*6.1.1. Genetic markers*

Lin et al. [75] described the link between integrated genomic signatures, the biological functions, and the background molecular pathways [76]. There were developed prediction models of activity for eight anticancer drugs [76], along with clinical responses to 5-FU (cDNA microarray analysis) [77] and resistance-related genes such as dihydropyrimidine dehydrogenase (DPD) and HB-EGF-like growth factor genes [78]. Also, it was reported that metallothionein-IG and heparin-binding epidermal growth factor-like growth factor (HB-EGF), glutathione-S-transferase, and cyclooxygenase-2 genes were cisplatin-resistance-related and genes such as ADAM22, CYR61, FN1, SPHK1, and GNAI1 were linked to doxorubicin response [79]. Furthermore, in some studies, the genetic polymorphism was linked to the response of 5-FU, cisplatin [80], and paclitaxel [81].

Cristescu et al. described four molecular subtypes of gastric tumors related to disease progression and prognosis: the mesenchymal-like type with highest recurrence frequency, microsatellite-unstable tumors that are hyper-mutated and are associated with the best overall prognosis, tumor protein 53 (TP53)-active and TP53-inactive types that have intermediate prognosis and recurrence [82].

As already mentioned, researchers from "The Cancer Genome Atlas (TCGA)" project proposed a molecular classification dividing gastric tumors into four subtypes, useful for stratifying patients and choosing the appropriate targeted treatment: (1) Epstein-Barr virus positive tumors: PIK3CA mutations, DNA hypermethylation, and amplification of JAK2, CD274, and PDCD1LG2; (2) microsatellite unstable tumors: increased mutation rates, including genes encoding targetable proteins involved in oncogenic signaling pathways; (3) genomically stable tumors: mutations of RHO-family GTPase-activating proteins; (4) tumors with chromosomal instability: marked aneuploidy and amplification of receptor tyrosine kinases [41].

#### *6.1.2. Epigenetic markers*

MicroRNA was linked by some studies to the resistance to trastuzumab [83], the pathologic response to neoadjuvant chemotherapy [81], and the chemotherapeutic response of cisplatin/ fluorouracil [84].

Long noncoding RNAs (lncRNAs) are potential biomarkers for gastric cancer especially using minimally invasive routes (blood, gastric secretions) [85]. The lncRNA MRUL (Multidrug resistence (MDR)-related and upregulated lncRNA) originated from tissue samples was associated with multidrug chemotherapeutic resistance [86].

Methylation-related biomarkers: methylation of Bcl-2/adenovirus E1B 19 kDa-interacting protein 3 and death-associated protein kinase (DAPK) correlate with poor response to fluoropyrimidine-based chemotherapy [87]; decreased methylation of the bone morphogenetic protein 4 (BMP4) correlates with cisplatin resistance, and the regain of treatment response may be achieved using targeted inhibition of BMP4 [88]. Also, the increased expression of Reprimo (a highly glycosylated cellular protein) due to methylation was associated with a lower response to cisplatin/5-FU chemotherapy [89].
