CAG regimen for refractory or relapsed adult T‐cell acute lymphoblastic leukemia: A retrospective, multicenter, cohort study

Abstract Adult patients with relapsed or refractory T‐cell acute lymphoblastic leukemia (R/R‐T‐ALL) have extremely poor prognosis, representing an urgent unmet medical need. Finding an optimal salvage regimen to bridge transplantation is a priority. The CAG (cytarabine, aclarubicin, and G‐CSF) regimen was initially used by one group in China, showing unexpectedly promising results in 11 R/R‐T‐ALL patients. Here, we report the multicenter results of 41 patients who received the CAG regimen as salvage therapy. After one cycle of the CAG regimen, complete remission and partial remission were achieved in 33 (80.5%) and two (4.9%) patients, respectively. Failure to respond was observed in six patients (14.6%). Early T‐cell precursor (ETP) (n = 26) and non‐ETP (n = 15) patients had a similar CR rate (80.8% vs 80.0%, P = .95). Among 41 patients, allo‐HSCT was successfully performed in 27 (66%) patients (22 in CR and 5 in non‐CR). With a median follow‐up time of 12 months, the estimated 2‐year overall survival and event‐free survival were 68.8% (95% CI, 47.3%‐83.0%) and 56.5% (95% CI, 37.1%‐71.9%), respectively. The CAG regimen was well‐tolerated, and no early death occurred. Our multicenter results show that the CAG regimen is highly effective and safe, representing a novel choice for adult patients with R/R‐T‐ALL and providing a better bridge to transplantation.

Until now, there has been no consensus about optimal salvage treatments for R/R-T-ALL, representing an urgent, unmet medical need. [4][5][6] Conventional chemotherapy options, such as FLAG (fludarabine, cytarabine, and granulocyte colony-stimulating factor), with or without idarubicin and clofarabine in combination with etoposide and cyclophosphamide, or even the only T-ALL target drug (nelarabine), have been used, but the overall response rates are only 30%-50%. 8,14,17,18 Therefore, a more effective and safer salvage regimen for R/R-T-ALL is highly awaited.
The CAG regimen [cytarabine (C), aclarubicin (A), and granulocyte colony-stimulating factor (G)], which is conventionally used in China in acute myeloid leukemia (AML), 19 showed unexpectedly promising preliminary results for the treatment of R/R-T-ALL in the pilot study performed by the Xue group. [20][21][22] In fact, 10 of 11 patients with R/R-T-ALL achieved CR. 21 The limited same size prompted us to further extend the study to validate these results of the CAG regimen for R/R-T-ALL in a multicenter setting. Here, we report the results of a multicenter study by the Chinese Leukemia Cooperation Group (CLCG) to evaluate the efficacy and safety of the CAG regimen for the treatment of adult R/R-T-ALL.

| Study design and data collection
This was a multicenter (six centers), retrospective analysis. The presence of R/R-T-ALL was confirmed. Eligibility criteria were age 16 years or older, and CAG regimen as salvage therapy from September 2012 to June 2019. The study was conducted in accordance with the Declaration of Helsinki, and approval was granted by the relevant ethics committees.
The CAG regimen (cytarabine, 10 mg/m 2 subcutaneously every 12 h on days 1-14 or 20 mg/m 2 subcutaneously every 12 h on days 1-7; aclarubicin, 20 mg intravenously daily on days 1-4; and granulocyte colony-stimulating factor (G-CSF), 200 mg/m 2 per day subcutaneously from day 1 until neutrophil recovery) was administered. After achieving CR, patients received the same CAG regimen or other regimens as subsequent consolidation and then allo-HSCT as soon as possible if a suitable donor was available.

| Statistical analysis
The measured outcomes were CR, partial response (PR), OS, and event-free survival (EFS). CR was defined as <5% blasts in the bone marrow (BM) with normalization of peripheral blood counts (ANC ≥ 1×10 9 /L and platelet count ≥100×10 9 /L) and complete resolution of extramedullary disease. PR was defined similarly, except for the presence of 6 to 25% BM blasts. 15,[25][26][27] Relapse was defined as the reappearance of blasts in the blood or in the bone marrow (>5%) or in any other extramedullary site after a CR. 25 Refractory disease was defined as the failure to achieve CR at the end of induction therapy. 25 OS was calculated as the interval between the diagnosis of refractory disease or relapse and death from any cause. EFS was defined as the time between the diagnosis of refractory disease or relapse and the day of the first adverse event (salvage failure, further relapse, or death from any cause). Survival probabilities were analyzed by the Kaplan-Meier method. Outcomes were 2-year overall survival and event-free survival were 68.8% (95% CI, 47.3%-83.0%) and 56.5% (95% CI, 37.1%-71.9%), respectively. The CAG regimen was well-tolerated, and no early death occurred. Our multicenter results show that the CAG regimen is highly effective and safe, representing a novel choice for adult patients with R/R-T-ALL and providing a better bridge to transplantation.

K E Y W O R D S
refractory or relapse, stem cell transplantation, T-cell acute lymphoblastic leukemia | 5329 compared between the different subgroups by the log-rank test. The independence of the categorical parameters was calculated using either a χ 2 test or Fisher's exact test. The distribution of the continuous variables was calculated using a Mann-Whitney U test. P < .05 were considered significant.

| RESULTS
The characteristics of all 41 patients are shown in Figure 1 and Table 1. There were 28 males and 13 females, and the median age was 31 years (range, 15-78 years). The numbers of patients with ETP-ALL and non-ETP-ALL were 26 (63.4%) and 15 (36.6%), respectively. There were 31 (75.6%) primary refractory patients who had undergone 1-3 prior salvage chemotherapy cycles including the VDLP, VDCP, or hypercyclophosphamide, vincristine, doxorubicin and dexamethasone (CVAD) regimens, and 10 relapsed patients (24.4%) (3 of 10 patients had undergone 1-2 prior salvage chemotherapy regimens) with a median relapse time of 14.5 months (4-48 months) ( Table 1). The last follow-up time was 1 July 2019, and the median follow-up time was 12 months (range, 1-63 months). There were 30 patients alive at the last contact, and 11 patients died of the disease.
As shown in Table 1, after one cycle of the CAG regimen among 41 patients, complete remission and partial remission were achieved in 33 (80.5%) and two (4.9%) patients, respectively. Failure to respond was observed in six patients (14.6%). ETP (n = 26) and non-ETP (n = 15) patients had a similar CR rate (80.8% vs 80.0%, P = .95). No patients experienced an early death within 1 month after the initiation of the CAG regimen. Among 41 patients, allo-HSCT was successfully performed in 27 (66%) patients (22 in CR and 5 in PR/NR), and 24 patients receiving allo-HSCT were still alive at the last follow-up. Thirty-three patients achieved CR after CAG with the intent of being transplanted, and 22 patients actually reached transplantation. Among 22 patients who received allo-HSCT after achieving CR from CAG, 10 patients directly bridged to allo-HSCT without consolidation, eight patients repeated the CAG regimen, and four patients received hyper-CVAD as consolidation before allo-HSCT.
Adverse events (AEs) were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03. As shown in Table 2, grade 3 to 4 neutropenia and thrombocytopenia occurred in 30 (73%) and 32 (78%) patients, respectively. The median duration of neutropenia during the first cycle of CAG therapy was 10 days (range 6-28 days). The most common nonhematologic AE was infection (grade 3-4, 58%). The infections that occurred during CAG chemotherapy were in the lung (55.6%), gastrointestinal tract (14.8%), skin (11.1%), bloodstream (11.1%), and unknown locations (7.4%). No cases of grade 3 to 4 nausea, vomiting, diarrhea, thrombosis/embolism, cardiac events, raised creatinine, or nervous system symptoms occurred. No early deaths occurred during the CAG regimen.
Among the 33 patients achieving CR after the CAG regimen, 22 patients successfully bridged to allo-HSCT, and 11 patients received chemotherapy-based postremission treatments. In the patients who successfully bridged to allo-HSCT

| DISCUSSION
Adult patients with R/R-T-ALL have poor outcomes. As such, a novel therapy is urgently needed. [4][5][6][7][8][9][10][11][12][13][14][15][16] Our multicenter results show that the CAG regimen is associated with a high CR rate of 80.5% and is well-tolerated, enabling most patients to bridge to allo-HSCT. Thus, this regimen represents a novel option for adult R/R-T-ALL patients. These promising results prompted a prospective multicenter trial that was initiated in June 2019 in China (Chinese Clinical Trial Registry number, ChiCTR1900023581).
The most interesting finding of this study, which included a relatively large sample size of patients from multiple centers in China, was the very high ORR rate (85.4%) associated with the use of the CAG regimen for R/R-T-ALL patients, confirming the previous results found in studies with smaller sample sizes from single centers. [20][21][22] The CAG regimen is commonly and exclusively used for R/R-AML or elderly AML patients in Asia because aclarubicin has not been launched in America and Europe. Empirically, one group in China found an unexpectedly promising primary result in which all six R/R-T-ALL patients diagnosed from April 2004 to July 2006 achieved CR. 20 Subsequently, the same group extended the treatment to another five patients, four of whom achieved CR. 21 Commonly used salvage regimens for R/R-T-ALL include FLAG-based or hyper-CVAD-based regimens, which have CR rates varying from 30% to 50%. [4][5][6]8,14,17,18 The only single-target drug for T-ALL is nelarabine, which has a CR rate varying from 31% to 36% as salvage monotherapy. 8,10,14 A recent study from the Children's Oncology Group showed that the proteasome inhibitor bortezomib plus chemotherapy could increase the CR rate to 68% among 22 patients with once-relapsed T-ALL. 16 According to the reported CR rates listed in Table 3, the CAG regimen seems to be a good salvage regimen for R/R-T-ALL, but the results should be interpreted cautiously for different eligible populations.
Good safety is another advantage of the CAG regimen for R/R-T-ALL according to its low hematological or nonhematological adverse events, which is in accordance with previous reports not only in T-ALL but also in AML. [19][20][21][22] FLAG and hyper-CVAD have early death rates of 5%-20% and 3%-15%, respectively, and infection rates of 20% to 60%. 17,[28][29][30][31] Infections of grade 3-4 occurred in 58% of patients during treatment with the CAG regimen in our study, which was similar to other regimens. 17,[28][29][30][31] However, no early deaths (within 1 month) occurred in our study. Novel treatments, such as nelarabine and bortezomib, have the adverse event of neuropathy at a rate of up to 37%, which limits their repetitive use. 8,10,12,16 However, no patients experienced neuropathy during treatment with the CAG regimen. The favorable safety of the CAG regimen is due to the low doses of araC and aclarubicin plus the use of G-CSF without using steroids. Previous studies have demonstrated that G-CSF has the potential to promote the recovery of neutrophils and an enhanced ability to kill phagocytized bacteria, 32,33 which may partially explain the safety of the CAG regimen. A high percentage (66%) of patients successfully bridged to allo-HSCT, translating into promising survival outcomes in our study, while previous studies have shown only 9.1% to 26.9% of patients bridging to allo-HSCT (Table 3). [7][8][9][10][11][12][13][14][15][16][17][18] The higher CR rate and lower toxicity of the CAG regimen allow more patients to rapidly proceed to HSCT; this is especially true for R/R-ETP-ALL patients, who would otherwise die within 6 months. 23,24 As shown in our study, of the 26 R/R-ETP-ALL patients, 14 successfully bridged to allo-HSCT after achieving CR, which contributed to the similarly promising outcomes of these ETP-ALL patients compared with those of the non-ETP-ALL patients. There were some limitations in our study. First, this was a one-arm trial rather than a randomized study, although we provided a mini-review of the historical literature. Given the rarity of R/R-T-ALL, it is unlikely that large randomized studies in adult patients will be undertaken to examine the question of which regimen is best for salvage therapy. We have initiated a prospective multicenter trial that will include 18 centers in China to validate this study. Second, the mechanism by which the CAG regimen affects T-ALL cells remains unclear, although a previous study found that T-ALL cells highly express G-CSF receptors, and G-CSF has a synergetic effect in eliminating T-ALL cells when administered simultaneously with araC and aclarubicin. 34 Moreover, 24 of 33 refractory patients in our study received only one prior induction regimen, which may introduce a bias and make it difficult to extend our findings to heavily treated patients. In summary, our multicenter results show that the CAG regimen is highly effective and well-tolerated in adult patients with R/R-T-ALL. The CAG regimen enabled most patients to successfully bridge to allo-HSCT and achieve improved outcomes; thus, it represents a novel option for this population with poor prognosis.