Characteristics and alterations of resistance in the overall STING cohort
Here, we performed ctDNA molecular profiling to assess the incidence of genomic aberrations involved in acquired resistance and their impact on the outcome of patients treated with targeted therapies in the advanced setting. To identify candidate mechanisms of acquired resistance, in each case, cell-free DNA (cfDNA) sequencing data were compared to sequencing data from pre-treatment tumor tissue to identify emergent alterations 3 (see Supplementary Methods).
Therefore, we evaluated a prospective cohort of 626 patients with advanced cancer who were treated with a targeted therapy and enrolled in an institutional molecular profiling program (STING, NCT04932525, sponsor: Gustave Roussy, Villejuif, France) to obtain a molecular profile based on ctDNA profiling (Fig. 1A). Their characteristics are described in Table 1. ctDNA profiling identified at least one previously validated resistance alteration in 193 patients out of 626 (31%). Notably, 86 patients (14%) exhibited > 1 detectable resistance alteration (range 2–16, median 3 per patient), suggesting frequent and profound tumor heterogeneity associated with acquired resistance (Fig. 1B and Supplementary Table 1). Patients with at least 1 emerging ctDNA alteration detected had significantly lower median overall survival (OS) than patients with no mechanism of resistance: 16.2 months (95% CI 14.8–18.3) vs. 10.2 months (6.5–13.9) (p < 0.001).
Incidence of ctDNA emerging alterations in patients with advanced cancer and treated with targeted therapy (A) 626 patients enrolled in the STING study and treated with a targeted therapy had ctDNA profiling. (B) Incidence of ctDNA emerging alterations in the whole cohort. (C) Incidence of ctDNA emerging alterations in patients with advanced colorectal cancer and receiving an anti-EGFR monoclonal antibody. (D) Kaplan-Meier curve of overall survival in colorectal cancer patients according to presence (red curve) or not (blue line) of at least one ctDNA emerging alteration. (E) Incidence of ctDNA emerging alterations in patients with advanced prostate cancer and receiving an anti-androgen hormonal therapy. (F) Kaplan-Meier curve of overall survival in prostate cancer patients according to presence or not of at least one ctDNA emerging alteration. (G) Incidence of ctDNA emerging alterations in patients with advanced breast cancer and receiving an anti-oestrogen hormonal therapy. (H) Kaplan-Meier curve of overall survival in breast cancer patients according to presence or not of at least one ctDNA emerging alteration (I) Incidence of ctDNA emerging alterations in patients with advanced non-small cell lung cancer and receiving an EGFR tyrosine kinase inhibitor. (J) Kaplan-Meier curve of overall survival in non-small cell lung cancer patients according to presence or not of at least one ctDNA emerging alteration
Emerging alterations in colorectal cancer patients: insights from subgroup analysis in the STING and BIP studies
We then analyzed the patterns and prognostic impact of ctDNA alterations in 4 specific subgroups of patients enrolled in the STING study: KRAS/BRAF wild-type (WT) colorectal cancer patients treated with anti-EGFR monoclonal antibodies, prostate cancer patients treated with antiandrogen therapy, hormone receptor-positive breast cancer patients treated with an aromatase inhibitor-based regimen and EGFR-mutated NSCLC patients treated with an anti-EGFR tyrosine kinase inhibitor. To confirm the findings observed in the STING study, we analyzed the same subgroups of patients enrolled in an independent precision medicine study (BIP, NCT02534649, sponsor: Institut Bergonié, Bordeaux, France). Their characteristics are described in Table 1.
In the STING study, ctDNA sequencing allowed the identification of genetic alterations involved in acquired resistance to EGFR-specific antibodies in 17 patients (38%) out of 45 with colorectal cancer (Fig. 1C and Supplementary Table 1). The most frequent emergent aberrations observed in our cohort of patients were RAS pathway mutations in 10 patients (22%). Among them, 5 patients (11%) harbored ≥ 2 variants (up to eight), and 2 also had KRAS amplification (13.3%). Other emergent alterations associated with resistance to anti-EGFR treatment observed in our cohort included EGFR mutations (n = 6, with 3 patients harboring ≥ 2 variants), MET amplification/mutation (n = 4), PI3KCA mutation (n = 3), HER3 mutation (n = 1), and BRAF fusion (n = 1). Overall, 9 patients (20%) exhibited polyclonal resistance. The median OS time was significantly lower in patients with at least one identified emerging ctDNA alteration: 6.7 (95% CI 4.2–9.1) vs. 10.8 months (95% CI 7.8–13.8), p = 0.04 (Fig. 1D). Similar patterns were observed in the BIP cohort, with 15 patients (24%) out of 62 harboring genetic aberrations involved in resistance; KRAS and EGFR gene alterations were the two most frequent. Resistance was polyclonal in 8 patients (12.9%) (Fig. 2A and Supplementary Table 2). The median OS time was like that observed in the STING study: 4.8 (95% CI 1.7–7.9) vs. 11.4 months (95% CI 5.8–17) in patients with no mechanism of resistance or at least one identified mechanism of resistance, respectively, p = 0.004 (Fig. 2B).
Incidence of ctDNA emerging alterations in patients with advanced cancer and treated with targeted therapy in the BIP study. (A) Incidence of ctDNA emerging alterations in patients with advanced colorectal cancer and receiving an anti-EGFR monoclonal antibody. (B) Kaplan-Meier curve of overall survival in colorectal cancer patients according to presence or not of at least one ctDNA emerging alteration. (C) Incidence of ctDNA emerging alterations in patients with advanced prostate cancer and receiving an anti-androgen hormonal therapy. (D) Kaplan-Meier curve of overall survival in prostate cancer patients according to presence or not of at least one ctDNA emerging alteration. (E) Incidence of ctDNA emerging alterations in patients with advanced breast cancer and receiving an anti-oestrogen hormonal therapy. (F) Kaplan-Meier curve of overall survival in breast cancer patients according to presence or not of at least one ctDNA emerging alteration (G) Incidence of ctDNA emerging alterations in patients with advanced non-small cell lung cancer and receiving an EGFR tyrosine kinase inhibitor. (H) Kaplan-Meier curve of overall survival in breast cancer patients according to presence or not of at least one ctDNA emerging alteration
Emerging alterations in prostate cancer patients: insights from subgroup analysis in the STING and BIP studies
At least one emerging ctDNA alteration involved in resistance to anti-androgen therapy was identified in 94 patients with prostate cancer (50%) out of 188 included in the STING study. AR amplifications were identified in 57 patients (60.6% of AR-altered patients and 30.3% overall) included in the STING cohort (Fig. 1E and Supplementary Table 1). A mutation or rearrangement in AR was identified in 59 patients (31.4%). Among them, 28 patients (14.9%) harbored ≥ 2 variants (up to five), and 25 harbored an AR amplification (13.3%). Hotspots are essentially located in the ligand binding domain and are involved in resistance to androgen receptor signaling inhibitors. The most frequent mutations included W742C/F (bicalutamide resistance) (n = 4), H875Y (n = 9), F877L (n = 3) and T878A (n = 25) (bicalutamide/enzalutamide/apalutamide resistance and promiscuous activation by progesterone), and L702H (resistance to abiraterone/enzalutamide, as well as the AR proteolysis-targeting chimera ARV-110, and activation by corticosteroids) (n = 25). Less common AR resistance mutations were V716M (n = 3), G689A (n = 1) and D891H (n = 2); such mutations have been previously described in ctDNA from patients with metastatic castration-resistant prostate cancer (mCRPC) who progressed on bicalutamide or abiraterone. Additional AR gene rearrangements that consisted of deletions or inversions impacting the genomic segment containing AR exons 5–7 and have already been reported in patient-derived xenograft models were identified in 5 patients. Overall, polyclonal resistance was observed in 40 patients (21%). The median OS time was significantly lower in patients with at least one identified emerging ctDNA alteration than in those with no identified alterations: 7.9 months vs. not reached (95% CI 5.5–10.3), p < 0.001 (Fig. 1F). In the BIP study, at least one emerging ctDNA alteration involved in resistance to anti-androgen therapy was identified in 28 patients (25%) out of 112. (Fig. 2C and Supplementary Table 2). AR amplifications were identified in 27 patients (61.4% of AR-altered patients and 24.5% overall). As observed in the STING study, the most frequent mutation observed was AR T878A. Among patients with mutation or rearrangement in AR, 5 patients (4.5%) harbored ≥ 2 variants (up to five) and 2 harbored an AR amplification (1.9%). Polyclonal mechanisms of resistance were identified in 10 patients (9%). Median OS was not reached in patients with no emerging mutation or patients with at least one emerging ctDNA alteration, but the survival time was significantly lower in patients with at least one emerging ctDNA alteration (p < 0.001) (Fig. 2D).
Emerging alterations in breastl cancer patients: insights from subgroup analysis in the STING and BIP studies
ESR1 mutations, which represent a key driver of endocrine therapy resistance, were identified in 31 patients (32%) out of 97 with advanced breast cancer who were treated with aromatase inhibitors in the STING study (Fig. 1G Supplementary Table 1). Among them, 12 patients (12%) harbored ≥ 2 variants (up to three). In the BIP study, ESR1 mutations were identified in 50 patients (20%) out of 250 (Fig. 2E and Supplementary Table 2). Eight patients (3.2%) harbored ≥ 2 variants (up to 7). In the STING and BIP studies, no survival difference was observed between patients harboring ctDNA emerging alterations and those who did not harbor such alterations (Fig. 1H F).
Emerging alterations in NSCLC patients: insights from subgroup analysis in the STING and BIP studies
In the STING study, ctDNA sequencing allowed the identification of genetic alterations involved in acquired resistance to anti-EGFR tyrosine kinase inhibitors in 19 patients (33.3%) out of 57 with EGFR-addicted NSCLC (Fig. 1I and Supplementary Table 1). The most frequent aberrations observed in our cohort of patients were the emergence of EGFR pathway mutations in 9 patients (15.8%). Other emergent alterations associated with resistance to anti-EGFR treatment observed in our cohort included PI3KCA mutation, HER3 amplification and fusions involving other key oncogene drivers, such as RET (n = 1), BRAF (n = 1) and NTRK1 (n = 1). Overall, the mechanisms of resistance were polyclonal in 8 patients (14%). In the BIP study, emerging alterations involved in resistance were observed in 4 patients (25%) out of 16 and included EGFR mutations, PI3KCA mutation and AKT amplifications (Fig. 2G and Supplementary Table 2). In both the STING and BIP studies, no survival difference was observed between patients harboring ctDNA emerging alterations and those who did not harbor such mutations (Fig. 1J H).
Add Comment