Illustration of the 3-step assessment on the main indirect comparisons of celta-cel against comparators. Bold cells indicate the main issues of the performed comparisons. The following 5 main confounders were considered and ranked as major confounders by experts: Refractory status, cytogenetic profile, R-ISS stage, plasmacytomas, and time to progression on last prior line
| Indirect comparison . | Step 1- Estimand . | Step 2- External source of data . | Step 3- Methods of comparison . | |||||
|---|---|---|---|---|---|---|---|---|
| Main objective . | Comparator . | Type . | Source . | Propensity score . | Method . | Balance diagnostics, common support . | Estimation of effect . | |
| Merz 2021 | ATT | Standard treatment heterogeneity | IPD | Retrospective German RWD database risk of bias | 9 confounders∗ cytogenetic and plasmacytomas missing | IPW | Undetailed “remaining imbalances” (SMD >0.20) | Weighted analyses with robust variance |
| Weisel 2022 | ATT | Physician choice heterogeneity | IPD | Follow-up of trial data (POLLUX, CASTOR, EQUULEUS) | 8 confounders† | IPW | Mean SMD reduced from 0.33 to 0.16 | Weighted analyses with robust variance |
| Costa 2022 | ATT | Conventional treatment heterogeneity | IPD | Retrospective study Risk of bias | 16 confounders‡ Plasmacytomas missing | Matching 1:1, no replacement, caliper 0.05 | SMD between 0.10 and 0.20 (ASCT, refractory to carfilzomib, penta-drug refractory) | Stratified/weighted analyses |
| Weisel 2022 | ATC not explicitly reported | Belantamab mafodotin | Aggregate | One-arm (2.5 mg/kg dose) of the 2-arm trial data (DREAMM-2) ECOG 0-2 | 4 confounders § Time to progression on the last regimen missing | Unanchored MAIC | ESS = 39 (60% reduction) no report of weight distribution | Weighted analyses |
| ATC not explicitly reported | Selinexor-DXM | Aggregate | RCT data (mITT of STORM-2) Penta-exposed ECOG 0-2 | 4 confounders § Time to progression on the last regimen missing | Unanchored MAIC | ESS = 73 (25% reduction) No report of weight distribution | ||
| ATC not explicitly reported | Melphalan-flufenamide-DXM | Aggregate | A subset of Single-arm trial data (HORIZON) received ≥2 prior LOTs ECOG 0-2 | 3 confounders refractory status missing | Unanchored MAIC | ESS = 85 (12% reduction) no report of weight distribution | ||
| Martin 2022 | ATC not explicitly reported | Ide-cel | Aggregate | Single-arm trial data (KarMMa) | 4 confounders § Time to progression on the last regimen missing | Unanchored MAIC | Skewed distribution of weights ESS: 46%-57% reduction | Weighted analyses failure times measured from cells infusion |
| Indirect comparison . | Step 1- Estimand . | Step 2- External source of data . | Step 3- Methods of comparison . | |||||
|---|---|---|---|---|---|---|---|---|
| Main objective . | Comparator . | Type . | Source . | Propensity score . | Method . | Balance diagnostics, common support . | Estimation of effect . | |
| Merz 2021 | ATT | Standard treatment heterogeneity | IPD | Retrospective German RWD database risk of bias | 9 confounders∗ cytogenetic and plasmacytomas missing | IPW | Undetailed “remaining imbalances” (SMD >0.20) | Weighted analyses with robust variance |
| Weisel 2022 | ATT | Physician choice heterogeneity | IPD | Follow-up of trial data (POLLUX, CASTOR, EQUULEUS) | 8 confounders† | IPW | Mean SMD reduced from 0.33 to 0.16 | Weighted analyses with robust variance |
| Costa 2022 | ATT | Conventional treatment heterogeneity | IPD | Retrospective study Risk of bias | 16 confounders‡ Plasmacytomas missing | Matching 1:1, no replacement, caliper 0.05 | SMD between 0.10 and 0.20 (ASCT, refractory to carfilzomib, penta-drug refractory) | Stratified/weighted analyses |
| Weisel 2022 | ATC not explicitly reported | Belantamab mafodotin | Aggregate | One-arm (2.5 mg/kg dose) of the 2-arm trial data (DREAMM-2) ECOG 0-2 | 4 confounders § Time to progression on the last regimen missing | Unanchored MAIC | ESS = 39 (60% reduction) no report of weight distribution | Weighted analyses |
| ATC not explicitly reported | Selinexor-DXM | Aggregate | RCT data (mITT of STORM-2) Penta-exposed ECOG 0-2 | 4 confounders § Time to progression on the last regimen missing | Unanchored MAIC | ESS = 73 (25% reduction) No report of weight distribution | ||
| ATC not explicitly reported | Melphalan-flufenamide-DXM | Aggregate | A subset of Single-arm trial data (HORIZON) received ≥2 prior LOTs ECOG 0-2 | 3 confounders refractory status missing | Unanchored MAIC | ESS = 85 (12% reduction) no report of weight distribution | ||
| Martin 2022 | ATC not explicitly reported | Ide-cel | Aggregate | Single-arm trial data (KarMMa) | 4 confounders § Time to progression on the last regimen missing | Unanchored MAIC | Skewed distribution of weights ESS: 46%-57% reduction | Weighted analyses failure times measured from cells infusion |
ASCT, allogeneic stem cell transplantation; DXM, dexamethasone; ECOG, Eastern Cooperative Oncology Group; ISS, international staging system; LOT, line of treatment; MM, multiple myeloma.
Age, sex, refractory status, R-ISS stage, time to progression on last prior line, number of prior LOTs, average duration of prior lines, years since diagnosis, ECOG status.
Age, refractory status, ISS stage, cytogenetic profile, time to progression on last regimen, plasmacytoma, number of prior LOTs, years since MM diagnosis.
Age, sex, race/ethnicity (white vs other), ISS stage 3 (vs 1, 2, or unknown), time from diagnosis to index date, number of prior LOT, prior autologous stem cell transplant, presence of high- risk cytogenetic abnormalities in any prior sample [t(4;14), t(14;16), del(17p)], refractoriness to bortezomib or ixazomib, refractoriness to carfilzomib, refractoriness to lenalidomide, refractoriness to pomalidomide, refractoriness to anti-CD38 monoclonal antibody, triple-class refractoriness, penta-drug exposure (to bortezomib or ixazomib plus carfilzomib plus lenalidomide plus pomalidomide plus anti-CD38 monoclonal antibody), and penta-drug refractoriness.
Refractory status, cytogenetic profile, R-ISS stage, plasmocytomas.