P973: attrition probability and predictors among multiple myeloma patients treated in an italian haematological tertiary care centre

Topic: 14. Myeloma and other monoclonal gammopathies - Clinical Background: Multiple Myeloma (MM) patients experience serial relapses and, although significant increased options for consecutive treatment lines, accumulating comorbidities with advanced age can make difficult treatment with loss of patients reaching subsequent lines of therapy (LOTs). In contrast with previous studies, we recently reported a low attrition rate (AR) across LOTs in a real-world MM population but it was found to be significantly higher in patients non-transplant-eligible (NTE) vs TE. Aims: Using a time-dependent analysis, we aimed to estimate AR probability at each subsequent LOT evaluating factors affecting this probability in a MM population treated in an Italian haematological tertiary care centre. Methods: Data from 419 patients receiving any MM treatment between 2011 and 2021 were retrospectively analysed. AR was defined as the fraction of patients, among those relapsing from a prior therapy, who did not receive a subsequent LOT for any reason. The Kaplan-Meier survival curve was used to determine the AR probability over follow-up time. Both univariate and multivariate Cox regression models were fitted and variables with a p-value < 0.5 in the multivariate analysis were considered a significant predictor of AR. Results: Median age of patients was 69 years (range 30-93 years), 39% of them aged ≤ 65 years and 61% were older. ECOG PS ≥ 2 was found in 22% of patients and a renal failure in 18%. Twenty nine percent of patients had more than 2 comorbidities and Simplified Frailty Score defined 30.5%, 26% and 43.5% as fit, intermediate-fit and frail patients, respectively. International Staging System (ISS) II and III was detected in 36% and 27% of patients, respectively. As frontline therapy, IMIDs were administered in 29.5% of patients, IMIDs+PIs in 36%, PIs in 33% and chemotherapy in 1.5% of cases. Thirty-three percent of patients received maintenance therapy; 16% of them had infectious complications and 5% thrombotic adverse events. Sixty-five percent of patients obtained a response higher than very good partial response. Median follow-up was 48.7 months (range 6–132 months). The AR probability of MM patients increased as follow-up time increased; it increased from 28% in patients who relapsed after LOT-1 to 43% for patients relapsing after LOT-2 but it resulted to be quite constant for subsequent LOTs, being 45% for patients who relapsed after LOT-3, 44% after LOT-4 and 48% after LOT-5. In univariate analysis age over 65 years, ISS II/III, two or more comorbidities, no transplant, response below very good partial remission, and no maintenance therapy were significantly associated with AR but duration of response was not. Using age and comorbidities, the two factors selected by stepwise Cox regression analysis, we constructed a risk model able to stratify patients with significantly different AR probability after LOT-1 and LOT-2. In patients > 65 years and with > 2 comorbidities AR probability was 56% after LOT-1 vs 32% for patients with one risk factor (age > 65 years or > 2 comorbidities) vs 5% for patients who no risk factors. After LOT-2 these values were 76%, 51% and 21%, respectively. Summary/Conclusion: Overall, in our experience, AR probability is lower than 50% across all subsequent LOTs. However, treatment of relapses in older patients with comorbidities remains a challenge since probability of undergoing a treatment after LOT-1 is lower than 50% and only a quarter of these patients could receive a LOT-3 after a second relapse. These results confirm the need to use the most effective therapies upfront to optimize their effectiveness. Keywords: Myeloma