Multi-Cancer Early Detection Requires a Rethink of Screening Clinical Trial Design
Dr. Eric Klein, Distinguished Scientist | June 16, 2023

Multi-cancer early detection (MCED) is a promising new technology to improve cancer detection rates. MCEDs are important because cancer is the second-leading killer of Americans expected to claim over 600,000 deaths this year. While MCED tests aren’t meant to replace recommended cancer screenings, the reality is there are only five cancer screening tests available in the US for the at-risk general population and most cancer deaths result from other cancer types for which there is no screening.

New approaches to reduce the profound toll of late-stage cancer are urgently needed and long overdue. 

MCED allows for efficient screening for many different types of cancer through a simple blood draw, in contrast to single-cancer screenings that screen for one cancer at a time. They are designed to detect cancer at early stages when outcomes are far more likely to be favorable and are optimized to minimize false-positive results and potential harms from unnecessary and costly diagnostic tests. For widespread adoption to be fully realized, there is an urgent need for clinical trials that move this new field forward.

Traditional clinical trials require large sample size, long duration, high cost and long latency for reporting results. Requiring large trials with mortality endpoints for newer screening modalities like MCED tests poses several challenges. One is the view that MCED test performance be evaluated for each cancer type separately, following the traditional approach of testing new interventions on single cancer types in separate trials. As MCED tests target multiple cancer types often through identifying a shared cancer signal, even in a relatively large trial, several cancer types would each have too few participants to produce reliable estimates of screening performance using cancer mortality as an endpoint. Two completed trials highlight these challenges. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial and the United Kingdom Collaborative Trial of Ovarian Cancer Screening Trial (UKCTOCS) included 78,216 and 202,638 women at average risk of ovarian cancer, respectively, randomized to ovarian cancer screening vs. standard of care. Final results of these trials were not published until 18 and 20 years after initial randomization, and both trials were negative. These trials used technology unlikely to advance over time, whereas genomic-based tests are continually advancing in AI and machine learning capabilities and would be obsolete in this duration of a trial.

We propose that modern approaches for cancer screening trials are needed to accelerate clinical evaluation of these promising technologies and their translation into clinical practice. My colleagues and I—including leaders representing several MCED test manufacturers—recently published a paper in Cancer Epidemiology, Biomarkers & Prevention that spotlights how alternative trial endpoints can be used to keep pace with this rapidly advancing technology.

In short, to realize the full potential of MCED, high-quality clinical trials with alternative trial designs can be used to establish efficacy at a fraction of the time and costs of trials powered for mortality. 

  • Biomarker vs cancer type trials: Modern tumor agnostic therapies that target a biomarker versus the cancer type are starting to change the traditional therapeutic clinical trial paradigm. Recently, trials for two investigational tumor agnostic therapies included multiple tumor types, several of which had less than 10 cases. Regulatory approvals have been granted based on all cancers combined without the requirement to show efficacy for each type, an approach that could be applied to MCED using Bayesian statistical methods to explore whether test performance is agnostic to different tumor types.
  • Reduction in incidence of late-stage cancer as a relevant surrogate for mortality: Alternative trial designs could be used to efficiently evaluate MCED tests, including those that use reduction in incidence of late-stage cancer as a relevant surrogate for mortality and mathematical modeling to project reductions in late-stage disease.
    • This approach has been adopted in several federally funded trials, including the Tomosynthesis Mammographic Imaging Screening Trial (TMIST), a breast cancer screening trial that is using reduction in incidence of late-stage cancer as the primary endpoint. The design was chosen primarily because of the length of time needed for mortality to be an endpoint.
  • The ongoing National Health Service (NHS)-Galleri trial uses this alternative endpoint —reduction in incidence of advanced stage (III and IV) cancers diagnosed in the intervention (blood tested by MCED) versus control (blood stored) arm 3-4 years following randomization. The trial completed enrollment at 140,000 participants in just 10 months, far exceeding traditionally designed trials.
  • Time to diagnosis endpoint: Time to diagnosis is another alternative endpoint that could be used for MCED, since delays in certain intervals (diagnostic and referral intervals, for example) are associated with advanced stage of cancer at diagnosis. A nested trial design is an additional option for retaining cancer mortality as the primary endpoint but enabling a reduction in the number of participants needed for study. With this design, blood would be collected from all participants but analyzed only for participants randomized to the screening arm. Modeling and sensitivity analyses could be explored to account for limitations inherent in this design.
  • Other alternative endpoints include candidacy for curative interventions at diagnosis, overall cancer detection rate, reduced treatment morbidity for early-stage cancers, increased treatment response rates, improved quality of life during and after treatment, utilization rates and rates of metastatic recurrence. New approaches can advance the scientific evaluation of MCED tests and allow crucial questions about this promising technology to be answered. For a cancer screening paradigm shift at the population level, clinical trials need to be conducted in a timely fashion. Requiring large trials with mortality endpoints creates barriers and delays for MCED development and adoption. The high mortality impact of cancer even with current screening recommendations means that we can and should rethink our clinical trial framework now, before these novel technologies become obsolete.

It is well understood and accepted that early detection of solid tumors is what matters to effectively treat and potentially cure them. We have an enormous opportunity to improve cancer care and outcomes with novel technology combining the advances of human genomics and machine learning, and MCED tests are poised to be the first public health application of the human genome project. Similar to rapid healthcare advancements during the COVID pandemic, realizing this potential will require innovative thinking. We need to move beyond the single cancer screening mindset into the new front in the war on cancer with MCEDs.