Fractionated total body irradiation in allogeneic bone marrow transplantation in leukemia patients: analysis of prognostic factors and results in 136 patients

Radiotherapy and Oncology 48 (1998) 267–276

Fractionated total body irradiation in allogeneic bone marrow transplantation in leukemia patients: analysis of prognostic factors and results in 136 patients Vittorio Donato a ,*, Valter Iacari a, Alfredo Zurlo c, Alberta Capua a, Vincenzo Tombolini e, Enzo Banelli a, Riccardo Maurizi Enrici a, Cinzia De Felice d, Giancarlo Giacco d, Anna Paola Iori e, William Arcese b, Carissimo Biagini a a

Department of Radiotherapy, Institute of Radiology, University of Rome ‘La Sapienza’, V. San Cipriano no. 60, 00136 Rome, Italy b Department of Haematology, University of Rome ‘La Sapienza’, Rome, Italy c Department of Radiotherapy, University of Rome ‘Tor Vergata’, Rome, Italy d Department of Medical Physics of Radiotherapy, University of Rome ‘La Sapienza’, Rome, Italy e Institute of Radiology, University of Aquila, Rome, Italy Received 1 August 1997; revised version received 10 June 1998; accepted 26 June 1998

Abstract Background and purpose: The results of a single-institution series of patients with chronic and acute leukemias are analyzed with regard to literature-reported predictor variables. Materials and methods: Between 1985 and 1994, 136 patients, 82 patients with chronic myeloid leukemia (CML) and 54 with acute leukemia (AL), received a uniform preparatory regimen of fractionated total body irradiation (TBI; 12 Gy in 3 days) plus different chemotherapy regimens before bone marrow transplantation. Eighty-six patients were considered to be in early phase of disease (CML in chronic phase or AL in first complete remission) and 50 in advanced phase (all those beyond first remission or first chronic phase). Ninety-five patients received unmanipulated allogeneic BM, and 41 T-lymphocyte-depleted BM. Results: The 5-year overall survival (OS) and disease-free survival (DFS) of the whole series were 43% and 31%, and median survival was 43 and 10 months, respectively. A Cox proportional hazard model identified variables related to overall and disease-free survival. For OS, graft versus host disease (GVHD) was the first independent variable (P , 0.0001), followed by age (P , 0.001), T-depletion (P , 0.01), disease status (P , 0.05) and type of leukemia (P , 0.05). With regard to DFS, only T-depletion (P , 0.0001), disease status (P , 0.01) and GVHD (P , 0.01) resulted predictor factors. Early complications after BMT were reported in 59 patients, TBIinduced delayed toxicity in 9 patients, and 16 patients suffered late complications. Conclusions: Our results confirm the curability of early phase leukemias with standard fractionated TBI-induced Allogeneic bone marrow transplantation (ABMT). With an homogeneous fractionated TBI schedule as employed in our series, T-cell depletion negatively affected the outcome.  1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: Fractionated total body irradiation; Allogeneic bone marrow transplantation; Leukemia

1. Introduction Bone marrow transplantation (BMT) is standard treatment for many hematological malignancies, including chronic myelogenous leukemia (CML), acute myelogenous leukemia (AML), and acute lymphoblastic leukemia (ALL).

* Corresponding author.

Worldwide, different regimens are used in bone marrow transplantation centers to prepare leukemia patients for allogeneic BMT. Given different disease stages, several authors have tested conditioning regimens using different combinations and high doses of chemotherapeutic agents such as cyclophosphamide (Cy), busulfan (BU) and etoposide (VP-16). Total body irradiation (TBI) is incorporated into a number of preparatory regimens, being administered either as a single exposure (STBI) at relatively low dose

0167-8140/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved PII S0167-8140 (98 )0 0069-3

268

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

rate (0.05-0.25), or through several fractions (FTBI), usually at higher total doses. The two main functions of TBI before allogeneic BMT are eradication of recipient immuno-competent cells to allow engraftment, and achievement of a significant leukemic cell killing, thus also providing physical space for allogeneic stem cells [14]. The influence of dose fractionation in attaining these objectives is controversial. Using FTBI, most radiobiological data suggest a better differential effect between normal and malignant cells, resulting in lower toxicity [8,26]. On the other hand, a less immunosuppressive effect has been reported with FTBI, that may lead to detrimental results in terms of graft failure or rejection, particularly with the concurrent use of T cell-depletion [9]. Some authors also reported a higher relapse rate with fractionated irradiation regimens, pointing out the need for either a higher total dose and/or dose rate [40]. Survival post-BMT has been linked to relapse by failure to achieve leukemic cell eradication and major preparatory regimen-related toxicity, mainly interstitial pneumonitis (IP). The status of the disease at the time of BMT is the strongest factor influencing the probability of relapse [11,15,20,41], whilst the age of patients and the occurrence of post-transplant acute graft versus host disease (a-GVHD) influence the rates of major non-leukemic complications [17,18,33,44]. Increases of chemotherapy and TBI dosage may result in a decrease in the number of relapses, but at the cost of more toxicity without improvements in disease-free and overall survival [6]. Fractionated total body irradiation (FTBI) and high-dose cyclophosphamide (CY) is certainly one of the most commonly used ablative therapies. With this regimen, relapse rates are reported to range from 10–25% for patients with good prognostic factors, to 50-80% for patients with advanced stages of leukemia [3]. Similar effectiveness has been reported with other combined TBI/chemotherapy regimens [3,4,28,39,49]. Hereafter, we report our experience in 136 leukemic patients who were treated with FTBI and high dose chemotherapy before receiving allogeneic BMT.

2. Material and methods Between January 1985 and December 1994, 162 patients with different hematological malignancies received allogeneic bone-marrow transplantation (BMT) with different conditioning regimens at the Department of Radiotherapy of the Radiology Institute and Department of Hematology of the University of Rome ‘La Sapienza’. In the present series we considered 136 leukemic patients, 82 with chronic myeloid leukemia (CML) and 54 with acute leukemia (AL), who received a uniform preparatory regimen of fractionated TBI (12 Gy over 3 days) before BMT (day 0). Patient characteristics are listed in Table 1. With regard to disease status, 86 patients were considered

to be in early phase (CML in chronic phase or AL in first complete remission), and 50 in advanced phase (all those beyond first remission or first chronic phase). Ninety-five patients received unmanipulated allogeneic BM, and 41 T-lymphocytes-depleted BM by the soybean agglutination and E-rosetting techniques. All patients received combined TBI/chemotherapy conditioning regimens. Different chemoterapeutic agents were administered either simultaneously or slightly before TBI: Cy (60 mg/kg on days −3, −2) was given in 68 patients, all with the early phase of the disease; VP-16 (20 mg/kg by continuous 24 h infusion on days −7, −6, −5) in 49 patients (31 in advanced phase and 18 in early phase); and Cy plus Ara-C (4 doses of 3 mg/m2 on days −5, −4) in 19 advanced phase patients. Post-transplant immunosuppression was adopted in order to prevent GVHD. Cyclosporine was administered to all patients, with doses of 1–3 mg/kg per day starting from day −1 to day +50, and thereafter slowly tapered until 5– 12 months after BMT in the absence of GVHD. Forty-seven patients also received Methotrexate administered i.v. at a dose of 15 mg/m2 on day +1, and then 10 mg/m2 on days +3, +6, +11. Prednisone was given to 14 patients, starting from 0.5 mg/kg on day +8 and therefore slowly tapered until day +180. Fractionated TBI was administered by a CGR ‘Neptune’ 6 MeV Linear Accelerator (LA), a Toshiba 4 MeV LA and, from 1989, a Siemens 6 MeV LA. TBI was perTable 1 Patient characteristics

Total number Male Female Mean age (±SD) Leukemias AL CML Disease status Early Advanced Age ,20 &great20 T-depletion Yes No No T-depletion Early Advanced CML T-depletion No T-depletion GVHD Acute No Chronic

n

%

136 82 54 26.8 (±12.4)

60.3 39.7

54 82

39.7 60.3

86 50

63.2 36.8

42 94

30.9 69.1

41 95

30.1 69.9

58 37

61.1 38.9

37 45

45.1 54.9

54 64 18

39.7 46.1 13.2

269

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

formed in 2 daily fractions, with single doses of 2 Gy spaced by an interval of 6 h; the treatment was repeated for 3 successive days, (from day −4 to −2) up to a total dose of 12 Gy at an instantaneous dose-rate with a range of 14–18 cGy/min. During treatment, patients were on a special couch with Plexiglas walls (cradle) and alloy trays to support tissue compensators. The source-patient distance was 290 cm; the unit head turned 45 degrees to obtain a ‘diamond’ shape of the field with the patient laying along the longitudinal axis of a 160 cm. field. Dose homogeneity was assured by the use of four fields: posterior–anterior, anterior–posterior and two lateral–laterals, with the patient alternatively assuming a supine and right or left decubitus position which was used at each fraction. To reduce interstitial pneumonitis (IP) occurrence, the dose was calculated on the midline of lung parenchyma, therefore resulting in a dose reduction of about 6-7% on mediastinum. For this reason we administered a boost to the mediastinum using two opposing fields, A.P. and P.A. in order to obtain a 12 Gy dose at the end of each fraction. Plexiglas tissue compensators were used for head, neck, and below hips. Dosimetry was performed with Capintec 194 A dosimeters, Farmer type ionization chambers of 0.6 cm3 and a 60 × 60 cm-wide Plexiglas phantom, made of sheets, each of 1cm width. During treatment, dosimeters were positioned on the wall of the cradle to allow real time evaluation and correction.

Statistical analysis was performed on a personal computer with an SPSS software package. Survival analyses were obtained using the Kaplan–Meier method [21], either for overall survival or disease-free survival. Overall survival (OS) was counted from the day of BMT until death from any cause. disease-free survival (DFS) was counted from the day of BMT until relapse or death from any cause. The statistical test for equality of survival distributions for single factor was the Logrank test [32]. To assess the influence of predictor variables on the survival times, the Cox proportional hazard model with multiple covariates was performed [10]. Predictor variables were selected in a forward fashion: the score statistic was used for entering variables into the model (if the probability of its score statistic was 0.05 or less) and the likelihood-ratio statistic based on the maximum partial likelihood estimates for variable removal (if the probability of its likelihood-ratio was 0.10 or greater). The model was tested for the hypothesis that all parameters were 0. For all patients, follow up ranged from 3 to 133 months (median 21.5 months).

3. Results All the 136 patients receiving transplants survived at least 1 month and showed signs of marrow engraftment. At the time of our analysis, with regard to OS, 76 (56%) out of 136 patients died. Forty-six out of 60 living patients are in continuous complete remission. Fourteen experienced

Table 2 Five-year overall (OS) and disease-free (DFS) survivals OS (%) Whole series Leukemias AL CML Disease status Early Advanced Age , 20 .20 T-depletion Yes No No T-depletion Early Advanced CML T-depletion No T-depletion GVHD Chronic No Acute

P

43

Med (m)

DFS (%)

43

31

.0.1 38 47

16 58

34 30

69 14

36 22

47 39

42 26

39 47

3 44

NR 14

55 28

39 NR

0 56

NR 39 13

53 25 31

med (m), median survival time in months; NR, not reached.

7 20 NR 8 ,0.001

,0.01 76 44 32

12 9

,0.001

,0.01 34 61

15 5

,0.001

,0.01 59 34

8 12

.0.1

,0.05 35 50

10

,0.01

.0.1 50 41

Med (m)

.0.1

,0.05 52 29

P

8 NR ,0.01 NR 8 8

270

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

Table 3 Cox proportional hazards model for overall survival (OS) Predictor variables

Coding scheme for categorical variables

GVHD

Chronic (0,0) Acute (1,0); No (0,1) Continuous (years) Yes (1); no (0) Early (0); advanced (1) AL (1); CML (0) Continuous (days)

Age T-depletion Disease status Leukemias Time Interval

Regression coefficient

Standard error

Significance level

Relative risk

1.6942 0.9818 0.0402 0.7964 0.5367 0.7859 (Not in the equation)

0.4588 0.446 0.0114 0.2785 0.2531 0.3182

,0.0001 ,0.001 ,0.05 ,0.001 ,0.01 ,0.05 ,0.05 .0.5

5.4421 2.6691 1.0410 2.2175 1.7104 2.1943

a relapse of the disease. For DFS, 90 out of 136 patients (66%) suffered the event. The 5-year OS and DFS are summarized in Table 2. Concerning GVHD occurrence, chronic GVHD was observed in 7 (17%) T-cell depleted patients and in 11 (11.5%) patients that received unmanipulated BM (P = 0.107). Acute GVHD occurred, respectively, in 11 (27%) and 43 (45%) patients (P = 0.007), and no GVHD was observed in 23 (56%) and 41 (43%) patients (P = 0.39). 3.1. Multivariate analysis To identify variables related to overall and disease-free survival that appear to be good predictors, we built a model based on the Cox regression procedure called the Cox proportional hazards model. For each categorical predictor variable, the assumption that hazards for two or more groups were proportional had been assessed by the ‘Logminus-log survival plot’. We considered six predictor variables either for overall or disease-free survival. Type of leukemia, T-depletion, GVHD and disease status were treated as categorical covariates, while age and time interval between diagnosis and BMT were considered as continuous variables. The main statistic parameters of the model are reported in Tables 3 and 4: coding scheme for categorical variables, regression coefficient with its standard error, significance level of the coefficient and the ‘relative risk’ associated

with the variable. For a continuous variable it gives the percentage change in the hazard rate for a unit increase of co-variate, while for categorical co-variate it means the ratio of the estimated hazard for a case with the characteristic to that of a case without the characteristic. For overall survival (Table 3) only the time interval between diagnosis and BMT is not relevant to the influencing of the survival function (P . 0.5). Ordered by their significance level for entry into the model, the first independent variable introduced was GVHD (P , 0.0001). Assuming as reference category chronic GVHD, acute GVHD was associated with a survival worsening of 5.44 times, while the absence of GVHD involved a relative risk of 2.67 (Fig. 3). The second independent variable was the age (P , 0.001), with a relative risk of 1.04 for each additional year of age, followed by T-depletion (P , 0.01) in which Tdepleted BMT suffers a survival worsening of 2.22 times. Disease status was the fourth independent variable entered into the model (P , 0.05), with an associated relative risk of 1.71 for advanced phase in comparison to early phase. In Fig. 4 separate hazard functions are plotted (the death rate per unit of time) for overall survival according to disease status. Type of leukemia was the last variable to be entered into the model (P , 0.05), with a survival worsening of 2.19 times for AL in comparison to CML. This model was tested for the null hypothesis that all parameters were 0, either by likelihood-ratio test or overall chi-square, with strongly significance level (P , 0.0001).

Table 4 Cox proportional hazards model for disease-free survival (DFS) Predictor variables

Coding scheme for categorical variables

Regression coefficient

Standard error

T-depletion Disease status GVHD

Yes (1); no (0) Early(0); advanced(1) Chronic (0,0) Acute (1,0); No (0,1); Continuous (years) AL (1); CML (0) Continuous (days)

1.1177 0.6734

0.2271 0.2166

1.1677 0.8406 (Not in the equation) (Not in the equation) (Not in the equation)

0.3960 0.3871 .0.1 .0.1 .0.1

Age Leukemias Time interval

Significance level , , , , ,

0.0001 0.01 0.01 0.01 0.05

Relative risk 3.0580 1.9609 3.2146 2.3178

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

In regard to the disease-free survival dependent variables (Table 4), the same model found only three independent variables as predictor factors: T-depletion (P , 0.0001) was the first, with a relative risk of 3.06 times greater for T-depleted BMT as against unmanipulated BMT. Disease status was the second independent variable to be entered into the model (P , 0.01), with a relative risk of 1.96 times greater for advanced phase versus early phase. The last factor was GVHD (P , 0.01), with a relative risk of 3.21 for acute GVHD and of 2.32 for no GVHD versus chronic GVHD. Type of leukemia, age and time interval between diagnosis and BMT did not influence the prognosis (P . 0.5). This model also reported a high significance level (P , 0.0001). Early complications after BMT were: oral mucositis in 28 cases and mild skin erythema in 31 patients. These complications are probably connected to the use of compensator filters which are used for the homogenization of the radiation range. We observed TBI-induced delayed toxicity in 9 patients (6 dead, 3 alive) with interstitial pneumonitis occurring between 10 and 120 days postirradiation, peaking at about day 40. Of these patients, 4 cases of CMV were found and treated with antiviral therapy (Gancyclovir). Other pulmonary complications were: Aspergillus in 6 cases (2 dead, 4 alive), Candida in 7 (all dead) and Klebsiella in 4 patients (all alive). In the patients who underwent T-depletion or who did not, differences in complication incidence were not found, especially in interstitial pneumonitis cases. Due to the low number of patients with interstitial pneumonitis, we cannot carry out a statistical analysis. We observed 10 other infections from Proteus (3 alive), Pseudomonas (3 alive) and viral (4 alive). A low number of complications, that should be considered in TBI induced late toxicity was observed in this series: veno-occlusive disease (VOD) in 3 cases (2 dead, 1 alive), thyroid abnormalities in 2 cases, cataracts developed in 5 patients and chronic keratoconjunctivitis in 1 case. Two patients reported late cardiac abnormalities; femoral aseptic osteonecrosis occurred in 3 cases. All irradiated women developed primary ovarian failure as indicated by elevated levels of follicle stimulating hormone (FSH); 13 (24%) patients below the age of 35 recovered ovarian function after 36 months from BMT.

4. Discussion Despite the widespread use of TBI in preparatory regimens for BMT, the optimal TBI schedule remains a controversial issue. Different TBI parameters such as total dose, dose rate, size of fraction, fractionation, influence the survival results, both in terms of relapse and major complication rates [29]. In fact, bone marrow transplantation is associated with significant and often life-threatening morbidity, resulting not only from infection and GVHD but also from highdose chemoradiotherapy. The patients receive TBI immedi-

271

ately before or simultaneously with chemotherapy administration, so it is difficult to separate toxicity induced from TBI or from chemotherapy. Several experiences have demonstrated that TBI related toxicity varies according to different irradiation parameters [29,44], such as total doses, fractionation and dose rate which all affect the incidence of lung complications. A higher TBI dosage or dose rate results in decreased relapse occurrence: this advantage is usually offset by increased morbidity, resulting in unchanged disease-free and overall survival [5,6]. Radiobiological and clinical data suggest that FTBI can improve the therapeutic index reducing late toxicity to organs at risk such as lung, lens, kidney and growing cartilage, which is a critical problem, particularly in pediatric patients [8,13,22,30,31,36,38,46]. Pulmonary complications stand among the most important post-BMT morbidities, and even though several factors contribute independently to lung toxicity [16,30,42] radiation doses to lung are crucially related to interstitial pneumonitis (IP) incidence, so that many BMT centers routinely use lung shielding [23,37]. Radiobiological studies and clinical experiences demonstrated that lung cells are able to recover at least partially from the sublethal damage between fractionated doses of TBI, and therefore can tolerate higher total dose and dose rate compared with STBI treatment [2,8,29,37,40,42,43,46], but recent experience with STBI at low dose rate (between 0.05 and 0.15 cGy/min) reported complication rates comparable to hyperfractionated techniques. With fractionation and lung shielding, the risk of IP with TBI in one comparative study did not result worse than that with chemotherapy ablative therapy alone [34]. Nevertheless, a recent retrospective study that excluded the influence of GVHD and chemotherapy effects on lungs, found that only dose rate and age of BMT recipients significantly influenced the incidence of IP, but not TBI fractionation. Again, the results of earlier reports favoring FTBI compared to STBI in AL [7], have recently been questioned by a new series of AL in early disease in which no influence of fractionation in any outcomes could be found [27,33]. Some authors, comparing recent results with similar historical series, considered the decreased influence of TBI schedules on outcomes to be strongly related to new improvements in the management of BM recipients, thus lowering the incidence and mortality rate of GVHD and IP [33]. With regard to potential additive effects on lung tissue of chemotherapy and TBI combination, a study based on a rat model concluded that the incidence of IP is only slightly enhanced by Cy [47]. As a matter of fact, the effects of several radiation parameters and the role of timing between TBI irradiation and chemotherapy remain rather controversial since all these parameters will affect the final outcome [23]. Furthermore, many authors try to extrapolate data in humans that is obtained from animal models [43,45–48]. The TBI/chemotherapy timing models that we employed were mainly used on the basis of common clinical practice [3,4,30,39].

272

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

The sparing effect of FTBI on normal tissues, compared with STBI of the same total dose, could be outweighed by an increased incidence of relapses, due to less efficient eradication of the recipient BM, with some authors underscoring the need for higher total doses or dose rate with fractionated regimens [40]. Recent studies suggest that leukemias have different sensitivity to fractionation, with ANLL cells being the less sensitive, and CML the most sensitive [9]. Whereas, in the former case, any fractionated scheme should obtain an effective leukemic cell killing with less toxicity on normal tissues, in the latter case, apart from turning back to STBI, the use of three to four large daily fractions has been proposed [49]. The possibility of the host BM cells repairing fractionated radiation damage, together with a less immunosuppressive effect observed with FTBI [12,35], in some clinical experiences actually showed detrimental results in terms of graft failure or rejection, particularly in cases of concurrent use of T cell-depletion, where a favorable ‘graft vs. host immunocompetent cells’ effect is lost [24,25]. As a matter of fact, T-depletion has been proposed with the aim of decreasing the incidence and severity of GVHD [19], but in most experiences this result was flawed by higher relapse and graft rejection rates [1,17]. In our series we did not observe any graft failure, which is generally contrary to most published data [9]. This result could be ascribed to differences in depleting methods between series, or more likely to our dose-rate that was .14 cGy/min in all patients. In fact, in a large series of T-cell depleted transplants [24], there was a highly significant reduction of graft

failure occurrence with dose-rate .14 cGy/min, whereas fractionation was not found to be a significant factor. Despite the encouraging engraftment data, T-cell depletion also had a negative impact on the outcome in our series. This is probably related to the loss of a beneficial graft vs. leukemia effect. The difference in outcome of the 41 Tdepleted patients compared to those that received unmanipulated BM is of statistical relevance both in terms of OS and DFS (Fig. 1). T-cell depletion was the strongest negative predictor variable to DFS. Further, the 5-year OS and DFS rates of 43% and 31% as obtained in our whole series, rise to 50% and 44%, respectively, when T-depleted patients are excluded. The results obtained by our 95 patients that received unmanipulated BM, stratified by disease status, are similar to those of other published series that did not use T-depletion (Fig. 2). The outcome of BMT recipients with CML, usually reported to have a more favorable prognosis, appears to be particularly affected by T-cell depletion. Finally, for a correct evaluation of these disappointing results, it must be underlined that in our case series the same FTBI schedule had been performed in all patients without taking BM manipulation into account. Other authors have already reported that in T-depleted patients the conditioning regimens need to be reinforced to take the loss of beneficial graft vs. leukemia effect into account Fig. 3 [9]. Whereas in our Institution, T-depletion is no longer employed, other centers are carrying out experiments with this technique using higher dose FTBI regimens and different chemotherapeutic agents [23]. As already reported by most authors [11,14,20,39,40] and

Fig. 1. DFS of all patients according to BM T-depletion.

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

273

Fig. 2. DFS of 95 patients transplanted with unmanipulated BM and stratified by disease status.

also in our series, the disease status proved to be one of the principal prognostic factors (Fig. 4). The impact on survival is related to a higher relapse rate, and a larger difference in disease-free survival than in overall survival is observed between early stage and advanced stage patients. On the basis of these results, some authors suggest that early BMT during the clinical course of leukemias could increase the probability of cure [33]. On the other hand, in our series the time interval between diagnosis and BMT considered as a continuous variable failed to result in a significant factor in predicting OS and DFS. Diagnosis of CML versus AL was associated with improved overall survival but not disease-free survival. The occurrence of GVHD clearly influenced overall and disease-free survival, as patients experiencing acute-GVHD had a worse outcome, as did those with chronic GVHD higher survival rates compared to patients without GVHD. In fact, a higher grade of acute GVHD is one of the most important factors influencing transplant mortality and disease-free survival in many published series of TBI [33,34] whereas chronic GVHD is thought to have a protective effect against leukemic relapses thanks to the graft-versusleukemia effect. The hazard function for overall survival stratified by GVHD patterns allows to clearly appreciate the different death rate per unit of time (Fig. 3). The negative effect of T-cell depletion on both diseasefree and overall survival has already been addressed. Our data underscore that the negative effect of T-cell depletion is not directly due to a decrease in chronic GVHD occurrence

in this patient subgroup. Instead, acute GVHD, with its negative impact on outcome, occurred significantly less frequently with T-cell depletion (P , 0.05). In fact, both Tcell depletion and GVHD resulted independent prognostic variables in OS and DFS in our series. Therefore, we can speculate that with T-cell depletion techniques there is a loss of beneficial graft versus leukemia effect even when there is not a decrease in chronic GVHD occurrence. In our series, age of transplant recipient did not result as being a significant factor for successful treatment outcome when it was differentiated into two classes (patients aged ,20 and those .20), with similar DFS and OS rates, whereas it seems to become a useful predictor factor for overall survival only if considered as a continuous variable. The patients in our series received either the ‘standard’ FTBI and CY preparatory regimen, or FTBI and VP-16. At the moment in our institution, FTBI and VP-16 is the more commonly used regimen for advanced stage patients, whereas patients with early phase CML are usually treated with chemotherapy only. Even though BMT appears to be the best post-remission treatment for leukemic patients reaching CR and having an HLA donor, efforts of investigators worldwide are still aiming to identify the ablative regimen with lower rates of rejection, relapse and toxicity. The outcome of BMT is affected by many variables such as, irradiation-, chemotherapy- and immunology- related ones. Therefore, only alterations of the conditioning regimen that lead to a very marked

274

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

Fig. 3. Hazard functions (i.e. death rate/unit of time) for OS according to GVHD occurrence. Cum Hazard, cumulative hazard.

improvement or worsening of the outcome, are likely to be detected in clinical studies [14]. This was the case of T-cell depletion in our series which comprised patients receiving a

rather homogeneous regimen of standard fractionated TBI. These results confirm previous observations that T-cell depletion as used in order to reduce the high toxicity level

Fig. 4. Hazard functions (i.e. death rate/unit of time) for OS according to disease status. Cum Hazard, cumulative hazard.

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

of combined TBI and chemotherapy, has a negative impact on survival of BMT recipients due to an increased relapse rate.

[17] [18]

References [1] Apperley, J.F., Mauro, F.R., Goldman, J.M., et al. Bone marrow transplantation for chronic myelogenous leukaemia in first chronic phase: importance of a graft-versus-leukaemia effect. Br. J. Haematol. 69: 239-245, 1988. [2] Bamberg, M., Beelen, D.W., Mahmoud, H.K., Molls, M. and Schaefer U.W. The incidence of interstitial pneumonitis: Comparison of total body irradiation schedules for bone marrow transplantation. Strahlenther. Onkol. 162: 218-222, 1986. [3] Blume, K.G., Forman, S.J., Donnel, M.R., et al. Total body irradiation and high-dose etoposide: A new preparatory regimen for bone marrow transplantation in patients with advanced hematologic malignancies. Blood 69:1015-1020, 1987. [4] Blume, K.G., Kopecky, K.J., Henslee-Downwy, J.P., et al. A prospective randomized comparison of total body irradiation – etoposide versus busulfan-cyclophosphamide as preparatory regimens for bone marrow transplantation in patients with leukaemia who were not in first remission: a Southwest Oncology Group Study. Blood 81: 2187-2193, 1993. [5] Buckner, C.D., Clift, R.A., Appelbaum, F.R., et al. Effects of treatment regimens in patients allografted for acute and chronic myelogenous leukaemia. Bone Marrow Transplant. 7(Suppl 2): 6-9, 1991. [6] Clift, R.A., Buckner, C.D., Appelbaum, F.R., et al. Allogenic marrow transplantation in patients with acute myeloid leukaemia in first remission: A randomized trial of two irradiation regimens. Blood 76: 1867-1872, 1990. [7] Clift, R.A., Buckner, C.D., Thomas, E.D., et al. The treatment of acute nonlymphoblastic leukaemia by allogenic bone marrow transplantation. Bone Marrow Transplant. 2: 243-258, 1987. [8] Cosset, J.M., Baume, D., Pico J.L., et al. Single dose versus hyperfractionated total body irradiation before allogenic bone marrow transplantation: A non-randomized comparative study of 54 patients at the Institute Gustave Roussy. Radiother. Oncol. 15: 151-160, 1989. [9] Cosset, J.M., Socie, G., Dubray, B., Girinsky, T., Fourquet, A. and Gluckman, E. Single dose versus fractionated total body irradiation before bone marrow transplantation: radiobiological and clinical indications. Int. J. Radiat. Oncol. Biol. Phys. 30: 477-492, 1994. [10] Cox, D.R. Regression models and life tables. J. Stat. Soc. (B) 34: 187-196, 1972. [11] Devergie, A., Reiffers, J., Vernant J.P., et al. Long-term follow up after bone marrow transplantation for chronic myelogenous leukaemia: Factors associated with relapse. Bone Marrow Transplant. 5: 379-386, 1990. [12] Down, J.D., Tarbell, N.J., Thames, H.D. and Mauch, P. Syngenic and allogenic bone marrow engrafment after total body irradiation: Dependance on dose, dose rate and fractionation. Blood 77: 661669, 1991. [13] Eifel, P.J., Sampson, C.M. and Tucker, S.L. Radiation fractionation sensitivity of epiphyseal cartilage in a weaning rat model. Int. J. Radiat. Oncol. Biol. Phys. 19: 661-664, 1990. [14] Gale, R.P., Butturini, A. and Bortin, M.M. What does total body irradiation do in bone marrow transplants for leukaemia? Int. J. Radiat. Oncol. Biol. Phys. 20: 631-634, 1991. [15] Gaynor, J., Chapman, D., Little, C., et al. A cause-specific hazard rate analysis of prognostic factors among 199 adults with acute lymphoblastic leukaemia: The Memorial Hospital experience since 1969. J. Clin. Oncol. 6:1014-1019, 1988. [16] Girinski, T., Socie, G., Ammarguellat, H., et al. Consequences of two different doses to the lungs during a single dose of total body irradia-

[19]

[20]

[21] [22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

275

tion: Results of a randomized study on 85 patients. Int. J. Radiat. Oncol. Biol. Phys. 30:821-824 1994. Goldman, J.M. Bone marrow transplantation for chronic myeloid leukaemia. Hematol. Oncol. 5: 265-279, 1987. Goldman, J.M., Gale, R.P., Horowitz, M.M., et al. Bone marrow transplantation for chronic myelogenous leukaemia in chronic phase: increased risk for relapse associated with T-cell depletion. Ann. Intern. Med. 108: 806-814, 1988. Hale, G., Bright, S., Chumbley, G., et al. Removal of T-cells from bone marrow for transplantation: A monoclonal anti-Lymphocytes antibody that fixes human complement. Blood 62: 888-893, 1983. Hoelzer, D., Thiel, E., Loffler, H., et al. Prognostic factors in a multicenter study for treatment of acute lymphoblastic leukaemia in adults. Blood 71: 123-129, 1988. Kaplan, E.L., Meier, P. Non parametric estimations for incomplete observations. J. Am. Stat. Assoc. 53: 457-481, 1958. Inoue, T., Ikeda H., Yamazaki, H., et al. Role of total body irradiation as based on the comparison of preparation regimens for allogenic bone marrow transplantation for acute leukaemia in first complete remission. Strahlenther Onkol. 169: 250-255, 1993. Latini, P., Aristei, C., Aversa, F., et al. Interstitial pneumonitis after hyperfractionated total body irradiation in HLA-matched T-depleted bone marrow transplantation. Int. J. Radiat. Oncol. Biol. Phys. 23: 401-405, 1992. Marmont, A.M., Horowitz, M.M., Gale, R.P., et al. T-cell depletion of HLA-identical tranplants in leukaemia. Blood 78: 2120-2130, 1991. Martin, P.J., Hansen, J.A., Bucker, C.D., et al. Effects on in vitro depletion of T-cell in HLA identical allogenic marrow grafts. Blood 66: 664-672, 1985. Mauch, P., Constine, L., Greenberger, J., et al. Hematopoietic stem cell compartment: Acute and late effects on radiation therapy and chemotherapy. Int. J. Radiat. Oncol. Biol. Phys. 31:1319-1339, 1995. McGlave, P.B., Haake, R.J., Bostrom, B.C., et al. Allogenic bone marrow transplantation for acute nonlymphocytic leukaemia in first remission. Blood 72: 1512-1517, 1988. Messner, H.A., Curtis, J.E. and Minden, M.M. The combined use of cytosine arabinoside, cyclophosphamide, and total body irradiation as preparative regimen for bone marrow transplantation in patients with AML and CML. Semin. Oncol. 12:184-192, 1985. O’Donoghue, J.A. Fractionated versus low-dose rate total body irradiation. Radiobiological considerations in the selection of the regimen. Radiother. Oncol. 7: 241-247, 1986. Ozsahin, M., Belkacemi, Y., Pene, F., et al. Interstitial pneumonitis following autologous bone-marrow transplantation conditioned with cyclophosphamide and total-body irradiation. Int. J. Radiat. Oncol. Biol. Phys. 34: 71-77, 1989. Ozsahin, M., Belkacemi, Y., Pene, F., Schlienger, M., Laugier, A. Total body irradiation and cataract incidence: a randomized comparison of two instantaneous dose rates. Int. J. Radiat. Oncol. Biol. Phys. 28: 343-347, 1993. Peto, R., Pike, M.C., Armitage, P., et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient II. Analysis and example. Br. J. Cancer 35: 1-39, 1977. Resbeut, M., Cowen, D., Blaise, D., et al. Fractionated or single-dose total body irradiation in 171 acute myeloblastic leukaemias in first complete remission: is there a best choice? Int. J. Radiat. Oncol. Biol. Phys. 31: 509-517, 1995. Ringden, O., Ruutu, T., Remberger, M., et al. A randomized trial comparing busulfan with total body irradiation as conditioning in allogenic marrow transplant recipients with leukaemia: A report from the nordic bone marrow transplantation group. Blood 83: 2723-2730, 1994. Salomon, O., Lapidot, T., Terenzi, A., Lubin, I., Rabi, I. and Reisner, Y. Induction of donor-type chimerism in murine recipients of bone marrow allografts by different radiation regimens currently used in treatment of leukaemia patients. Blood 76: 1872-1878, 1990.

276

V. Donato et al. / Radiotherapy and Oncology 48 (1998) 267–276

[36] Sanders, J.E., Ritchards, S., Mahonet, P., et al. Growth and development following marrow transplantation for leukaemia. Blood 68: 1129-1135, 1986. [37] Shank, B., Chu, F.C., Dinsmore, R., et al. Hyperfractionated total body irradiation for bone marrow transplantation. Results in seventy leukaemia patients with allogenic transplants. Int. J. Radiat. Oncol. Biol. Phys. 9: 1607-1611, 1983. [38] Shank, B., O’Reilly, R.J., Cunningham, I., et al. Total body irradiation for bone marrow transplantation: the Memorial Sloan-Kettering cancer center experience. Radiother. Oncol. 18(Suppl. 1): 68-81, 1990. [39] Snyder, D.S., Chao N.J., Amylon M.D., et al. Fractionated total body irradiation and high-dose etoposide as a preparatory regimen for bone marrow transplantation for 99 patients with acute leukaemia in first complete remission. Blood 82: 2920-2928, 1993. [40] Socie, G., Devergie, A., Girinsky, T., et al. Influence of the fractionation of total body irradiation on complications and relapse rate for chronic myelogenous leukaemia. Int. J. Radiat. Oncol. Biol. Phys. 20: 397-404, 1991. [41] Sokal, J.E., Baccarini, M., Russo, D. and Tura, S. Staging and prognosis in chronic myelogenous leukaemia. Sem Hematol 25: 49-61, 1988. [42] Tait, R.C., Burnett, A.K., Robertson, A.G., et al. Subclinical pulmonary function defects following autologous and allogenic bone marrow transplantation: Relationship to total body irradiation and graft-

[43]

[44] [45]

[46]

[47]

[48]

[49]

versus-host disease. Int. J. Radiat. Oncol. Biol. Phys. 20: 1219-1227, 1991. Tarbell, N., Amato, D., Down, J., Mauch, P. and Hellman, S. Fractionation and dose rate effects in mice. A model for bone marrow transplant in man. Int. J. Radiat. Oncol. Biol. Phys. 13: 1065-1069, 1987. Thomas, E.D. and Clift, R.A. Indications for marrow transplantation in chronic myelogenous leukaemia. Blood 73: 861-864, 1989. Van Roengen, E., Kuijpers, W.C. and Madhuizen, H.T. Fractionation effects and repair kinetics in rat kidney. Int. J. Radiat. Oncol. Biol. Phys. 9: 1607-1611, 1990. Van Roengen, E., Thames, HD. and Travis, E.L. Recovery from radiation damage in mouse lung: Interpretation in terms of two rates of repair. Radiat. Res. 133:225-233, 1993. Varekamp, A.E., de Vries, A.J., Zurcher, C. and Hagenbeek, A. Lumg damage following bone marrow transplantation: II. The contribution of cyclophosphamide. Int. J. Radiat. Oncol. Biol. Phys. 13: 1515-1521, 1987. Vriesendorp, H.M., Hermann, M.G. and Saral, R. Future analysis of total body irradiation. Int. J. Radiat. Oncol. Biol. Phys. 20: 635-637, 1991. Zander, A.R., Keating, M., Kanojia, M., et al. Allogenic bone marrow transplantation in acute leukaemia: Piperazinedone and fractionated total body irradiation as an alternative conditioning regimen. Transplant. Proc. 15: 1392-1398, 1983.