Both ALL and AML groups had higher levels of SOX2 expression as compared to control (P < 0.001 for both) and OCT3/4 expression (P = 0.005 and P = 0.003 respectively) as compared to controls. No significant difference between ALL and AML patient groups was found with regard to SOX2 and OCT3/4 levels (P = 0.128, P = 0.832 respectively) (Table 2).
Table 2 Comparison of SOX2 and OCT3/4 expression levels at diagnosis among studied groupsLevels of SOX2 and OCT3/ expression at remissionAt bone marrow remission the expression levels of both SOX2 and OCT3/4 were significantly downregulated as compared to diagnosis levels and the differences were statistically significant in AML [for SOX2; P = < 0.001 and for OCT3/4; P = < 0.001] and in ALL group [for SOX2; P = < 0.001 and for OCT3/4; P = < 0.001] (Tables 3 and 4) (Fig. 3).
Table 3 Comparison of SOX2 expression at diagnosis, remission, and relapse among leukemic patientsTable 4 Comparison of OCT3/4 expression at diagnosis, remission, and relapse among leukemic patientsFig. 3Kinetic expression of SOX2 in ALL and AML patient groups (A, B) and OCT3/4 in ALL and AML patient groups (C, D). Both ALL and AML patients achieved CR showed significantly lower levels of SOX2 and OCT3/4 compared to levels detected at diagnosis. In relapsed cases, levels of SOX2 and OCT3/4 were significantly higher than levels at CR in both ALL and AML cases
Levels of SOX2 and OCT3/4 expression at relapseIn the AML group; both SOX2 and OCT3/4 were elevated in relapse compared to their level at remission (P = 0.005, P = 0.005 respectively), also in the ALL group, relapsed cases had significantly higher SOX2 and OCT3/4 levels compared to remission (P = 0.008, P = 0.008 respectively) (Tables 3 and 4) (Fig. 3).
Impact of OCT3/4 and SOX2 expressions on acute leukemia patients’ outcomeROC curve was used to evaluate the cut-off value of OCT3/4 and SOX2 as prognostic markers for survival in both ALL and AML. In ALL, ROC curve analysis demonstrated that SOX2 ≥ 16.3% and OCT3/4 ≥ 1.55% identified patients likely to experience death with AUC (area under the curve) of 0.763 (P value = 0.006) sensitivity 70.6% and specificity 66.7% and AUC of 0.758, (P value = 0.007), sensitivity 70.6% and specificity 71.4% respectively (Fig. 4A).
Fig. 4ROC curve (A) to identify the optimal SOX2 and OCT3/4 expression cutoff value that predicts death in ALL (best cut-off for SOX2 > 16.3% with AUC 0.763, P = 0.006, sensitivity 70.6% and specificity 66.7%) and (best cut-off for OCT3/4 > 1.55% with AUC of 0.758, P = 0.007, sensitivity 70.6% and specificity 71.4%). B To identify the optimal SOX2 and OCT3/4 expression cutoff value that predicts death in AML (best cut-off for SOX2 > 21.7% with AUC 0.854, P = < 0.001, sensitivity 88.0% and specificity 82.4%) and (best cut-off for OCT3/4 > 1.65% with AUC of 0.692, P = 0.037, sensitivity 68.0% and specificity 70.6%)
In AML, SOX2 ≥ 21.7% and OCT3/4 ≥ 1.65% identified patients likely to experience death with AUC of 0.854 (P value = < 0.001), sensitivity 88.0% and specificity 82.4% and AUC of 0.692 (P value = 0.037), sensitivity 68.0% and specificity 70.6% respectively (Fig. 4B).
Evaluation of OCT3/4 and SOX2 expression in relation to clinicopathological featuresIn ALL cases; both SOX2 and OCT3/4 markers were significantly elevated in unfavorable cytogenetic risk stratification compared to favorable risk stratification, in non-CR cases compared to CR cases, in relapsed cases compared to non-relapsed cases, and in dead patients compared to living patients. SOX2 was significantly elevated in T-ALL cases compared to B ALL cases (Table 5).
Table 5 Association of SOX2 and OCT3/4 expression with demographic, clinical, and laboratory data of ALL patientsIn AML cases, poor and intermediate cytogenetic risk stratification had higher SOX2 and OCT3/4 levels compared to favorable risk stratification, also in non-CR patients compared to patients achieved CR, in relapsed cases compared to non-relapsed cases, and in dead patients compared to living patients (Table 6).
Table 6 Association of SOX2 and OCT3/4 expression with demographic, clinical, and laboratory data of AML patientsImpact of OCT3/4 and SOX2 expression on OS and DFSBased on ROC curve analysis the acute leukemia patients were classified into 2 groups, ALL patients with high SOX2 expression (16.3%) and OCT3/4 (1.55%) expression. Also, AML patients were classified as those with high SOX2 expression and low SOX2 expression and also those with high OCT3/4 expression and low OCT3/4 expression according to ROC curve cut-off value of SOX2 (21.7%) and a cut-off value of OCT3/4 (1.65%) expression.
Adult ALL patients expressed high SOX2 and OCT3/4 had significantly shorter OS and DFS as compared to those with low expression (P = 0.007; P = 0.016 respectively and P = 0.006 for SOX2) (Fig. 5). Moreover, AML, patients expressing high SOX2 and OCT3/4 had significantly shorter OS as compared to those with low SOX2 (P = < 0.001) and OCT3/4 expression (P = 0.022). As regards DFS, those with high SOX2 and OCT3/4 also had significantly shorter DFS than those with low SOX2 (P = 0.003) and OCT3/4 (< 0.001) (Fig. 6).
Fig. 5Kaplan-Meier curve for the impact of SOX2 expression and OCT3/4 expression on OS and DFS of ALL patients. A, B Patients with low SOX2 expression (< 16.3%) and low OCT3/4 expression (< 1.55%) had longer cumulative OS survival as compared to those with high expression (P = 0.007, and P = 0.016 respectively). C, D Patients with low SOX2 expression (< 16.3%) had longer cumulative DFS as compared to those with high expression (P = 0.006). While DFS was not significantly different between those with high and low OCT3/4 (P = 0.284)
Fig. 6Kaplan-Meier curve for the impact of SOX2 expression and OCT3/4 expression on OS and DFS of AML patients. A, B Patients with low SOX2 expression (< 21.7%) and low OCT3/4 expression (< 1.55%) had longer cumulative OS survival as compared to those with high expression (P < 0.001), and (P = 0.022) respectively). C, D Patients with low SOX2 expression (< 21.7%) and OCTA3/4 had longer cumulative DFS as compared to those with high expression (P = 0.003, P = < 0.001 respectively)
COX regression analysis to identify biological factors that could predict CR, OS, and DFSCOX regression analysis was conducted to predict factor (s) affecting OS and DFS using age, gender, laboratory data, type, risk stratification, response, SOX2, and OCT3/4 as covariates. In ALL cases; non-CR (P = 0.021), high SOX2 expression (P = 0.029) and high OCT3/4 expression (P = 0.018) were significant independent factors for shorter OS in ALL patients in multivariate analysis. Moreover, both high SOX2 expression (P = 0.001) and high OCT3/4 expression (P = 0.025) were significant independent factors for shorter DFS in ALL patients in multivariate analysis (Tables 7 and 8).
Table 7 Cox regression analysis to predict the hazardous factor that could predict OS in ALL patientsTable 8 Cox regression analysis to predict the hazardous factor that could predict DFS in ALL patientsHowever, in AML cases; high SOX2 expression (P = 0.001) and high OCT3/4 expression (P = 0.045) were significant independent factors for shorter OS in AML patients in multivariate analysis. Moreover, both high SOX2 expression (P = 0.010) and high OCT3/4 expression (P = 0.023) were significant independent factors for shorter DFS in AML patients in multivariate analysis (Tables 9 and 10).
Table 9 Cox regression analysis to predict the hazardous factor that could predict OS in AML patientsTable 10 Cox regression analysis to predict the hazardous factor that could predict DFS in AML patients
Comments (0)