In our study, we analyzed 55 patients with aggressive progression from a cohort of 1,073 cases, which were classified into two groups based on the PFS duration of first-line TKI treatment. According to the depicts of the OncoPrint, 20% of patients (2 out of 10) in the aggressively progressive group respectively carried AXIN2, PLCG2 or RAD51C mutation. To our knowledge, AXIN2, a component of the Wnt signaling pathway, plays an important role in tumor suppression. AXIN2 mutations may contribute to abnormal activation of the Wnt signaling pathway and enhance the proliferation and invasion of tumor cells4. Accumulated evidence suggests that AXIN2 has close relations with the development of colon cancer, breast cancer, liver cancer and so on5–7. Zhang et.al8 conducted targeted sequencing with the 1000-gene panel in NSCLC patients treated with EGFR TKIs and found that AXIN2 might be one of the key genes to the prognosis of NSCLC treated with EGFR TKIs. We also found AXIN2 mutation in only two patients whose PFS for first-line treatment lasted only 1–2 months. We assumed that AXIN2 might be one of the factors contributing to the aggressive progression of first-line TKI therapy.
Phospholipase Cγ2 (PLCG2) is a critical signaling molecule and associated with cancer, neurodegeneration and immune disorders9. Ma et al.10 concluded that PLCG2 might promote hepatocyte proliferation in rat liver cell via ERK and NF-κB pathway. In small cell lung cancer (SCLC), Chan et al. 11 found a phenotype with high PLCG2 expression, which was stem-like and significantly correlated with metastases and worse overall survival. In recent years, clinical researchers have identified PLCG2 as a tumor microenvironment (TME)-related gene in soft tissue sarcoma12. They argued that high expression of PLCG2 might help form anti-tumor TME via IL-6/JAK/STAT3 signaling pathway, and then promote the proliferation, invasion and metastasis of tumor cells. However, few studies have described the roles of PLCG2 in NSCLC. Based on the above reports and our study, it is reasonable to believe that PLCG2 may stimulate rapid proliferation and metastasis of tumor cells by remodeling tumor microenvironment or activating specific signaling pathways, which contributes to poor efficacy of EGFR TKI treatment.
RAD51C, a newly identified gene highly associated with breast and ovarian cancer, is involved in the DNA double-strand break repair pathway and plays an important role in DNA damage response. Chen et al.13 investigated the clinical significance of RAD51C in NSCLC and suggested that high RAD51C might be an independent predictor of poor efficacy in NSCLC patients receiving chemotherapy and/or radiotherapy. However, the effect of RAD51C in NSCLC patients receiving EGFR TKI has not been reported. More and more in-depth studies are needed in the future to clarify the relationship between RAD51C and efficacy of EGFR TKI in NSCLC.
In addition to compound mutations, we were looking for other predictors. Ki-67 is widely considered as a marker of proliferation because it is expressed in all phases of the cell cycle except the G0 phase14. Ki-67 has significant prognostic value in breast cancer15 and prostate cancer16, and high Ki-67 is associated with poor prognosis in early breast cancer and prostate cancer. Currently, plenty of studies have analyzed the potential role of high Ki-67 index in NSCLC. Most studies have focused on patients with early-stage NSCLC (stageⅠ-Ⅲ), and a meta-analysis have shown that high Ki-67 index predicts worse DFS and OS in patients with early-stage NSCLC17. At the same time, because NSCLC is a cancer with different histological subtypes, different histological subtypes of patients may have different clinical outcomes. Ki-67 might be an independent prognostic factor for lung adenocarcinoma18, but not all analyses have statistical differences.
In terms of predicting efficacy, Ki-67 index might predict the efficacy of chemotherapy. Wang et al.19 reported that the high expression of Ki-67 may also be an indicator of the shortened PFS of chemotherapy. Nevertheless, studies that have used Ki-67 index to predict the efficacy of targeted therapy in advanced NSCLC are rare. In our study, we found that high expression of Ki-67 (Ki-67 ≥ 45%) may be an indicator of poor efficacy of first-line EGFR TKI treatment and its specificity is in the range of 77–84%. We argued that Ki-67 might be used as a biomarker to predict first-line EGFR TKI curative treatment in advanced EGFR-mutant NSCLC patients.
Compared with the first-generation EGFR TKI, the second-generation EGFR TKI can inhibit multiple ErbB family members (EGFR/ErbB1, HER2/ErbB2, ErbB3 and ErbB4) simultaneously and effectively interrupt downstream information transmission20, while the third-generation EGFR TKI can irreversibly and selectively inhibit both EGFR-TKI-sensitizing and EGFR p.Thr790Met (T790M) resistance mutations. The efficacy of single EGFR TKI is various in different populations and different gene mutation types, requiring more refined management of patients. FLAURA study compared the third-generation EGFR TKI with the first-generation EGFR TKI and demonstrated a significant benefit in OS, but unfortunately there was no statistical difference between the Asian population and exon 21 L858R mutation21. In the ARCHER 1050 study, on the contrary, Asian populations and patients with exon 21 mutations benefited from the second-generation drug22. On the other hand, if we focused on the survival data of the exon 21 L858R mutation subgroup alone, the OS in this subset of patients in the ARCHER 1050 was nearly 10 months, which is longer than the first-generation EGFR TKI 22. In FLAURA, patients with exon 21 mutations had the third-generation and the first-generation EGFR TKI with no statistical difference between the two groups21. Therefore, for advanced NSCLC patients with EGFR mutations, especially exon 21 L858R mutation, the choice of the second-generation EGFR TKI may be more reasonable for the first-line therapy.
The exploration of treatment strategies after EGFR TKI aggressive progression is still ongoing. In EGFR-mutant patients with systemic progression, the option of second-line treatment depends on EGFR T790M status. At the time of progression of erlotinib, gefitinib or afatinib, the standard treatment is osimertinib for patients with EGFR T790M mutation23. Combinations of EGFR-TKIs with different drugs (including other TKIs, monoclonal antibodies and chemotherapy) are currently under investigation. An analysis showed that a combination of osimertinib with pemetrexed or cisplatin can delay the emergence of resistance24. But some clinical trials argued that compared with EGFR-TKI plus chemotherapy, the PFS of single TKI was no significant difference, and the OS is even lower when compared with chemotherapy in T790M patients25,26. Immunotherapy with PD-1/PD-L1 inhibitors has become a new standard for second-line treatment of NSCLC. In the IMpower 150 trial, benefit was observed from the addition of atezolizumab and bevacizumab to platinum doublet chemotherapy (the ABCP regimen) in the EGFR-mutant patients. Overall survival and PFS were significantly improved in the ABCP arm, suggesting immunotherapy combined with antivascular therapy or combined chemotherapy are potential therapeutic strategies27.In our study, we found that for the aggressively progressive patients, PFS of chemotherapy in combination with other treatments was better than single TKI during the first 10 months in the second-line therapy. Although the difference was not significant between two groups, the tendency remained. Therefore, the combination therapy to overcome first-line-TKI aggressive resistance needs further confirmation. Genetic profiling of the tumor at aggressive progression could help to identify of resistance mechanisms and select the most appropriate combination approach.
Several limitations must be noted in our study. The samples we selected for this study were limited. We delimited the cutoff value of Ki-67 according to ROC curves, but since many individual studies have defined different cutoff value of Ki-67, it may lead to the heterogeneity of observation results. Future prospective studies are needed to confirm the cutoff value of Ki-67 and the relationship between Ki-67 and prognosis in advanced NSCLC. Last but not least, gene variation is one of the important factors affecting the efficacy of first-line targeted therapy in NSCLC patients. Since most patients in our current study did not complete the 448-gene panel after aggressive progression, we could not identify the possible mechanism leading to aggressive progression.
In conclusion, we observed that high expression of Ki-67 might be a possible indicator of short-term survival on first-line targeted therapy. Additionally, for those patients with high Ki-67 expression (≥ 45%) or combined with AXIN2, PLCG2 or RAD51C mutation, we may predict that they are likely to show aggressive progression on first-line EGFR TKI therapy. And for these patients, second-generation TKI may benefit more from first-line targeted therapy. When they show aggressive progression on first-line TKI therapy, we argued that combination therapy might be considered instead of sing TKI therapy.