INSURE: a pooled analysis of ixazomib-lenalidomide-dexamethasone for relapsed/refractory myeloma in routine practice

Abstract

Aim: We pooled data from three observational studies (INSIGHT MM, UVEA-IXA and REMIX) to investigate the real-world effectiveness of ixazomib-lenalidomide-dexamethasone (IRd) in relapsed/refractory myeloma. Materials & methods: INSIGHT MM was a prospective study conducted in countries across Europe, Asia and North/Latin America while UVEA-IXA and REMIX were multicenter, retrospective/prospective studies conducted in Europe. Patients who had received IRd as ≥2nd line of therapy were analyzed. Primary outcomes were time-to-next treatment (TTNT) and progression-free survival (PFS). Results: Overall, 564 patients were included (median follow-up: 18.5 months). Median TTNT and PFS were 18.4 and 19.9 months; both outcomes were numerically longer for earlier versus later lines. Median treatment duration was 14.0 months. Overall response rate was 64.6%. No new safety concerns were noted. Conclusion: The effectiveness of IRd in routine practice appears similar to the efficacy observed in TOURMALINE-MM1. IRd benefit in earlier versus later lines was consistent with previous reports.

Keywords: :

• effectiveness
• ixazomib
• ixazomib-lenalidomide-dexamethasone
• IRd
• multiple myeloma
• pooled analysis
• progression-free survival
• proteasome inhibitor
• relapsed/refractory
• routine clinical practice
• time-to-next treatment

Ixazomib is the first oral proteasome inhibitor (PI) to be licensed to treat multiple myeloma (MM). It is approved in combination with lenalidomide and dexamethasone (IRd) as an all-oral regimen to treat relapsed/refractory multiple myeloma (RRMM) in patients who have received ≥1 prior therapy, based on the results of the TOURMALINE-MM1 phase III study [1]. In this pivotal trial, IRd demonstrated superior progression-free survival (PFS) which was the primary end point (median, 20.6 vs 14.7 months; hazard ratio [HR], 0.74, 95% CI: 0.59–0.94; p = 0.01) and improved response rates (overall response rate [ORR], 78 vs 72%; ≥ very good partial response [VGPR] rate, 48 vs 39%) versus placebo plus lenalidomide and dexamethasone (placebo-Rd), with limited added toxicity in patients with RRMM [2].

Clinical trials are the gold standard for establishing the efficacy and safety of new therapies, yet their results may not always be reproducible in routine clinical practice [3]. Reasons for this include the rigorous eligibility criteria, resulting in a highly selected population of patients that are enrolled into clinical trials, which is not reflective of the broader patient populations that are cared for in routine clinical practice. It has been estimated that up to 72% of real-world patients with RRMM would not meet the eligibility criteria for randomized controlled trials [4]. Real-world studies with less stringent eligibility criteria may permit inclusion of a more diverse patient population, including elderly and frail patients, and those with comorbidities or reduced performance status, and may better inform on the effectiveness of therapies when used in routine clinical practice. Thus, it is important to supplement clinical trial data with real-world evidence.

Several retrospective and prospective observational studies have now shown the comparable real-world effectiveness of IRd in second or later lines of therapy (LoT) to the efficacy observed in TOURMALINE-MM1, with median PFS ranging from 17.5 to 27.6 months [5–8]. To confirm and further investigate these reports, the current analysis evaluated real-world treatment patterns and clinical outcomes for patients with RRMM treated with IRd in routine clinical practice, overall, by LoT, and by frailty status, using a large, global dataset pooled from three observational studies [8–11].

Materials & methods

Study design

INSURE is a pooled, global analysis of data from three observational studies that actively recruited patients with RRMM receiving IRd as part of routine clinical practice across multiple countries worldwide: INSIGHT MM, UVEA-IXA and REMIX (Figure 1) [12]. INSIGHT MM (NCT02761187) was a large, prospective study, which enrolled 4307 MM patients from 15 countries across Europe, Asia and North and Latin America (2337 with newly diagnosed MM [NDMM] and 1939 with RRMM) [10]. Patients were enrolled over a period of 5 years; the study ended when all patients completed at least 2 years of follow-up. Baseline information on patient characteristics, diagnosis and previous treatment were collected from hospital or clinic records. Data on MM-specific disease management, safety and outcomes were then obtained quarterly [10]. The second study, UVEA-IXA, was a multicenter, longitudinal, retrospective/prospective cohort study of 309 RRMM patients receiving ixazomib-based treatment via an early access program in eight countries across Europe (Czech Republic, Greece, Hungary, Italy, Slovakia, Slovenia, Spain and UK) [11]. The study comprised two phases: a retrospective chart review, and a prospective follow-up with quarterly data capture. The enrollment period for the prospective phase was approximately 9 months and follow-up was 12 months from chart review initiation. The third study, REMIX, was a multicenter, retrospective/prospective study of 197 RRMM patients receiving IRd via a compassionate use program in France [8]. In REMIX, baseline data were collected from medical records, examination results and patient interviews. After IRd treatment had commenced, assessments were conducted quarterly for 24 months and every 6 months thereafter per standard practice [8]. Adverse event (AE) data were collected quarterly with a particular focus on certain AEs, including thrombocytopenia, gastrointestinal events and peripheral neuropathy, on the case report form. Patients were followed for a minimum of 24 months for the last patient enrolled and a maximum period of 48 months for the first patient enrolled.

Figure 1. Summary of studies included in the INSURE pooled analysis.

*Retrospective chart review was conducted.

^†Patients included per eligibility criteria outlined in the ’Patients’ section of the ’Materials and methods’ in the main body of the manuscript.

DOT: Duration of treatment; IRd: Ixazomib-lenalidomide-dexamethasone; LoT: Line of therapy; ORR: Overall response rate; OS: Overall survival; PFS: Progression-free survival; RRMM: Relapsed/refractory multiple myeloma; TTNT: Time-to-next treatment.

Reprinted from [12], Copyright © 2021 The Author(s).

Data taken from [8–11].

For this combined INSURE analysis, data from the three source studies were pooled for adult patients with ≥2 prior lines of therapy for MM who had received IRd as ≥2nd LoT. If the same patient appeared to exist in multiple databases, and follow-up started from the same LoT across databases, only the observation with the longest follow-up period was used. If follow-up started from different LoTs across databases, only the observation from the earlier LoT was used. Patients were followed up until the end of each study, loss to follow-up, or death, whichever occurred first. Data cutoff dates for each of the data sources pooled in the combined INSURE analysis were 1 March 2021 for INSIGHT MM, 30 September 2019 for UVEA-IXA and 4 June 2020 for REMIX.

Patients

Adult patients (aged ≥18 years) with a diagnosis of MM were included if they had evidence of RRMM (defined as initiation of at least a 2nd LoT for MM after the diagnosis date) and had initiated a 2nd, 3rd, or later LoT with IRd in routine clinical practice. Only patients treated with the IRd triplet regimen were included; any patients who had received other ixazomib-based regimens were excluded. Assignment of LoT was physician-assessed. Patients were excluded from the analysis if IRd had been received during any prior LoT (including as maintenance or consolidation); IRd treatment had started >90 days prior to signing the informed consent form (INSIGHT MM only); they had been enrolled in a clinical trial at the time of receiving IRd; they had received a stem cell transplant in the same LoT as IRd; or if there was >2 months (≥60 days) difference in the start dates for ixazomib or lenalidomide.

Outcomes & assessments

The primary outcomes in INSURE were time-to-next treatment (TTNT) and PFS. Secondary outcomes included ORR, rates of complete response (CR) and VGPR, duration of treatment (DOT), overall survival (OS), safety (discontinuations or dose reductions due to AEs), and demographic and clinical characteristics. Effectiveness outcomes are defined in Supplementary Table 1. For time-to-event outcomes, the index date and start of follow-up was defined as the date of initiation of IRd treatment in the 2nd or later LoT in patients with RRMM. For assessment of cytogenetic risk at baseline, high risk was defined as the presence of del[17p], t[4;14], and/or t[14;16]. Frailty was defined based on a simplified International Myeloma Working Group (IMWG) [13] frailty scale based on age, Charlson Comorbidity Index (CCI) and Eastern Cooperative Oncology Group performance status (ECOG PS), assessed at the start of IRd (score 0–1: non-frail; score ≥2: frail; see supplementary methods for further details on the frailty definition) [14].

Due to differences in how safety data were captured in each study, AEs and discontinuations/dose reductions due to AEs were reported separately for each study.

Statistical analysis

No statistical hypotheses were tested as part of this study. Consequently, no sample size calculations were performed; all available and eligible patients in the pooled dataset were included in the analysis.

Analyses were performed for the overall population and by frailty status (non-frail or frail) and LoT (2nd, 3rd or ≥4th). Time-to-event outcomes were analyzed using Kaplan–Meier methodology. Univariable and multivariable analyses by Cox proportional hazards modeling were used to investigate the association between patient characteristics/treatment history and TTNT, PFS, DOT and OS, as well as adjust for the heterogeneity associated with the data from each individual study. A p-value of <0.05 was considered statistically significant and no corrections were made for multiplicity. The multivariable analysis included adjustments for patient and treatment history characteristics. The variables included in univariable and multivariable analyses were line of IRd therapy, sex, age, comorbidities, ECOG PS, cytogenetic risk, initial ixazomib dose, frailty score, estimated glomerular filtration rate (eGFR), prior exposure to immunomodulatory drugs (IMiDs), lenalidomide history, PI history, time since diagnosis to IRd therapy and data source as a random effect.

Results

Patients & treatment history

In total, 564 patients (INSIGHT MM, n = 181; UVEA-IXA, n = 195; REMIX, n = 188) enrolled in 17 countries were included in INSURE; most patients were enrolled in France (n = 204, 36.2%) and the UK (n = 144, 25.5%; Supplementary Table 2).

Patient baseline characteristics and disease characteristics by study are detailed in Supplementary Table 3, and the pooled baseline data at the start of IRd therapy, overall, by line of IRd therapy, and by frailty status are summarized in Table 1 [12]. Median age in the overall population was 68 years. Among 242 patients with cytogenetic data available (missing, n = 322), 15.3% were considered high risk at diagnosis. A total of 17.5% of 492 patients with available data had an ECOG PS of ≥2 at the start of treatment with IRd, and 30.2% of 520 patients had an eGFR of ≤60 ml/min/1.73 m². The median time from diagnosis to start of IRd was 39.3 months in the overall population (N = 564), and 29.7, 43.4 and 71.1 months among patients who received IRd as 2nd (n = 230), 3rd (n = 215) and ≥4th LoT (n = 119), respectively (Table 1).

Table 1. Patient baseline and disease characteristics overall, by line of IRd therapy and by frailty status.

Characteristic	All (N = 564; 100%)	2nd LoT (n = 230; 40.8%)	3rd LoT (n = 215; 38.1%)	≥4th LoT (n = 119; 21.1%)	Non-frail (n = 242; 42.9%)	Frail (n = 164, 29.1%)
Male, %	51.8	46.5	57.7	51.3	53.7	46.3
Race^†, n	254	81	121	52	136	89
White/Caucasian, %	91.3	91.4	90.1	94.2	91.2	92.1
Age at start of IRd, n	553	227	210	116	242	164
Median, years (range)	68 (36–92)	69 (36–91)	68 (36–92)	68 (39–87)	67 (36–80)	76 (37–92)
Aged ≤65 years, %	36.7	36.1	37.1	37.1	47.1	17.1
Aged 66–75 years, %	42.0	37.4	43.3	48.3	50.8	29.3
Aged >75 years, %	21.3	26.4	19.5	14.7	2.1	53.7
Time from MM diagnosis to start of IRd therapy, n	550	225	209	116	238	158
Median, months (range)	39.3 (0–396.8)	29.7 (0–162.8)	43.4 (3.6–172.4)	71.1 (7.2–396.8)	39.3 (3.3–387.4)	34.4 (3.6–396.8)
Follow-up time from IRd onset, months, n	564	230	215	119	242	164
Median, months (range)	18.5 (0–51.8)	20.5 (1.0–51.8)	18.1 (0–50.6)	15.1 (0–42.0)	20.6 (0–47.2)	16.4 (0–46.5)
M-protein type at diagnosis, n	384^‡	172	132	80^‡	169^‡	126
IgG, %	54.7	59.9	48.5	53.8	55.0	51.6
IgA, %	19.0	20.3	19.7	15.0	16.6	20.6
Light chain only^‡, %	21.1	17.4	23.5	25.3	22.0	23.8
Cytogenetic risk at diagnosis^§, n	242	118	82	42	116	71
High, %	15.3	16.1	19.5	4.8	17.2	12.7
Standard, %	84.7	83.9	80.5	95.2	82.8	87.3
ECOG PS at the start of IRd^¶, n	492	196	194	102	242	164
0, %	32.3	32.7	33.0	30.4	47.1	7.3
1, %	50.2	50.5	51.5	47.1	52.9	47.6
≥2, %	17.5	16.8	15.5	22.5	0	45.1
CCI at the start of IRd^¶, n	445	177	179	89	242	164
0, %	64.5	65.0	67.6	57.3	78.9	42.1
1, %	12.1	14.1	10.1	12.4	14.9	9.8
≥2, %	23.4	20.9	22.3	30.3	6.2	48.2
eGFR at the start of IRd^#, n	520	211	199	110	235	156
≥60 ml/min/1.73 m^2, %	69.8	73.5	66.3	69.1	82.1	45.5
30–60 ml/min/1.73 m^2, %	22.5	20.9	25.6	20.0	12.8	41.7
<30 ml/min/1.73 m^2, %	7.7	5.7	8.0	10.9	5.1	12.8
Frailty score at the start of IRd^††, n	406	161	168	77	242	164
0–1, %	59.6	59.0	60.7	58.4	100	0
≥2, %	40.4	41.0	39.3	41.6	0	100

† Race was not collected for the REMIX study.

‡ n = 383 (overall), 79 (≥4th LoT) and 168 (non-frail) for light chain only assessment.

§ Defined as the presence of del[17p], t[4;14], and/or t[14;16]; patients for whom an abnormality was not detected but were not assessed for all abnormalities were classified as missing.

¶ From 1 year prior to until ≤90 days after the start of IRd therapy for INSIGHT MM and REMIX patients; date of assessment was not available for UVEA-IXA patients and was replaced by the date of IRd initiation.

# As recorded for UVEA-IXA, for INSIGHT MM and REMIX, the values were estimated according to serum creatinine, age and race.

†† Defined by a simplified IMWG frailty score (based on age, CCI and ECOG PS; 0–1 [non-frail] vs ≥2 [frail]) [13,14], assessed at the start of IRd.

CCI: Charlson Comorbidity Index; ECOG PS: Eastern Cooperative Oncology Group performance status; eGFR: Estimated glomerular filtration rate; Ig: Immunoglobulin; IMWG: International Myeloma Working Group; IRd: Ixazomib-lenalidomide-dexamethasone; LoT: Line of therapy; MM: Multiple myeloma.

Reprinted from [12], Copyright © 2021 The Author(s).

Patients had received a median of 2 prior LoTs (range, 1–12), with approximately 80% receiving IRd in either 2nd (40.8%) or 3rd (38.1%) LoT (Table 2). The percentages of patients who were naive, exposed, and refractory to PIs, IMiDs and lenalidomide are shown in Table 2. Overall (n = 562; data missing for two patients), the percentage of patients exposed or refractory in any prior line to PIs (bortezomib, carfilzomib, or ixazomib) was 92.7%; and IMiDs (lenalidomide, pomalidomide, or thalidomide) was 71.9% (lenalidomide, 29.9%). In 2nd LoT, percentages of patients who were exposed or refractory to a PI, to an IMiD, or to lenalidomide were 92.6, 48.0 and 13.1%, respectively; in 3rd LoT were 92.5, 85.0 and 28.0%, respectively; and in ≥4th LoT were 93.3, 94.1 and 65.5%, respectively. In all 564 patients, induction therapies received prior to IRd included bortezomib (89.4%), high-dose chemotherapy and autologous stem cell transplantation (ASCT; 50.2%), thalidomide (48.8%), lenalidomide (27.7%), daratumumab (10.1%) and carfilzomib (7.3%; Table 2).

Table 2. Summary of prior therapy (any LoT), including prior exposure/refractoriness to lenalidomide, PIs and IMiDs, overall and by line of IRd therapy.

	All (N = 564; 100%)	2nd LoT (n = 230; 40.8%)	3rd LoT (n = 215; 38.1%)	≥4th LoT (n = 119; 21.1%)
Median number of prior therapies, % (range)	2.0 (1–12)	–	–	–
Any prior exposure to lenalidomide, n	562	229	214	119
Naive, %	69.6	86.9	72.0	31.9
Exposed, %	17.8	10.0	19.2	30.3
Refractory^†, %	12.1	3.1	8.9	35.3
Any prior exposure to a PI^‡, n	562	229	214	119
Naive, %	6.6	7.4	6.5	5.0
Exposed, %	73.1	85.6	69.6	55.5
Refractory^†, %	19.6	7.0	22.9	37.8
Any prior exposure to an IMiD^§, n	562	229	214	119
Naive, %	28.1	52.0	15.0	5.9
Exposed, %	51.1	44.1	62.6	43.7
Refractory^†, %	20.8	3.9	22.4	50.4
Drug exposure in any LoT prior to IRd, %
Bortezomib	89.4	88.7	88.8	91.6
Dexamethasone	89.4	77.8	96.3	99.2
High-dose chemotherapy and ASCT	50.2	43.0	52.1	60.5
Thalidomide	48.8	32.2	63.3	54.6
Cyclophosphamide	46.8	24.8	64.7	57.1
Lenalidomide	27.7	11.7	25.1	63.0
Melphalan	23.8	26.1	19.5	26.9
Prednisone	19.9	22.2	16.7	21.0
Daratumumab	10.1	3.0	5.1	32.8
Carfilzomib	7.3	2.2	6.5	18.5
Pomalidomide	6.4	0.4	0.5	28.6

† Defined as patients who progressed while receiving or within 60 days of discontinuing a PI- or IMiD-containing regimen, or if the treatment-free interval between discontinuation of that regimen and the next index regimen (not containing a PI or IMiD) was ≤60 days.

‡ Bortezomib, carfilzomib, or ixazomib.

§ Lenalidomide, pomalidomide, or thalidomide.

ASCT: Autologous stem cell transplantation; IMiD: Immunomodulatory drug; IRd: Ixazomib-lenalidomide-dexamethasone; LoT: Line of therapy; PI: Proteasome inhibitor.

Outcomes: overall & by LoT

Median duration of follow-up from the start of IRd therapy for all patients was 18.5 months. Time-to-event results by line of IRd therapy and frailty status are summarized in Supplementary Table 4.

While time-to-event results by cytogenetic risk status are not available due to missing data, this variable was included in the univariable and multivariable analyses (Supplementary Tables 5–8). High versus standard cytogenetic risk was found to have a significant effect (p < 0.05) in the univariable analysis of PFS (HR >1; Supplementary Table 6).

TTNT

Median TTNT was 18.4 months (95% CI: 15.3–20.8) overall and was numerically longer in patients receiving IRd in earlier versus later LoTs (Figure 2A) [12].

Figure 2. Kaplan–Meier analyses.

(A) TTNT, and (B) PFS, in the overall analysis population and by line of IRd therapy.

IRd: Ixazomib-lenalidomide-dexamethasone; LoT: Line of therapy; NE: Not estimable; PFS: Progression-free survival; TTNT: Time-to-next treatment.

Reprinted from [12], Copyright © 2021 The Author(s).

Univariable & multivariable analyses of TTNT

Supplementary Table 5 shows univariable and multivariable analyses of TTNT by Cox proportional hazards. In the univariable analysis, receiving IRd in ≥4th versus 2nd LoT was significantly associated with a shorter TTNT (HR >1). In the multivariable analysis, variables statistically significantly associated with a shorter TTNT (HRs >1) among patients treated with IRd were ECOG PS 1 or ≥2 versus 0, and refractory to lenalidomide versus lenalidomide naive in prior LoTs. By contrast, time from diagnosis to initiation of IRd of >5 years or 2–3 years versus <1 year were statistically significantly associated with a longer TTNT (HRs <1).

PFS

280 (49.6%) had progressed or died at the time of analysis; median PFS was 19.9 months (95% CI: 16.6–23.6) overall, and numerically longer in patients receiving IRd in earlier versus later LoTs (Figure 2B) [12]. At 24 months, PFS rates were 44.8, 46.1 and 43.1% for 2nd, 3rd and ≥4th LoT, respectively (Supplementary Table 4).

Univariable & multivariable analyses of PFS

Supplementary Table 6 shows univariable and multivariable analyses of PFS by Cox proportional hazards. In the multivariable analysis, variables statistically significantly associated with a higher risk of progression (HRs >1) among patients treated with IRd included: ≥4th LoT versus 2nd LoT, age 60–64 years versus <55 years, ECOG PS 1 or ≥2 versus 0, refractory versus naive to lenalidomide. Conversely, >5 years versus <1 year from diagnosis to IRd therapy was significantly associated with lower risk of progression (HR <1).

DOT

Median DOT was 14.0 months overall and was numerically longer in patients receiving IRd as 2nd or 3rd (16.9 and 14.8 months, respectively) versus ≥4th LoT (7.5 months; Figure 3A & Supplementary Table 7) [12]. In the univariable analysis, receiving IRd in ≥4th versus 2nd LoT was significantly associated with a shorter DOT (HR >1; Supplementary Table 7).

Figure 3. Kaplan–Meier analyses.

(A) DOT, and (B) OS, in the overall analysis population and by line of IRd therapy.

DOT: Duration of treatment; IRd: Ixazomib-lenalidomide-dexamethasone; LoT: Line of therapy; NE: Not estimable; NR: Not reached; OS: Overall survival.

Reprinted from Leleu X, Boccadoro M, Lee HC, et al. Poster 2701 presented at the 63rd Annual Meeting of the American Society of Hematology (ASH) 2021, Copyright © 2021 The Author(s) [12].

OS

184 patients (32.6%) had died at data cutoff; OS data were not mature, with median OS not reached in the overall study population or in patients receiving IRd as 2nd or ≥4th LoT and in patients receiving IRd as 3rd LoT, median OS was 34.9 months (95% CI: 27.2–not estimable; Figure 3B & Supplementary Table 8) [12]. In the overall population, survival probability at 6, 12, 18 and 24 months was 89.9, 80.0, 71.7 and 65.8%, respectively (Supplementary Table 4). At all timepoints, the survival probability was numerically higher in earlier versus later LoTs (Supplementary Table 4). In the univariable analysis, receiving IRd in ≥4th versus 2nd LoT was significantly associated with lower survival (HR >1; Supplementary Table 7).

Best response

Best ORR to IRd therapy among the 404 response-evaluable patients was 64.6%. ORR was 70.5, 63.1 and 52.8% among patients receiving IRd as 2nd, 3rd, and ≥4th LoT, with a median time to best response of 3.9, 5.2 and 3.5 months, respectively (Figure 4) [12].

Figure 4. Best response to IRd therapy overall and by line of IRd therapy at the time of IRd initiation.

Response data were missing for 160 patients overall: 47, 66 and 47 for patients receiving IRd as 2nd, 3rd and ≥4th LoT, respectively. Percentages may not sum due to rounding. ORR: PR + VGPR + CR + sCR.

CR: Complete response; IRd: Ixazomib-lenalidomide-dexamethasone; LoT: Line of therapy; MR: Minimal response; ORR: Overall response rate; PD: Progressive disease; PR: Partial response; sCR: Stringent complete response; SD: Stable disease; VGPR: Very good partial response.

Reprinted from Leleu X, Boccadoro M, Lee HC, et al. Poster 2701 presented at the 63rd Annual Meeting of the American Society of Hematology (ASH) 2021, Copyright © 2021 The Author(s) [12].

Safety

Dose reductions and discontinuations are shown in Table 3 [12]. Across all three studies, ixazomib and lenalidomide dose reductions occurred in 13.5 and 23.9% of patients, respectively, and drug discontinuation occurred in 49.5 and 48.0% of patients, respectively. In INSIGHT MM, the most common AE leading to ixazomib dose reduction was peripheral neuropathy (24.0%). In UVEA-IXA, the most common AE leading to ixazomib dose reduction was diarrhea (38.9%). The most common AE leading to ixazomib dose discontinuation was thrombocytopenia in both studies (INSIGHT MM, 18.5%; UVEA-IXA, 24.2%). Rash-related AEs were recorded as a reason for ixazomib dose reduction and discontinuation in 16.0 and 9.3% of patients, respectively, in INSIGHT MM, and 5.6 and 6.0%, respectively, in UVEA-IXA. Notably, safety data for REMIX were recorded differently than for INSIGHT MM and UVEA-IXA, hence, most common AEs leading to ixazomib dose reduction for REMIX were not reported. Rates of AEs and serious AEs and most common AEs reported in REMIX are shown in Table 4 [12]. Most common AEs were diarrhea (14.9%) and thrombocytopenia (14.4%); rash-related AEs were reported in 4.3% of patients. No new safety signals were identified in any of the three studies [2].

Table 3. Dose modifications in the INSURE pooled analysis, and common AEs associated with ixazomib dose reduction or discontinuation in INSIGHT MM and UVEA-IXA^†.

	All (N = 564)	2nd LoT (n = 230; 40.8%)	3rd LoT (n = 215; 38.1%)	≥4th LoT (n = 119; 21.1%)
Dose reductions, %
Ixazomib	13.5	13.9	13.5	12.6
Lenalidomide	23.9	24.3	27.9	16.0
Dexamethasone	13.8	15.2	14.0	10.9
Discontinuations, %
Ixazomib	49.5	48.7	47.4	54.6
Lenalidomide	48.0	44.8	50.2	50.4
Dexamethasone	45.7	43.9	48.8	43.7
	INSIGHT MM (n = 181)		UVEA-IXA (n = 195)
Most common AEs associated with ixazomib dose reduction^‡, n	25		18
Diarrhea, %	20.0		38.9
Peripheral neuropathy, %	24.0		16.7
Thrombocytopenia, %	20.0		11.1
Nausea/vomiting, %	8.0^§		16.7
Fatigue, %	16.0		5.6
Most common AEs associated with ixazomib discontinuation^‡, n	54		33
Thrombocytopenia, %	18.5		24.2
Diarrhea, %	9.3		18.2
Infections and infestations, %	14.8		6.1^¶
Peripheral neuropathy, %	7.4		12.1
Fatigue, %	11.1		0

† Most common AEs associated with ixazomib dose reduction were not reported for REMIX.

‡ Occurring in >10% of patients with dose reduction or discontinuation in at least one study. For dose reductions in INSIGHT MM, 8% of patients had a rash, 4% of patients had a vesicular rash and 4% of patients had a grade 1 upper extremities skin rash, which totals to 16%; however, these were recorded as separate AEs and are thus not reported in the table, as individually they did not occur in >10% of patients.

§ Nausea only.

¶ Infection only.

AE: Adverse event; LoT: Line of therapy.

Reprinted from [12], Copyright © 2021 The Author(s).

Table 4. REMIX Safety Summary^†.

	REMIX (n = 188)
AEs, %	64.9
Serious AEs, %	37.8
Most common AEs^†, %
Diarrhea	14.9
Thrombocytopenia	14.4
Most common serious AEs^‡, %
Thrombocytopenia	5.3

† Occurring in >10% of patients.

‡ Occurring in >5% of patients.

AE: Adverse event.

Reprinted from [12], Copyright © 2021 The Author(s).

Results by frailty status

Patients & treatment history

Patient baseline characteristics and disease characteristics at the start of IRd therapy by frailty status are summarized in Table 1. Among non-frail (n = 242) and frail (n = 164) patients, the median time from diagnosis to start of IRd was 39.3 and 34.4 months (Table 1). Patient baseline characteristics and disease characteristics by frailty status and line of IRd therapy are summarized in Supplementary Table 9.

Among both non-frail and frail groups, patients had received a median of 2 prior LoT (range, 1–8 and 1–11, respectively). In the non-frail and frail groups, the percentages of patients exposed or refractory to PIs or IMiDs in any prior line were: PIs (bortezomib, carfilzomib, or ixazomib), 92.6 and 92.7%; IMiDs (lenalidomide, pomalidomide, or thalidomide), 74.4 and 64.0% (lenalidomide, 29.8 and 27.4%).

Outcomes

In the non-frail and frail groups, median TTNT was 21.4 months (95% CI: 16.1–27.5) and 12.6 months (95% CI: 9.9–15.3; Figure 5A) [12], and median PFS was 21.6 months (95% CI: 18.2–30.0) and 11.8 months (95% CI: 9.0–15.6), respectively (Figure 5B) [12]. Median DOT was 16.1 months (95% CI: 12.7–20.4) in non-frail patients and 9.7 months (95% CI: 7.2–11.9) in frail patients (Supplementary Figure 1A) [12]. At data cutoff, 61 non-frail and 85 frail patients had died. Median OS was not reached in non-frail patients and 20.1 months (95% CI: 17.5–30.4) in frail patients (Supplementary Figure 1B). In the non-frail and frail groups, ORR was 67.1 and 59.2%, respectively, with a median time to best response of 4.7 and 3.2 months. The univariable analysis showed that having frailty scores of ≥2 versus 0–1 was significantly associated with worse outcomes for TTNT, PFS, DOT and OS (HRs >1; Supplementary Table 5–8).

Figure 5. Kaplan–Meier analyses.

(A) TTNT, and (B) PFS, by frailty status.

LoT: Line of therapy; PFS: Progression-free survival; TTNT: Time-to-next treatment.

Reprinted from [12], Copyright © 2021 The Author(s) [12].

Ixazomib dose reduction occurred in 12.8% non-frail and 12.2% frail patients, and 48.8 and 58.5%, respectively, discontinued ixazomib therapy (Supplementary Table 10). In INSIGHT MM, the most common AEs leading to ixazomib dose discontinuation were thrombocytopenia in non-frail patients (26.1 vs 14.3% of frail patients) and infections and infestations in frail patients (21.4 vs 17.4% of non-frail patients). In UVEA-IXA, the most common AEs leading to ixazomib dose discontinuation were thrombocytopenia in non-frail patients (35.7 vs 12.5% of frail patients) and diarrhea in frail patients (25.0 vs 14.3% of non-frail patients). Ixazomib dose discontinuations due to rash were reported for non-frail patients only (17.4%) in the INSIGHT MM trial; ixazomib dose discontinuations due to all rash-related AEs occurred in non-frail patients only (21.7%) in INSIGHT MM. In UVEA-IXA, ixazomib dose reductions due to cutaneous rash occurred in non-frail patients only (10.0%). In REMIX, thrombocytopenia and diarrhea were the most commonly reported AEs in both non-frail and frail groups (both 16.4% of non-frail and 17.4 of frail patients; Supplementary Table 11).

Discussion

The findings from INSURE, an analysis of a large, global, pooled dataset of 564 patients, show that the effectiveness of IRd in routine clinical practice (median PFS, 19.9 months, ORR, 64.6%) is similar to the efficacy of IRd observed in the registrational TOURMALINE-MM1 trial (median PFS, 20.6 months; ORR, 78%) [2]. As real-world studies often provide limited data on the occurrence of disease progression, TTNT is a real-world proxy for PFS [15,16]. Therefore, it is unsurprising that in INSURE, similar median values were obtained for each outcome measure, with the slightly shorter median TTNT likely due to some patients switching treatment for reasons other than progression. Both the assessments of TTNT and PFS were feasible in INSURE, indicating the robustness of the observed effectiveness of IRd. The safety profile of IRd was also generally comparable to that of TOURMALINE-MM1 [2], with no new, additional safety concerns.

Similar outcomes were observed in the present analysis and the TOURMALINE-MM1 trial [2] despite INSURE patients tending to have less favorable baseline characteristics. A qualitative, indirect comparison of patients’ characteristics between the two studies shows that, compared with TOURMALINE-MM1 patients, INSURE patients were slightly older (median age, 68 years [INSURE] vs 66 years [MM1]) with a worse PS (ECOG PS ≥2 [INSURE] or 2 [MM1], 17.5 vs 6%) and poorer renal function (eGFR [INSURE] or creatinine clearance [MM1] <30 ml/min/1.73 m², 7.7 vs 1%). INSURE patients were also more heavily pretreated than TOURMALINE-MM1 patients and more likely to have received a PI or IMiD in a prior LoT. Compared with TOURMALINE-MM1 patients, a higher proportion of INSURE patients had received a greater number of prior therapies (3rd/4th LoT, 38.1/21.1% [INSURE] vs 29/10% [MM1]), a prior PI (prior bortezomib, carfilzomib, or ixazomib, 92.7% [INSURE] vs 70% [MM1]) and prior lenalidomide (29.9 [INSURE] vs 12% [MM1]). The median number of prior LoTs was 2 in INSURE versus 1 in TOURMALINE-MM1. Furthermore, a lower proportion of INSURE patients had received a previous ASCT (50.2 [INSURE] vs 57% [MM1]). INSURE also included 19.6% of patients who were refractory to PIs and 12.1% of patients who were refractory to lenalidomide; both of these patient subtypes were excluded from the TOURMALINE-MM1 trial [2]. Consequently, INSURE patients comprised a more diverse, advanced, and difficult-to-treat real-world, RRMM patient population than the TOURMALINE-MM1 trial population [2]. The results from this and other real-world population studies [17] complement data from clinical trials to provide a more complete picture of the outcomes that can be expected to be obtained with IRd across a broad patient population.

These results from INSURE confirm the overall real-world effectiveness of IRd, and also indicate a greater benefit of treatment when IRd was used in earlier versus later lines: TTNT, PFS, DOT, ORR and OS were improved in earlier versus later lines, although statistical comparisons were not undertaken. The difference in OS outcomes between LoTs was likely due to salvage therapies received following progression on treatment. Future studies should investigate the association between OS and subsequent therapies. Furthermore, in the univariable analysis, receiving IRd in ≥4th versus 2nd LoT was significantly associated with worse outcomes for TTNT, DOT and OS. This apparent greater treatment benefit seen with IRd in earlier versus later lines is consistent with reports from previous, smaller real-world studies of IRd and other regimens in RRMM patients [6,7,18,19]. It is worth noting, however, that the improved benefits surrounding earlier LoT may be impacted by patient or disease characteristics. An analysis by LoT indicated that PFS probability at 24 months was comparable across all LoTs (44.8, 46.1 and 43.1% for 2nd, 3rd, and ≥4th LoT, respectively). Interestingly, compared with patients in the 2nd LoT group, fewer patients in the ≥4th LoT group were aged >75 years (26.4 vs 14.7%, respectively) and fewer patients had a high cytogenetic risk status (16.1 vs 4.8%, respectively). Despite this finding, the overall results align with current treatment goals of optimizing patient persistence on therapy or utilizing a treat-to-progression approach to maximize patient outcomes [20–22]. Due to the convenience of the all-oral nature of IRd therapy, this regimen is a suitable option for continuous therapy/using a treat-to-progression approach. A recent real-world study of daratumumab-containing treatments showed a similar pattern of incremental reductions in effectiveness with each LoT [23]. Furthermore, a previous report suggests that the proportion of patients reaching ≥4th LoTs in clinical practice is as low as 15% [19]. Together, these findings indicate the importance of providing effective treatment regimens, such as IRd, in each and every LoT in the RRMM treatment course to optimize clinical outcomes.

This combined analysis of three real-world studies provides valuable data from a larger patient pool across different countries and regions; however, it may be impacted by typical limitations inherent to real-world studies. These include selection bias due to convenience sampling (since healthcare providers selected the patients in the three source studies), confounding bias (introduced by differences in routine data collection across countries and studies), missing data and inconsistent data reporting across study sites [3,24]. For example, data on International Staging System (ISS) or Revised-ISS classification, new primary malignancies and AEs by grade were not captured. Additionally, in REMIX, the incidence and number of AEs reaching the >10% threshold were low compared with safety data reported in the TOURMALINE-MM1 study [2], which is likely related to the infrequent follow-up and limited data collection that are common in real-world studies [3,24]. Nonetheless, the use of three separate data sources and a large, diverse group of patients and prescribers, along with implementation of statistical adjustments to address bias and confounding in the multivariable analyses, should have helped to minimize the impact of potential bias on the data. For the primary outcome of PFS, patients without an adequate response assessment during the study period were censored at the last date they were known to be alive within their current line of treatment. While censoring can affect the estimation of PFS, it is a well-accepted method for estimating time-to-events in the absence of complete data [25]. To accurately determine PFS in this study, one would need to know the proportion of patients without regular response assessments who had progressed and the length of survival prior to progression for each of these patients. As nearly three-quarters of the INSURE patients (72%) were evaluable for response, the impact of these particular patients on the PFS estimates is unlikely to be high. It should be noted that, as the analysis set comprised mainly European patients, caution should be used when extrapolating these data to other geographies. Future research is warranted to examine the effectiveness of IRd in real-world clinical practices in different countries and regions of the world.

The findings from INSURE will advance the treatment of MM by increasing our understanding of achievable outcomes and expected tolerability in patients who are typically excluded from clinical trials, providing prescribers with confidence that the outcomes observed in IRd clinical trials can be replicated in the more heterogeneous patient populations typically seen in clinical practice. This is supported by a previous study that demonstrated the utility of the IRd regimen in intermediate-frail and frail patients [26]. In the present study, frail patients achieved worse outcomes than non-frail patients, which has been reported elsewhere [13]; this might be due to the slightly higher rate of ixazomib dose discontinuations due to adverse events among frail versus non-frail patients; however, the difference in outcomes observed among frail and non-frail patients is likely to be expected given, by definition, frail patients were older, had more comorbidities and worse functional status (as defined by CCI score and ECOG PS) at the start of IRd. Despite this, median TTNT with IRd exceeded 12 months in the present study, indicating a positive benefit in this vulnerable patient group. The all-oral IRd regimen may be particularly suitable for this patient group due to reduced need for visits to healthcare facilities versus parental PI regimens, which can be associated with substantial time commitment, inconvenience and economic burden for patients [3], as well as increased risk of exposure to infections. In INSURE, similar proportions of frail and non-frail patients had their ixazomib dose reduced due to AEs, suggesting that IRd therapy is manageable across patient populations.

Conclusion

The findings from this large, global dataset pooled from three observational studies indicate that the real-world effectiveness of IRd therapy in patients with RRMM is similar to that seen in the registrational phase III TOURMALINE-MM1 trial, despite the real-world patients having less favorable baseline characteristics. The data also indicate a greater benefit of IRd treatment, in terms of PFS, TTNT and DOT, when it is used in earlier versus later LoTs, and provide insight into the outcomes that can be achieved with IRd in a frail MM patient population. Thus, these findings should help to increase the understanding of achievable outcomes with IRd therapy in patients with RRMM in routine clinical practice.

Summary points

• Real-world studies may permit inclusion of a more diverse patient population, due to less stringent eligibility criteria, and may better inform on the effectiveness of therapies when used in routine clinical practice.

• INSURE is a pooled, global analysis of data from three observational, multicenter studies (INSIGHT MM, UVEA-IXA and REMIX) that recruited patients with relapsed/refractory multiple myeloma receiving ixazomib-lenalidomide-dexamethasone (IRd) as part of routine clinical practice.

• Patient-level data on demographics, disease characteristics, treatment history, effectiveness and safety collected from all three studies were pooled and analyzed.

• In total, 564 patients (INSIGHT MM, n = 181; UVEA-IXA, n = 195; REMIX, n = 188) enrolled in 17 countries were included in INSURE.

• The median time from diagnosis to start of IRd was 39.3 months overall, and 29.7, 43.4 and 71.1 months among patients who received IRd as 2nd (n = 230), 3rd (n = 215) and ≥4th line of therapy (LoT; n = 119), respectively.

• Median duration of follow-up from the start of IRd therapy for all patients was 18.5 months; median time-to-next treatment and progression-free survival were 18.4 and 19.9 months; both outcomes were numerically longer for earlier versus later lines.

• Median treatment duration was 14.0 months and was numerically longer in patients receiving IRd as 2nd or 3rd versus ≥4th LoT.

• Best overall response rate to IRd therapy among the 404 response-evaluable patients was 64.6%. Overall response rate was 70.5, 63.1 and 52.8% among patients receiving IRd as 2nd, 3rd, and ≥4th LoT.

• Across all three studies, ixazomib and lenalidomide dose reductions occurred in 13.5 and 23.9% of patients, respectively, and drug discontinuation occurred in 49.5 and 48.0% of patients, respectively. No new safety concerns were noted.

• The effectiveness of IRd in routine practice appears similar to the efficacy observed in TOURMALINE-MM1. IRd benefit in earlier versus later lines was consistent with previous reports.

Author contributions

X Leleu, HC Lee, D Cherepanov, DM Stull, E Terpos: conception or design of the work; X Leleu, HC Lee, JA Zonder, M Macro, K Ramasamy, C Hulin, J Silar, M Kuhn, K Ren, N Bent-Ennakhil, D Cherepanov, DM Stull, E Terpos: acquisition, analysis or interpretation of data of the work; X Leleu, HC Lee, JA Zonder, M Macro, K Ramasamy, C Hulin, J Silar, M Kuhn, K Ren, N Bent-Ennakhil, D Cherepanov, DM Stull, E Terpos: drafting the work or revising it critically for important intellectual content; all authors provided final approval of the manuscript to be published and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

X Leleu, HC Lee and JA Zonder were steering committee members for the INSIGHT MM study. K Ramasamy, N Bent-Ennakhil and E Terpos were steering committee members for the UVEA-IXA study. X Leleu, M Macro and C Hulin were steering committee members for the REMIX study. J Silar, M Kuhn, K Ramasamy, N Bent-Ennakhil, D Cherepanov, DM Stull and E Terpos were steering committee members for INSURE.

Financial disclosure

These studies were funded by Takeda Development Center Americas, Inc. (TDCA), Lexington, MA. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Competing interests disclosure

X Leleu: Honoraria: AbbVie, Amgen, BMS, Janssen-Cilag, Novartis, Roche, Sanofi, Takeda, Pfizer; Non-financial support: Takeda. HC Lee: Consultancy: BMS, Celgene, Genentech, Karyopharm, Legend Biotech, GSK, Sanofi, Oncopeptides, Pfizer, Takeda, Allogene Therapeutics, Janssen Pharmaceutical; Research funding: BMS, Janssen, GSK, Takeda, Regeneron, Amgen. JA Zonder: Consultancy: Prothena, BMS, Takeda, Janssen, Regeneron, Research funding: BMS, Janssen; Member of board of directors/advisory committees: Takeda, BMS. MM: Honoraria: Takeda, Janssen, GSK, Sanofi, BMS/Celgene; Travel, accommodation, and research funding: Takeda, Janssen. K Ramasamy: Honoraria and member of board of directors/advisory committees: Amgen, Janssen, Takeda, BMS/Celgene, GSK, Oncopeptides, Adaptive Biotech, Karyopharm, Pfizer, Sanofi; Research funding: Amgen, Janssen, Takeda, BMS/Celgene, GSK; Travel and conference registration: Amgen, Janssen, Takeda, BMS/Celgene. C Hulin: Honoraria: Janssen, BMS/Celgene, Sanofi, Takeda, AbbVie. J Silar, M Kuhn: None. KRen, NB-E, DC, DMS: Employee: Takeda (D Cherepanov and DM Stull also currently hold equity in Takeda). E Terpos: Research funding, honoraria, and consultancy: Amgen, BMS, GSK, Janssen-Cilag, Takeda, Sanofi; Honoraria only: Eusa Pharma, Novartis. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing disclosure

Editorial support was provided by Victoria Enwemadu of Takeda Pharmaceuticals U.S.A., Inc., Lexington, MA. Medical writing support for the development of this manuscript, under the direction of the authors, was provided by Jenny Wilkinson, PhD, and Laura Webb, PhD, of Ashfield MedComms, an Inizio company, funded by Takeda Pharmaceuticals U.S.A., Inc., Lexington, MA, and complied with the Good Publication Practice (GPP) guidelines (DeTora LM, et al. Ann Intern Med 2022;175:1298-304).

Ethical conduct of research

Each of the three studies was conducted in accordance with the Declaration of Helsinki, and any applicable local regulations. The research was approved by local independent review boards or independent ethics committees at each site and informed consent was obtained from all patients.

Open access

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/

Acknowledgments

The authors would like to thank all patients and their families, as well as all investigators for their valuable contributions to this study.

Supplementary data

To view the supplementary data that accompany this paper please visit the journal website at: www.tandfonline.com/doi/suppl/10.2217/fon-2023-0604

Data availability statement

The datasets, including the redacted study protocol, redacted statistical analysis plan, and individual participants’ data supporting the results reported in this article will be made available within 3 months from initial request to researchers who provide a methodologically sound proposal. The data will be provided after its de-identification, in compliance with applicable privacy laws, data protection, and requirements for consent and anonymization.