Hypothesis: Different patients have different biomarkers, if doctors know about the biomarkers of patients; they may be able to prescribe a regimen that is better suited to the patient's specific needs. This is a pilot study. Here, we used whole exon sequencing and Integrated genomic network analysis to identify the biomarker or gene. We aimed to learn if the drug chosen based on biomarkers can help to control metastatic gastrointestinal cancer who had failed from all standard and available regimens.

Rational: Cancer sequencing (CS) promises to become the centerpiece of personalized oncology by informing on treatments targeted to each tumor's unique genetic constitution. This data can be critical to making an informed decision for disease management, though this may not be the case for all patients. CS identifies variations or differences in the DNA and/or RNA of the cells in an individual's tumor by comparison to that of his/her normal cells. These somatic variations may, on further interpretation, be identified as key drivers of carcinogenesis. Such information may predict a patient's prognosis, response to currently available treatments or prompt the development of novel therapeutics. Though CS has the potential to personalize and optimize cancer care, it may produce a vast amount of data and unique changes in the DNA/RNA that may be difficult to interpret at the present time.

Using the Integrated genomic network analysis, we could have better understanding of the underlying processes and pathways involved in tumor onset and progression. And then we could choose a specific treatment regimen and develop personalized cancer therapies

Inclusion Criteria:

1. Pathologic diagnosis of Gastrointestinal cancer
2. The subject has a diagnosis metastatic gastrointestinal cancer, and failed from standard treatment, and no other regimen is available.
3. The subject has measurable lesion of gastrointestinal cancer.
4. The subject's The Eastern Cooperative Oncology Group (ECOG) performance status of 0-2.
5. The subject has adequate hematologic function as defined by an absolute neutrophil count (ANC) >/= 1,500/mm3, platelet count >/= 100,000/mm3, White Blood Count (WBC) >/= 3,000/ mm3, and hemoglobin >/= 9 g/dL.
6. The subject has adequate hepatic function as defined by a total bilirubin level </= 1.5 * the upper limit of normal (ULN) (bilirubin >/= 1.5 * ULN with known Gilbert's disease is allowed), and alkaline phosphatase, aspartate aminotransferase/alanine aminotransferase (AST/ALT) </= 2.5 * the upper limit of normal or </= 5.0 * ULN if liver metastases are present.
7. Serum creatinine clearance >50ml/min, either by Cockcroft-Gault formula or 24-hour urine collection analysis
8. The subject is >/=18 years of age.
9. The subject has signed informed consent.
10. Women of childbearing potential must agree to use adequate contraception (hormonal or barrier method of birth control; abstinence) prior to study entry and for the duration of study participation. Childbearing potential will be defined as women who have had menses within the past 12 months, who have not had tubal ligation, hysterectomy or bilateral oophorectomy. Should a woman become pregnant or suspect that she is pregnant while participating in this study, she should inform her treating physician immediately.

Exclusion Criteria:

1. pregnant or breast-feeding.
2. Subjects will be excluded for other concurrent severe and/or uncontrolled medical disease which could compromise participation in the study
3. without enough tumor sample for analysis.
4. Refuse to sign the informed consent.