Tyrosine kinase inhibitors for solid tumors in the past 20 years (2001–2020)

First-generation reversible EGFR-TKIs (e.g., gefitinib [14, 15], erlotinib [16] and icotinib [17, 18]) have yielded significant survival benefits for patients with advanced NSCLC harboring EGFR-sensitizing mutations. Additionally, efforts to investigate them in adjuvant setting have also been made [19‐21]. Second-generation EGFR-TKIs (e.g., afatinib, dacomitinib) bind irreversibly to EGFR and typically belong to pan-HER inhibitors. Dacomitinib yielded an improved median progression-free survival (mPFS) (14.7 vs 9.2 months; hazard ratio (HR) 0.59; p < 0.0001) and median overall survival (mOS) (34.1 vs 26.8 months; HR 0.76; p = 0.044) compared to gefitinib in first-line treatment of advanced EGFR-mutant NSCLC [22, 23]. However, both afatinib and dacomitinib have increased toxicities, which may limit their use in clinical practice.

Approximately 50% of resistance to first- and second-generation EGFR-TKIs are due to EGFR T790M mutation, in which the significantly bulkier methionine residue replaces the small polar threonine at position 790 of EGFR exon 20. As a gatekeeper to the adenosine triphosphate (ATP) binding pocket of EGFR, T790M could cause conformational change resulting in the development of steric hindrance; it could also increase the ATP affinity; all of these reduce binding ability and access of first- and second-generation EGFR inhibitors to the EGFR ATP binding pocket [24].

Osimertinib is a third-generation irreversible EGFR-TKI that inhibits both EGFR-sensitizing and EGFR T790M mutations, and was initially approved for NSCLC with EGFR T790M mutation [25]. Later, it also demonstrated superiority over gefitinib or erlotinib in the first-line treatment of EGFR-mutant NSCLC [26, 27]. Along with its favorable safety, osimertinib is likely to surpass other approaches in the standard of care [28]. In addition, osimertinib as adjuvant therapy for stage IB–IIIA EGFR-mutant NSCLC after complete tumor resection also achieved meaningful survival results [29].

Up to 40% of NSCLC patients with EGFR mutation develop central nervous system (CNS) metastases on or after first- or second-generation EGFR-TKIs treatment due to their poor penetration of the blood–brain barrier (BBB). Notably, osimertinib demonstrated favorable efficacy for patients harboring CNS metastases with or without prior EGFR-TKIs treatment [30] or those with EGFR T790M mutation [25]. Besides, EGFR-mutated NSCLC patients with leptomeningeal metastases could also benefit from osimertinib after progression on previous EGFR-TKIs [31].

Almonertinib, another third-generation EGFR-TKI, was approved by China National Medical Products Administration (NMPA). In its phase II trial, an objective response rate (ORR) of 68.9% was observed in patients with previously treated EGFR T790M-positive NSCLC along with a CNS ORR of 60.9% [32]. Other third-generation EGFR-TKIs including furmonertinib (AST2818) [33, 34], lazertinib (YH25448) [35], BPI-7711 [36], and nazartinib (EGF816) [37] have shown promising efficacies and acceptable safeties in advanced NSCLC with EGFR T790M mutation. In a phase IIb trial of furmonertinib, an ORR of 73.6% was observed in patients with EGFR T790M mutated NSCLC [34]. In safety analysis, skin and gastrointestinal disorders as well as interstitial lung disease (ILD) related to furmonertinib seem to be less common than osimertinib [33].

Long-term responses of third-generation EGFR-TKIs are often compromised by acquired resistant mutations, with EGFR exon 20 C797S mutation as the predominant cause [10]. The prime therapeutic strategy after resistance of osimertinib remains unclear. Though patients harboring C797S in trans with T790M (at different alleles) may respond to the combination of first- and third-generation EGFR-TKIs [38, 39], patients harboring C797S in cis with T790M (at the same allele), which is more common, are likely to show no response [40]. Fourth-generation EGFR-TKIs are under development to overcome osimertinib resistance mediated by EGFR-dependent mutation mechanisms, such as EAI045 against T790M and C797S mutations [41], and TQB3804 against osimertinib-resistant EGFR triple mutant (d746-750/T790M/C797S, L858R/T790M/C797S) or double mutant (d746-750/T790M, L858R/T790M) [42]. In addition, the combination of brigatinib with cetuximab has showed preliminary efficacy in patients with EGFR/T790M/cis-C797S triple mutation [43].

Targeted therapy for rare EGFR mutations remains an unmet need in NSCLC. Osimertinib showed efficacy against NSCLC with uncommon mutations including L861Q, G719X, or S768I substitutions [44]. Several agents such as TAK-788 (mobocertinib), poziotinib, and tarloxotinib are under investigations for a refractory variant: EGFR exon 20 insertion. A phase I/II study of TAK-788 demonstrated an ORR of 54% in previously treated NSCLC patients harboring EGFR exon 20 insertions [45] and was granted a breakthrough therapy designation by FDA. Poziotinib failed to meet its primary endpoint (ORR 14.8%), but induced tumor reduction in 65% of NSCLC patients with EGFR exon 20 insertion mutants in a phase II trial [46].

EGFR-TKIs are effective treatment approaches for EGFR-sensiting-mutant NSCLC. T790M mutation has been the most common mechanism of resistance to first- or second-generation EGFR-TKIs, which fortunately can benefit from third-generation EGFR-TKIs. Novel inhibitors for uncommon EGFR mutations have been emerging. Besides, fourth-generation EGFR-TKIs are under development for resistance of third-generation ones caused by C797S mutation. Moreover, combination treatments have been under investigations. Unlike the concern of toxicities with EGFR-TKIs in combination with the programmed death ligand 1 (PD-L1) antibody [47], combining EGFR-TKIs with anti-VEGF antibody (e.g., ramucirumab) or chemotherapy has shown survival benefit in patients with EGFR mutations.

First-generation ALK-TKI crizotinib, which targets ALK, ROS1, and c-MET, showed superiority for ALK-positive NSCLC over chemotherapy [50]. However, its unsatisfactory PFS benefits and limited control of brain metastases pushed the development of second-generation ALK-TKIs (ceritinib, alectinib and brigatinib) which are characterized as higher selectivity and CNS penetration, and they are generally effective after failure of crizotinib [51]. As a new ALK-TKI, ensartinib potently inhibits wild-type ALK and common crizotinib-resistant mutations, demonstrating an ORR of 52% in patients who were progressed on crizotinib [52, 53].

Second-generation ALK-TKIs have shown favorable efficacies after progression on crizotinib in clinical practice. A mOS of 89.6 months has been reported in 84 ALK-positive NSCLC patients by the sequencing treatment of second-generation ALK-TKIs after crizotinib resistance in a real-world setting [54]. Moreover, attempt to set second-generation ALK-TKIs as first-line therapy in ALK-positive NSCLC has achieved surprising outcomes, and its standard place has been established. Alectinib is considered as a preferred choice, while ceritinib and brigatinib can serve as other recommended options. In a phase III study of alectinib, it significantly prolonged mPFS compared to crizotinib in treatment-naïve advanced ALK-positive NSCLC patients (34.8 vs 10.9 months; HR 0.43) [55, 56]. Despite the superiority of ceritinib over crizotinib predicted by adjusted indirect comparison in front-line setting [57], no comparative study has been prospectively conducted yet. For brigatinib, a superior mPFS over crizotinib was also observed in the first-line setting (24.0 vs 11.0 months; HR 0.49; p < 0.0001) [58, 59]. It has received an approval from both European Medicines Agency (EMA) and FDA as a first-line approach for metastatic ALK-positive NSCLC. Second-generation ALK-TKIs such as alectinib and brigatinib also demonstrated CNS benefits over crizotinib [59, 60]. Other second-generation ALK-TKIs including WX-0593 [61] and CT707 [62] are under clinical investigations and have produced promising outcomes.

Generally, 56% of patients treated with second-generation ALK-TKIs develop acquired resistance due to secondary ALK mutations [63]. When such resistance occurs, lorlatinib, a third-generation ALK/ ROS1-TKI with potency against most known ALK mutations, is a therapeutic option. In a phase II trial, lorlatinib demonstrated meaningful activity as both first-line and subsequent therapies for ALK-rearranged NSCLC, with an ORR of 90% in treatment-naïve patients, 47% in patients with one previous ALK-TKI treatment, and 38.7% in patients with two or more previous ALK-TKIs treatment [64]. In respect of safety profile, apart from common adverse events (AEs) such as hypercholesterolemia and hypertriglyceridemia, grade ≥ 3 neurological toxicity including peripheral neuropathy (2%), cognitive effect (1%), and dizziness (1%) should be taken with caution. With complex mechanisms of resistance to lorlatinib being identified, future tailored approaches for such patients are warranted [65]. Another third-generation ALK-TKI CT-3505 is under investigation (ChiCTR1900025619).

Treatment strategies for ALK-rearranged NSCLC patients have advanced considerably with the development of crizotinib and newer generations of ALK-TKIs. Acquisition of resistance to ALK-TKIs ultimately occurs; the best sequencing strategy of first- to third-generation ALK-TKIs warrants further investigations. Additionally, combination treatment of ALK-TKIs and immune checkpoint inhibitors (ICIs) is associated with higher morbidity in many cases, combination with other classes of agents are ongoing.

Larotrectinib and entrectinib are two approved first-generation TRK inhibitors for adult and pediatric (12 years of age and older) patients with solid tumors harboring NTRK gene fusions which are unresectable, resistant, or lack of satisfactory standard therapy.

Larotrectinib is a selective inhibitor of TRKA/B/C which obtained its approval based on the combined analysis of three phase I/II trials involving 17 unique TRK fusion-positive tumor types [95]. In an expanded data set, patients treated with larotrectinib achieved an ORR of 79%, with manageable toxicities: the most common grade 3–4 TRAEs included increased ALT (3%), anemia (2%), and decreased neutrophil count (2%) [96]. Later analysis showed patients who had received more lines of treatment tend to have less effective response to larotrectinib; response rate dropped more sharply as the Eastern Cooperative Oncology Group (ECOG) performance status (PS) got worse, but a ECOG PS score of 1–2 still can benefit from larotrectinib treatment [97].

Entrectinib is a multi-kinase inhibitor targeting TRK, ROS1, and ALK. The pooled analysis revealed an ORR of 57.4% in patients with TRK fusion across 10 tumor types [98]. In a recent study, entrectinib produced favorable responses in children and adolescents with refractory CNS and extracranial solid tumors harboring NTRK, ROS1, or ALK fusions, as well as those with ALK-mutated neuroblastoma [99].

Other multi-kinase inhibitors including crizotinib, cabozantinib, ponatinib, nintedanib, lestaurtinib, altiratinib, foretinib, merestinib, MGCD516, PLX7486, DS-6051b, and TSR-011 have varying degrees of activity against TRKA/B/C in vitro, but their clinical activities have not been as robust as those of larotrectinib and entrectinib [93, 100].

On-target or off-target mechanisms would disappointedly result in resistance to first-generation TRK inhibitors. On-target resistance mechanisms mainly refer to NTRK kinase domain mutations involving amino acid substitutions of the solvent front, the gatekeeper residue, or the xDFG motif [100].

Two major developing second-generation TRK inhibitors selitrectinib (LOXO-195) and repotrectinib (TPX0005) are designed to overcome the acquired on-target resistance of first-generation ones and possess enhanced activities against wild-type TRKA/B/C. Selitrectinib (LOXO-195) selectively targets multiple TRK kinase domain mutations including solvent front and xDFG substitutions [101]. The largest data set of LOXO-195 till now enrolled 31 TRK-fusion patients with 11 cancer types progressing or being intolerant to at least one prior TRK inhibitor: the ORR was 54% in patients with on-target TRK mutations [102]. Repotrectinib (TPX-0005), another next-generation ROS1/TRK/ALK inhibitor, has shown promising anti-tumor activity, a confirmed partial response (reduced by 82%) in an entrectinib-resistant patient with a salivary gland tumor and a tumor regression (reduced by 33%) in a patient with larotrectinib-resistant cholangiocarcinoma were reported [103].

Mechanisms of off-target resistance to first- or second-generation TRK inhibitors include KRAS and BRAF V600E mutations, MET amplifications, IGF1R activation, etc. The convergent activation of mitogen-activated protein kinase (MAPK) pathway was also proposed to mediate the resistance of TRK inhibition. Second-generation TRK inhibitors are ineffective against off-target resistance, whereas a single targeted agent for off-target mutation or combined with TRK inhibition might re-established disease control in this situation. For example, simultaneous inhibition of TRK and MEK (belongs to MAPK pathway) was found to successfully manage some off-target TRK resistance in vitro and vivo [104, 105].

Table 4 summarizes advances of TRK inhibitors.

The potency of VEGFR-associated multi-targeted TKIs was supported by robust evidence in HCC [108]. Sorafenib, targeting VEGFR, PDGFR, FGFR, and other signaling targets, is recommended for front-line therapy for unresectable HCC [109, 110]. When compared with sorafenib, lenvatinib demonstrated a superior mPFS (7.4 vs 3.7 months; p < 0.0001) and a non-inferior mOS (13.6 vs 12.3 months; HR 0.92). Lenvatinib produced fewer grade ≥ 3 palmar–plantar erythrodysaesthesia but with higher incidence of hypertension and proteinuria [111]. More recently, the combination of lenvatinib with anti-programmed cell death protein-1 (PD-1) antibody pembrolizumab showed encouraging anti-tumor activity in patients with untreated/unresectable HCC with an ORR of 36%, a mPFS of 8.6 months, and a mOS of 22.0 months [112]. This combination has been granted a breakthrough therapy designation by FDA. In a phase II/III trial involving advanced HCC, another VEGFR-associated multi-targeted TKI donatinib achieved a superior OS over sorafenib (12.1 vs 10.3 months; HR 0.83; p = 0.0363) [113]. In addition, several other VEGFR-associated multi-targeted TKIs including regorafenib [114], cabozantinib [115], and apatinib [116] are applied as subsequent-line therapies of HCC.

RCC is another cancer type deriving great benefit from VEGFR-associated multi-targeted TKIs. Sorafenib, sunitinib, pazopanib, cabozantinib, the combination of axitinib and pembrolizumab/avelumab were successively approved as first-line treatment options for metastatic RCC.

Anlotinib has yielded favorable outcomes in lung cancer and was approved for third-line or further-line therapy for both NSCLC [117] and SCLC [118] by NMPA. In its phase II trial for patients with relapsed SCLC, a mPFS of 4.1 months was reported [119]. Similarly, apatinib presented a mPFS of 5.4 months in patients with extensive-stage SCLC after failure of two or more lines of chemotherapy [120].

Furthermore, VEGFR-associated multi-targeted TKIs also demonstrated survival benefits in patients with thyroid carcinoma, soft tissue sarcoma (STS), and other solid malignancies [121] (see Table 5).

Preclinical and clinical studies suggested that the combination of anti-angiogenesis inhibitors with ICIs could provide superior anti-tumor activity over either single agent. VEGFR inhibitors might potentially improve immunotherapeutic activity of PD-1/PD-L1 antibodies by enhancing tumor infiltration of immune cells and reducing immunosuppressive effects of myeloid-derived suppressor cells [122]. Investigational combinations of VEGFR-associated multi-targeted TKIs and anti-PD-1/PD-L1 antibodies are summarized in Table 6.

METex14 occur in approximately 3% of lung adenocarcinoma, 2% of other lung neoplasms, and rare in other tumors [133]. Intriguingly, crizotinib was initially developed as a MET inhibitor and later on made great achievements in ALK and ROS1 inhibition. But it still showed meaningful activity against MET amplification and METex14 [134]. In a recent study, crizotinib demonstrated an ORR of 32% in NSCLC patients with METex14-mutation [135]. Other multi-targeted TKIs such as cabozantinib, merestinib, glesatinib, and sitravatinib also showed meaningful activities against MET [136].

Apart from multi-targeted TKIs, selective MET inhibitors like tepotinib, camaptinib, and savolitinib have emerged with promising survival benefits. Tepotinib has received a breakthrough therapy designation by FDA for treatment of metastatic NSCLC after failure of platinum-based therapy with an ORR of 47% and a mPFS of 11 months. 27% of patients experienced grade 3–4 AEs, with peripheral edema being the most common AE (7%) [137‐139]. Recently, tepotinib was approved by Japanese Ministry of Health, Labor and Welfare (MHLW) for the treatment of unresectable, advanced or recurrent NSCLC with METex14 mutation, making it the first approved MET-TKI worldwide.

Another MET inhibitor, capmatinib, was approved for the treatment of adult NSCLC patients with METex14 mutation regardless of treatment history. In a phase II study, the efficacy of capmatinib was evaluated in advanced NSCLC patients with METex14 mutation or MET amplification across 6 cohorts. The ORRs were 41% and 68% among cohort 4 (previously treated METex14 mutation) and cohort 5b (treatment-naïve METex14 mutation), respectively. Its safety profile was acceptable across all cohorts (n = 315), with peripheral edema (49.2%), nausea (43.2%), and vomiting (28.3%) as the most common AEs [140]. Other cohorts also demonstrated the efficacy of capmatinib in advanced NSCLC with high-level MET amplification [141, 142].

Savolitinib (also called volitinib) demonstrated promising anti-tumor activity and manageable toxicity in pulmonary sarcomatoid carcinoma (PSC) and other types of NSCLC with METex14-mutation, with an ORR of 47.5%. Notably, 14.3% of patients discontinued savolitinib due to TRAEs, with liver injury and hypersensitivity being the most common AEs (each 2.9%) [143].

MET amplification is an important resistant mechanism of EGFR-TKIs for NSCLC treatment, accounting for 6.25–22%. More importantly, this patient population is unlikely to respond well to osimertinib or other third-generation EGFR-TKIs. Preclinical and clinical data suggest the combination of EGFR-TKIs with MET-TKIs could be an alternative option to overcome MET-driven acquired resistance of NSCLCs who have progressed on a previous EGFR-TKI [144, 145]. For instance, tepotinib plus gefitinib significantly prolonged mPFS (19.3 vs 5.5 months; HR 0.18), mOS (37.3 vs 13.1 months; HR 0.08), and ORR (75.0 vs 42.9%; OR 4.00) compared to chemotherapy for such patient population. In terms of safety, tepotinib plus gefitinib combination treatment significantly increased grade ≥ 3 amylase and lipase, while anemia, neutrophil, or white blood cell count decrease was less common compared to chemotherapy [146]. In a phase Ib/II trial, capmatinib plus gefitinib yielded an ORR of 47% in EGFR-mutant NSCLC patients with high MET amplification. The most common grade 3–4 AEs were also elevated amylase and lipase levels [147, 148]. Now with increasing use of osimertinib in the front-line treatment of EGFR-mutant NSCLC, combining MET-TKIs with osimertinib has also been explored. Savolitinib plus osimertinib presented an ORR of 48% (with or without a previous third-generation EGFR-TKI) along with acceptable tolerability [149]. The exciting results suggest it may be necessary to identify MET status before starting osimertinib treatment in patients who failed previous former-generation EGFR-TKI treatment.

Fluorescence in situ hybridization (FISH), next generation sequencing (NGS), immunohistochemistry (IHC), and droplet digital PCR (ddPCR) are methods to detect MET-mediated resistance, each with its own advantages and disadvantages. The results of different testing methods do not overlap completely, and a single assay might overlook suitable patients. Therefore, applying more than one method is recommended in future clinical practice and scientific research. Besides, other biomarkers like phosphorylated MET (p-MET) should be explored to help predict response and tailor treatment.[144, 149].

KIT proto-oncogene takes part in fertility, homeostasis, and melanogenesis, while the dysregulation of KIT has been found to participate in the occurrence of leukemia, gastrointestinal stromal tumor (GIST), melanoma, and other cancers [169]. KIT usually presents in multi-targeted TKIs as an inconspicuous target (Table 5) since a single selective KIT-TKI failed to cure most cancers. However, the aberrant activation of KIT is particularly responsible for the tumorigenesis of GIST, making it a pivotal target in this disease entity. Besides, KIT inhibition also showed efficacy in KIT-positive melanoma.

Platelet-derived growth factor (PDGF) is a family of a multi-functional polypeptide involved in cellular growth, proliferation, differentiation, and angiogenesis. PDGFR is found to play a crucial role in angiogenesis by promoting the maturation of new blood vessels and up-regulating the expression of VEGF [170]. Most VEGFR-associated multi-kinase inhibitors target PDGFR as well to augment anti-angiogenesis effect and suppress tumor growth (Table 5). Additionally, the inhibition of PDGFR plays an important role specifically in the treatment of GIST (Table 11).

GIST generally resists to conventional chemotherapy. Fortunately, since GIST has high frequency of KIT and PDGFR mutation (approximately 80% of GISTs harbor KIT mutation, 10% involve PDGFR mutations), KIT and PDGFR inhibition has been recognized as the primary therapeutic modality for unresectable or metastatic GIST [171]. Table 8 summarizes advances of KIT/PDGFR TKIs.

Imatinib remains as first-line treatment of KIT-positive unresectable GIST. Though more than half of GISTs respond to imatinib, resistance inevitably occurs. Sunitinib and regorafenib are the standard second- and third-line treatment for advanced GIST, respectively [172, 173]. Sunitinib greatly improved median time to tumor progression than placebo (27.3 vs 6.3 weeks; HR 0.33; p < 0.0001) in patients with advanced GIST after failure of imatinib, but with a low ORR of 6.8% [174]. Studies indicated the inconsistent activity of sunitinib in imatinib-resistant populations, with higher response in patients harboring ATP-binding pocket mutations than those with mutations in KIT activation loop [175].

Recently, two selective TKIs targeting KIT and PDGFRA mutants, avapritinib and ripretinib, were approved as fourth-line treatment for GIST. A phase I trial of avapritinib demonstrated an ORR of 86% in GIST patients with PDGFRA exon18 mutation and an ORR of 22% in those who have failed ≥ third-line treatment. Toxicity was generally manageable with anemia, fatigue, hypophosphatemia, hyperbilirubinemia, neutropenia, and diarrhea being the most common grade 3–4 AEs [176]. A phase III trial of ripretinib demonstrated an improved mPFS (6.3 vs 1.0 months; HR 0.15; p < 0.0001) and mOS (15.1 vs 6.6 months; HR 0.36; p = 0.0004) compared to placebo [177, 178]. Besides, in a phase I study, an investigational KIT inhibitor PLX9486 alone or in combination with pexidartinib presented preliminary efficacy against resistant GIST [179].

KIT mutations occur in 35–40% of mucosal and acral melanoma, and 28% of melanomas on chronically sun-damaged skin [180]. Imatinib and nilotinib demonstrated ORRs of 17–30% and disease control rates (DCRs) of 35–57% in metastatic melanoma patients with KIT mutation/amplification [181‐183]. However, most of the reported response only had short duration, and no further phase III trials have been conducted. Until now, none KIT-TKI has received an approval for KIT-mutant melanoma.