Oral Decitabine and Tetrahydrouridine as Epigenetic Priming for Pembrolizumab-Mediated Immune Checkpoint Blockade in Patients With Inoperable, or Unresectable Locally Advanced or Metastatic Non-Small Cell Lung Cancers and Esophageal Carcinomas

Background: * Non-small cell lung cancers (NSCLC) esophageal cancer (EsC) and malignant pleural mesotheliomas (MPM) account for approximately 185,000 deaths annually in the United States, with over two thirds of patients presenting with advanced, incurable disease. 1st-line platinum-based chemotherapy for advanced NSCLC, esophageal cancer (ESCs) or malignant pleural mesothelioma (MPM) produces transient responses at best, with most patients succumbing to disease within 12-16 months following diagnosis. * Recent randomized clinical trials have demonstrated response rates approximating 20% in unselected patients with advanced NSCLC or EsC, and nearly 45% in patients with tumors exhibiting high level expression of programmed death ligand 1 (PD-L1) following administration of pembrolizumab, a humanized monoclonal anti-Programmed cell death protein 1 (PD-1) antibody. * Approximately 17% of unselected MPM patients have exhibited objective responses following administration of pembrolizumab or other PD-1 inhibitors. * Preclinical studies have demonstrated that epigenetic drugs such as deoxyribonucleic acid (DNA) demethylating agents and histone deacetylase (HDAC) inhibitors can prime cancer cells and tumor microenvironments thereby enhancing efficacy of immune checkpoint inhibitors. * Although a potent DNA demethylating agent, Decitabine has poor bioavailability and inconsistent distribution in solid tumors due to rapid inactivation by cytidine deaminase (CDA) which is present in high levels in many organs. * Recent studies in rodents and nonhuman primates, as well as a Phase 1 clinical trial (NCT#01685515) in patients with sickle cell disease have demonstrated that oral tetrahydrouridine (THU), an inhibitor of CDA, significantly enhances bioavailability/solid-tissue-distribution of low dose oral Decitabine (DAC), thereby enhancing systemic DNA demethylation with acceptable toxicities. * Preliminary results of recent clinical trial suggest that oral DAC-THU can increase the frequency and magnitude of responses to immune checkpoint inhibitors in lung cancer patients with low or absent intra-tumoral PD-L1 expression. * These data support further evaluation of oral DAC-THU in combination with immune checkpoint inhibitors for therapy of thoracic malignancies. Objectives: Phase I -To define pharmacokinetics, toxicities and maximum tolerated dose of oral DAC-THU in combination with pembrolizumab in patients with inoperable, or unresectable locally advanced or metastatic NSCLC, EsC, or MPM. Phase II -To determine clinical response by Response Evaluation Criteria in Solid Tumors (RECIST) criteria to oral DAC-THU in combination with pembrolizumab in patients with inoperable, or unresectable, locally advanced or metastatic NSCLC, EsC, or MPM. Eligibility: Inclusion Criteria * Male or female, 18 years or older with histologically or cytologically proven, inoperable, or unresectable locally advanced, or metastatic NSCLC, EsC, or MPM. * Measurable disease. * Patients with high PD-L1 expression ((Bullet) 50%) and low PD-L1 expression (0-49%) in cancer cells by immunohistochemistry are eligible. * NSCLC patients with no prior systemic treatment, or those with prior first line treatment including an immune checkpoint inhibitor are eligible for study. * MPM patients who have received, refused, or are ineligible for first line chemotherapy are eligible. * Patients with EsC including Seiwert-Stein Type I and Type II gastro-esophageal junction (GEJ) carcinomas who have received or refused standard of care first line therapy and/or targeted therapy are eligible. * Patients who received DNA demethylating agents or PD-1/PD-L1 inhibitors for another malignancy may be eligible for study if there were no dose-limiting immune related events, and there has been either no clinical evidence of disease or minimal residual disease that has been stable for at least three years. * Willingness to undergo tumor biopsies if safely accessible per PI discretion before and after treatment. * Eastern Cooperative Oncology Group (ECOG) performance status 0 2. * No evidence of unstable or decompensated myocardial disease; adequate pulmonary reserve. * Adequate renal, hepatic and hematopoietic function. Exclusion Criteria * Patients with any targetable mutation for which there is approved first, or second line therapy. * Serious cardiovascular conditions. * Active Hepatitis A, Hepatitis B or Hepatitis C. * Human immunodeficiency virus (HIV) or acquired immunodeficiency syndrome (AIDS)-related illness. * Other active infection requiring systemic therapy. * Pregnant or breastfeeding women. * Patients who are receiving systemic corticosteroids. * Patients receiving another investigational agent. * Another malignancy. Design: * The Phase I component will be a standard 3+3 design combining high and low PD-L1 expressers with incremental dose escalation of oral DAC-THU to define maximum tolerated dose (MTD). * Simon 2-stage design for Phase II studies will be used to determine clinical response at the MTD. * Patients will receive oral DAC-THU- on T-W for two weeks out of every 3 for 9 weeks * Pembrolizumab will be administered on Wednesday, Thursday or Friday at a fixed intravenous dose of 200 mg every 3 weeks. * One cycle is three weeks; one course is 9 weeks. Treatment evaluation using RECIST 1.1 every 10 +/- 1 weeks. * Those patients exhibiting disease progression or unacceptable toxicities will be removed from study. Patients exhibiting stable disease or disease regression will be offered an additional course of therapy followed by treatment evaluation. Treatment will continue in this manner until off-study criteria have been met. * Once the MTD for DAC/THU has been identified, the MTD dose level will be expanded by 4 patients to confirm its safety. Then, including these 10 patients at the MTD, a total of 10 NSCLC patients with high (50% or greater) intratumoral PD-L1 expression and 10 NSCLC patients with low (0-49%) intratumoral PD-L1 expression will be accrued to the first stage of each of two separate Phase II cohorts using individual Simon optimal designs. If 5 or more patients of the 10 first stage NSCLC patients in the high PD-L1 cohort respond to treatment, the cohort will be expanded to 23 patients. If 11 of 23 of these patients respond to treatment, the trial will be deemed positive for NSCLC with high PD-L1 expression. If 2 or more of the 10 first stage NSCLC patients in the low PD-L1 expression cohort respond to treatment, the cohort will be expanded to 29 patients. If 6 or more of these 29 patients experience a response, the trial will be deemed positive for NSCLC with low PD-L1 expression. Up to 10 EsC patients, including those considered to be part of the Phase I component after the MTD has been identified, will be enrolled into a separate cohort to examine responses to DAC-THU/pembrolizumab at the MTD. If 2 or more of these 10 EsC patients respond to treatment, these findings may warrant an amendment or a separate Phase II trial to determine response rates to DAC-THU/pembrolizumab in EsC patients. Similarly, if 2 or more of 10 MPM patients respond to treatment, these findings may warrant an amendment or a separate Phase II trial to determine response rates to DACTHU/pembrolizumab in MPM. * Biopsies of index lesions will be obtained at baseline and at treatment evaluation following the first course of therapy for analysis of pharmacodynamic endpoints. * Patients will be followed for toxicity for at least 30 days after treatment has been discontinued, start of new anti-cancer treatment or until death, whichever occurs first.