Increasing Frequency and Marked Stability of Complete Molecular Response Is Observed in Imatinib-Treated CML Patients with Long-Term Follow Up.

The degree of reduction of BCR-ABL in imatinib-treated patients with chronic phase CML is an important indicator of prognosis. The IRIS trial established that with first-line therapy patients with a major molecular response (MMR, 3 log reduction from a standardized baseline value for untreated patients) have a significantly more favorable progression free survival. Although 40% achieved a MMR by 12 months, very few had undetectable BCR-ABL according to strict PCR sensitivity criteria. We measured peripheral blood BCR-ABL levels by quantitative PCR at 3 to 6 month intervals in 155 patients with chronic phase CML enrolled in clinical trials of imatinib for up to 6 years. We aimed to (i) determine if BCR-ABL levels continued to decline over time, and (ii) evaluate the stability and significance of undetectable BCR-ABL. The patients included the Australasian subset of IRIS trial patients treated with 400mg of imatinib; 29 first-line patients evaluated for a median of 69 months (25^th to 75^th percentile range (pr) 58–72) and 24 second-line patients for a median of 54 months of imatinib (pr 38–60). 102 de-novo patients enrolled in the TIDEL trial of 600mg imatinib were evaluated for a median of 39 months (pr 30–42). Complete molecular response (CMR) was defined as undetectable BCR-ABL at a PCR sensitivity of at least 4.5 logs below the standardized baseline value confirmed on subsequent analysis after at least 3 months. The BCR control transcript level determined sensitivity and was dependent on RNA quality and reverse transcription efficiency. Of note CMR may not indicate eradication of leukemic cells, rather a reduction of BCR-ABL below the detection limit. CMR occurred in 34 patients who had 178 analyses after achieving CMR (median 4 tests per patient) and a median follow up of 15 months (pr 9–24). Very low level BCR-ABL was detected in 3 patients, the remaining 31 had undetectable BCR-ABL on every subsequent assay. Of the IRIS trial patients treated with first-line imatinib, 41% achieved a CMR by 69 months, a frequency significantly higher than occurred in these patients at 24 months (7%, p=0.006). The rate of CMR appeared to increase substantially beyond the 3 year time point (7%, 24% and 34% at 3, 4 and 5 years). 75 patients achieved MMR but not CMR and were followed for a median of 24 months after achieving MMR (pr 17–33). Six of 75 patients (8%) lost MMR as defined by >2-fold rise in BCR-ABL and loss of MMR on 2 consecutive analyses. The median fold rise was 18-fold (4 to 1900-fold), of whom 1 went on to blast crisis. Four of the 6 patients had BCR-ABL mutations detected at the time of the rise and 1 of the remaining patients had duplicate Ph. MMR was lost in these 6 patients within 18 months of its achievement. The overall rate of CMR and MMR (including patients with CMR) did not differ significantly between the 3 treatment groups at the 3 year time point (CMR 7%, 8% and 18%; MMR 66%, 71% and 70% for first-line 400mg, second-line 400mg and first-line 600mg respectively). In conclusion at a median follow-up of 5.75 years of 400mg first-line imatinib, CMR was achieved in 41% of patients. Importantly, of all patients who achieved a CMR in this study using strict criteria to define the sensitivity of analysis, none have lost MMR and 91% have maintained CMR. The slow acquistion and marked stability of CMR favour the notion that the leukemic stem cell pool is steadily declining with prolonged exposure to imatinib.