Clinical Trial Design

Clinical Trial DesignsSpecial topicsResearchResidents
Written By Sydney Oesch

Episode Notes

  1. Types of Trials

    1. Screening trials

      1. Designed to improve the discovery of early, asymptomatic disease

      2. Should both identify individuals and prescribe treatment

    2. Prevention Trials

      1. Designed to reduce the effects of a disease

      2. Primary prevention

        1. Investigate ways to reduce the risk of disease occurring

        2. I.e. vaccine trials

      3. Secondary prevention

        1. Designed to identify treatment of early-stage disease, thereby reducing risk of progression to later stage disease

      4. Tertiary prevention

        1. Designed to identify interventions that decrease the morbidity and/or mortality associated with a disease after people have been diagnosed

          1. This is the majority of oncology trials

  2. Phases of Drug Development

    1. Preclinical

      1. Studying the drug in a basic science lab - tumor cell lines, tissue cultures, animals

    2. Phase I

      1. “Safety phase”

      2. Aimed to identify dosages of drugs - specifically, maximum tolerated dose (MTD)

        1. Goal is to identify recommended dose level (RDL) OR recommended phase II dose (RP2D) OR no recommended dose if none of the effective doses are tolerable

      3. Enroll small groups of patients who are administered increasing doses of the developing drug

      4. Clinical responses also recorded but not primary aim

    3. Phase IB Expansion

      1. Take a specific dose or specific subgroup from the initial portion of the phase I trail, then expand its use in that group

      2. Often randomized to two different doses

        1. If no difference in efficacy, then FDA will require the lower dose for next phases

    4. Phase II

      1. “Efficacy Phase”

      2. Investigates whether an agent produces a sufficiently robust response to move to phase 3

      3. How to measure efficacy?

        1. RECIST criteria

        2. Time to first progression or recurrence

      4. Smaller than phase 3 trials and typically not placebo-controlled

      5. Primary outcome usually overall response rate (not PFS or OS)

    5. Phase IIa

      1. Compare response rates to comparative historical control

      2. Multi-arm compares multiple experimental regimens against SOC response rate for that disease site/stage/setting

    6. Phase IIb

      1. Compare response rates of experimental drug to response in patients given SOC in real time

      2. Commonly used for biomarker-specific/targeted therapy trials (often there isn’t an accepted historical response rate)

    7. Safety Lead-In

      1. Small cohort

      2. Allows for quicker movement from phase I to phase II

      3. Can be used for new drug combinations sometimes when toxicities are not anticipated to be that severe

      4. If not severe, then can expand to phase II trial

      5. Example is chemo + IO combination trials

    8. Phase III

      1. Designed to determine efficacy compared to SOC in a specific patient population

        1. Aim is to change SOC and to get FDA approval

      2. Types of Phase III

        1. Superiority: is the intervention better than the SOC?

          1. Accounts for majority

        2. Non-inferiority: is the intervention not worse than the SOC?

          1. Designed to demonstrate that there is something better (i.e. side effects, QOL, cost) with non-inferior oncologic outcomes (i.e. PFS, OS)

        3. Equivalency: rarely used in oncology

      3. “Best”: randomized, blinded, placebo-controlled

        1. Randomization

          1. Assigning participants randomly to the control arm vs experimental arm

            1. Reduces selection bias

          2. Examples: stratified randomly permutated blocks, dynamic treatment allocation

        2. Blinding

          1. Non-disclosure of control vs experimental arm

          2. Single - just the patient

          3. Double - patient and administering physician

          4. Triple - patient, physician, statistician

        3. Placebo

          1. Another way to reduce bias and strengthen integrity of trial

    9. Adaptive Trials

      1. Allows quicker movement from Phase II to Phase III

      2. Allow trial design to change during the trial, based on interim data

      3. Saves resources, allow the trial to progress quicker w/o having to design, accrue, and get approval for two separate trials

    10. Phase IV

      1. Conducted after regulatory approval to evaluate long-term effectiveness, rare adverse events, and real-world outcomes

  3. Clinical Trial Outcomes/Endpoints

    1. Typical Primary Endpoints

      1. Phase I/II: Overall Response Rate

      2. Phase III+: Progression free survival (PFS), overall survival (OS), or progression free survival 2 (PFS2)

    2. Definitions

      1. PFS2: time from randomization (or start of initial therapy) to progression on the next line of therapy or death, whichever occurs first

        1. captures both: duration of disease control on the initial (first-line) treatment (PFS1), AND duration of control on the subsequent therapy after progression

      2. Time to progression (TTP) does not include deaths

      3. Progression free survival (PFS) does include deaths; includes stable disease

      4. Duration of response (DOR): only includes patients with an objective response to therapy

    3. PFS Controversies

      1. Used to be viewed as a good surrogate for OS; not always true now with multiple lines of therapy

      2. Conversely, forcing wait for OS may delay approval

      3. Possible biases:

        1. Assessment of progression bias: due to potential treatment effects, true blinding may not be possible

          1. Blinded Independent Central Review (BICR) of imaging can help overcome this bias

        2. Assessment-time bias: the time point at which you discover a progression is when you will say progression occurs

          1. Assessments in both SOC and experimental arms must be the same

  4. Toxicities/Adverse Events

    1. Standardly graded 1-5

      1. Grade 1: mild; Grade 5: death

      2. Grade 3+ considered “significant”

        1. Treatments adjusted or discontinued at this level

    2. Should compare AE rate of investigational drug to SOC therapy

  5. Patient Reported Outcomes (PROs) / Patient Reported Outcome Measures (PROMs)

    1. PROs are measured by PROMs - validated questionnaires

    2. Help identify and characterize symptoms and toxicities, in addition to measurement of graded adverse events

    3. Best practices

      1. Be efficient and thoughtful about which PROMs you use - pts can experience respondent fatigue and clinical work flow can be burdened by over-collecting

      2. Complement of PROMs should include a general QOL tool and a disease-specific PROM

    4. General QOL PROMs

      1. FACT-G (27 questions) and EORTC QLQ C30 (30 questions) best studied and most commonly used

      2. Can add disease-specific scales - usually increases to ~60 questions

      3. FACT-G7 (7 questions) over 3 domains - physical, emotional, functional well-being

        1. Good for clinical use outside of trials

        2. See Figure 1 for example

      4. FACT-GP5: “I am bothered by the side effects of treatment”

        1. SGO recommends the FACT-G7 + GP5 for use in routine clinical settings

    5. Disease Sites

      1. Ovarian Cancer PROMs

        1. Disease-specific subscales from FACT or EORTC

        2. NCCN/FACT Ovarian Symptom Index (NFOSI): 18-item measure

      2. Uterine Cancer PROMs

        1. FACT-G/FACT-EN (15 items)

        2. EORTC-EN24 (24 items)

        3. SGO states uterine disease is understudied in terms of PROs

      3. Cervical Cancer PROMs

        1. Heterogeneous group when comes to impact on QOL

        2. Anxiety and sexual function concerns very relevant for this group even w/o disease recurrence

        3. FACT and EORTC with cervical cancer add-ins

      4. Vulvar/Vaginal Cancer

        1. Stigma around these diseases so symptoms may not be spontaneously reported by patients

        2. FACT-V (15-items) to add to general QOL PROM

        3. EORTC is developing their scale

    6. Additional Scales

      1. Sexual Function

        1. Up to 80% of female cancer pts report distressing sexual, vaginal, and/or menopausal concerns related to cancer and treatment; 90% say their oncologists don’t ask them about these things!

        2. Female Sexual Function Index (FSFI) - 19 items

          1. Modifiable for sexually diverse population and those who aren’t currently sexually active

        3. PROMISE sexual function and satisfaction measures version 2.0 (PROMIS SexFS v2.0)

          1. Appropriate for patients who are sexually active with or without a partner

        4. EORTC - currently validating a 22 item QLQ-SHQ22

        5. SGO recommended directly asking “would you like to discuss your sexual health?”

  6. Biomarkers

    1. Can be used as outcomes (biomarker change) or as inclusion/selection criteria for clinical trials

    2. Predictive vs Prognostic Biomarkers

      1. Prognostic: refers to the cancer outcome, independent of the therapy received

      2. Predictive: whether the presence of a specific biomarkers impacts the chance of response to therapy

        1. Quantitative Effect: both groups response to therapy but some get a better impact

        2. Qualitative Effect: response vs no response

      3. Proving biomarker is predictive can be challenging - need at least two comparison groups available, need to perform a statistical test for an interaction, and the test for interaction needs to be statistically significant

    3. Clinical Trial Designs

      1. Definitions

        1. Integral biomarker: biomarker-defined criteria for enrollment into a trial

        2. Integrated biomarker: used to test the hypothesis during the study, but isn’t a requirement for enrollment

      2. Biomarker enrichment trials: only include patients with the biomarker expressed

        1. Can reduce required sample size while maintaining power

        2. Helpful if biomarker positivity prevalence is low

      3. Randomized umbrella trials: target one disease of interest, with several different biomarkers being evaluated

        1. Usually hierarchical/prioritized - one biomarker is the “prioritized” biomarker, and if a patient has that one, they’re triaged to that arm; if no, then check next biomarker

          1. Example in endometrial cancer - RAINBO trial - for the different molecular subtypes of endometrial cancer

      4. Basket trials: disease agnostic, test a specific targeted therapy on people w/ many types of tumors, as long as they have a specific biomarker

        1. Example: DESTINY-PanTumor02

  7. Clinical Trial Eligibility

    1. Exists primarily to reduce the influence of confounding variables

      1. Aims to identify the target population and make them similar at baseline

      2. Aims to exclude people in whom toxicities are expected to outweigh benefits

      3. Aims to follow regulatory guidelines regarding eligible populations

      4. Aims to protect personal privacy

      5. Aims to exclude patients who would not be able to comply with the planned intervention

    2. Real-world Issues

      1. Creating “ideal” populations can reduce the generalizability 

      2. Historically contributed to preferential enrollment of white patients from high SES who have less comorbidities or other reasons for exclusion

      3. Comorbidity criteria

        1. Tracking the relationship between accrual and comorbid conditions in clinical TRial enrollment (TRACE), Oluloro et al 2024

          1. Aimed to characterize the comorbidity profile of patients with uterine cancer by race and compare w/ expected clinical trial enrollment patterns based on standard/historic Comorbidity Exclusion Criteria (CEC)

          2. >284,000 patients included - 73% white, 14% black, 3% asisan

          3. Odds of clinical trial exclusion based on comorbidities were 2x higher for black vs white patients w/ OR 2.09

      4. Language

        1. Often an exclusion criteria for enrollment

          1. Even when not, still lower rates of enrollment for patients with limited English proficiency

          2. Jorge et al. 2023

            1. Enrollment rate 7.5% for English-speaking vs 2.2% for limited English proficiency

  8. Best Practices for Discussing Clinical Trial Enrollment w/ Patients

    1. Consider enrollments for all patients

      1. Documented belief amongst medical providers that patients of color do not want to participate in research studies

        1. Langfor et al. 2013 (Cancer)

          1. No difference along racial or ethnic lines in clinical trial refusal rates or “no desire to participate in research” as a reason to refuse clinical trial

    2. Common reasons pts decline enrollment

      1. Discomfort w/ randomization and possibly receiving placebo

      2. Fear of experimentation

      3. Concerns about privacy

      4. Time cost of the consent and monitoring process

    3. Highlight values

      1. Clinical benefit for them

      2. Improving our knowledge of how to treat patients in the future

      3. Increased knowledge about their own cancer/genetics

    4. Core Principles

      1. Trials should be offered to all eligible patients without preconceived assumptions about who might or might not be interested in participation

      2. Clinicians should be forthcoming about both the limitations of their knowledge and the hopes informed by earlier phases of investigation

      3. Transparency throughout the process is essential; while the ultimate outcomes of a trial cannot be predicted, patients should be provided with a clear framework of what to expect along the way

Figure 1.

FACT-G7 (7 questions) over 3 domains - physical, emotional, functional well-being; QOL survey

Reference List

1. Sisodia RC, Dewdney SB, Fader AN, et al. Patient reported outcomes measures in gynecologic oncology: A primer for clinical use, part I. Gynecol Oncol 2020;158(1):194–200; doi: 10.1016/j.ygyno.2020.04.696.

2. Sisodia RC, Dewdney SB, Fader AN, et al. Patient reported outcomes measures in gynecologic oncology: A primer for clinical use, Part II. Gynecol Oncol 2020;158(1):201–207; doi: 10.1016/j.ygyno.2020.03.022.

3. Anonymous. Vulva Cancer | EORTC – Quality of Life. n.d. Available from: https://qol.eortc.org/questionnaire/qlq-vu34/ [Last accessed: 4/27/2025].

4. Anonymous. Patient-Reported Outcomes Measurement Information System (PROMIS) | NIH Common Fund. n.d. Available from: https://commonfund.nih.gov/patient-reported-outcomes-measurement-information-system-promis [Last accessed: 4/27/2025].

5. Ballman K V. Biomarker: Predictive or prognostic? Journal of Clinical Oncology 2015;33(33):3968–3971; doi: 10.1200/JCO.2015.63.3651.

6. Oluloro A, Pike M, Moore A, et al. Tracking the relationship between accrual and comorbid conditions in clinical TRrial enrollment (TRACE). Gynecol Oncol 2024;190:S72; doi: 10.1016/J.YGYNO.2024.07.106.

7. Oluloro A, Temkin SM, Jackson J, et al. What’s in it for me?: A value assessment of gynecologic cancer clinical trials for Black women. Gynecol Oncol 2023;172:29–35; doi: 10.1016/j.ygyno.2023.03.002.

8. Jorge S, Masshoor S, Gray HJ, et al. Participation of Patients with Limited English Proficiency in Gynecologic Oncology Clinical Trials. JNCCN Journal of the National Comprehensive Cancer Network 2023;21(1):27–32; doi: 10.6004/jnccn.2022.7068.

9. Montes de Oca MK, Howell EP, Spinosa D, et al. Diversity and transparency in gynecologic oncology clinical trials. Cancer Causes and Control 2023;34(2):133–140; doi: 10.1007/S10552-022-01646-Y/METRICS.

10. Langford AT, Resnicow K, Dimond EP, et al. Racial/ethnic differences in clinical trial enrollment, refusal rates, ineligibility, and reasons for decline among patients at sites in the National Cancer Institute’s Community Cancer Centers Program. Cancer 2014;120(6):877–884; doi: 10.1002/cncr.28483.

 


Sydney Oesch
Next

Surgical Management of Placenta Accreta