This Phase I dose-escalation trial is designed to evaluate the safety of autologous CD19CAR T-cells. Dose-limiting toxicity is defined as development of NCI grade 3 - 4 non-hematologic toxicity that can be attributed to the treatment. Dose escalation is guided by the modified continual reassessment method (mCRM) in order to determine the maximum tolerated dose (MTD) of transduced T-cells. For this study, MTD is defined as the dose which causes DLT in at most 20% of eligible cases. Three dose levels are being evaluated namely, 2x10exp7 cells/m2, 1x10exp8 cells/m2, and 2x10exp8 cells/m2 with prior probabilities of toxicity estimated at 5%, 12%, and 25%, respectively. Based on our previous trials, the investigators expect a shallow dose-toxicity curve for the doses proposed in this trial. In this trial, mCRM is implemented based on an exponential model with a cohort of size 2. To reduce the probability of treating patients at unacceptable toxic dose levels, the investigators employ modifications to the original CRM. Specifically, there will be more than one subject treated in each cohort, dose escalation is limited to no more than one dose level, and patient enrollment starts at the lowest dose level.

Two patients will initially be allocated to the lowest level cohort and followed for 6 weeks post IV injection of transduced EBV-CTLs and peripheral blood T cells for evaluation of DLTs. If any or both patients initially enrolled develop DLTs, the trial will be stopped. Otherwise, two patients will be enrolled at the next dose level.

For the intermediate dose level, if any or both patients develop DTLs, the mCRM will guide the next dose-level by estimating the toxicity probability. The dose will be the same or de-escalated depending on the current estimated probability of MTD. Two more patients will be enrolled at the recommended dose level by the CRM. If the lowest dose level is selected and any of the patients develops DLTs, the trial is stopped; otherwise, two patients will be enrolled at the next dose level. If the intermediate dose level is selected and none of the patients develops DLTs, two patients will be enrolled at the next dose level.

If the trial has not stopped, from this point on, if the rate of DLTs for any dose level (taking into account all patients treated at that dose level) becomes greater than 20%, the dose is de-escalated and the algorithm is repeated. Otherwise, the dose will be increased or kept at the maximum level.

The trial will continue until a minimum of 10 patients is treated, with six patients accrued at the current MTD. Depending on patient availability, a maximum 18 patients will be accrued into this Phase I trial. The final MTD will be the dose with probability closest to the target toxicity rate at these termination points. The investigators therefore expect to enroll between 10-18 patients into this trial

Simulations with 10,000 replications were performed to determine the operating characteristics of the proposed design and compared with the standard 3+3 dose-escalation design. The proposed design provides better estimates of the MTD based on a higher probability of declaring the appropriate dose level as the MTD, afforded smaller number of patients accrued at lower and more likely ineffective dose levels, and maintained a slightly lower average total number of patients required for the trial. As indicated above, the first dose level is safe in previous studies for CTL in lymphoma patients and the investigators expect a shallow dose-toxicity curve over the range of doses proposed in our study. Therefore, the investigators feel comfortable with slightly more accelerated dose-escalations without compromising patient safety. In fact, our simulations indicate that the investigators will not observe increased toxicities with this design and estimate the average number of toxicity to be comparable between the modified CRM and standard 3+3 design.