Incidence and outcome of BCR-ABL mutated chronic myeloid leukemia patients who failed to tyrosine kinase inhibitors

Abstract

Purpose: To assess the incidence of BCR-ABL kinase domain (KD) mutation detection and its prognostic significance in chronic phase chronic myeloid leukemia (CP-CML) patients treated with tyrosine kinase inhibitors (TKIs).

Patients and methods: We analyzed characteristics and outcome of 253 CP-CML patients who had at least one mutation analysis performed using direct sequencing. Of them, 187 patients were early CP (ECP) and 66 were late CP late chronic phase (LCP) and 88% were treated with Imatinib as first-line TKI.

Results: Overall, 80 (32%) patients harbored BCR-ABL KD mutations. A BCR-ABL KD mutation was identified in 57% of patients, who progressed to accelerated or blastic phases (AP-BP), and 47%, 29%, 35%, 16% and 26% in patients in CP-CML at the time of mutation analysis who lost a complete hematologic response, failed to achieve or loss of a prior complete cytogenetic and major molecular response, respectively. Overall survival and cumulative incidence of CML-related death were significantly correlated with the disease phase whatever the absence or presence of a mutation was and for the latter the mutation subgroup (T315I vs P-loop vs non-T315I non-P-loop) (P<.001). Considering patients who were in CP at the time of mutation analysis, LCP mutated patients had a significantly worse outcome than ECP-mutated patients despite a lower incidence of T315I and P-loop mutations (P<.001). With a median follow-up from mutation analysis to last follow-up of 5 years, T315I and P-loop mutations were not associated with a worse outcome in ECP patients (P = .817).

Conclusion: Our results suggest that early mutation detection together with accessibility to 2nd and 3rd generation TKIs have reversed the worst outcome associated with BCR-ABL KD mutations whatever the mutation subgroup in CP-CML patients.

Keywords: BCR-ABL kinase domain mutation; chronic myeloid leukemia; tyrosine kinase inhibitors.

Figure 1

BCR‐ABL tyrosine kinase domain mutations (n = 80 patients) according to status at TKI start (early chronic phase (ECP) vs late chronic phase (LCP)) disease phase (chronic phase (CP) vs accelerated and blastic phases (AP‐BP)) at the time of mutation analysis (1A); mutation subgroups patients according to disease status (AP‐BP; failure to achieve or loss of complete hematologic response (CHR); failure to achieve or loss of complete cytogenetic response (CCyR); failure to achieve or loss of major molecular response (MMR)) at the time of mutation analysis; sustained MMR with BCR‐ABL increase of one log (1B)

Figure 2

Overall survival (2A) and cumulative incidence of CML‐related death (2B) from the time of mutation analysis to the last follow‐up according to the phase of the disease at the time of mutation analysis (chronic phase, CP; accelerated and blastic phase, AP‐BP) and to the absence or presence of a BCR‐ABL tyrosine kinase domain mutation (n = 253 patients; CP without mutation, n = 155 pts; CP with mutation, n = 56 pts; AP‐BP without mutation, n = 18 pts; AP‐BP with mutation, n = 24 pts).The probability of OS and rates of CI of CML‐related death at 5 y were 91% (95% CI: 86‐97.9) and 3.2% (95% CI: 0‐6.9), 79% (95% CI: 68.5‐91.7) and 14.8% (95% CI: 4.4‐25.1), 33.9% (95% CI: 16.7‐68.7) and 52% (95% CI: 27‐76.9), 32% (95% CI: 17.3‐60.1) and 58% (95% CI: 36.4‐80.6) for CP without mutation, CP with mutation, AP‐BP without mutation and AP‐BP with mutation, respectively (OS, P < .001; CI CML‐related deaths, P < .001)

Figure 3

Overall survival (3A) and cumulative incidence of CML‐related death (3B) from the time of mutation analysis to the last follow‐up according to status at TKI start ((early chronic phase (ECP) vs late chronic phase (LCP)) and absence or presence of a BCR‐ABL tyrosine kinase domain mutation (n = 211 patients; ECP without mutation, n = 128 pts; ECP with mutation, n = 33 pts; LCP without mutation, n = 27 pts; LCP with mutation, n = 23 pts).The probability of OS and rates of CI CML‐related death at 5 y were 90% (95% CI: 83.1‐97.7) and 4% (95% CI: 0‐8.8), 91% (95% CI: 81.4‐100) and 8% (95% CI: 0‐19.4), 96% (95% CI: 89‐100) and 0% (95% CI: 0‐0), 64% (95% CI: 46.4‐87.7) and 23% (95% CI: 4.7‐40.7) for ECP without mutation, ECP with mutation, LCP without mutation and LCP with mutation, respectively (OS, P<.001; CI CML‐related deaths, P<.001)

Figure 4

Overall survival (4A) and cumulative incidence of CML‐related death (4B) of early chronic phase patients (ECP) (n = 161 patients) from the time of mutation analysis to the last follow‐up according to absence of mutation (n = 128 patients) or presence and to type of mutation (n = 33 patients; T315I mutation, n = 8 pts; P‐loop mutation, n = 7 pts; non‐T315I non‐P‐loop mutations, n = 18 pts). Indeed, the probability of OS and rates of CI CML‐related deaths at 5 y were 91% (95% CI: 86‐97.6) and 4% (95% CI: 0‐8.8), 71% (95% CI: 48.8‐100) and 12% (95% CI: 0‐37), 100% (95% CI: 100‐100) and 0% (95% CI: 0‐0), 79% (95% CI: 66.3‐94.3) and 9% (95% CI: 0‐26.9) for ECP without mutation, ECP with T315I mutation, ECP with P‐loop mutation and ECP with non‐T315I non‐P‐loop mutation, respectively (OS, P = .817; CI CML‐related deaths, P = 0.524)