USE OF PATIENT-DERIVED TUMOR XENOGRAFT MODEL TO GUIDETREATMENT DECISIONS FOR PATIENTS

USE OF PATIENT-DERIVED TUMOR XENOGRAFT MODEL TO GUIDETREATMENT DECISIONS FOR PATIENTSYanping Zhu, Chun-Mei Zhao, Yanpeng Hao, Zhiqiang Wang, Li Wang, Duan ChenBackground/aim:Patient-derived tumor xenograft (PDTX) models are established by directengraftment of patient-derived tumor fragments into immunocompromised mice. Approx.20, 000 PDTXs have been so far created in academic and industry labs in US and Europefor basic and preclinical cancer research, such as biomarker discovery, drug screening forpersonalized medicine, understanding of drug-resistance mechanisms and novel therapydevelopment. The aim of our work was to use PDTXs to guide clinical decisions for patients.Materials and Methods:From January 2015 to November 2018 at 5 clinical centers inChina, we have collected tumor samples by surgery, biopsy and/or endoscopy from 723patients with esophageal cancer (12 cases), gastric cancer (84), colorectal cancer (65), livercancer (294), pancreatic cancer (46), sarcomas (94), gynecological cancer (63), lung cancer(26), and others (34). In 434 out of 723 PDTXs, the primary tumors (P0) grew to 800 mm3in mice, and then re-implanted into mice to produce the first passage (P1). When the P1tumors grew to 100–150 mm3, clinical tests of treatment regimens were conducted. It took3-5 months from tumor sampling to clinical report. The cost was about 200 000 RMB(approx. 30,000 USD) per patient. In the remaining 289 PDTX, the P0 tumors grew in a“tumor-microenvironment matrix” that was implanted in mice, named as superPDTX (patentapplication no. 201810916601.5). Biomarkers of superPDTX included Ki67 index andpathological scores. It took 2 weeks at a cost of 30 000 RMB (or 5 000 USD) per patients.In addition, whole exome sequencing (WES) was performed in 8 superPDTXs.Results:The overall engraftment rate of PDTX (P1) was 56.2%, e.g. esophageal cancer was 45.5%,gastric cancer was 63.9%, colorectal cancer was 62.2%, liver cancer was 53.8%, and pancreaticcancer was 52.6%. A cohort study of 21 PDTXs/patients showed that the consistency oftreatment effects was 85.7% (18/21) in which 15 out of 17 was validated in co-clinical trialand 3 out of 4 was in predictive testing. A total of 289 superPDTXs were used for pharmaceuti-cal drug screening (61 cases) and clinical testing (228 patients, including 160 GI cancers).The successful rate of superPDTX (P0) clinical testing was 92%. Another cohort of 22superPDTXs/patients receiving 28 treatment regimens showed a consistency rate of 82.1%(23/28) in which 8 out of 13 was validated in co-clinical trial and 15 out of 15 was inpredictive testing. Up-to-date, we have banked the PDTX tissue samples (P0-P2) of > 35000. WES data showed that >90% mutations identified in primary tumors were present inthe corresponding superPDTX samples (P0).Conclusion:Based on our results and clinicalpractice, we suggested that superPDTX testing could be used routinely to enhance clinicaldecision making for personalized cancer care. Gastroenterology, Volume 156, Issue 6, Supplement 1, May 2019, Pages S-695 https://doi.org/10.1016/S0016-5085(19)38662-7