Page 241 - 2022_01-Haematologica-web

ASCT for fit, older patients in Myeloma XI
pathway were categorized by age group <65, 65-69 and 70-75 and their baseline characteristics, treatment and harvest data summa- rized. PFS and OS, measured from baseline trial randomization, were compared between age groups using the Kaplan-Meier method. Comparisons were made between the allocated groups using the Cox proportional hazards model stratified by the mini- mization stratification factors, excluding center, and hazard ratios (HR) and 95% confidence intervals (95% CI) were estimated. The frequencies of reported serious adverse events and patients’ deaths were examined to compare transplant-related morbidity and mor- tality between age groups.
Relative survival estimates were obtained using flexible para- metric survival models on the hazard scale with four degrees of freedom17 with the same covariates included in the model. Relative survival was defined as the observed survival divided by the expected survival where the expected survival was obtained from national life tables stratified by age at diagnosis, sex and cal- endar year. UK life-time risk was estimated from data available from the Office for National Statistics.18
In order to compare outcomes between patients undergoing ASCT or not, a second set of analyses was performed using an age-matched group of patients in the TE and TNE pathways. This was defined by overlap of the age distributions in each pathway and comprised the older patients within the TE pathway and younger patients within the TNE pathway but excluded the extremes in both pathways. The optimal overlap was of patients aged 64-70 within each pathway, which was chosen with the aim of maximizing the number of patients in the analysis while achieving a balance between the patients receiving ASCT (52.5%) and those who did not receive ASCT (47.5%). Three groups were considered: (i) TE patients who underwent ASCT (TE-ASCT); (ii) TE patients who did not undergo ASCT (TE-noASCT) and (iii) TNE patients (who did not undergo ASCT). The characteristics of the patients in the three groups at entry into the trial were sum- marized. Patients were scored according to the UK Myeloma Research Alliance Myeloma Risk Profile (MRP)19 and the propor- tion of patients in each of the groups compared. When analyzing time-to-event outcomes in order to avoid a survivor or immortal time bias that would be incurred by comparing all patients within these groups from baseline, patients were only included if they remained eligible to continue in the trial at the end of induction ± intensification and their outcomes were measured from the end of induction ± intensification therapy. This time-point was close to the ASCT date for the patients in the TE-ASCT group, but repre- sents a common time-point that could be identified in all patients to enable comparison.
To estimate the treatment effect of ASCT as compared to no ASCT in this subgroup, propensity score weighting using inverse probability of treatment weights was implemented. Propensity score weighting is a useful tool to account for imbalances in observed confounders between groups when estimating treat- ment effects from non-randomized data. A propensity score is a single score that represents the probability of receiving a treat- ment, conditional on a set of observed covariates. The goal of cre- ating a propensity score is to balance covariates between individ- uals who did and did not receive a treatment, making it easier to isolate the effect of a treatment. The propensity score was based on a patient’s age, sex, World Health Organization (WHO) per- formance status, International Staging System (ISS) stage, induc- tion treatment and response after completing induction treat- ments. The propensity score was applied using normalized inverse probability of treatment weighting (IPTW). In IPTW, each treated subject (ASCT) receives a weight equal to the inverse of the propensity score, and each control subject (no ASCT) receives a weight equal to the inverse of one minus the propensity score.
For the IPTW analysis, probability weights were applied to the individual participants’ data for calculation of the survivor func- tion estimate and partial likelihood in the Cox model. Statistical analysis was performed using SAS v9.4 (SAS Institute Inc., Cary, NC, USA) and Stata IC v16 (StataCorp. College Station, TX, USA).
PFS was defined as the time from the point stated above to the date of confirmed disease progression or death from any cause. OS was defined as the time from the point stated above to the date of death from any cause. Cytogenetic profiling was per- formed using multiplex ligation-dependent probe amplification and quantitative real-time polymerase chain reaction analysis.20,21 Cytogenetic risk was defined as standard when there were no adverse lesions, high in the presence of t(4;14), t(14;16), t(14;20), del(17p), or gain(1q), and ultra-high when more than one adverse lesion was present.22,23 The data cut-off for inclusion in this analysis was May 31, 2019.
Results
Outcomes for transplant-eligible pathway patients by age
The 2,042 patients enrolled in the TE pathway had a median age of 61 (28-75) years; 546 (27%) were aged 65-69 and 101 (5%) were aged 70-75 (Table 1). Older patients were more frequently categorized as ISS stage III and had a lower performance status than younger patients. There was no significant difference in the proportion of patients in each of the cytogenetic risk groups or the number of patients in each arm of the induction treatment randomiza- tion between age groups. Response at the end of induction was similar across age groups.
Older patients in the TE pathway were less likely than younger patients to undergo stem cell harvest at the end of induction. The percentage of patients undergoing stem cell harvest fell from 73.5% in those aged <65 years, to 62.2% among those aged 65-69 and only 57.4% in the group aged 70 or older (Table 1, Online Supplementary Figure S1). The reason given for not proceeding to stem cell harvest and subsequent transplantation was more likely to be due to the clinician/patient not considering that they were fit enough to proceed in the older age groups than in the younger patients (Table 1).
Older patients had a lower median harvested CD34+ cell count but still had a high rate of achieving the standard cut- off of 2x106 CD34+ cells/kg needed for one ASCT. The per- centage of patients achieving this target was 95.0% in those aged <65, 90.0% in those aged 65-69 and 88.7% among patients aged 70 or older. Conversely, fewer patients in the older age groups achieved the cut-off of 4x106 CD34+ cells/kg considered adequate for two ASCT, with reduc- tions from 63.4% to 45.6% to 32.1% in the respective age groups.
Of the 2,042 patients in the TE pathway, 1,370 (67%) received melphalan. Most patients (84.7%) received 200 mg/m2 with only 10.5% reported to have received 140 mg/m2. The proportion receiving the lower dose increased in the older age groups with the 140 mg/m2 dose being given to only 5.5% of patients aged <65 years, but 19.9% of those aged 65-69 and 45.5% of those aged >70 years. This appeared to be due to both an increased incidence of elevated serum creatinine (>200 mmol/L) in the older age groups and the systematic use of the lower dose for older patients in some centers.
Response to transplant, PFS and OS outcomes for patients of different ages within the TE pathway who underwent
haematologica | 2022; 107(1)
233