Genomic landscapes of glioma at various stages under first-line treatment
Sequencing data from primary tumors and TISF-DNA in glioma patients, The genome of primary tumor (n = 60) represents the characteristics of glioma before surgery, 0-35d after surgery (n = 20) represents the genomic characteristics of minimal residual disease in tumor lumen at the early stage after surgical resection, 36-120d after surgery (n = 19) represents the genomic characteristics of glioma undergoing chemotherapy (low grade) and chemoradiotherapy (high grade), and 120d after surgery (n = 38) represents genomic characteristics of further chemotherapy during treatment, and radiographic tumor recurrence was found in 26 patients at the time of sampling.
In general, the genomes are different at different stages of the treatment process, but there is some similarity. For the primary tumor (Figure 1A), the mutation rates were TP53 (44%), IDH1 (39%), PTEN (24%), CIC (17%), EGFR (17%), NF1 (13%), ATRX (11%), NOTCH1 (11%), PIK3CA (11%), PIK3R1 (11%). At the stage of minimal residual disease 0-35d after surgery (Figure 1B), mutation rates were higher in NF1 (45%), SETD2 (45%), CIC (40%), TP53 (40%), BRCA2 (30%), GNAS (30%), PIK3R1 (30%), PTEN (30%), TSC1 (30%), TSC2 (30%). After treatment 36-120 days after surgery (Figure 1C), mutation rates were higher in NF1 (53%), TP53 (53%), TSC2 (47%), PTCH1 (42%), BRCA1 (37%), FAT1 (37%), BRCA2 (32%), EGFR (32%), NOTCH1 (32%), SETD2 (32%). 120d after surgery (FIG. 1D), the mutation rates were TP53 (37%), SETD2 (39%), IDH1 (26%), NF1 (26%), NOTCH1 (26%), RELA (24%), EGFR (16%), FAT1 (21%), GNAS (21%), TP53 (37%), SETD2 (39%), IDH1 (26%), NF1 (26%), NOTCH1 (26%). TSC2 (21%). For radiographic recurrence (FIG. 1E), the mutation rates were TP53 (58%), NF1 (38%), SETD2 (38%), EGFR (35%), FAT1 (31%), PTEN (31%), RELA (31%), TSC2 (31%), BCOR (27%), CIC (27%). Gliomas at different stages were highly heterogeneous during postoperative treatment, with only 15.75% shared mutation rate and 84.25% private mutation rate (Figure 1F).
Genomic VAF changes under first-line treatment
We found that VAF(Variant Allele Fraction, In the primary tumor tissue, there was such a high VAF(Figure 2A),97.37% of the clones had a VAF value greater than 1%, and only 2.27% of the low-frequency clones. However, in the early stage after tumor resection (within 35d), 87.84% of the low-frequency clones had a VAF value less than 1% (Figure 2B), and only 12.16% of the clones with a VAF value greater than 1% With further postoperative treatment (36-120 days), mutations with VAF greater than 1% increased slightly (Figure 3C) to 14.29% 120 days after surgery, clone mutations with VAF greater than 1% were significantly increased, accounting for 30.42%(Figure D). In TISF samples from patients with recurrence, clone mutations with VAF greater than 1% were found to be slightly higher overall, accounting for 30.84% (Figure E, Figure G, p < 0.0001). Overall, tumor recurrence was associated with an increased frequency of genomic cloning after glioma surgery (Figure F, G). In addition, we also found that patients with higher VAF within 35 days after surgery may be associated with postoperative residual. Imaging showed that patients with significant residual VAF were larger than those with complete radiographic resection (Figure 2H, p = 0.016), while patients with radiographic recurrence during postoperative treatment had higher VAF than those without recurrence (Figure 2H, p < 0.0001, p < 0.0001).
Changes in specific gene types in glioma under first-line treatment
Polyclonal mutations, MMR-related mutations of mismatch repair genes (MSH2, MSH3, MSH6, MLH1, MLH3, PMS1 and PMS2) were summarized. The accumulation of G:C>A:T (transitions at non-CPg sites in hypermutated gliomas after exposure to alkylating Agents17,18), and VAF greater than 5% of the genome changes (Figure 3). In general, Polyclonal mutation, MMR related mutation and gene clone with VAF > 5% increased gradually in gene clone with postoperative treatment and tumor progression after glioma resection, and the proportion of patients also increased gradually. Temozolomide-associated hypermutation did not detect secondary phenomena in gene clones and patients (Figure 3C, 3D). Therefore, MMR and Polyclonal mutations may be associated with glioma recurrence after treatment, which is consistent with previous studies.
Genomic characterization of glioma patients was continuously monitored
TISF samples from 15 patients in this study were obtained by continuous monitoring, TISF-1 (postoperative 35 days),TISF-2 (postoperative 36-120 days), and TISF-3 (postoperative 120 days), and they were divided into two groups: non-recurrence group (n = 4) and recurrence group (n = 11). Similar to the above conclusions, low frequency distribution was observed in both groups at the early stage after glioma resection (Figure 4A,4B), but increased frequency of some genes in TISF-3 was observed in the recurrence group, while not in the non-recurrence group With the postoperative progression of glioma, there were more private mutations in TISF-DNA than in the primary tissue (Figure 4C), with a maximum of 156 mutations in TISF-3 Mutations shared between TISF and primary tissue decreased gradually, with the highest in TISF-1 and decreased with tumor progression (Figure 4D). A lower proportion of mutations shared between TISF-1 and TISF-3 was also found in TISF-DNA samples (Figure 4E). This suggests that the minimal residual disease in the early postoperative stage may be closer to the primary tumor, while the tumor at recurrence is more heterogeneous with the primary tumor.
Six patients underwent surgical resection after radiographic recurrence, and 129 gene loci in the recurrent TISF were not detected in the recurrent tissue (Figure 5A), and only 32 loci were detected in the recurrent tissue. There were only 17 identical mutation loci in the primary tumor and the recurrent tumor, with a consistency rate of only 30.36%. There was no significant correlation between the frequency of mutations in the genes consistent with the primary and recurrent tissues (Figure 5B, P = 0.5912, R2 = 0.01468), and even there was no significant correlation between the frequency of mutations shared between TISF samples from 26 recurrent patients and their tissues (Figure. 5C, P = 0.8987, R2 = 0.0004102), but a positive correlation was found between the frequency of shared mutations in relapsed tissues and relapsed TISF (Figure 5D, P < 0.0001, R2 = 0.8737). Shared mutant genes (Figure 5F1-F6), most of which were up-regulated and only 3 loci were down-regulated (Figure 5E). Patients with private mutated genes in primary and recurrent TISF (n = 5, Figure. 5H1-H5) were significantly different in gene frequency (Figure 5G), indicating the heterogeneity of tumor evolution under therapeutic pressure.
Clinical relevance of genomic changes in glioma
Based on the results of this study, we summarized the correlation between genomic changes after glioma surgery and clinical practice. We found that the elevation of VAF predates radiographic findings (Figure 6A, Patient 24, patient 31, and patient 34), indicating that tumor DNA-relapse may be present when radiographic findings are not positive and is an ultra-early manifestation of relapse. We performed a second tumor resection for 6 patients with radiological manifestations of recurrence, and only 3 patients were found to have pathological manifestations of tumor recurrence, with polyclonal mutations and high VAF values in their recurrent TISF. Two patients presented with extensive inflammatory and necrotic tissue, and no clones were detected in their TISF (Figure 6B). Finally, we summarized the evolution model in the case of glioma in the first-line treatment (Figure 6C), the first primary tumors is surgical resection, and its leading cloning is cleared, the postoperative residual disease stages exist a large number of low frequency and cloning of mutation, in the treatment of postoperative line pressure part gene mutation in cloning advantage frequency increased, the imaging may not appear at this time a positive result, As the tumor evolved further, it became more and more obvious that the dominant clone would lead to tumor recurrence.