Phd in IINNOVATION IN IMMUNO-MEDIATED AND HEMATOLOGICAL DISORDERS

Thesis title: PREDICTIVE ROLE OF DIGITAL DROPLET PCR AT THE TIME OF TKI DISCONTINUATION IN CHRONIC PHASE CHRONIC MYELOID LEUKEMIA PATIENTS

Discontinuation has become a new therapeutic endpoint in chronic myeloid leukemia (CML). Unfortunately, about 50% of patients experience loss of major molecular response (MMR). Sustained treatment-free remission (TFR) is reached only in patients with a longer median duration of TKI treatment and sustained stable deep molecular response evaluated by RQ-PCR. We evaluated a series of CML patients in deep molecular response (MR) who attempted a TFR, with the aim of finding a droplet digital PCR (ddPCR) cut-off, predictive of a low rate of molecular recurrence free survival (MRFS) after discontinuation. Fifty-seven patients, with male prevalence (65%) and a median age of 47 years (range 21-82) at diagnosis, discontinued treatment in a real-life setting. Sokal score stratification identified 31 patients (54%) as low risk, 22 (38.5%) as intermediate, and 4 (7%) as high risk. Thirty-eight patients (67%) carried the b3a2 type of transcript, whereas 19 (33%) carried the b2a2 type. Nineteen patients received previously interferon-alpha (IFN) (median duration 25 months). Median duration of TKI treatment was 11.2 years (range 2.4-20.4). The median duration of deep MR was 2.7 years. Forty-six patients (80.7%) attempted discontinuation after first-line treatment, and 11 (17.5%) after subsequent lines (39 patients had received imatinib, 6 patients dasatinib, 12 nilotinib). Fifteen patients (26%) relapsed at a median time of 4.4 months with an MRFS of 68.6% at 30 months. All 57 cases were tested by ddPCR prior to treatment discontinuation: 31 (54.4%) were ddPCR-negative and only 4 of them relapsed (13%), while 26 (45.6%) proved ddPCR-positive and 11 (73%, p=0.002) relapsed. To optimize the use of ddPCR as a possible tool to predict the success of treatment discontinuation, we applied a ROC curve analysis to detect a cut-off of molecular residual disease that can better identify candidates to TFR. We obtained an AUC of 68.7% (95%CI 53.1-84.4) and identified a cut-off value of 0.001 copies/µl (cp/µl). Sensitivity, specificity, positive and negative predictive values were 73.3%, 64.3%, 42% and 87%, respectively. Before stopping the TKIs, 4/15 (27%) relapsed patients had a molecular residual disease ≤0.001 cp/µl vs 11/15 (73%) with >0.001 cp/µl (p<0.001). In conclusion, ddPCR can be a more sensitive and accurate method for the detection of molecular residual disease prior to treatment discontinuation. The proposed cut-off value should be validated in a large cohort of CML patients considered eligible to stop TKI treatment.