The median number of losses per tumor was 6.9% clones (range 0.1–24%) and the median number of gained clones was 9.8% (range 1.1–21%). Recurrent gains identified in at least 60% of the tumors involved chromosome regions 5p, 7p, 7q, 8q, 10q, 11q, 12p, 14q, 16p, 17p, 19p, 19q, and 20q. Likely target genes include TERT, EGFR, MYC, MYEOV, CCND1, FGF4, FGF3, CTTN and AKT1. Loss of genetic material in at least 40% of the samples affected 3p, 5q, 8p, 9p and 11q (Table ). A homozygous deletion of 9p21.3, corresponding to the CDKN2A locus, was identified in one case and was verified with PCR using CDKN2A-specific primers (sequences available from the authors upon request, data not shown). High-level amplifications were observed in 33 regions and recurrently involved 11q22 (harbouring PDFG in 3 tumors), 11q13 (in 11 tumors, encompassing MYEOV, CCND1, FGF4, FGF3, PPFIA, FAD, TMEM16A, CTTS and SHANK2) and 7p11 (including EGFR in 4 tumors). Overexpression of EGFR was validated by immunohistochemistry, which revaled a highly positive (3+) staining in 19 tumors, 12 of which had copy number gain of 7p11. All 4 tumors with HLA of the EGFR locus showed 2+ or 3+ EGFR staining.
Table 1 Summary of the most frequent copy number gains and losses. Table listing the most frequent copy number gains and losses, sorted in decreasing order, in relation to tumor stage. Columns give the cytoband, start reporter, end reporter, size (Mbp), and candidate (more ...)
To evaluate whether genes from a same signaling pathway were correlated, a complementary analysis of pairs of genomic loci harboring genes (CDKN2A, MDM2, RB, MDM4, CDKN1A, PIK3CA, PTEN, AKT1, TP53, MYC, CCND1, CCNE1, BCL2, CDK4, E2F3) from central tumorigenic pathways was performed. Significant correlations (P < 0.05) were identified between gain of the MDM4 locus and loss of the RB locus, gain of MDM2 and gain of TP53, gain of BCL2 and loss of CDKN1A, gain of BCL2 and loss of CDKN2A, gain of CCND1 and loss of CDKN2A. Significant negative correlations were found between gain of MDM2 and gain of PIK3CA, gain of PIK3CA and loss of E2F3, loss of PTEN and gain AKT1, and concordant gain of AKT1 and gain of CDK4. Pairwise correlation analysis between members of the EGFR pathway identified significant (P < 0.05) correlations between gains of AKT1/HER3, MAPK1/PIK3CA, and MAPK1/AKT1 and between gain of EGFR and loss of PTEN.
Despite genetic complexity (Figure ), the genomic profiles were found to correlate with tumor stage, differentiation, and development of metastases with a lower (mean 13% versus
25%) number of changes in highly differentiated ESCC than in poorly differentiated tumors (see Additional file 2
). A Chi2 test identified ~400 clones that mapped to 6 genomic regions (2p11.1-2q11.2, 2q35, 3p21.31, 4q12, 4q21.3-4q28.3, 5q12.1) that were significantly (P < 0.05) more often found in poorly differentiated tumors. The chromosomal regions most frequently affected by copy number gains differed between stage I and stage II-IV tumors. Gains of 6p25.3, 12p13.33 and 17p13.1 were present in all stage I tumors. Stage II-IV tumors showed a higher number of changes, with the most common gains affecting 12p13.33 (100%), 5p15.33 (85%), 20q13.33 (85%) and 11q13.3 (77%). Overall, copy number losses were more common in stage II tumors and the most frequent losses were shared by different tumor stages without a significant difference in frequency (see Additional file 3
Figure 1 A) Genome-wide frequency plot of DNA copy number gains (red) and losses (green) for all 30 ESCC tumors. B) Genome-wide copy number profile for a Stage I tumor (ESCC 33). Characteristic alterations for a Stage I tumors are gains on 6p, 12p, and 17p and (more ...)
Smaller (pT1) tumors had the lowest number of copy number changes (mean 12% of the clones affected), whereas pT2-T4 tumors showed alterations of 27%, 16%, and 18% of the clones. When gains/losses were correlated to presence of lymph node metastases, 838 clones that mapped to 21 genomic regions were identified (see Additional file 5
). Copy number gain in 7p22.3 was significantly (P
= 0.01) associated with lymph node metastases, and correctly predicted nodal metastases in 63% of the patients, as represented by an area under ROC curve of 0.73. Changes in 19 regions, corresponding to 1074 clones, correlated to metastasis at diagnosis (see Additional file 6
) with copy number gain of 8q21.3 being significantly associated with the presence of metastasis (P
= 0.03). Gain of 8q21.3 could classify 60% of the patients with distant metastasis with 40% sensitivity and 70% specificity.
Univariate analysis verified that stage and tumor size were associated with prognosis; stage HR 1.6, P
= 0.04 and pT HR 1.8, P
= 0.05. When copy number gains and losses were correlated to prognosis, Cox proportional hazards analysis identified 1284 clones, with a p-value <0.05, mapping to 30 regions (see Additional file 7
). When gains and losses were separately considered, 7 regions remained significantly associated to prognosis in univariate analysis. When these regions were entered with stage into multivariate analysis, gain of 1p36.32 and gain of 19p13.3 independently predicted poor prognosis (Table and Figure ).
Table 2 Regions correlated to prognosis in univariate and multivariate Cox analysis. Table listing the regions significantly correlated to prognosis in univariate and multivariate analysis. Columns give the type of aberration, cytoband, hazard ratio (HR), P-value (more ...)
Figure 2 Kaplan-Meier survival plots of the two prognostic regions in multivariate analysis. A) Highly significant difference in survival between patients without gain and with gain of 1p36.32 (P = 0.005). B) Difference in survival between patients without gain (more ...)