Results

Patient characteristics

Among the 4,991 patients with mCRC enrolled in GOZILA between March 2018 and February 2022, tissue RAS and BRAF V600E MT status was confirmed in 1,563 patients before the initiation of mCRC treatment. Among these, 172 patients for whom ctDNA was measured before the administration of first-line chemotherapy were excluded, resulting in 1,391 patients whose data were evaluated. The consolidated standards of reporting trials (CONSORT) flow diagram for the study is shown in Figure 1. The median age at the time of blood sampling was 61.0 (range, 25.0–85.0) years, and 800 (57.5%) patients were male. Overall, 347 (24.9%) had right-sided tumors, and 810 (58.2%) had metastases involving multiple organs (Table 1). The liver was the most frequent site of metastasis (62.0%), followed by the lungs (47.4%), lymph nodes (32.6%), and peritoneum (26.5%). The median number of therapy lines between tissue assessment and ctDNA testing was 2 (range, 1–13). Based on baseline tissue and subsequent ctDNA testing results, patient cohorts were grouped as follows: Patients were initially classified into three cohorts: tissue RAS/BRAF WT (n = 788, 56.7%), tissue RAS MT (n = 478, 34.3%), and tissue BRAF MT (n = 125, 9.0%). Subsequent liquid biopsy (ctDNA) testing revealed dynamic shifts within these groups: persistent RAS WT, n = 567 (40.8%); persistent RAS MT, n = 387 (27.8%); acquired RAS MT_, n_ = 221 (15.9%); Neo_RAS_ WT, n = 91 (6.5%); persistent BRAF MT, n = 102 (7.3%); and Neo_BRAF_ WT, n = 23 (1.7%).

Figure 1: CONSORT diagram of the studyctDNA: circulating tumor DNA, RAS: rat sarcoma viral oncogene homolog, WT: wild-type, MT: mutant, CH: clonal hematopoiesis.

Characteristics | Persistent RAS WT (n = 567) | NeoRAS WT (n = 91) | Acquired RAS MT (n = 221) | Persistent RAS MT (n = 387) | Persistent BRAF MT (n = 102) | NeoBRAF WT (n = 23)
Age at enrollment, years
Median [range] | 60.0 [26.0–85.0] | 60.0 [32.0–79.0] | 62.0 [25.0–85.0] | 62.0 [25.0–85.0] | 59.0 [27.0–82.0] | 63.0 [41.0–75.0]
Sex
Male | 335 (59.1) | 46 (50.5) | 136 (61.5) | 203 (52.5) | 65 (63.7) | 15 (65.2)
Female | 230 (40.6) | 45 (49.5) | 85 (38.5) | 184 (47.5) | 37 (36.3) | 8 (34.8)
Unknown | 2 (0.4) | 0 (0) | 0 (0.0) | 0 (0) | 0 (0) | 0 (0)
Primary site
Right-sided colon | 112 (19.8) | 27 (29.7) | 42 (19.0) | 126 (32.6) | 35 (34.7) | 6 (26.1)
Left-sided colon | 451 (79.5) | 64 (70.3) | 178 (80.5) | 261 (67.4) | 66 (65.3) | 17 (73.9)
Unknown | 4 (0.7) | 0 (0) | 1 (0.5) | 0 (0) | 1 (1.0) | 0 (0)
Cecum | 20 (3.5) | 8 (8.8) | 7 (3.2) | 50 (12.9) | 14 (11.2) | 3 (13.0)
Ascending colon | 46 (8.1) | 15 (16.5) | 13 (5.9) | 58 (15.0) | 13 (10.4) | 2 (8.7)
Transverse colon | 46 (8.1) | 4 (4.4) | 22 (10.0) | 18 (4.7) | 14 (11.2) | 1 (4.3)
Descending colon | 38 (6.7) | 4 (4.4) | 11 (5.0) | 14 (3.6) | 13 (10.4) | 2 (8.7)
Sigmoid colon | 212 (37.4) | 17 (18.7) | 66 (29.9) | 83 (21.4) | 27 (21.6) | 8 (34.8)
Rectum | 201 (35.4) | 43 (47.3) | 101 (45.7) | 164 (42.4) | 43 (34.4) | 1 (4.3)
Unknown | 4 (0.7) | 0 (0) | 1 (0.5) | 0 (0) | 1 (0.8) | 0 (0)
Metastatic site
Single organ | 255 (45.0) | 51 (56.0) | 95 (43.0) | 118 (30.5) | 50 (49.0) | 8 (34.8)
Multi-organ | 312 (55.0) | 39 (42.9) | 126 (57.0) | 267 (69.0) | 52 (51.0) | 15 (65.2)
Unknown | 0 (0.0) | 1 (1.1) | 0 (0) | 2 (0.5) | 0 (0) | 
Liver
Present | 367 (64.7) | 21 (23.1) | 137 (62.0) | 251 (64.9) | 67 (66.3) | 19 (82.6)
Absent | 200 (35.3) | 70 (76.9) | 84 (38.0) | 136 (35.1) | 34 (33.7) | 4 (17.4)
Lung
Present | 229 (40.4) | 51 (56.0) | 102 (46.2) | 239 (61.8) | 30 (29.4) | 9 (39.1)
Absent | 338 (59.6) | 40 (44.0) | 119 (53.8) | 148 (38.2) | 72 (70.6) | 14 (60.9)
Lymph node
Present | 199 (35.1) | 15 (16.5) | 76 (34.4) | 119 (30.7) | 35 (34.3) | 9 (39.1)
Absent | 368 (64.9) | 76 (83.5) | 145 (65.6) | 268 (69.3) | 67 (65.7) | 14 (60.9)
Peritoneum
Present | 136 (24.0) | 38 (41.8) | 62 (28.1) | 95 (24.5) | 32 (31.4) | 5 (21.7)
Absent | 431 (76.0) | 53 (58.2) | 159 (71.9) | 292 (75.5) | 70 (68.6) | 18 (78.3)
Bone
Present | 45 (7.9) | 2 (2.2) | 14 (6.3) | 35 (9.0) | 5 (4.9) | 2 (8.7)
Absent | 522 (92.1) | 89 (97.8) | 207 (93.7) | 352 (91.0) | 97 (95.1) | 21 (91.3)
Other | 66 (11.6) | 12 (13.2) | 30 (13.6) | 38 (9.8) | 12 (11.8) | 3 (13.0)
Treatment history
Anti-VEGF antibody | 496 (87.5) | 75 (82.4) | 201 (91.0) | 259 (66.9) | 84 (82.4) | 19 (82.6)
Anti-EGFR antibody | 503 (88.7) | 0 (0) | 193 (87.3) | 4 (1.0) | 60 (58.8) | 10 (43.5)
Treatment lines at the time of sampling
First | 190 (33.5) | 36 (39.6) | 46 (20.8) | 104 (26.9) | 49 (48.0) | 13 (56.5)
Second | 115 (20.3) | 28 (30.8) | 44 (19.9) | 107 (27.6) | 24 (23.5) | 4 (17.4)
Third | 113 (19.9) | 14 (15.4) | 58 (26.2) | 68 (17.6) | 16 (15.7) | 1 (4.3)
Fourth | 63 (11.1) | 6 (6.6) | 38 (17.2) | 56 (14.5) | 6 (5.9) | 4 (17.4)
Fifth | 42 (7.4) | 3 (3.3) | 21 (9.5) | 24 (6.2) | 6 (5.9) | 0 (0.0)
Sixth or later | 44 (7.9) | 4 (4.4) | 35 (15.9) | 28 (7.3) | 7 (6.9) | 0 (0.0)
First, second, and third line | 418 (73.7) | 78 (85.7) | 148 (67.0) | 279 (72.1) | 89 (87.3) | 19 (82.6)
Fourth line or later | 149 (26.3) | 13 (14.3) | 73 (33.0) | 108 (27.9) | 13 (12.7) | 4 (17.4)
cfDNA fraction
≥1.0% | 442 (78.0) | 10 (11.0) | 202 (91.4) | 309 (79.8) | 85 (83.3) | 10 (43.5)
<1.0% | 125 (22.0) | 81 (89.0) | 19 (8.6) | 78 (20.2) | 17 (16.7) | 13 (56.5)

Association of clinicopathological characteristics with molecular profiles

In the full Neo_RAS_ WT cohort (Group A), the prevalence of liver (23.1%, p < 0.001), lymph node (16.5%, p < 0.001), and multi-organ metastasis (42.9%, p < 0.001) was lower than that in the other cohorts. However, lung (56.0%, p < 0.001) and peritoneal (41.8%, p = 0.004) metastases were more common than those in the other cohorts (Table 1).

Similarly, in the Neo_RAS_ WT (Group B), the prevalence of liver (35.7%, p < 0.001), lymph node (21.4%, p < 0.001), and multi-organ metastasis (45.2%, p < 0.001) was lower; lung (61.9%, p < 0.001) and peritoneal (31.0%, p = 0.004) metastases were more common than those in other cohorts. (Table S1). Treatment with anti-EGFR antibody was most common among patients without RAS MT on initial tissue testing (Table 1). Notably, the frequency of sampling during first-to third-line treatment was significantly higher in patients with BRAF MT at initial tissue testing (Table 1). Furthermore, the proportion of patients with a maximum ctDNA variant allele frequency (VAF) of ≥1% was significantly lower in the Neo_RAS_ group, followed by the Neo_BRAF_ WT group (Table 1).

Incidence of gene alteration in each cohort

The incidences of specific baseline (tissue) RAS MT and acquired (ctDNA) RAS MT are shown in Figure S2A. The most common baseline KRAS MT was in codons 12 and 13, detected in 71.3 and 16.3% of patients, respectively. At disease progression, KRAS codon 12 and 13 mutations were detected in 51.4 and 10.5% of samples, respectively. The most commonly acquired RAS MT was KRAS codon 61 (55.5%), followed by NRAS codon 61 (18.2%). At the time of disease progression, the maximum VAF of any RAS MT in patients with acquired RAS MT was significantly lower than in those with persistent RAS MT (0.79% vs. 6.9%, p < 0.0001; Figure S2B). Furthermore, the relative clonality of RAS MT in the acquired RAS MT cohort was significantly lower than that in the persistent RAS MT cohort (8.9% vs. 80.2%, p < 0.0001; Figure S2C). However, no significant difference was observed in terms of the proportion of patients with a maximum ctDNA VAF (any gene) of ≥1% between the acquired RAS MT and persistent RAS MT groups (Table 1). The proportion of samples with multiple KRAS MT detected in ctDNA was also significantly higher in patients with acquired RAS MT (42.7% vs. 4.1%, p < 0.0001).

The incidence of non-RAS genetic alterations in each group is presented in Table 2. The acquired RAS MT cohort had a significantly higher rate of genetic alterations than that in the other groups, especially in genes related to receptor tyrosine kinases (EGFR, FGFR2, and MET), the RAS-RAF-MAPK pathway (MAPK_1-3), and the PI3K-AKT-mTOR pathway (PTEN, mTOR, TSC1, and PDGFR). Furthermore, the persistent BRAF MT cohort exhibited a significantly higher rate of genetic alterations (BRCA1, SMAD4, and TP53) and tended to have a higher rate of genetic alterations (ARIDA1, PIK3CA, and MYC amplification) than those in the Neo_BRAF WT group.

Gene alteration | Persistent RAS WT (n = 567) | NeoRAS WT (n = 91) | Acquired RAS MT (n = 221) | Persistent RAS MT (n = 387) | Persistent BRAF MT (n = 102) | NeoBRAF WT (n = 23)
Mutations
Receptor tyrosine kinase
EGFR | 112 (19.8) | 6 (6.6) | 85 (38.5) | 44 (11.4) | 60 (58.8) | 10 (43.5)
ERBB2 | 54 (9.5) | 3 (3.3) | 18 (8.1) | 24 (6.2) | 7 (6.9) | 0 (0)
FGFR2 | 26 (4.6) | 2 (2.2) | 25 (11.3) | 32 (8.3) | 5 (4.9) | 0 (0)
MET | 47 (8.3) | 4 (4.4) | 28 (12.7) | 34 (8.8) | 6 (5.9) | 2 (8.7)
RAS-RAF-MEK pathway
MAP2K1 | 45 (7.9) | 1 (1.1) | 46 (20.8) | 5 (1.3) | 3 (2.9) | 1 (4.3)
MAP2K2 | 5 (0.9) | 1 (1.1) | 9 (4.1) | 1 (0.3) | 2 (2.0) | 0 (0)
MAPK3 | 6 (1.1) | 0 (0.0) | 8 (3.6) | 5 (1.3) | 1 (1.0) | 0 (0)
PI3K-AKT-mTOR pathway
PIK3CA | 81 (14.3) | 2 (2.2) | 48 (21.7) | 98 (25.3) | 22 (21.6) | 2 (8.7)
PTEN | 28 (4.9) | 1 (1.1) | 19 (8.6) | 20 (5.2) | 8 (7.8) | 1 (4.3)
mTOR | 0 (0.0) | 2 (2.2) | 3 (1.4) | 1 (0.3) | 0 (0) | 0 (0)
TSC1 | 11 (1.9) | 0 (0.0) | 10 (4.5) | 6 (1.6) | 2 (2.0) | 0 (0)
PDGFR | 37 (6.5) | 0 (0.0) | 37 (16.7) | 41 (10.6) | 0 (0) | 0 (0)
Other pathways
SMAD4 | 58 (10.2) | 3 (3.3) | 38 (17.2) | 65 (16.8) | 26 (25.5) | 1 (4.3)
MYC | 18 (3.2) | 0 (0.0) | 16 (7.2) | 16 (4.1) | 5 (4.9) | 2 (8.7)
FBXW7 | 20 (3.5) | 0 (0.0) | 19 (8.6) | 38 (9.8) | 7 (6.9) | 0 (0)
ARID1A | 53 (9.3) | 1 (1.1) | 33 (14.9) | 44 (11.4) | 13 (12.7) | 0 (0)
TP53 | 403 (71.1) | 29 (31.9) | 183 (82.8) | 289 (74.7) | 86 (84.3) | 6 (26.1)
BRCA1 | 41 (7.2) | 3 (3.3) | 36 (16.3) | 29 (7.5) | 16 (15.7) | 0 (0)
Amplifications
EGFR | 169 (29.8) | 1 (1.1) | 101 (45.7) | 76 (19.6) | 18 (17.6) | 5 (21.7)
ERBB2 | 66 (11.6) | 1 (1.1) | 9 (4.1) | 3 (0.8) | 7 (6.9) | 0 (0)
FGFR1 | 64 (11.3) | 1 (1.1) | 40 (18.1) | 21 (5.4) | 5 (4.9) | 0 (0)
MET | 29 (5.1) | 0 (0.0) | 19 (8.6) | 1 (0.3) | 6 (5.9) | 0 (0)
KRAS | 32 (5.6) | 0 (0.0) | 22 (10.0) | 10 (2.6) | 2 (2.0) | 1 (4.3)
NRAS | 7 (1.2) | 0 (0.0) | 15 (6.8) | 5 (1.3) | 1 (1.0) | 0 (0)
MTOR | 14 (2.5) | 2 (2.2) | 16 (7.2) | 12 (3.1) | 5 (4.9) | 0 (0)
MYC | 38 (6.7) | 1 (1.1) | 24 (10.9) | 17 (4.4) | 13 (12.7) | 0 (0)

Association of RAS and BRAF profile with OS

The median OS was 45.3 months (42.5–47.3 months) among all patients included in the study. Among patients stratified by tissue RAS/BRAF status, those with tissue RAS/BRAF WT (n = 788) had a median OS of 49.4 months, which served as the reference. Those with tissue RAS MT (n = 478) and BRAF MT (n = 125) were associated with a significantly shorter OS (median 41.0 and 25.4 months, respectively; hazard ratio (HR), 1.30 and 2.52, respectively; 95% CI, 1.11–1.52 p Log rank = 0.001 and 1.95–3.24, p Log rank < 0.0001 respectively Figure 2).

Figure 2: Kaplan-Meier plots of overall survival based on gene alteration (tissue RAS and BRAF) statusRAS: rat sarcoma viral oncogene homolog, BRAF: Raf murine sarcoma viral oncogene homolog B, MT: mutant, WT: wild type.Statistical text for Figure 2: Data are evaluated using the Kaplan-Meier method, and survival outcomes between cohorts are compared using the two-sided log rank test. represents the number of individual patients in each cohort. The exact values are reported in the figure. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using univariable Cox proportional hazard regression models.

Among patients stratified by treatment exposure, those treated with anti-EGFR mAb-containing regimens had a median OS of 47.1 months, which served as the reference. Patients treated with regimens without anti-EGFR mAb had a significantly shorter OS (median 41.1 months; HR, 1.25; 95% CI, 1.08–1.45; p Log rank = 0.003). Similarly, patients treated with anti-VEGF mAb-containing regimens had a median OS of 45.9 months, which served as the reference. Patients treated with regimens without anti-VEGF mAb had a significantly shorter OS (median 34.5 months; HR, 1.62; 95% CI, 1.27–2.06; p Log rank < 0.0001).

Among patients stratified by the line of therapy at the time of ctDNA measurement, those treated in lines 1–3 had a median OS of 39.0 months, which served as the reference. Outperforming expectations, patients treated in line 4 or later demonstrated a significantly longer OS, with a median OS of 60.2 months (HR, 0.56; 95% CI, 0.47–0.66; p Log rank < 0.0001).

Based on post-treatment ctDNA testing results, those in the persistent RAS WT cohort had a median OS of 51.7 months (95% confidence interval [CI] 47.6–55.1 months), which was used as a reference for statistical comparisons. Patients with Neo_BRAF_ WT had a shorter median OS, although not significant, than those with persistent RAS WT (49.4 months; HR, 1.08; 95% CI, 0.80–1.55; p Log rank = 0.79; Figure 3); Similar outcomes were observed for Neo_RAS_ WT Group A (45.6 months; HR, 1.12; 95% CI, 0.80–1.55; p Log rank = 0.49; Figure 3) and Neo_RAS_ WT Group B (47.1 months; HR, 1.12; 95% CI, 0.71–1.77; p Log rank = 0.61; Figure S3). In contrast, OS was significantly shorter in patients in the acquired RAS MT cohort (median 43.0 months; HR, 1.23; 95% CI, 1.00–1.54; p Log rank = 0.0497), the persistent RAS MT cohort (median 41.0 months; HR, 1.45; 95% CI, 1.20–1.76; p Log rank < 0.0001; Figure 3), and the persistent BRAF MT cohort (median 22.5 months; HR, 3.05; 95% CI, 2.03–4.56; p Log rank < 0.0001; Figure 3).

Figure 3: Kaplan-Meier plots of overall survival based on gene alteration (RAS and BRAF in pretreatment tissue and post treatment ctDNA) statusNeoRAS and NeoBRAF cohorts were defined as those with tissue-confirmed pretreatment RAS and BRAF MTs and no post-treatment these mutations detected by ctDNA analysis. RAS: rat sarcoma viral oncogene homolog, MT: mutant, WT: wild-type, BRAF: Raf murine sarcoma viral oncogene homolog B.Statistical text for Figure 3: Data are evaluated using the Kaplan-Meier method, and survival outcomes between cohorts are compared using the two-sided log rank test. represents the number of individual patients in each cohort. The exact values are reported in the figure. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using univariable Cox proportional hazard regression models.

Compared with the other cohorts, Neo_RAS_ WT Group A showed no significant difference in OS relative to the acquired RAS MT (HR, 0.92; 95% CI, 0.66–1.28; p Log rank = 0.61) and persistent RAS MT (HR, 0.79; 95% CI, 0.59–1.07; p Log rank = 0.13) cohorts. However, OS was significantly longer than for patients with BRAF MT (HR, 0.47; 95% CI, 0.32–0.67; p Log rank < 0.0001). Neo_RAS_ WT Group B also showed no significant difference in OS relative to the acquired RAS MT (HR, 0.91; 95% CI, 0.59–1.43; p Log rank = 0.66) and persistent RAS MT (HR, 0.78; 95% CI, 0.52–1.16; p Log rank = 0.19) groups. However, OS was significantly longer than that for patients with persistent BRAF MT (HR, 0.48; 95% CI, 0.32–0.73; p Log rank = 0.0016).

Patients with Neo_BRAF_ WT had a significantly longer OS than those with persistent BRAF MT (49.4 months vs. 22.5 months; HR, 0.31; 95% CI, 0.18–0.55; p Log rank = 0.0037). In contrast, acquired BRAF V600E was detected in 9.5% (54/567) of patients with RAS WT, and these patients had a shorter median OS than those with RAS WT whose tumors remained BRAF V600 WT (46.1 vs. 51.9 months; HR, 1.46; 95% CI, 0.96–2.23; p Log rank = 0.04).

To assess the robustness of our findings, a sensitivity analysis was performed using multiple ctDNA fraction cutoff values (0.5%, 1%, 2%, 5%, and 10%). All cutoffs demonstrated significant OS discrimination (HR:1.43, 1.39, 1.35, 1.47, and 1.64, respectively). A threshold of 1% was selected for subsequent analyses as it aligned with the previously reported analytical sensitivity of the Guardant360® assay (97.5% at VAF ≥1%)23 and showed meaningful prognostic discrimination (median OS: 51.1 months vs. 43.7 months; HR, 1.39; 95% CI, 1.17–1.64; p < 0.0001; Figure S4).

Multivariable analysis results for OS

In univariable Cox proportional hazards analysis, BRAF or RAS MT in tissue prior to treatment, Neo_BRAF_ WT, ctDNA fraction, treatment lines at the time of sampling, and history of anti-EGFR or anti-VEGF antibodies influenced OS. In multivariable analysis, the following factors were independently associated with shorter OS: BRAF MT in tissue before treatment (HR, 2.75; 95% CI, 2.10–3.61; p < 0.001); ctDNA fraction ≥1.0% (HR, 1.58; 95% CI, 1.31–1.91; p < 0.001); RAS MT in tissue before treatment (HR, 1.31; 95% CI, 1.00–1.70; p = 0.049); and treatment lines at the time of sampling (HR, 1.77; 95% CI, 1.50–2.10; p < 0.001). History of anti-VEGF antibodies (0.70, 95% CI, 0.55–0.90, p < 0.001) and Neo_BRAF_ WT (HR, 0.37; 95% CI, 0.16–0.85; p = 0.019) were associated with longer OS (Table 3).

Overall survival | Univariable analysis | Multivariable analysis
HR | Lower 95% CI | Upper 95% CI | p value | HR | Lower 95% CI | Upper 95% CI | p value
Sex (Female∗ or Male) | 0.9 | 0.78 | 1.05 | 0.21 |  |  |  | 
Age (<65∗ or ≥65 years) | 0.95 | 0.61 | 1.5 | 0.84 |  |  |  | 
Primary tumor location (left∗ or right) | 1.12 | 0.94 | 1.32 | 0.2 |  |  |  | 
Liver metastasis (negative∗ or positive) | 1.04 | 0.90 | 1.21 | 0.59 |  |  |  | 
Lung metastasis (negative∗ or positive) | 0.99 | 0.86 | 1.16 | 0.99 |  |  |  | 
Lymph node metastases (negative∗ or positive) | 0.86 | 0.73 | 1.01 | 0.059 |  |  |  | 
Peritoneal metastasis (negative∗ or positive) | 0.91 | 0.77 | 1.08 | 0.28 |  |  |  | 
Multiple organ metastases (negative∗ or positive) | 0.96 | 0.83 | 1.12 | 0.63 |  |  |  | 
RAS mutation in tissue prior to therapy initiation (negative∗ or positive) | 1.17 | 1.006 | 1.37 | 0.042 | 1.31 | 1 | 1.70 | 0.049
Acquired RAS mutation (negative∗ or positive) | 1.07 | 0.87 | 1.32 | 0.51 |  |  |  | 
Neo RAS WT (negative∗ or positive) | 0.78 | 0.56 | 1.08 | 0.13 |  |  |  | 
BRAF V600E mutation in tissue prior to therapy initiation (negative∗ or positive) | 2.28 | 1.79 | 2.92 | <0.001 | 2.75 | 2.1 | 3.61 | <0.001
Neo BRAF WT (negative∗ or positive) | 0.31 | 0.13 | 0.72 | 0.006 | 0.37 | 0.16 | 0.85 | 0.019
Treatment history of anti-VEGF antibody (no∗ or yes) | 0.61 | 0.48 | 0.79 | <0.001 | 0.7 | 0.55 | 0.9 | <0.001
Treatment history of anti-EGFR antibody (no∗ or yes) | 0.8 | 0.69 | 0.93 | 0.0028 | 0.92 | 0.72 | 1.19 | 0.53
Treatment lines at the time of sampling (1st, 2nd, and 3rd∗ or 4th or later) | 1.8 | 1.52 | 2.12 | <0.001 | 1.77 | 1.5 | 2.1 | <0.001
ctDNA fraction (<1.0%∗, ≥1.0%) | 1.39 | 1.15 | 1.67 | <0.001 | 1.58 | 1.31 | 1.91 | <0.001