Similar Survival for Patients Undergoing Reduced-Intensity Total Body Irradiation (TBI) Versus Myeloablative TBI as Conditioning for Allogeneic Transplant in Acute Leukemia

International Journal of

Radiation Oncology biology

physics

www.redjournal.org

Clinical Investigation: Leukemia

Similar Survival for Patients Undergoing Reduced-Intensity Total Body Irradiation (TBI) Versus Myeloablative TBI as Conditioning for Allogeneic Transplant in Acute Leukemia John L. Mikell, MD,* Edmund K. Waller, MD, PhD,y Jeffrey M. Switchenko, PhD,z Sravanti Rangaraju, MBBS,y Zahir Ali, MS,y Michael Graiser, PhD,y William A. Hall, MD,* Amelia A. Langston, MD,y Natia Esiashvili, MD,* H. Jean Khoury, MD,y and Mohammad K. Khan, MD, PhD* Departments of *Radiation Oncology, yHematology and Oncology, and zBiostatistics and Bioinformatics, Winship Cancer Institute, Emory University, Atlanta, Georgia Received Oct 9, 2013, and in revised form Jan 28, 2014. Accepted for publication Feb 21, 2014.

Summary Reduced intensity total body irradiation (riTBI) is often given to patients who cannot tolerate myeloablative TBI (mTBI) as conditioning for stem cell transplantation in hematologic malignancies. Here we report our institutional experience, comparing outcomes of riTBI with those of mTBI for patients with acute leukemia. Our data indicate that rates of recurrence and overall survival for riTBI-treated patients were similar to those for mTBI patients and that the riTBI cohort experienced reduced toxicity.

Purpose: Hematopoietic stem cell transplantation (HSCT) is the mainstay of treatment for adults with acute leukemia. Total body irradiation (TBI) remains an important part of the conditioning regimen for HCST. For those patients unable to tolerate myeloablative TBI (mTBI), reduced intensity TBI (riTBI) is commonly used. In this study we compared outcomes of patients undergoing mTBI with those of patients undergoing riTBI in our institution. Methods and Materials: We performed a retrospective review of all patients with acute leukemia who underwent TBI-based conditioning, using a prospectively acquired database of HSCT patients treated at our institution. Patient data including details of the transplantation procedure, disease status, Karnofsky performance status (KPS), response rates, toxicity, survival time, and time to progression were extracted. Patient outcomes for various radiation therapy regimens were examined. Descriptive statistical analysis was performed. Results: Between June 1985 and July 2012, 226 patients with acute leukemia underwent TBI as conditioning for HSCT. Of those patients, 180 had full radiation therapy data available; 83 had acute lymphoblastic leukemia and 94 had acute myelogenous leukemia; 45 patients received riTBI, and 135 received mTBI. Median overall survival (OS) was 13.7 months. Median relapse-free survival (RFS) for all patients was 10.2 months. Controlling for age, sex, KPS, disease status, and diagnosis, there were no significant differences in OS or RFS between patients who underwent riTBI and those who underwent mTBI (PZ.402, PZ.499, respectively). Median length of hospital stay was shorter for patients who received riTBI than for those who received mTBI (16 days vs 23 days, respectively; P<.001), and intensive care unit admissions were

Reprint requests to: John L. Mikell, MD, 1365 Clifton Rd, NE, Ste T104, Atlanta, GA 30322. Tel: (912) 507-1503; E-mail: jmikell@emory. edu Int J Radiation Oncol Biol Phys, Vol. 89, No. 2, pp. 360e369, 2014 0360-3016/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2014.02.032

Conflict of interest: none.

Volume 89  Number 2  2014

Reduced intensity TBI

361

less frequent following riTBI than mTBI (2.22% vs 12.69%, respectively, PZ.043). Nonrelapse survival rates were also similar (PZ.186). Conclusions: No differences in OS or RFS were seen between all patients undergoing riTBI and those undergoing mTBI, despite older age and potential increased comorbidity of riTBI patients. riTBI regimens were associated with shorter length of hospital stay, fewer intensive care unit admissions, and similar rates of nonrelapse survival, which may reflect reduced toxicity. Prospective trials comparing riTBI and mTBI are warranted. Ó 2014 Elsevier Inc.

Introduction Adult patients with acute leukemia have poor outcomes after conventional chemotherapy, and allogeneic hematopoietic stem cell transplantation (allo-HSCT) is often used as a second therapy after relapse or for patients whose disease is refractory to primary therapy (1). Total body irradiation (TBI) was the first modality to be used as part of the conditioning regimen prior to allo-HSCT, and TBI continues to be used as part of HSCT conditioning regimens. Despite the emergence of chemotherapeutic conditioning regimens used in lieu of TBI, TBI offers a relatively favorable toxicity profile, and patients treated with TBI-based conditioning regimens and allo-HSCT have disease-free and overall survival (OS) rates similar to those in patients who receive chemotherapy-based conditioning regimens (2-8). In an effort to extend the benefits of allo-HSCT to patients whose Karnofsky performance status (KPS) and comorbidities might preclude them from the procedure, reduced intensity conditioning (RIC) regimens were developed (9, 10). As part of RIC, reduced intensity TBI (riTBI) is often used. riTBI commonly entails a single fraction of TBI therapy with a dose ranging from 2 to 6 Gy. With these reduced doses, radiation therapy serves to suppress the immune system and deplete enough host hematopoietic elements within the marrow to allow for donor stem cell engraftment. Rather than using larger doses of radiation therapy to eliminate disease, RIC allo-HSCT relies on the graft-versus-leukemia effect of donor T cells to achieve disease control. Currently, it is unclear whether myeloablative TBI (mTBI) and reduced intensity regimens result in similar disease control and OS rates in leukemia patients undergoing allo-HSCT. Supplemental radiation (SR) therapy treatments are often included along with TBI in an effort to reduce local failure. These treatments include testicular or cranial boosts, craniospinal irradiation, or boosts to known areas of local disease. Although some of these treatments are commonplace, such as testicular boosts for acute lymphoblastic leukemia, these regimens add toxicity and are of questionable benefit when combined with TBI (11, 12). Here we report our institutional outcomes using TBI as conditioning for HSCT for patients with acute leukemia. This study explored the outcomes for patients who specifically received TBI as part of their conditioning regimen, either as

part of RIC or full myeloablative regimens. We compared the outcomes of patients who underwent riTBI with those of patients who had full mTBI. Furthermore, we examined the effect of SR therapy regimens within RIC and full mTBI regimens to determine what benefit SR therapy affords.

Methods and Materials Patient data We conducted our analysis using patientedata collected within an institutional review board-monitored, prospectively acquired database of patients undergoing HSCT at our affiliated hospitals. Following institutional review board approval, patient data including details of the HSCT procedure, diagnosis, disease status (DS) at time of HSCT, KPS, survival time, and time to progression were extracted from an institutional database. In an effort to indirectly measure regimen toxicity, we recorded patients’ length of hospital stay (LOS) following transplantation and intensive care unit (ICU) admissions within 30 days of HSCT. Information regarding the patients’ radiation therapy regimens, including dose, fractionation, and use of supplemental treatments (such as focal boosts) was recorded.

Patient population We included all patients 18 years of age and older who were treated for acute leukemia. Diseases included acute myelogenous leukemia (AML), acute lymphoblastic leukemia (ALL), acute biphenotypic leukemia, and other undifferentiated acute leukemias. For the main analysis, patients who underwent haploidentical transplantation were excluded. The database includes patients treated from June 1985 through July 2012. We included only those patients who received TBI as part of their conditioning regimens. Patients from both RIC and myeloablative conditioning regimens were included.

Conditioning regimens RIC regimens included riTBI regimens, defined as singlefraction treatments of 6 Gy or less. mTBI included fractionated regimens of 12 Gy or more. Within the riTBI and mTBI groups, we identified patients who underwent some

362

Mikell et al.

SR therapy beyond standard regimens. We defined standard regimens as 2 Gy for riTBI and 12 Gy for mTBI. SR therapy included boosts to either testes, cranium or elsewhere, craniospinal irradiation, total skin electron beam, or escalated TBI dose beyond standard. For the subset analyses of SR, we included haploidentical patients to allow for adequate sample sizes within the subgroups. Chemotherapeutic drugs used in conjunction with riTBI or mTBI based conditioning regimens were recorded.

TBI technique During the study period, patients at our institution were treated in the lateral decubitus position. Treatment was delivered using anteroposterior and posteroanterior fields. For myeloablative radiation therapy, treatment was delivered twice a day, with an interval of at least 6 hours between fractions of 1.65 to 2 Gy. For riTBI regimens, treatment was delivered in a single fraction. For all patients, a dose heterogeneity limit of 5% was allowed. To achieve this dose constraint in myeloablative regimens, various sections of the body were blocked during exposure to the final 1 to 2 fractions. Dose was prescribed to the midplane. Lungs were shielded with blocks of Wood’s metal (Cerrobend). The TBI technique at our institution has evolved slightly over the course of the study. Initially, for patients undergoing mTBI, the radiation dose to the lungs was limited to 8 Gy, and a 3-Gy supplemental boost was given to the anterior and posterior chest walls (1.5 Gy  2 fractions to each chest wall). This was the technique used until 2009, when the technique was altered to limit the total lung dose to 10 Gy. This recent technique obviated the need for separate chest wall boosts.

Statistical analysis End points included OS, relapse-free survival (RFS), nonrelapse survival (NRS), LOHS, and rates of ICU admission. Descriptive statistics were reported for each variable. The c2 test was used to compare categorical variables, and analysis of variance (ANOVA) was used to compare numerical variables across conditioning regimens and ICU admission. Mean LOSs were compared across categorical variables, using ANOVA, whereas the Pearson correlation coefficient was used to assess the relationship between numerical variables and LOS. OS, RFS, and NRS curves for conditioning regimens were generated using the Kaplan-Meier method, and curves were compared using a log-rank test. We performed unadjusted univariate and multivariate analyses for each end point by using either Cox proportional hazards models for the survival end points, logistic regression for ICU admission, or linear regression for LOHS. The multivariate analysis included KPS, sex, diagnosis, and DS, whereas other covariates were entered in the model subject to a backward variable selection method with an alpha value of .05 removal criteria. Further subset analysis was

International Journal of Radiation Oncology  Biology  Physics

performed separately for AML and ALL patients. Statistical analysis was conducted using SAS, version 9.3 software, and the significance level for all tests was set at a P value of .05.

Results Patients We identified 226 patients who underwent TBI as part of their conditioning regimen for HSCT. Of those patients, 201 had full radiation therapy data. After we excluded haploidentical patients, 180 patients were eligible for analysis. Patients included in the analysis were treated between June 1996 and July 2012. Of the 180 patients included in the analysis, 45 underwent riTBI, and 135 underwent mTBI. Of the analyzed patients, 83 had ALL, and 94 had AML. Patient characteristics are included in Table 1. There were significant differences between the patients who received riTBI and those who received mTBI. Patients undergoing riTBI were generally older than mTBI patients (median, 56.96 years for riTBI vs 35 years for mTBI, P<.001). A higher percentage of AML patients received riTBI than ALL patients (35.11% vs 14.46%, respectively, PZ.002). With regard to transplantation type, patients undergoing matched unrelated donor (MUD) HSCT were more likely than those undergoing matched related donor (MRD) transplants to receive riTBI (44.57% vs 4.65%, respectively, P<.001). Also, the percentage of patients receiving riTBI increased significantly over the course of the study period compared with those receiving mTBI, from 0% before 1999 to 35% to 39% by 2012 (P<.001). Patients with 1 or 2 degrees of human leukocyte antigen (HLA) disparity were more likely to have been treated with riTBI than those with 0 degrees of disparity (52.38% vs 24.64%, respectively, PZ.009). Patients undergoing riTBI had lower mean CD34 counts in their grafts than those undergoing mTBI (8.28 vs 12.48, respectively, PZ.021). KPS and sex were similar across conditioning regimens. Differences between patients in each conditioning regimen group are outlined in Table 1. Details of the chemotherapy regimens and TBI doses within each conditioning regimen are included in Table 2.

Survival: All patients Of the 180 patients included in the analysis, 118 were deceased at last follow up. The median OS for all patients was 13.7 months. The median OS was 10.9 months for all riTBI patients and 15.8 months for all mTBI patients (log rank PZ.807) (Fig. 1 A). HLA disparity and CD34 cell counts were not significant on univariate analysis. Controlling for age, sex, KPS, DS, and diagnosis, there were no significant differences in OS rates between patients who underwent riTBI and those who underwent mTBI (hazard ratio [HR] 1.23, 95% confidence interval [CI] 0.76-2.02, PZ.402) (Table 2 and Fig. 1 A). Within the mTBI cohort, there were no differences in OS between patients who

Volume 89  Number 2  2014 Table 1

Reduced intensity TBI

363

Patient characteristics and univariate analysis of patient characteristics

Covariate Sex Diagnosis Status

Type KPS ICU Admission Death HLA disparity Age (years) Length of stay (days) TBI dose CD34 cell content of the graft (106 cells/kg) Year of transplantation

No. of patients undergoing riTBI (%) (NZ45)

Statistics

Level

n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) N Median N Median N Median N Median n (%) n (%) n (%) n (%)

Female Male ALL AML CR1 CR2 Other MRD MUD Low (80) High (>80) No Yes No Yes 0 1/2

1995-1999 2000-2004 2005-2009 2010-2012

20 25 12 33 26 11 8 4 41 13 28 44 1 16 29 34 11

(25.64) (24.51) (14.46) (35.11) (30.23) (30.56) (14.55) (4.65) (44.57) (28.89) (29.79) (27.33) (5.56) (25.81) (24.58) (24.64) (52.38) 45 56.96 45 16 45 200 28 8.28 0 (0) 10 (17.54) 23 (38.98) 12 (35.29)

No. of patients undergoing mTBI (%) (NZ135) 58 77 71 61 60 25 47 82 51 32 66 117 17 46 89 104 10

(74.36) (75.49) (85.54) (64.89) (69.77) (69.44) (85.45) (95.35) (55.43) (71.11) (70.21) (72.67) (94.44) (74.19) (75.42) (75.36) (47.62) 135 35.06 133 23 135 1200 13 12.48 30 (100) 47 (82.46) 36 (61.02) 22 (64.71)

Parametric P value .862 .002 .083

<.001 .913 .043 .856 .009 <.001 <.001 <.001 .021 <.001

Abbreviations: ALL Z allogeneic transplant (matched sibling/relative); AML Z acute myelogenous leukemia; CR Z complete remission; HLA Z human leukocyte antigen; KPS Z Karnofsky performance status; MRD Z matched related donor; MUD Z matched unrelated donor; SYN Z syngeneic transplant. Bold text indicates statistically significant values.

received SR and those who did not on multivariate analysis (95% CI 0.86-3.15, PZ.129). Within the riTBI cohort, there appeared to be improved survival for those who received standard riTBI compared to those who received riTBI plus SR on multivariate analysis (HR 0.23, 95% CI 0.07-0.73, PZ.013). Younger age and diagnosis of ALL were associated with improved survival within the riTBI cohort (HR 1.05, 95% CI 1.01-1.09, PZ.011, and HR 0.12, 95% CI 0.03-0.43, PZ.001, respectively).

Relapse rate and relapse-free survival: All patients Within the study period, 62 patients had a documented relapse. Of these, most relapses were documented in peripheral blood or in bone marrow. The overall relapse rate was 34.4%. The relapse rate for all riTBI patients was 24.4% compared to 37.8% for mTBI (PZ.103). The median RFS for all patients was 10.2 months. The median RFS rate for all riTBI patients was 9.5 months, compared to 10.3 months for all mTBI patients (log-rank PZ.957) (Fig. 1 B). On univariate analysis, first complete

remission (CR1) CR1 status was associated with improved RFS (HR 0.548, 95% CI 0.365-0.824, PZ.014). This effect was not maintained on multivariate analysis. On multivariate analysis, there were no differences in RFS between those who received mTBI and those who received riTBI (HR 1.18, 95% CI 0.73-1.91, PZ.499, table 3). Within the mTBI cohort, there was a strong trend toward improved RFS rate in those who received SR compared with those who did not (HR 1.89, 95% CI 1.00-3.59, PZ.051). Within the riTBI group, standard riTBI was associated with improved survival compared to riTBI plus SR therapy (HR 0.24, 95% CI 0.07-0.78, PZ.018). Younger age and diagnosis of ALL were also associated with improved RFS within the riTBI cohort (HR 1.04, 95% CI 1.00-1.09, PZ.028, and HR 0.19, 95% CI 0.060.62, PZ.006, respectively).

Nonrelapse survival: All patients The median NRS time was 23.5 months. The median NRS time was 17.5 months for all riTBI patients and 28.3

364

International Journal of Radiation Oncology  Biology  Physics

Mikell et al.

Table 2

Detailed description of TBI doses and chemotherapy regimens used in both conditioning regimens

Regimen Myeloablative TBI

Variable

Level

Chemotherapy regimen

Cyclophosphamide ATG, fludarabine, thiotepa ATG, cyclophosphamide cyclophosphamide, etoposide cyclophosphamide, fludarabine thiotepa fludarabine, thiotepa ATG, melphalan etoposide 1110 1200 1225 1295 1302 1320 1400 1110 cyclophosphamide fludarabine ATG, fludarabine cyclophosphamide, fludarabine Ara-C, ATG, cyclophosphamide, fludarabine, etoposide ATG, clofarabine, cyclophosphamide, fludarabine, etoposide Missing 200 400 550 600

TBI dose (cGy)

Reduced-Intensity TBI

Chemotherapy regimen

TBI dose (cGy)

No. of patients (NZ135)

%

115 7 6 1 2 1 1 1 1 1 107 1 5 3 11 7 1 4 27 7 1 3

85.2 5.2 4.4 .7 1.5 .7 .7 .7 .7 .7 79.3 .7 3.7 2.2 8.1 5.2 .7 9.3 62.8 16.3 2.3 7.0

1

2.3

2 34 6 4 1

75.6 13.3 8.9 2.2

Abbreviations: ATG Z anti-thymocyte globulin; Ara-C Z arabinofuranosyl cytidine; TBI Z total body irradiation.

months for all mTBI patients. This difference was not statistically significant (log-rank PZ.829) (Fig. 1 C). On multivariate analysis, there were no differences in NRS rates between riTBI and mTBI patients (HR 1.56, 95% CI 0.81-3.03, PZ.186 Table 3).

Toxicity: All patients The median length of stay post transplant (LOS) was significantly shorter for patients who underwent riTBI than for those who underwent mTBI (16 vs 23 days, respectively, P<.001). On univariate analysis, riTBI was associated with shorter LOHS than mTBI (95% CI 13.11 to 4.39, P<.0001). riTBI was also associated with decreased ICU admission within 30 days of transplantation (2.22% vs 12.69%, respectively, PZ.043).

Acute myelogenous leukemia Among AML patients in our series, 33 had riTBI and 61 had mTBI. The median OS for patients with AML was 11.2

months. The median OS for riTBI patients was 8.1 months versus 19.9 months, respectively (log-rank PZ.170) (Fig. 1 A). On multivariate analysis, mTBI was associated with improved OS compared to riTBI (HR 1.90, 95% CI 1.023.54, PZ.042) (Table 4). Within the mTBI group, addition of SR therapy did not appear to affect survival (95% CI 0.66-4.04, PZ.291). Within the riTBI group, standard riTBI appeared to be associated with improved OS compared to riTBI plus SR therapy (HR 0.21, 95% CI 0.070.66, PZ.008). The median RFS period for AML patients undergoing allo-HSCT with TBI-based conditioning was 9.7 months. The median RFS period for riTBI patients was 7.3 months compared to 12 months for mTBI patients (log-rank PZ.334) (Fig. 1 B). There were no significant differences between RFS in patients who received riTBI and that for patients who received mTBI on univariate or multivariate analysis (multivariate HR 1.72, 95% CI 0.93-3.17, PZ.085) (Table 4). The overall relapse rate for AML patients was 37.2%. The relapse rates were 27.3% for riTBI patients and 42.6% for mTBI patients (PZ.142). Within the mTBI arm, addition of SR therapy was not associated with a difference in RFS (95% CI 0.78-4.73, PZ.157). Within the riTBI arm,

Volume 89  Number 2  2014

Reduced intensity TBI

365

Fig. 1. Kaplan-Meier curves for all overall survival (A), relapse-free survival (B), and nonrelapse survival (C). AML Z acute myelogenous leukemia; ALL Z acute lymphoblastic leukemia. the standard riTBI was associated with improved RFS compared to riTBI plus SR therapy (HR 0.12, 95% CI 0.030.45, PZ.002). The median NRS for all AML patients was 17.5 months. The NRS was 10.4 months for riTBI patients. For mTBI patients, the median NRS was not reached. This difference was not statistically significant (log-rank PZ.296) (Fig. 1 C). On multivariate analysis, however, mTBI was associated with improved NRS compared with riTBI (HR 2.92, 95% CI 1.14-7.49, PZ.026, Table 4).

Acute lymphoblastic leukemia Eighty-three patients in our series underwent allo-HSCT with a diagnosis of ALL. Of those patients, approximately 66% were male. Among ALL patients, 12 received riTBI,

and 71 had mTBI-based conditioning regimens. The median OS was 17 months. The median OS for riTBI was not reached, whereas the median OS for mTBI was 15.5 months (log-rank PZ.073) (Fig. 1 A). On multivariate analysis, riTBI was associated with improved survival compared to mTBI (HR 0.25, 95% CI 0.07-0.86, PZ.028, (Table 5). Lower age was also associated with improved OS on multivariate analysis (HR 1.04, 95% CI 1.01-1.08, PZ.025). Within the mTBI arm, the addition of SR therapy was not associated with improved OS (95% CI 0.76-5.75, PZ.151). Within the riTBI arm, the addition of SR therapy was not associated with a significant effect on OS (95% CI 0.26-16.39, PZ.487). The median RFS for ALL patients was 10.2 months. The median RFS for riTBI patients was not reached, whereas

366

International Journal of Radiation Oncology  Biology  Physics

Mikell et al.

Table 3

Multivariate analysis of overall, relapse-free, and nonrelapse survival for patients with acute leukemias

Outcome Overall survival

Covariate

Level

HR

95% CI Low

95% CI Up

HR P value

Type 3 P value

XRT Intensity (riTBI vs mTBI) KPS

riTBI mTBI Low (80) High (>80) Female Male ALL AML CR1 CR2 Other riTBI mTBI Low (80) High (>80) Female Male ALL AML CR1 CR2 Other riTBI mTBI Low (80) High (>80) Female Male ALL AML CR1 CR2 Other

1.23 1.14 0.88 0.89 0.78 0.75 1.18 1.25 0.92 1.01 0.72 0.88 1.56 1.14 0.86 1.01 0.74 0.77 -

0.76 0.71 0.56 0.54 0.46 0.38 0.73 0.79 0.58 0.63 0.42 0.47 0.81 0.57 0.45 0.51 0.35 0.28 -

2.02 1.84 1.40 1.44 1.35 1.46 1.91 1.99 1.43 1.62 1.22 1.65 3.03 2.29 1.63 2.01 1.59 2.12 -

.402 .592 .593 .625 .379 .395 .499 .347 .700 .968 .222 0.683 .186 .707 .637 .967 .447 0.611 -

.402

Sex Diagnosis Status

Relapse-free survival

XRT Intensity (riTBI vs mTBI) KPS Sex Diagnosis Status

Nonrelapse survival

XRT intensity KPS Sex Diagnosis Status

.592 .593 .625 .603

.499 .347 .700 .968 .461

.186 .707 .637 .967 .738

Abbreviations: ALL Z allogeneic transplant; AML Z acute myelogenous leukemia; ATG Z anti-thymocyte globulin; CI Z confidence interval; CR Z complete remission; HR Z hazard ratio; KPS Z Karnofsky performance status; mTBI Z myeloablative TBI; riTBI Z reduced intensity TBI; TBI Z total body irradiation; XRT Z radiation therapy.

the median RFS for mTBI patients was 9.5 months (logrank PZ.113) (Fig. 1 B). On multivariate analysis, RFS trended toward improvement with riTBI compared to mTBI (HR 0.33, 95% CI 0.11-1.01, PZ.052, Table 5). Increasing age was associated with worse RFS (HR 1.04, 95% CI 1.00-1.07, PZ.029). The relapse rate for all ALL patients was 31.3%. The relapse rate for riTBI patients was 16.7% compared to 33.8% for mTBI patients (PZ.342). Within the mTBI cohort, the addition of SR therapy was not associated with a difference in RFS (95% CI 0.82-6.12, PZ.113). Within the riTBI cohort, the addition of SR therapy was not associated with a difference in RFS (95% CI 0.41-28.14, PZ.259), although, again, small sample size limits this analysis (nZ12). The median NRS for all ALL patients undergoing the TBI-based conditioning regimens and allo-HSCT was 61.6 months. The median NRS for riTBI patients was not reached. The median NRS for mTBI patients was 28.3 months. This difference was not statistically significant

(log-rank PZ.194) (Fig. 1 C). On multivariate analysis, there were no significant differences in NRM between riTBI and mTBI patients (multivariate analysis HR 0.72, 95% CI 0.20-2.58, PZ.616, Table 5).

Discussion In a large retrospective review using a database compiled by the Center for International Blood and Marrow Transplant Research (CIBMTR), Marks et al (13) compared the outcomes of patients undergoing full myeloablative HSCT with that of patients undergoing RIC (13). The authors reported similar age-adjusted OS but higher rates of relapse in the RIC cohort and no differences in treatment-related mortality (TRM). Of the 98 patients in the RIC arm, only 36 received TBI-based conditioning (the remainder received chemotherapy-based conditioning). On multivariate analysis, use of TBI conditioning was associated with improved OS

Volume 89  Number 2  2014 Table 4

Reduced intensity TBI

367

Multivariate analyses of overall, relapse-free, and non-relapse survival for patients with acute myelogenous leukemia

Outcome Overall survival

Covariate

Level

HR

95% CI Low

95% CI Up

HR P Value

Type 3 P Value

XRT intensity

riTBI mTBI Low (80) High (>80) Female Male CR1 CR2 Other riTBI mTBI Low (80) High (>80) Female Male CR1 CR2 Other riTBI mTBI Low (80) High (>80) Female Male CR1 CR2 Other

1.90 1.68 1.12 0.96 0.94 1.72 1.60 1.07 0.85 0.94 2.92 2.01 1.30 0.82 0.82 -

1.02 0.89 0.60 0.48 0.41 0.93 0.86 0.58 0.42 0.42 1.14 0.78 0.53 0.30 0.22 -

3.54 3.16 2.08 1.94 2.16 3.17 2.98 1.96 1.70 2.08 7.49 5.17 3.18 2.20 3.11 -

.042 .110 .722 .910 .883 .085 .136 .838 .642 .873 .026 .148 .559 .689 .776 -

.042

KPS Sex Status

Relapse-free survival

XRT intensity KPS Sex Status

Nonrelapse survival

XRT Intensity KPS Sex Status

.110 .722 .988

.085 .136 .838 .896

.026 .148 .559 .917

Abbreviations: CR Z complete remission; HR Z hazard ratio; KPS Z Karnofsky performance status; mTBI Z myeloablative TBI; riTBI Z reduced intensity TBI; TBI Z total body irradiation; XRT Z radiation therapy. Bold text indicates statistically significant values.

compared with chemotherapy-based regimens. However, the question of how TBI-based RIC would compare with mTBI conditioning regimens remains, as the previous study was confounded by its inclusion of chemotherapy-based regimens along with TBI in the RIC cohort. Our data, the first of its kind to our knowledge, demonstrate OS, RFS, and relapse rates that are similar for riTBI and mTBI across all patients with acute leukemias. On subgroup analysis, survival for patients with ALL appeared to be improved with riTBI, whereas survival for patients with AML was better with mTBI. The differences in OS for AML patients appears to be driven by worse NRS, as RFS was similar for the 2 regimens. In the case of ALL patients, the improvement in OS appears to be driven by nonsignificant improvements in RFS and median NRS. For RIC to produce similar age-adjusted RFS and relapse rates across all groups might come as a surprise, as those patients undergoing RIC were older, generally more ill, and received lower radiation doses than their counterparts undergoing myeloablative conditioning for the same level and type of disease. These similarities in RFS and relapse rate are paired with reduced acute toxicity, as reflected by reduced LOS and rates of ICU admission following riTBI transplantation. Although NRS was worse with riTBI in AML patients, likely due to the older age and

higher comorbidity in the riTBI group. The similar RFS in the ALL group may indicate that the overall toxicity burden imparted by mTBI in ALL patients results in a similar NRS, as patients treated with riTBI, who are generally older and less fit prior to transplantation. Some of the differences in the observed versus expected outcomes in the riTBI cohort, compared to the mTBI cohort, may be secondary to the fact that a large proportion of ALL patients in our riTBI cohort were in CR1, which has previously been identified as a low-risk treatment group with low rates of relapse following nonmyeloablative conditioning (14). In addition, reduced toxicity (as indicated by reduced LOS, ICU admissions, and surprisingly similar NRS in the ALL patients) may bolster OS in our riTBI cohort. Last, it may also be possible that the graftversus-leukemia effect is more important for determining outcome after any TBI-based conditioning regimen than the antileukemic effect of more toxic conditioning regimens. The differences in OS and NRS seen between AML and ALL patients appear to reflect differences in patient selection for riTBI. ALL patients appear to be selected for riTBI predominately secondary to positive prognostic factors (CR1), whereas AML patients appear to be routed toward riTBI if they are poor candidates for mTBI. Ultimately, the 2 conditioning regimens produced similar

368

International Journal of Radiation Oncology  Biology  Physics

Mikell et al.

Table 5

Multivariate analysis of overall, relapse-free, and nonrelapse survival for patients with acute lymphoblastic leukemia

Outcome Overall survival

Covariate

Level

HR

95% CI Low

95% CI Up

HR P value

Type 3 P value

XRT intensity

riTBI mTBI Low (80) High (>80) Female Male CR1 CR2 Other

0.25 0.72 0.66 0.39 0.46 1.04 0.33 0.97 0.79 0.47 0.93 1.04 0.72 0.51 0.69 0.55 0.59 -

0.07 0.31 0.31 0.17 0.14 1.01 0.11 0.45 0.39 0.20 0.32 1.00 0.20 0.14 0.24 0.17 0.10 -

0.86 1.66 1.40 0.92 1.56 1.08 1.01 2.09 1.62 1.08 2.72 1.07 2.58 1.92 1.97 1.86 3.32 -

.028 .444 .280 .032 .215 .025 .052 .935 .526 .075 .897 .029 .616 .319 .484 .339 .546 -

.028

KPS Sex Status

Relapse-free survival

Age (yrs) XRT intensity KPS Sex Status

Nonrelapse survival

Age (yrs) XRT intensity KPS Sex Status

riTBI mTBI Low (80) High (>80) Female Male CR1 CR2 Other riTBI mTBI Low (80) High (>80) Female Male CR1 CR2 Other

.444 .280 .094

.025 .052 .935 .526 .142

.029 .616 .319 .484 .626

Abbreviations are as in Table 4.

relapse rates and RFS, indicating that both are effective at controlling disease. Multiple reports have demonstrated lower toxicity and TRM in patients undergoing RIC compared to those undergoing myeloablative conditioning (15-18). Analysis of toxicity indicators measured in our study, LOS, ICU admissions, and NRS, appears to reflect significantly less toxicity associated with riTBI. Although these outcomes are indirect measures of toxicity, they are readily defined and clinically relevant end points. Although the previous CIBMTR comparison of conditioning regimens revealed no differences in TRM, we believe that shorter hospital stays and fewer ICU admissions may indicate improved rates of TRM (at least in the acute setting) when treating with riTBI. We also believe that the finding of similar NRS for riTBI and mTBI in ALL patients reflects a reduction in the overall toxicity burden, as the older and less fit riTBI patients would be expected to have worse NRS than mTBI, if the toxicities of the regimens were equal. Given the significant heterogeneity of radiation therapy regimens used prior to HSCT at our own institution, we sought to measure the possible benefit of additional sitespecific SR therapy added to standard courses of 2-Gy riTBI or 12-Gy mTBI. At our institution, the use of TBI plus SR

regimens are often markers for more advanced disease, as clinicians increase TBI dose or add focal regimens in an effort to maximize disease control in areas considered at high-risk for local failure. In our review, we found mixed evidence for the benefit of SR therapy. In an examination of all patients with acute leukemia, we found that the addition of SR appeared to be associated with worse outcomes (OS and RFS) for patients receiving riTBI but similar outcomes for patients receiving mTBI. Analysis within specific diseases yielded similar results. The benefit of SR remains difficult to interpret from our data, given small sample sizes and the heterogeneity of the SR regimens. Our analysis does carry some limitations. In our attempt to extend the analysis to the largest patient population, we generated a heterogeneous cohort of all patients with acute leukemia. We attempted to control this heterogeneity by performing subset analyses within AML and ALL as well as by including diagnosis within our multivariate analyses. As a consequence, some of our subset analyses suffer from small sample sizes, especially in the SR analyses. Additionally, the retrospective nature of our review limits its immediate applicability. We lack prospectively acquired toxicity scores, which limited our ability to perform a more detailed analysis of TRM and regimen-related toxicity. The

Volume 89  Number 2  2014

retrospective nature of our study also carries the inherent risk of patient selection bias. Another limitation is the disparity in the frequency of riTBI throughout the study period, which introduces some limitation in the interpretation of long-term survival and relapse data. Supportive measures for transplantation patients have also improved over the years, which limits the interpretation of our acute toxicity measures (LOS and ICU admissions). Given these limitations, we intend for this hypothesis-generating data to serve as the basis for future prospective work to determine for which mTBI may still be appropriate.

Reduced intensity TBI

6.

7.

8.

Conclusions 9.

Overall, our experience indicates that riTBI-based conditioning regimens produced similar OS, RFS, and relapse rates for all patients with acute leukemias compared to mTBI-based regimens. The use of riTBI-based conditioning regimens may be associated with improved survival in patients with ALL. For AML, survival seemed to improve with mTBI, likely as a result of worsened NRSin riTBI patients, but RFS and relapse rates remained similar. These favorable results with riTBI are paired with shorter hospital stays following transplantation and fewer ICU admissions, reflecting improved toxicity profile of riTBI-based conditioning regimens. Our findings provide a strong basis for protocols designed to extend the use of riTBI into younger, more fit patients, potentially obviating the need for more toxic myeloablative therapy.

References 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013;63:11-30. 2. Davies SM, Ramsay NK, Klein JP, et al. Comparison of preparative regimens in transplants for children with acute lymphoblastic leukemia. J Clin Oncol 2000;18:340-347. 3. Dusenbery KE, Daniels KA, McClure JS, et al. Randomized comparison of cyclophosphamide-total body irradiation versus busulfancyclophosphamide conditioning in autologous bone marrow transplantation for acute myeloid leukemia. Int J Radiat Oncol Biol Phys 1995;31:119-128. 4. Kroger N, Zabelina T, Kruger W, et al. Comparison of total body irradiation vs busulfan in combination with cyclophosphamide as conditioning for unrelated stem cell transplantation in cml patients. Bone Marrow Transplant 2001;27:349-354. 5. Ringden O, Labopin M, Tura S, et al. A comparison of busulphan versus total body irradiation combined with cyclophosphamide as

10.

11.

12.

13.

14.

15.

16.

17.

18.

369

conditioning for autograft or allograft bone marrow transplantation in patients with acute leukaemia. Acute leukaemia working party of the European Group For Blood And Marrow Transplantation (EBMT). Br J Haematol 1996;93:637-645. Santos GW, Tutschka PJ, Brookmeyer R, et al. Marrow transplantation for acute nonlymphocytic leukemia after treatment with busulfan and cyclophosphamide. N Engl J Med 1983;309:1347-1353. Shi-Xia X, Xian-Hua T, Hai-Qin X, et al. Total body irradiation plus cyclophosphamide versus busulphan with cyclophosphamide as conditioning regimen for patients with leukemia undergoing allogeneic stem cell transplantation: A meta-analysis. Leuk Lymphoma 2010;51: 50-60. Copelan EA, Hamilton BK, Avalos B, et al. Better leukemia-free and overall survival in aml in first remission following cyclophosphamide in combination with busulfan compared with TBI. Blood 2013;122: 3863-3870. Giralt S, Estey E, Albitar M, et al. Engraftment of allogeneic hematopoietic progenitor cells with purine analog-containing chemotherapy: harnessing graft-versus-leukemia without myeloablative therapy. Blood 1997;89:4531-4536. McSweeney PA, Niederwieser D, Shizuru JA, et al. Hematopoietic cell transplantation in older patients with hematologic malignancies: Replacing high-dose cytotoxic therapy with graft-versus-tumor effects. Blood 2001;97:3390-3400. Alexander BM, Wechsler D, Braun TM, et al. Utility of cranial boost in addition to total body irradiation in the treatment of high risk acute lymphoblastic leukemia. Int J Radiat Oncol Biol Phys 2005;63:11911196. Quaranta BP, Halperin EC, Kurtzberg J, et al. The incidence of testicular recurrence in boys with acute leukemia treated with total body and testicular irradiation and stem cell transplantation. Cancer 2004;101:845-850. Marks DI, Wang T, Perez WS, et al. The outcome of full-intensity and reduced-intensity conditioning matched sibling or unrelated donor transplantation in adults with Philadelphia chromosome-negative acute lymphoblastic leukemia in first and second complete remission. Blood 2010;116:366-374. Kahl C, Storer BE, Sandmaier BM, et al. Relapse risk in patients with malignant diseases given allogeneic hematopoietic cell transplantation after nonmyeloablative conditioning. Blood 2007;110:2744-2748. Baron F, Maris MB, Sandmaier BM, et al. Graft-versus-tumor effects after allogeneic hematopoietic cell transplantation with nonmyeloablative conditioning. J Clin Oncol 2005;23:1993-2003. Diaconescu R, Flowers CR, Storer B, et al. Morbidity and mortality with nonmyeloablative compared with myeloablative conditioning before hematopoietic cell transplantation from HLA-matched related donors. Blood 2004;104:1550-1558. Sorror ML, Maris MB, Storer B, et al. Comparing morbidity and mortality of HLA-matched unrelated donor hematopoietic cell transplantation after nonmyeloablative and myeloablative conditioning: Influence of pretransplantation comorbidities. Blood 2004;104:961968. Sorror ML, Sandmaier BM, Storer BE, et al. Long-term outcomes among older patients following nonmyeloablative conditioning and allogeneic hematopoietic cell transplantation for advanced hematologic malignancies. JAMA 2011;306:1874-1883.