Comparative Analysis of Autografting in Chronic Myelogenous Leukemia: Effect of Priming Regimen and Marrow or Blood Origin of Stem Cells

To the Editor:

The report of Verfaillie et al1 compares their results to those of our group in attempting to harvest Philadelphia (Ph) chromosome-negative progenitors from chronic myelogenous leukemia (CML) patients for use in autologous transplant. This is not a reasonable comparison, in view of the differences in the patients studies, the type of drugs for mobilization, and the timing of harvesting. From our experience in developing this strategy over 9 years in 176 patients, it is evident that (1) a combination with idarubicin, intermediate-dose cytarabine, and etoposide (ICE/mini-ICE protocols)2-5 or modifications of it6,7 is the most effective therapy to mobilize Ph-negative hematopoietic progenitor cells (HPC). Few patients received high-dose hydroxyurea on the basis of the above criteria, but the small number of patients treated does not allow us to state any definitive conclusions.8 These data have been confirmed also in advanced highly pretreated patients cytogenetically refractory to interferon-α (IFN-α).9 It must be underlined that ICE is the acronym of idarubicin-containing regimen and not ifosfamide containing regimen, as wrongly reported by Verfaillie et al.1 I do not understand the reason why the investigators decided to use Cy alone or in combination with low doses of mitoxantrone and cytarabine instead of the most effective mini-ICE or IDAC protocols. The differences between their results and our results may be due to the different myeloreductive approaches. It has been clearly demonstrated that cyclophosphamide is not myeloreductive enough and poorly effective on leukemic patients; besides, it has been well demonstrated that the duration of aplasia is not important but rather the quality of cells killed by the oncolytic drugs. Perhaps the therapy used by Verfaillie et al1 did not determine an adequate myeloreduction in the marrow and the growth factors stimulated the residual leukemic cells instead of the arising diploid residual HPC. In addition, we deem that very strict criteria are needed to start the collections. As we recently published, we have learned in these past few years that the best collections in terms of CD34⁺ cells and diploid cells were achieved when the white blood cell count (WBC) was greater than 0.8 × 10⁹/L combined with ≥10/μL CD34⁺ cells on peripheral blood. The investigators say that they started their collections when the absolute WBC was between 500 and 1,000/μL without CD34⁺cell control in peripheral blood, but in Table 1, the WBC at the time of mobilization was incredibly high, ie, 2.9 to 84 × 10⁹/L (median, 7.7 × 10⁹/L; Cy/GM/BM); 2.4 to 27.2 × 10⁹/L (median, 9.8 × 10⁹/L; MAC/G/PB), and 2.6 to 36.1 × 10⁹/L (median, 4.7 × 10⁹/L; Cy/G/PB). With respect to our experience, this may be the reason why they obtained such different results from ours. Only 15 patients received this approach during early chronic phase, and the short duration and intensity of chemotherapy administration could have had a negative impact on the type of cells mobilized in the peripheral blood when they were recovering. Another difference between our experience and that of Verfaillie et al derives from the timing of starting the procedure. In our experience, the median time between diagnosis and HPC mobilization was 2 months, versus 8 months for Verfaillie et al. As is known, the best results have been reported for HPC mobilization performed as soon as possible after diagnosis when the WBC reached ≤20 × 10⁹/L after hydroxyurea. The explanation for an early approach derives from biological reports demonstrating that the benign hematopoietic clone decreases with time, and it is possible that the normal HPC pool has decreased considerably. The data reported in the report by Verfaillie et al seem to confirm that an early approach combined with more intensive therapy (MAC/G/PB) does better in terms of cytogenetic response than the protocols with cyclophosphamide alone.

In conclusion, our work over these last 9 years with idarubicin-containing regimens (ICE/mini-ICE) allows us to state that the results with this procedure are as follows: (1) poor in accelerated/blastic phase, probably because diploid cells are no longer available in marrow and patients cannot withstand toxic therapy; (2) compromised by longer hospitalization, less good cell phereses, and longer recovery in highly pretreated patients with interferon; (3) the high rate of diploid collections is achieved in untreated patients in the first few weeks from diagnosis and the response is correlated with Sokal score (excellent results in low and intermediate risk and less good results in high-risk patients); and (4) ICE or mini-ICE protocols determine the same Ph-positive HPC collections, but mini-ICE has more potential benefits because it may be administered in outpatient facility and is less GI toxic and because shorter hospitalizations were needed. Diploid cells are most likely in pheresis 1, but at least 2 to 3 phereses are required to collect more than 3 to 4 × 10⁶CD34⁺ cells. We have treated 45 patients in the earlier phase of disease (at a median of 2 months from diagnosis) and not pretreated with interferon. Thirty-two patients were autografted, and 3 other patients are now receiving the transplant. Ten patients could not be autografted because of very high number of Ph-positive cells (6 patients), low HPC collected (3 patients), and MUD (1 patient). Thirty of 32 patients are alive, with 18 patients still Ph-negative (10 patients) or less than 34% Ph-positive (8 patients).

In summary, I deem that no correlation between the study of Verfaillie et al and our study may be performed. This is due mainly to the fact that Verfaillie et al treated very few patients in early chronic phase and with three different protocols. Moreover, other discrepancies between the two studies arise from mobilization therapy, timing, and choice of the best day to start the collections.