTK/
RAS/RAF pathway were observed in RASA1, SOS1 in the updated TCGA analysis
which analyzed a much larger cohort of samples. 9 Overall, despite showing a
few similarities between LUAD and SQCC, investigators of TCGA reported
prominent differences between the genomic landscapes of these subtypes.
These subtypes have more of their alterations in common with other cancers
than with one another. SQCCs more closely resembled head and neck
squamous cell and bladder cancer, while LUAD resembled glioblastoma
multiforme and colorectal cancer in this regard. 9 Immunotherapies The vast
majority of lung cancers do not harbor alterations that are targetable by TKIs.
1,2
Immune checkpoint inhibitors are approved for use in patients with
metastatic NSCLC. There is a clear need to develop optimal predictive
biomarkers to identify those who are likely to respond to immune checkpoint
inhibitors. Mutational burden has been correlated with better response to
checkpoint inhibitors. Furthermore, using exome and transcriptome
sequencing and sophisticated bioinformatics, it is now possible to identify
mutated and expressed genes that could potentially serve as a trigger for
immune response (so called neoantigens) once immune checkpoints like
programmed death-1 or programmed death ligand-1 are inhibited.. Swanton
and colleagues performed a neoantigen and clonality analysis on TCGA
samples to examine characteristics such as neoantigen burden and
intratumor heterogeneity (ITH), and their impact on survival. In LUAD, a
higher neoantigen burden was significantly associated with longer survival.
Although not statistically significant, there was a trend towards longer
survival in molecularly homogeneous tumors (