A New Method To Evaluate the Prognostic Value of Response to Prednisone Pre-Treatment in Adult (15–60 Yrs) Patients with Acute Lymphoblastic Leukemia: Results of the GIMEMA LAL 2000 Study.

The prognostic significance of the response to initial prednisone treatment in adult ALL has been recently emphasized. Prednisone response is usually defined on the basis of the peripheral leukemic blast count. The threshold value for the defintion as good or poor prednisone response is 1000 blasts/mmc on day 8 of prednisone pre-treatment. The drawback of this definition is the difficulty of classifying patients with less than 1000 blasts at diagnosis. In the LAL2000 GIMEMA study we therefore evaluated whether the blast reduction rate, which is not affected by the initial blast level, could be a factor with comparable prognostic value. The protocol design provided a 7-day (−6 to 0) pre-treatment phase with an escalating dose of prednisone up to 60 mg/sqm. On day 1 before starting the induction the response was assessed both according to the absolute blast count (< versus ≥ 1000/mmc) (criterion 1) and according to the blast reduction rate ≥ 75% (criterion 2) in the peripheral blood. The induction included high dose Daunorubicin; for patients in complete remission (CR) this was followed by consolidation with high dose ARA-C, chemo and radio prophylaxis of the central nervous system, and periodical reinduction over a three years maintenance period. Patients with adverse cytogenetic features [i.e. t(9;22), t(4;11), t(1;19)] who achieved a CR were treated according to the HAM protocol that included high dose ARA-C and Mitoxantrone followed by Imatinib for Ph+ ALL and by allogeneic or autologous hemopoietic stem cells transplantation for the others. Between September 2000 and December 2003 a total of 368 patients were evaluable for response to induction. The median age was 35 years (15–60) and median WBC count 15′10⁹/L (0.3–872); 72 (20%) were T ALL and 121 (33%) had cytogenetic high risk features (104 (86%) Ph+, 4 (3%) t(4;11) and 13 (11%) t(1;19)). Eighty-seven percent of the patients were evaluable for response to steroid pre-treatment: ’responders’ were 75% according to criterion 1 (blast <1′10⁹/L on day 0), and 80% according to criterion 2 (blast reduction rate ≤75% on day 0). The overall CR rate was 83%. The probability of response was significantly higher in prednisone responders with respect to non responders according to both criteria: 87% versus 63% (p<0.0001) according to criterion 1, 85% versus 68% (p=0.001) according to criterion 2. Also the post CR outcome was better in steroid responders, regardless of the definition. Using criterion 1, median disease-free survival (DFS) was 24 months in responders and 11 months in non responders; using criterion 2, median DFS was 23 months in responders and 12 months in non responders. Both criteria were significantly related to DFS in a multivariate analysis adjusted for cytogenetic risk and WBC count at diagnosis (>=/<50). In conclusion, our study confirms that the sensitivity to steroids is an independent prognostic factor for the outcome of adult ALL; moreover, we propose an alternative method of its evaluation with respect to the one currently used. This method has the advantage of allowing to classify all patients, regardless of the initial blasts level, and shows a comparable prognostic value.