ROC: Receiver operating characteristic; PPV: positive predictive value; NPV: negative predictive value. ALL: acute lymphoblastic leukemia; BCP-ALL: B-cell precursors ALL; AML: acute myeloid leukemia; GI: Group I; GII: Group II; PBMC: peripheral blood mononuclear cell. * Value of AC133 expression to predict cases of poor prognosis. With the aim of highlighting a functional role of Smad 7, transforming growth factor β1 (TGF-β1), and miR-181a in clinical diagnosis and prognosis of pediatric-ALL, Nabhan M et al. evaluated miR-181 expression and Smad 7 and TGF-β1 protein levels in the serum of 60 children with ALL [16]. The authors reported lower TGF-β and higher levels of Smad7 expression in controls compared to patients. Smad 7 showed a correlation with hepatomegaly, splenomegaly, and lymphadenopathy [16]. The authors reported high AUC values, especially when they were combined (sensitivity = 100%; specificity = 93.3%; cutoff values: miR-181a ≤ 0.97, Smad7 ≥ 400.1 pg/mL, and TGF-β1 ≤ 0.25 ng/mL). Regarding cell proliferation, the TGF-β1 signaling pathway is regulated at different levels and can act as tumor suppressor or promotor depending on the cell type [16]. So, TGF-β1 regulates cell proliferation, differentiation, apoptosis, adhesion, and migration. To activate Smad proteins, a family of receptors is required to undergo phosphorylation and activation. These receptors belong to the kinase domain type I TGF-β, which play a crucial role in controlling the TGF-β signaling pathway. The results showed resistance to the growth inhibitory and apoptotic effects of the TGF-β1 signaling pathway due to elevated Smad 7 levels, thus meaning that the TGF-β1 signaling pathway showed to be involved in ALL pathogenesis [16]. The presence of autoantibodies VDAC1 (voltage-dependent anion-selective channel 1) (p < 0.05) and α-enolase (p < 0.05) has also been reported as overexpressed in the serum of patients with BCP-ALL [14]. Researchers have postulated that the presence of these proteins could be immunogenic at the early stages of the disease (although their mechanism remains unclear). Because of their functions, α-enolase and VDAC1 may be used as auxiliary tools for the diagnosis and immunological surveillance, treatment, and prognosis of pediatric BCP-ALL [14]. A valuable method for evaluating multiple proteins is “proteomics”, which evaluates the proteome. The proteome is a group of proteins expressed in a cell, characterized by their distinct localization, post-translational modifications, interactions, and turnover ^[35][48]. In leukemias, aberrant protein expression patterns can identify the pathogenesis of the disease, post-translational modifications, origin, and the differentiation of leukemic cells ^{[7][35][36][56]}[7,48,60,78]. If successful, the study of one or a small number of proteins contributes to the understanding of the pathogenesis of disease; understanding the dynamics of the human proteome is crucial for developing biomarkers to be used as measurable indicators for disease severity and progression, patient stratification, and drug development. To distinguish between hematological malignancies with a proteome profile, researchers evaluated plasma from 107 patients older than 18 years with AML, ALL, or APLL, and 144 patients with lymphomas by using a proximity extension assay (PEA) ^[56][78]. PEA is a technology that translates protein information into actionable knowledge by linking protein-specific antibodies to DNA-encoded tags ^[57][79]. Regarding AL, the authors reported von Willebrand factor (vWF), SYND1, TNF-RSF6B, MPO, VIM, TNF-R1, IL-6, CTSD, and FURIN levels were remarkably higher compared to controls (fold change [FC] > 2; p-values < 1.0 × 10⁻⁴⁰); meanwhile, ADAM-TS15 (FC = 0.43; P = 3.34 × 10⁻⁴⁰) levels were lower. Moreover, the most differentiating proteins between ALL, AML, APL, and controls were TCL1A, CD27, and CD48. They did not find significant protein expression differences between of AML and ALL, perhaps due to the small number of patients. Therefore, those proteins can distinguish between leukemia, healthy patients, and other hematological malignancies. The functions of altered proteins were associated with hemostasis, inflammation, and cell-to-matrix integration ^{[56][58][59][60]}[78,80,81,82]. Speaking of hemostasis, vWF helps bind platelets to vascular injury sites after a proteolytic reaction regulated by FURIN ^[56][60][78,82]. These data are consistent with increased vWF and FURIN levels because thrombocytopenia and blood coagulation disturbances are crucial in acute leukemia. Similarly, the expression of the members of the TNF-R superfamily is involved in the inflammatory response and contributes to cell differentiation and survival ^[56][61][78,83]. In a study using high throughput technologies for proteins, Zhu et al. evaluated exosomal proteins in plasma from the blood samples of patients with BCP-ALL ^[45][68]. After isolating exosomes, they performed proteomic analysis by mass spectrometry and identified 342 differentially expressed proteins associated with Notch and the autophagy pathways, confirmed by Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. Because of this analysis and their potential role in BCP-ALL, the authors selected ATG3 and ADAM17 as candidate biomarkers for BCP-ALL progression. In the study, both proteins were overexpressed in patients with BCP-ALL, as verified by Western blotting. A high AUC value calculated for ADAM17 (AUC = 0.989, 95% confidence interval (CI): 0.960–1.018, p < 0.0001) and for ATG3 (AUC = 0.956, 95% CI: 0.885–1.026, p < 0.0001), indicating their potential as diagnostic biomarkers for BCP-ALL ^[45][68]. In another study, the authors evaluated the peripheral lymphocytes of 15 children diagnosed with ALL by using 2D electrophoresis, mass spectrometry, and identification in silico [4]. Among the 15 differentially expressed proteins, they could only identify eight. Glutathione S-transferase P (GSTP) (p < 0.01) and prohibitin (PHB) (p < 0.01) were upregulated. Peroxiredoxin 4 (PRDX4) (p < 0.01), the 60S acidic ribosomal protein P0 (p < 0.01), pyridoxine-5′ -phosphate oxidase (p < 0.01), triosephosphate isomerase 1 (p < 0.05), cytoplasmic actin (p < 0.01), and hypothetical protein FLJ26567 (p < 0.01), were downregulated; it is theorized that the expression levels of this last protein depends on the cell type. Although the authors were not able to speculate on the specific role of each protein in leukemogenesis, the whole panel of proteins demonstrated diagnostic value [4]. Similarly, in children with ALL, Shi et al. evaluated protein differentiation in serum blood samples. Mass spectrometry, bioinformatic analysis, and additional validation with immunoassays revealed that platelet factor 4 (PF4) (p = 1.54 × 10⁻⁷) and connective tissue activating peptide III (CTAP-III) (p = 7.19 × 10⁻⁸) were downregulated, and the fragments C3a-8137 (p = 5.35 × 10⁻⁵) and C3a-8937 (p = 5.13 × 10⁻⁵) were upregulated. Together, these proteins showed a sensitivity of 96% and a specificity of 98% for ALL diagnosis, implicating them as potential biomarkers for pediatric-ALL [7].

2.2. RNA Type BCP-ALL Biomarkers

RNA can generate numerous proteins through protein-coding regions. Most genomic sequences can be transcribed into protein-coding RNA. In contrast, the non-coding parts are transcribed to produce non-coding RNAs (ncRNAs) ^[62][63][84,85]. Based on their size, location, and interacting partners, ncRNAs are classified into several types: transfer RNAs (tRNAs), ribosome RNAs (rRNAs), small nucleolar RNAs (snoRNAs), small nuclear RNAs (snRNAs), microRNAs (miRNAs), small interfering RNAs (siRNAs), long non-coding RNAs (lncRNAs), circular RNA (circRNAs), Piwi-interacting RNAs (piRNAs), and enhancer RNAs (eRNAs) ^[62][84]. It has been discovered that numerous microRNAs are essential in the initiation, progression, and metastasis of cancer ^[64][86]. miRNAs are stable in serum and display distinct expression patterns in healthy individuals and those with cancer. Hence, they are suitable biomarkers for cancer detection and prognosis [32]. Furthermore, one miRNA can modulate the expression of several genes, and there are many miRNA expression patterns in patients with ALL ^[65][87]. In a study carried out in Mexico, the authors evaluated the diagnostic usefulness of the circulating miRNA expression profile in plasma samples from patients with BCP-ALL. miR-511 showed the highest mean overexpression (FC: 159.5, p = 0.002), while miR-199a-3p was the most under-expressed (FC: −13.48, p < 0.001). ROC curves analysis provided good values for miR-511 (cut-off = 9.458, specificity 1, sensitivity 1, AUC = 1), miR-34a (cut-off = 7.179, specificity 1, sensitivity 0.92, AUC = 0.98), miR-222 (cut-off = −0.1325, AUC = 0.91), miR-26a (cut-off = 2.073, AUC = 0.91), miR-221 (cut-off = −0.1861, AUC = 0.92), and miR-223 (cut-off −4.309, specificity 1, sensitivity 0.89, AUC = 0.93). Moreover, pathway analysis revealed that mainly Wnt, MAPK, TGF-beta, p53, Jak-STAT, NOTCH, and B- and T-cell receptor signaling pathways are activated by the evaluated miRNAs (miR-511, miR-19b, miR-195, miR-565, miR-34a, miR-222, miR-363, miR-181a, miR-181c, miR-199a-3p, miR-340, miR-335, miR-99b, miR-221, miR-744, miR-223, miR-26a, miR-224, and miR-151-3p). These miRNAs have value as markers for diagnosis and to identify therapeutic targets for BCP-ALL ^[50][73]. CircRNAs are a type of non-coding RNA and represent a recent research hotspot in the field of RNA. circRNAs form covalently closed loop structures with neither 5′–3′ polarities nor polyadenylated tails and lengths between 100 to thousands of nucleotides ^[66][67][88,89]. circRNAs can modulate miRNA-target expression, acting like miRNA axes. Moreover, they can interact with RNA-binding proteins and regulate cellular processes ^[68][90]. Circular RNA (circRNA) expression in PBMCs of pediatric patients with BCP-ALL has also been evaluated ^[68][90]. After the comparison of circRNA expression profiles in B- and T-cells, and monocyte populations, which was validated by reverse transcriptase quantitative polymerase chain reaction (RT-qPCR), four circRNAs were more highly expressed in healthy donors than in patients with BCP-ALL: circIL4R (p < 0.0001), circZCCH7 (p = 0.0307), and circX (intergenic) (p = 0.076). Furthermore, circPVT1 (p = 0.0002), circHIPK3 (p < 0.0001), circPAX5 (p < 0.0001), and circAFF3 (p = 0.0115) were overexpressed in patients with BCP-ALL. The expression of the target set of circRNAs according to BCP-ALL cytogenetic subtype showed circAFF2 was highly expressed in TCF3-PBX1 (p = 0.0252), BCP-ALL, and to a lesser extent in ETV6- RUNX1 BCP-ALL (p = 0.021). CircBCL2 (intronic) was upregulated in ETV6-RUNX1 fusions (p = 0.0166), circSETBP1 and circX (intergenic) were significantly reduced in MLL rearranged samples (p = 0.0274 and p = 0.0472, respectively), and circIKZF1 was lower in BCP-ABL and hyperdiploid leukemias than in the ETV6-RUNX1 subtype (p = 0.0154), in which the expression was conserved at levels comparable to B-cells. Overall, circRNAs have the potential to regulate specific cell functions, cell differentiation, maturation stages, and the growth of leukemic cells. As circRNA deregulation has been observed in patients with BCP-ALL, based on their cytogenetic subtype and previous cancers, circRNAs could potentially serve as markers for BCP-ALL. However, further research is necessary to determine the exact role that circRNAs play in leukemogenesis and their potential as markers for the disease ^[68][69][90,91]. Figure 1 shows a summary of the biomarkers associated with BCP-ALL grouped as protein markers (Figure 1A), nucleic acid molecules (Figure 1B,C), metabolites (Figure 1D), and according to the different stages of disease.

2.3. Metabolites Type BCP-ALL Biomarkers

Metabolites are the final products of gene expression and the direct outcome of enzymatic and protein activity. Metabolite profiles provide information about tumor microenvironments for ALL ^[30][70][30,92]. These metabolite profiles can be studied by metabolomics. “Metabolomics” involves the quantitative measurement of time of the metabolic responses of multicellular systems to pathophysiological stimuli or genetic modification [17]. Perturbation in the metabolism of patients at different disease stages could provide a unique metabolic signature to monitor treatment outcomes and disease progression ^[71][93]. Metabolism in cancer is a major research area in cancer biology. It examines how metabolic activities are changed in cancer cells compared to normal cells [30]. As shown in Table 1 and Figure 1D, researchers have reported metabolic differences in patients with ALL compared to healthy controls. Musharraf et al. ^[70][92] compared the metabolite profiles in the serum of 96 patients with ALL, AML, and APA, obtained by using the nuclear magnetic resonance spectroscopy technique. The authors reported high lactate levels and low levels of alanine, glutamine, histidine, lysine, valine, and proline. They theorized that patients adopt a secondary metabolic pathway to generate glucose because cancer cells require more glucose than can be provided by glycolysis. These findings are in contrast with those reported by Morad et al. [30], who evaluated the plasma of patients with ALL (n = 14), AML (n = 16), and breast cancer (n = 25). They reported high threonine, proline, glycine, alanine and lysine levels and low lactate levels in the ALL group. These changes could be due to tumor tissue competing for nitrogen compounds found in the amino acid structure ^[30][70][30,92]. Another group found that fatty acids were elevated as reservoirs, corresponding with an accumulation of carnitine, which plays a role in fatty acid metabolism ^[70][92]. Upon closer analysis, there seem to be notable variations in the approaches employed, which could explain the discrepancies in the outcomes. A crucial factor to consider is that the patient selection was comprised of diverse cancer types, like breast cancer and aplastic anemia, which may have influenced the results. Furthermore, though both studies utilized nuclear magnetic resonance spectroscopy, the validation techniques differed. One study used variable importance in projection values, while the other used principal component analysis. To gain a comprehensive understanding of the implications of these disparities, providing further context and specific examples would be highly beneficial. Overall, with a more thoughtful and informed examination, people can better appreciate the nuances of these findings. In a recent study focused on developing a novel photo-diagnostic strategy using fluorescence emission spectra, the authors used plasma or red blood cell extract from 45 patients with AL. They found four metabolites that were altered and correlated: nicotinamide adenine dinucleotide (NAD) + hydrogen (H) (NADH), FAD, tyrosine kinase, and tryptophan (p < 0.05). NADH and tryptophan were decreased in patients with ALL, and there was a connection between tyrosine kinase and tryptophan. However, FAD was increased in patients with AL. Thus, the AL diagnosis can be made based on the FAD, NADH, tyrosine kinase, and tryptophan levels, but the ratios can help to discern between AML and ALL [8].