Circulating Tumor DNA (ctDNA) as a Prognostic Tool in Patients With Advanced Lung Adenocarcinoma

Overview

Lung cancer is the leading cause of cancer death in the U.S. and throughout the world. Lung cancers are broadly divided histologically into small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). About 25% of patients with NSCLC have stage I or II disease. The primary treatment modality is surgical resection,2 and 5-year survival rates are 65% for stage I and 41% for stage II disease. However, more than 70% of patients with NSCLC present with stage III or IV disease. Patients with stage III disease are most commonly treated with chemoradiation, and 5-year survival rate is 26%. Chemotherapy and targeted therapy are often used for stage IV disease, which has a 5-year survival rate of 4%. Tyrosine kinase inhibitor (TKI) is a targeted therapy against specific molecules in critical cell-signaling pathways involved in lung carcinogenesis. The currently available FDA approved TKIs for advanced NSCLC include afatinib, gefitinib, and erlotinib that inhibit epidermal growth factor receptor (EGFR) signaling 6 and crizotinib that inhibits anaplastic lymphoma kinase (ALK) signaling. However, only tumors that carry the corresponding oncogenic mutations (e.g., sensitizing EGFR mutations) would respond well to these TKIs. Meta-analyses of clinical trials evaluating the efficacy of gefitinib and erlotinib have demonstrated that NSCLC patients who are EGFR mutation-positive have a lower risk of disease progression when treated with an EGFR-TKI as compared to those treated with chemotherapy (HR = 0.43, 95% confidence interval, CI=0.38-0.49). EGFR-TKI, however, confers no benefits to patients who are EGFR wildtype (HR = 1.06, 95% CI=0.94-1.19). A phase III trial of crizotinib has also demonstrated the superiority of crizotinib to standard chemotherapy in ALK-positive NSCLC patients (HR = 0.49; 95% CI=0.37-0.64). In Hong Kong, as in other parts of Asia like in China and in Taiwan, other than the majority of lung cancer patients being smokers, there is also a prominence of non-smokers in lung cancer. Compared with Caucasians, there is also a relatively higher incidence of EGFR mutation in lung adenocarcinomas. The prevalence of EGFR mutation in Asian population with lung adenocarcinomas can reach up to 60% compared to at most 30% in the Caucasian population. These EGFR mutant tumors will demonstrate better response to the drug EGFR-TKI, boosting up the response rate to almost 70% compared to 30% with conventional chemotherapy for lung cancer. Even with this remarkable response, however, EGFR-TKI will eventually fail in EGFR mutant lung cancer. There is an imminent need to look for newer therapeutic targets or agents that can overcome this acquired resistance to anti-cancer drugs and to explore alternative molecular signaling pathways that could interact or enhance EGFR signaling pathways to modulate the therapeutic response in lung cancer.

Although EGFR- and ALK-TKIs can achieve a response rate as high as 70%, all patients treated with TKIs invariably develop resistance to the therapy. The median progression-free survival is 10-16 months. The most common mechanism of acquired resistance to TKIs is the therapy-induced clonal selection of a minor subpopulation of resistant cancer cells that were present in the original tumor. Emergence of the EGFR mutation T790M occurs in about 50-70% of patients with acquired resistance to EGFR-TKIs. Other EGFR mutations and mutations in phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) and B-Raf Proto-Oncogene (BRAF) are also associated with EGFR-TKI resistance, but they occur at low frequencies. Resistance to ALK-TKI is more complex and involves various resistant mutations. TKI resistance remains a major problem in clinical management of NSCLC. Patients with acquired resistance can be treated with second generation TKIs, though none are FDA approved yet, or by combination therapy strategies. Therefore, molecular characterization of tumor throughout the course of disease is helpful to match new drugs to the tumor's evolving genomic profile and guide effective personalized therapies. However, serial tissue sampling to monitor molecular signatures of tumor is invasive, impractical, and not a routine clinical practice. Obtaining sufficient tissue materials for genotyping is also a major hurdle in tissue sampling. There is a need to develop a technology that permits non-invasive serial analysis of the tumor genomic profiles. Cell-free circulating DNA is fragmented DNA found in circulation that is not associated with cells or cell fragments. When tumor cells die, they release tumor DNA into the bloodstream. The cell-free circulating DNA derived from tumors, known as circulating tumor DNA (ctDNA), carries mutations present in the tumor and hence can be distinguished from cell-free circulating DNA derived from normal cells. It has been shown that the detection of ctDNA and its concentration correlate with tumor stage and cancer survival. Moreover, ctDNA in plasma can be used to detect genomic alterations in solid cancers, and that there is a high concordance in detected mutations between paired formalin-fixed paraffin-embedded (FFPE) and plasma DNA samples. In a study of acquired resistance to EGFR blockade in colorectal cancer patients, repeated serum samples were collected at 4-week intervals until disease progression. Using mathematical modeling, this study had the following important findings: resistant mutations were present in a clonal subpopulation within the tumors prior to the initiation of treatment, it took a fairly consistent period of time (about 5-6 months) for the subclone to expand and repopulate the lesion, and circulating resistant mutations could be detected several months before radiographic evidence of disease progression. This seminal study demonstrated the potential of using a ctDNA test to track genomic evolution and selection in tumors in a non-invasive manner in order to facilitate individualized therapies and hence to prolong remission. The investigators have demonstrated plasma detection of EGFR mutations in patients with advanced stage lung adenocarcinoma bearing EGFR mutations, correlating with prognosis of subjects on EGFR-TKI. One prospective study had used real-time polymerase chain reaction (RT-PCR) to detect EGFR mutations in ctDNA from patients with advanced NSCLC. Among patients who were EGFR mutation + at baseline (pre-treatment), those who lost the EGFR mutation at cycle 3 of treatment (chemotherapy +/- erlotinib) had better progression free survival; median survivals were 7.2 vs. 12.0 months in patients who were EGFR mutation (+,+) and (+,-) at baseline and cycle 3, respectively. Other studies have also demonstrated the feasibility of other oncogenic mutations especially KRAS mutation. Although these studies demonstrated the feasibility of detecting tumor mutations in ctDNA, they were limited to examining a single gene (e.g., EGFR or KRAS). Other studies had applied next generation sequencing in patients with NSCLC. Max Diehn's lab at Stanford University has developed a method to quantify ctDNA by deep sequencing of \>130 genes. In 17 patients with paired plasma DNA and tumor tissue samples, they were able to detect all mutations previously identified in tissue plus many additional somatic variants. They also found levels of ctDNA to be highly correlated with tumor volume. Their study examined multiple genes, but did not have a prospective component to track ctDNA mutations and correlate specific mutations with treatment outcome. Testing of ctDNA in patients who receive chemotherapy has never been done. Genomic profiling can identify mutations associated with resistance and response to chemotherapy. The investigators therefore propose a longitudinal study in patients with advanced NSCLC treated with first-line TKI or chemotherapy to collect serial blood samples prospectively and, using next-generation sequencing of ctDNA, to examine the evolutionary genomic profiles. This study aims to evaluate utilities of the ctDNA test in identifying genomic markers to predict treatment response and survival in patients with advanced NSCLC. This proposed study will examine the utilities of ctDNA in identifying genomic markers for NSCLC prognosis in patients treated with first-line TKI or chemotherapy. After diagnosis, patients will be followed at 3-month intervals. At each study visit, plasma samples will be collected and, whenever clinically indicated, tissue samples will also be obtained. Genomic profiling of tumors will be done in the FFPE tissue sample and in ctDNA extracted from the prospectively collected plasma samples. The aims are: 1. To determine concordance and discordance of somatic mutations found in ctDNA and tumor tissue DNA. 2. To identify mutations in ctDNA that are associated with prognosis (treatment response and progression-free survival) in patients who receive (a) EGFR-TKI treatment, (b) ALK-TKI treatment, or (c) chemotherapy. 3. To track the molecular time course in terms of (a) variation of total ctDNA concentration over time, (b) when the mutations associated with resistance/recurrence are first detectable in plasma, and (c) how the mutation fractions of the resistance-associated mutations vary over time. 4. To combine information from Aims 2 and 3 to develop prediction models for prognosis in patients who receive (a) EGFR-TKI treatment, (b) ALK-TKI treatment, or (c) chemotherapy.