Sokal and Hasford Scores Are Able To Predict Cytogenetic Response and Survival in CML Patients Who Failed Prior Therapy and Are Receiving Salvage Therapy with Imatinib or Other Tyrosine Kinase Inhibitors.

Abstract

Sokal and Hasford scores have been used to estimate the survival of CML patients following alkylator-based or interferon-based therapy, and may also predict the outcome of patients receiving first-line imatinib. These scores were designed for patient characteristics at the time of diagnosis and they have been assumed to have no clinical value at other times during the course of the disease (eg. second-line or subsequent therapy). We seek to determine the applicability of the Sokal and Hasford scores in predicting outcomes in patients receiving second or subsequent-line therapy with imatinib and other novel tyrosine kinase inhibitors (TKI). This analysis is based on 379 CML patients in chronic phase (90%) or accelerated phase as defined by cytogenetic clonal evolution (10%), who were enrolled on clinical trials of imatinib following interferon failure (n=294), or nilotinib or dasatinib following imatinib failure (n=85). Significance of relationship between Sokal and Hasford risk categories and response were tested using the Chi-Square test, and differences in survival assessed using the log-rank method. P-values are summarised below:Imatinib following Interferon FailureNilotinib/Dasatinib Following Imatinib FailureSokalHasfordSokalHasfordLow Risk219 (74%)176 (60%)39 (46%)45 (53%)Intermediate Risk65 (22%)111 (38%)34 (40%)38 (45%)High Risk10 (3%)6 (2%)12 (14%)2 (2%)Any Cytogenetic Responsep<0.001p<0.001p=0.01p=NSMajor Cytogenetic Responsep<0.001p<0.001p=0.002p=0.04*Complete Cytogenetic Responsep<0.001p<0.001p=0.005p=0.07*Survivalp<0.0001p<0.0001Insuff EventsInsuff Events*comparison between low and intermediate risk groups onlyAmong patients receiving imatinib following failure of interferon therapy, both Sokal and Hasford scores were highly predictive of cytogenetic response and overall survival. Due to the low number of deaths (n=5) on the nilotinib and dasatinib trials, the effect of Sokal and Hasford scores on predicting survival could not be assessed but they remain predictive of cytogenetic response. We conclude that the Sokal and Hasford scores calculated at the time of therapy failure are useful predictors of subsequent response to salvage therapy with imatinib and other TKIs.