This is a multicenter, open label, phase I/II safety study that will enroll 30 subjects with
      metastatic esophageal and gastroesophageal junction adenocarcinomas post first line systemic
      chemotherapy. Subjects will be treated with oral Abemaciclib 150 mg PO daily bid given with
      ramucirumab 8mg/kg every 2 weeks iv until evidence of disease progression or unacceptable
      toxicities.

      A total of 30 subjects will be enrolled. The primary goal is to describe the safety profile
      of Abemaciclib in combination with Ramucirumab among all enrolled subjects.

      If grade 3 or higher treatment-related adverse events occur in 20 subjects, the upper bound
      of 95% Wilson confidence interval for the adverse event rate would be below 81% (16.7% -
      47.9%).

      The safety analysis will be performed in all treated subjects. Adverse event data will be
      listed individually and graded according to the National Cancer Institute Common Terminology
      Criteria, version 4.03.

      Summary statistics will include counts and proportions as well as rates with 95% confidence
      intervals. Toxicities will be reported as a tabulated table by type and grade.

      Objective response rate is defined as the percentage of subjects who achieve an objective
      response by RECIST1.1 criteria (i.e. Complete response or Partial Response) to Abemaciclib in
      combination with Ramucirumab. We will estimate the objective response rate, along with the
      Wilson 95% confidence interval, for the population of subjects.

      Overall survival will be defined as the time from study enrollment to death. This will be
      summarized using a Kaplan-Meier curve.

      The proportion of subjects with grade 4 or higher treatment-related adverse events will be
      monitored continuously throughout the trial using a Bayesian stopping guideline. A Beta (1,
      19) prior, representing a toxicity rate of 5%, slightly lower than the expected rate of 6%,
      was used in the development of our guidelines. The therapy will be re-evaluated if the
      posterior probability that the toxicity rate exceeds 10% is greater than 75%. Table 3
      summarizes the stopping boundaries starting with the initial cohort of 3 subjects through the
      maximum sample size of 30 subjects.

      The probability of triggering the stopping guidelines was assessed for a range of possible
      toxicity rates using simulations with 5000 replicates. The probability of stopping to
      re-evaluate was 1% if the true proportion with an unacceptable toxicity was 5%. In
      comparison, the probability of stopping early was 99.6% if the true proportion with an
      unacceptable toxicity was 40%