Page 102 - 2019_08-Haematologica-web

C. Fava et al.
2nd generation imatinib
Figure 2. Tyrosine kinase inhibitor (TKI)-treatment- free remission (TFR) curves adjusted for age at dis- continuation, Sokal score, line of therapy, and dura- tion of disease.
Table 4. Median and Interquartile Range of duration of treatment in patients who discontinued treatment in first line or in second or further lines of therapy.
Lines of treatment at discontinuation
Duration of total treatment P [median (IQR)]
1st Line
82 (60; 105)
128 (86;169)
<0.001
≥2nd Line
IQR: Interquartile Ranges.
Table 5. Multivariate Cox regression analysis for restarting therapy. Figures reported are Hazard Ratios and 95% confidence intervals.
HR 95%CI
Age at discontinuation (10 yrs difference) 0.84 0.73
Sokal score
Intermediate vs. low 0.92 0.54
High vs. low 2.07 1.16
Line of therapy: 2nd vs. 1st line 0.80 0.50
2nd generation TKIs vs. imatinib 0.43 0.20
Duration of total therapy (one yr increase) in patients treated with imatinib* 1.00 0.94
Duration of total therapy (one yr increase) in patients treated with 2nd generation TKIs** 0.78 0.65
P
0.97 0.02
1.57 0.76
3.71 0.01
1.30 0.37
0.91 0.03
1.07 0.90
0.93 0.01
*HR =1 expresses no risk increase associated to the increase of 1 year of the duration of therapy in patients treated with imatinib. ** HR < 1 expresses the risk reduction asso- ciated to the increase of 1 year of the duration of therapy in patients treated with 2nd generation tyrosine kinase inhibitor (TKI); yr: years; HR: Hazard Ratios; CI: Confidence Intervals.
variable), time to DMR and DMR duration (continuous variables), depth of MR at stop (MR4.5 vs. MR4, MR5 vs. MR4 and Undetectable vs. MR4), reasons for discontinua- tion (pregnancy, ISAV study and toxicity vs. shared deci- sion with medical doctor). The only statistically signifi- cant risk factors that affected TFR were age at discontinu- ation (with a higher risk for younger patients) and line of treatment (Table 3). When we assessed the duration of total treatment for patients who discontinued TKI in front line versus second line, we observed that patients who dis- continued treatment front line had a significantly shorter duration of treatment (P<0.001) (Table 4).
Multivariate analysis - The line of treatment lost statisti- cal significance in a multivariate analysis including age at discontinuation, Sokal score, duration of total treatment, line of treatment, and type of TKI at discontinuation (Table 5). Patients treated with second generation TKI showed a better TFR (HR 0.43; 95%CI: 0.20-0.91) (Table 5 and Figure 2). Duration of total treatment was positively
associated with TFR among patients treated with second generation TKI with a 22% risk reduction for one addi- tional year of treatment (HR: 0.78; 95%CI: 0.65-0.93).
Discussion
Although at present no guidelines explicitly recommend treatment discontinuation, this study showed that many physicians have already experienced TKI cessation in their clinical practice because of intolerance, toxicity, and patient desire to stop the treatment. This multi-center observational study has confirmed that treatment cessa- tion was safe as no progression occurred and the overall TFR was 69% at 12 months, consistent with data reported in previous studies.6-25 After discontinuation, patients were monitored with the same frequency as in the EURO-SKI study: most of the patients had a molecular evaluation every month for the first six months, every six weeks for
1594
haematologica | 2019; 104(8)