Table 1 Types of CID trials.

From: Effective delivery of Complex Innovative Design (CID) cancer trials—A consensus statement

 

(i) Evaluation of one experimental treatment (E), common to all cohorts

(ii) Comparative evaluation of multiple experimental treatments (E1, E2, E3, …), common to all cohorts

(iii) Non-comparative evaluation of single (E) or multiple experimental treatments (E1, E2, E3, …), each specific for a biomarker-defined cohort

Population defined by a single disease

No biomarker i.e., single cohort defined by disease

“Standard” RCT investigating E vs Control (C)

Multi-arm multistage (MAMS) design comparing E1 vs E2 vs E3 etc. vs C

e.g., STAMPEDE29

Not relevant for trials with no biomarker

Population stratified by single biomarker (B) i.e., two cohorts: B + and B−

Stratified biomarker design investigating E vs C within each cohort

Biomarker stratified MAMS design comparing E1 vs E2 vs E3 etc. vs C separately in B + and B−

Biomarker strategy design evaluating the strategy of using E in B + cohort and C in B− cohort

Population stratified by multiple biomarkers

i.e., biomarkers B1 + , B2 + , B3 + … and B-

Stratified biomarker design investigating E vs C within each cohort

Biomarker stratified MAMS design comparing E1 vs E2 vs E3 etc. vs C within each biomarker cohort

e.g., I-SPY2,30 BATTLE31

Umbrella trial evaluating treatment E1 in B1 + cohort, E2 in B2 + cohort etc. Experimental treatments could be allocated to a basket of biomarker cohorts e.g., PROFILE 1001,14 FOCUS4,32 National Lung Matrix Trial21

Population defined by multiple diseases

No biomarker i.e., multiple cohorts defined only by disease

Stratified RCT investigating E vs C across a basket of disease types

Stratified MAMS design comparing E1 vs E2 vs E3 etc. vs C across a basket of disease types

Not relevant for trials without biomarkers

Population stratified by single biomarker i.e., two cohorts B + and B- within each disease

Stratified biomarker design investigating E vs C within each biomarker cohort across a basket of disease types

Stratified MAMS design comparing E1 vs E2 vs E3 etc. vs C separately in B + and B− across a basket of disease types

Biomarker strategy design evaluating the strategy of using E in B + cohort and C in B− cohort across a basket of disease types

Population stratified by multiple biomarkers i.e., biomarkers B1 + , B2 + , B3 + , … and B- within each disease

Stratified biomarker design investigating E vs C within each biomarker cohort across a basket of disease types

Stratified MAMS design comparing E1 vs E2 vs E3 etc. vs C within each biomarker cohort across a basket of disease types

Umbrella trial evaluating treatment E1 in B1 + cohort, E2 in B2 + cohort etc. across a basket of disease types e.g., NCI-MATCH33

  1. (i) Descriptions specify the comparison of experimental arms to a control arm C, but designs can include experimental arms that would be compared to historical controls; (ii) all designs described that include multiple cohorts or multiple experimental treatments can also be dynamic platform trials; (iii) all designs can be adaptive and incorporate seamless transition from one phase to the next
  2. Examples of studies shown in Bold type
Back to article page