Precision Prevention Studies: A Targeted Approach to Cancer Prevention

Abstract

Precision prevention trials are biologically driven interception studies conducted in high cancer risk groups. These are smaller, potentially faster, cheaper, and more commercially attractive than traditional large-scale population prevention studies. In this article, we discuss the key challenges of conducting precision prevention research and their mitigations.

Introduction

The incidence of cancer is rising and, of the 18 million cancers currently diagnosed annually worldwide, it is estimated that more than 60% are preventable, not only with lifestyle adaptations such as smoking cessation but also with active interceptions like vaccines or targeted medications (1). An example is the human papillomavirus vaccine which, in the United Kingdom, has already led to an 87% population level reduction in cervical cancer among those vaccinated during childhood through the National Health Service (NHS) national vaccination program (2). Effective prevention has the potential to reduce cancer’s societal burden and is increasingly a priority for funders and governments (3). However, it is challenging for cancer researchers to efficiently evaluate the safety and efficacy of preventive agents using existing clinical trial methodologies. Traditional population-based phase III cancer prevention trials, when not targeted to very high–risk populations, are expensive and slow, taking an average of 10 to 15 years (4–7). They have tended to evaluate repurposed drugs or diet-derived micronutrients and vitamins in nonbiologically selected participants and have limited scope for exploring biological hypotheses. With a few notable exceptions, these studies have largely failed to change practice or prompt positive licensing decisions, factors that have made cancer prevention unattractive to pharmaceutical and biotechnological investors (8). Precision prevention studies (PPS) have the potential to overcome these limitations as they differ from traditional population prevention studies in multiple aspects, summarized in Table 1. Although PPSs have been acknowledged as necessary to advance cancer prevention, there is a lack of practical guidance around their design (9, 10). In this article, we define the central tenets of PPSs and their risks and mitigations.

Table 1.

Comparison of key features between traditional population prevention studies and PPSs.

Feature	Traditional population prevention	Precision prevention
Phase	Phase 3 or 4	Mostly phase 1/2 but also phase 3
Focus	Evaluating risk reduction across an entire population	Evaluates risk reduction of an intervention in a specific group of people with a matched biological vulnerability
Biological endpoints	Not required	Serial biosamples from all participants to explore biological impact and mechanism of action of intervention and identify or validate markers of tumorigenesis
Primary endpoints	Incidence of a specific cancer(s) (primary), incidence of other cancers, and/or overall survival (secondary) behavioral changes and impact on other diseases	Short-term surrogate endpoints: cancer incidence and change to biomarkers.
Useful for	A repurposed or repositioned intervention or dietary-derived agent (e.g., vitamin, micronutrient) that is cheap, easy to administer, and has minimal adverse effects	A novel, repurposed, or repositioned intervention. Greater than minimal side effects tolerable for higher-risk individuals
Goals	Benefit wider population to a lesser degree	Benefit smaller population to a larger degree
Risks	Lack of biological endpoints prohibits detailed understanding of mechanisms of action in study participants and future biologically informed stratification. Long duration of studies is unattractive to commercial investment and difficult to finance through other government/charitable funders.	Too small to give information on behavioral impacts of interventions or fully explore effects on cancer subtypes. Requires recurrent sampling from participants. Can introduce bias by selecting centers able to collect and process samples. Risk of overestimating or underestimating benefits of an intervention if an unreliable surrogate is used.
Followed by	Further confirmatory trials or introduction into clinical practice	Further targeted and confirmatory studies

Defining the aims of PPSs

The goal of PPSs is to study preventive interventions that are matched to biological vulnerabilities in specific high cancer risk settings. In this respect, they can precede but do not supersede population prevention studies. PPSs are designed to investigate the biological impact of an intervention, define its toxicity and efficacy, and discover or validate biomarkers to take forward. Given their smaller size and use of surrogate endpoints, PPSs can potentially speed the clinical development of cancer preventives.

PPSs can be conducted in primary, secondary, or tertiary settings, provided the intervention being evaluated is biologically targeted and measurable (defined in Fig. 1). At a conceptual level, they draw on the principles of precision oncology, in which drugs are designed to target specific biological features of established cancers, such as inhibiting proteins produced by mutated driver genes (11, 12). One example is the tyrosine kinase inhibitor osimertinib which specifically inhibits mutant EGFR protein and has now replaced chemotherapy alone in the primary treatment of patients with EGFR-mutated non–small cell lung cancer (NSCLC; ref. 13). These inhibitors were discovered by conducting compound library screens in EGFR-mutant cell lines (14), an approach that cannot easily be replicated for identifying preventive therapies as there are very few precancer cell lines currently available.

Figure 1.

Alignment of the molecular and cellular changes occurring during cancer development and opportunities for intervention with preventive therapies during an individual’s life course at both the cellular (top) and clinical (bottom) levels. (Created in BioRender. Brown, K. (2025) https://BioRender.com/l09tjrs.)

Identifying precancers

Precancers represent a tantalizing research conundrum as, unless directly visualized (e.g., colonic polyps via colonoscopy or cervical intraepithelial neoplasia via colposcopy), the majority are too small to be seen using current radiologic imaging methods such as CT or MRI and have no proven circulating biomarkers to assist in their detection. For cancer types lacking effective screening strategies, their precancer equivalents are generally discovered incidentally and retrospectively by microscopic pathologic examination of surgically resected tissues. Examples of precancers include prostatic intraepithelial neoplasia, cervical intraepithelial neoplasia, and ductal carcinoma in situ, which are believed or shown to precede invasive prostate, cervical, and breast cancers, respectively (15–17). Due to the difficulty of clinically identifying precancers, their tumorigenesis trajectory remains poorly understood. For example, until recently, the commonest form of ovarian cancer—high-grade serous cancer—was believed to arise from the ovarian surface epithelium. Detailed histopathologic studies revealed, however, that the precursor lesions called serous tubal intraepithelial carcinomas in fact arise in the adjacent fallopian tube and predate invasive ovarian cancer by approximately 6 years (18).

Insights into the evolution of precancer from an initial founder mutation within a single cell to an established multicellular cancer have been provided through various cancer-specific genetic mapping studies, for example in NSCLC (19). This allows estimates of the duration of precancer development, which is often years or even decades before becoming invasive disease. Precancers seem to be distinct biological entities when compared with their corresponding invasive counterparts and do not share all the Hanahan and Weinberg hallmarks of cancer (20). Multiple investigators, including from the NCI’s PreCancer Atlas collaboration, have identified common precancer hallmarks across a variety of tissue types (21–23). For example, they have shown that precancers display age-related genetic alterations, epigenetic changes resulting from aberrant methylation at specific sites, metabolic alterations, hijacking of regenerative cell states, disruption of immune surveillance, and remodeling of the tissue microenvironment mediated by senescent fibroblasts. Importantly most precancers spontaneously regress or fail to progress to invasive cancer. Hence, a major challenge in the field is to identify the point of reversibility of a precancer and the optimal timepoint for an intervention once beyond it. Just as an understanding of cancer biology has enabled cancer chemotherapy to be superseded by targeted therapies, insights into the biology of precancers could pave the way for the discovery of rationally designed targeted preventives.

The challenge of targeting precancers

Currently, preventive interventions are often repurposed naturally derived drugs. An example is aspirin (acetylsalicylate) which, in the CAPP2 study, was shown to reduce the incidence of colorectal cancer in individuals with Lynch syndrome (24). Aspirin was originally extracted as salicin from willow bark and later meadowsweet flowers and subsequently was synthetically acetylated to improve tolerability. It was first identified as having cancer-protective effects in the 1980s (25, 26). Aspirin has multiple mechanisms of action. Not only is it anti-inflammatory via the inhibition of COX enzymes, but it also binds and activates the enzyme AMP-activated protein kinase, inhibits mTOR, induces autophagy, and alters cellular metabolism (27). Which of these aspects, if any, contribute to cancer prevention is under debate. Other repurposed agents with cancer-preventive actions that were similarly developed empirically include the naturally occurring polyphenol resveratrol, the antidiabetic drug metformin, and anticholesterol statins (28–30). An understanding of the molecular basis for their preventive effects will require careful biological deconvolution so that future, more targeted versions can be synthesized.

Newer preventive agents are now being rationally designed to target precancer hallmarks, such as therapies to activate immune cells surrounding precancers (31). Proteomic technologies such as immunopeptidomics have led to the identification of peptides called neoantigens presented by HLA molecules on the surface of cells after carcinogen exposure, some remaining present on precancers and throughout the transition to invasive disease (31). These aberrant neoantigens are caused by genetic alterations within the cell, such as hotspot mutations, frameshifts, or altered mRNA splicing. Cancer-preventive vaccines are being developed to stimulate T cells to recognize and neutralize precancer cells expressing these neoantigens (32). In addition, targeted therapies are in development against cell signaling pathways that are aberrantly active in precancers such as WNT/β-catenin inhibitors to target colorectal polyps and prevent bowel cancer (33).

Importantly, there is considerable heterogeneity between precancers; for example, colonic polyps can be broadly divided into adenomas and sessile types, each with different underlying molecular drivers (34). Molecular characteristics may also differ across the precancer trajectory; for example, there are at least three pancreatic intraepithelial neoplastic stages with different transcriptomic features in each (35). For these reasons, it is unclear whether a single targeted agent or rational combinations are required to prevent cancer in certain disease contexts. It is currently challenging to design a PPS without these insights.

Using surrogate markers

Given the length of time needed for cancer incidence and survival to be determined in traditional cancer prevention trials, surrogate endpoints are commonly used to determine the potential efficacy of an intervention in PPSs. For trials in very high–risk populations, cancer incidence can be used as the primary surrogate endpoint, as a proportion of participants will already be harboring precancers. This includes people with heritable predisposition syndromes, such as Li–Fraumeni syndrome (linked to TP53 pathogenic variants), breast and ovarian cancer syndromes (linked to BRCA1/2 variants), or Lynch syndrome (variants in mismatch repair genes such as MLH1, MSH2, MSH6, and PMS2), in which lifetime cancer risk can be as high as 90% (36–38). Even in these contexts, relatively large sample sizes are needed to demonstrate decreases in cancer incidence. For example, if a two-arm PPS were designed in a high-risk population with an estimated 20% cancer risk over 5 years, 500 participants would need to be recruited to have an 89% chance of finding a significant (at two-sided α of 0.05) difference between arms if the intervention reduced cancer incidence by 50% (from 20% to 10%). For a smaller 30% reduction (from 20% to 14%), the sample size required to maintain the same power would be 1,600. This creates recruitment challenges for studies in high-risk but niche populations. Study size can be reduced by using adaptive and model-based methodologies, such as Bayesian frameworks that incorporate prior information (such as rates of polyp regression and progression) for control arm estimates, early stopping rules for efficacy or futility, and the incorporation of longitudinal biomarker changes and other surrogate endpoints. Other methods to reduce accrual sizes include refining inclusion criteria to capture those at the highest risk of cancer within these high-risk settings, for example by narrowing age inclusion to around known age-incidence peaks, by recruiting only carriers with higher penetrance variants and participants who have already had cancer, or by incorporating polygenic risk scores to further refine risk stratification. Cancer risk can also peak in individuals without heritable risk, such as those exposed to cancer predisposing environmental factors such as long-term smoking, pollutants, or carcinogens (e.g., asbestos). In this case too, the studies need to be positioned around known peaks in incidence as the estimated cancer risk over the duration of the study will inform recruitment.

If the precancer is detectable and easy to biopsy, lesion regression can be a suitable surrogate PPS endpoint. Examples include mammographic breast density as a surrogate marker of ductal carcinoma in situ and visual regression of actinic keratosis (a precursor to squamous cell cancer) or oral leucoplakia (a precursor to oral cancer; refs. 39–41). However, using lesion regression as a surrogate endpoint has limitations, for example, if sampling of the lesion causes its regression or it naturally fluctuates over time, e.g., breast density diminishes with age. In the colorectal context, many small or nonadvanced colonic adenomas never become malignant, whereas truly precancerous forms—such as sessile serrated lesions—can be flat, subtle, and hard to detect at colonoscopy, leading to difficulties in relying on polyp burden quantitation as a surrogate marker. To some extent, this can be countered by including a robust (no treatment) control arm, using artificial intelligence–enhanced colonoscopy to measure polyps with greater consistency, stratifying polyps by size and type, and powering the study for both precancer progression and regression. Overall, although precancer regression can be used as a surrogate endpoint for PPSs, it is not generally reliable enough to replace cancer incidence as an endpoint for regulatory approval.

Unlike cancer studies in which tumor tissue can be measured radiographically and biopsied directly or by using “liquid biopsy” biomarkers such as ctDNA, the field of cancer prevention—in which tumors are not yet present and precancer lesions are hard to biopsy—should be reliant on surrogate markers, but these have yet to be discovered and validated. This challenge also faced the field of cardiovascular medicine before the identification of the circulating biomarker “low density lipoprotein cholesterol” as a surrogate marker of atherosclerosis. Low density lipoprotein cholesterol transformed cardiovascular prevention by giving a 20- to 30-year lead time warning of future coronary heart disease and stroke (42, 43). The lack of a reliable precancer biomarker reflects a limited understanding of precancer biology which is morphologically and molecularly distinct from that of invasive cancer. However, ongoing research to identify such markers—e.g., using cell-free DNA or circulating RNA, miRNA, or peptides—is underway. For such markers to be reliable, they would ideally accurately and quantitatively reflect the transition to invasive disease and play a role in tumorigenesis. This distinguishes them from one-off risk prediction tools (like germline genetic tests) or early cancer detection (multi-cancer early detection) tests which often lack precancer specificity and sufficient sensitivity.

In addition, imaging methods are being developed to identify precancer characteristics, such as their altered metabolism, that may also be useful surrogate markers. However, to ensure that the long-term impact of an intervention is considered within a PPS, especially at phases IIb and III, cancer incidence should be included as a principal endpoint. In addition, measurement of overall survival should be considered, either as part of the main PPS or as a separate observational extension.

The possibility of natural precancer regression necessitates the use of a randomized control arm within a PPS. This can challenge recruitment, particularly if participants are expected to maintain long-term adherence to the agent/placebo (e.g., for more than 5 years), attend for multiple visits, and/or undergo tests and complete questionnaires that impose logistic, financial, or time commitments. An “open-label” or nonblinded control arm is an alternative but carries the risk of disproportionate study discontinuation amongst those in the control arm, especially if participants are not fully engaged in the overarching goal of the study or if it is not embedded within a standard-of-care pathway. In the context of rare genetic cancer predisposition syndromes, in which accrual is already challenging, an unappealing study design can prove disastrous for recruitment, and hence it is useful to engage participant representatives early to mitigate this alongside considering participant-friendly measures such as decentralized designs, guaranteeing access to active drug on study completion, and providing compensation for travel expenses.

Optimal timing of intervention

PPSs should be designed to enroll participants in the years before the peak in their cancer risk so that they are most likely to benefit from the intervention. However, timing the intervention also requires an understanding of the precancer trajectory which can commence up to 30 years before invasive cancer. Not only are there multiple types of precancers, but the process of carcinogenesis is comprised of sequential stages, each dominated by different molecular processes (Fig. 1). For example, (i) an initiating genetic mutation or epigenetic event is followed by (ii) promotion or the clonal expansion of abnormal cells alongside proinflammatory signaling and then (iii) the acquisition of additional mutations and evasion of immune surveillance before (iv) undergoing malignant transformation. It is important to time the delivery of a targeted intervention before the acquisition of additional molecular abnormalities eclipses the original target. As an example, targeting BCR-ABL during the chronic phase in chronic myelogenous leukemia (which can be considered a premalignant state) with tyrosine kinase inhibitors such as imatinib is highly effective for extended periods of time, but the same agent given during the accelerated phase or in blast crisis (the latter being progression to acute leukemia and marked by acquisition of multiple new mutations) shows decreasing effectiveness (44). The challenge, of course, is to identify the point along the carcinogenesis pathway when an intervention is optimal in an at-risk individual. In epithelial cancers, progressive immune evasion is one of the hallmarks of precancer and cancer, limiting an effective antitumor response. It is possible that interventions, such as cancer-specific vaccines, may be more effective if administered during early carcinogenesis when immunosurveillance mechanisms are intact (45).

Embedding biological endpoints

Routine serial biological sampling of participants in a PPS provides a unique longitudinal dataset to explore the impact of an intervention (comparing with baseline samples and those from participants in the control arm of the study), identify or validate biomarkers of precancer, and explore the molecular trajectory of tumorigenesis. Unlike many studies in which translational research is absent or included as a nonessential endpoint, biological studies should be included as endpoints in PPSs, using validated assays where possible. In addition to biospecimens from the at-risk epithelium (e.g., premalignant lesions, where appropriate), samples can include serial collection of whole blood, which can be fractionated to generate plasma, peripheral blood mononuclear cells, cell-free DNA, etc. Agreement to collect and analyze tumor and adjacent histologically normal tissue from those participants who develop cancers during the study is also useful to understand escape mechanisms (if in the active arm) or provide molecular comparators (if in the control arm). Other trial-relevant biospecimens, such saliva or fecal samples for microbiome analysis, should also be collected, if possible. Ideally, leftover samples should be stored and made available to other researchers with oversight for the biobank provided by an impartial sample access committee.

Defining a dose and schedule

PPSs can be used to test the preventive impact of new therapies, those that have an established role in treating advanced cancer or are repurposed from treating an unrelated condition. However, it remains uncertain whether the dosing schedules used for a drug’s primary indication are the same when the drug is repurposed for prevention. It is possible that lower, higher, or intermittent dosing is more suitable for targeting a precancer than the continuous dosing designed to combat chronic disease, especially if the preventive mechanism of action differs from that of the agent’s primary therapeutic indication (46). In the absence of a response biomarker or visible precancer regression, currently the main driver of preventive dosing decision-making is toxicity. Often, there are reservations about acceptable toxicity profiles, the assumption being that a drug can only be acceptable as a cancer-preventive agent if it has no serious toxicities and high tolerability. Although legitimate for a low-risk general population, higher-risk individuals may be more willing to accept an agent with marginally higher toxicity burden (47). Methods to reduce risk include administration of agents topically or regionally (e.g., inhalational vs. oral or intramuscular delivery) over a shorter treatment course or by intermittent dosing.

For therapies under development in which a safe, well-tolerated efficacious dose has not previously been established, PPSs can and should incorporate a dose-finding element to identify the lowest effective dose to minimize toxicity in healthy populations. An example is the COLO-PREVENT polyp prevention trial platform (ISRCTN13526628), which involves high-risk individuals recruited through the English and Welsh NHS Bowel Cancer Screening Program in which the diet-derived compound resveratrol is being tested at two doses compared with the placebo following preclinical studies (48).

Defining “effective” is challenging in the early phases of drug development if preventive studies have not yet been conducted. In this case, a surrogate efficacy marker is necessary. For example, if conducting dose-finding studies of a preventive vaccine, a desirable dose is one with the least toxicity but the strongest immunogenicity (e.g., measured using ELISpot assays), assuming that stronger T-cell activation will later correspond to more robust cancer-preventive activity.

Redefining participants and embedding participant/public involvement

High-risk individuals, whether from genetic predisposition, the presence of precancer, or other (e.g., environmental) exposures, may not have had first-hand experience of cancer and therefore are not considered “patients” in the traditional sense. The term “cancer” has upsetting connotations as the individuals may have witnessed family members experiencing treatment and dying from the disease. If possible, studies should be designed to minimize or avoid the requirement to attend cancer clinics for procedures or follow-up. For studies in those with genetic cancer predisposition syndromes, cancer incidence peaks at lower age groups than those typically enrolled in cancer trials, and thus study information and recruitment materials should be tailored toward these younger participants, e.g., through age-appropriate social media content. Many are in school or employment and may lack the time and financial resources to seek out—and repeatedly visit—trial clinics. Trials conducted during the COVID-19 pandemic have shown that many visits can be conducted remotely, and this option should be maximized in PPSs.

Input from people with relevant lived experience is critical for a successful PPS, such as individuals with the same specific genetic risk and who are within a similar age range as future study participants. Advisors can be sought from people participanting in cancer screening programs or undergoing surveillance for premalignant lesions. This engagement is not only essential for designing the PPS but also in considering the risks versus benefits of an intervention, publicizing the trial to encourage enrollment, and disseminating the findings amongst specific high-risk communities.

Achieving a broad geographic reach

Recruiting people at high genetic cancer risk into PPSs can be challenging, as their rarity means the numbers of known affected individuals will be low. As well as working with charities and support agencies to publicize studies, it is important that study sites are colocated in population hotspots, or at national referral centers. For example, a study targeting BRCA variant carriers should include trial centers serving Jewish communities, as BRCA mutations are more prevalent among individuals of Ashkenazi Jewish descent (49). This approach is also relevant for identifying communities at high cancer risk due to environmental exposures; for example, a PPS in long-term smokers should enroll from regions and/or countries in which tobacco use is most prevalent (50).

Health literacy, access to the Internet, and trial participation can be challenging in underserved communities and require close working with community leaders, local pharmacies, and primary care physicians to enhance enrollment. Preplanning to ensure that information sheets are translated to languages compatible with the likely study population is essential. It is also important to collect participant demographics so that efforts to widen catchment can be made. A major challenge, however, is that an early-phase PPS requires substantial infrastructure (e.g., ability to obtain and process invasive biopsies and investigational pharmacies to dispense unapproved interventions) that limit their reach outside of specialized clinical research centers. In this context, as before, decentralization of studies can help enrollment, particularly of younger, ethnically diverse, and/or economically challenged participants. This can include direct-to-participant enrollment and provision of consent materials, learning from the United Kingdom’s COVID-19 PRINCIPLE trial, in which potentially eligible participants with active COVID-19 infection were identified from their NHS data and contacted directly by the trials team and medication was posted directly to their homes (51, 52). These and other efforts should be made to maximize access and prevent the longer-term magnification of cancer inequalities.

Ensuring real-world utility

To facilitate future adoption into clinical practice it is important, where possible, to design a PPS pragmatically around existing standard-of-care services. This has the advantage of minimizing additional burden on clinical staff working within stretched healthcare systems, enabling participants to remain in their existing healthcare pathways during the PPS, and paves the way for future adoption of that intervention into those pathways. However, it incurs risk if surveillance imaging is not conducted or reported within required PPS timeframes at individual centers.

Conducting studies within studies

PPSs can be conducted as standalone studies or as substudies within a single overarching protocol. A natural partnership is with early cancer detection, e.g., multi-cancer early detection studies in which at-risk participants can be identified and guided toward a specific PPS. Platform PPSs could investigate multiple interventions in specific high-risk groups, adopting a similar rationale as the UK Lung Matrix trial or the US Lung Cancer Master Protocol trial in which patients with advanced NSCLC were allocated targeted treatments according to their tumor genetic profiles (53, 54).

A basket trial design is particularly well-suited for genetic cancer syndromes such as Lynch syndrome or BRCA-related cancers, in which multiple preventive strategies may be integrated within a single, adaptive protocol. Alternatively, a feedforward trial sequence could be used, in which findings from an initial study inform the design and execution of subsequent trials, with biospecimens and data collected prospectively to support iterative hypothesis refinement. Another viable approach is a seamless phase II/III trial design, wherein preliminary efficacy is evaluated in a defined subset of participants using validated surrogate biomarkers during the phase II component, while treatment continues uninterrupted across the entire cohort. The definitive phase III endpoint—typically cancer incidence—is then assessed at a later stage in all participants. For example, a chemoprevention trial for oral cancer might assess short-term surrogate endpoints, such as resolution of oral leukoplakia in a subset of participants during phase II, while the entire cohort continues therapy to allow for evaluation of cancer incidence in phase III. Key challenges associated with such trial designs include their inherent complexity, the need for substantial upfront financial investment (particularly to secure phase III funding), and the difficulty of establishing acceptable toxicity thresholds in the absence of robust efficacy data from early-phase cohorts. This design is not appropriate in scenarios in which a phase II trial is conducted exclusively in a high-risk genetic population and the phase III cohort comprises a broader, lower-risk population selected through biomarker stratification or risk prediction models.

Attracting commercial partnerships and appealing to regulatory bodies

Whereas traditional population prevention studies were often unattractive to commercial investors, the shorter timelines and strong translational components of PPSs make them potentially more appealing. With innovations in early diagnostics, including new approaches such as vaccines, the future cancer market is significant, both in protecting those with genetic risk factors of cancer and protecting the population from age- or lifestyle-associated (e.g., obesity and smoking) sporadic cancers. Academia has led the way thus far, but a hybrid funding mechanism could be designed whereby cancer surrogate/incidence outcomes are collected early (e.g., at 2 years) with support from a pharmaceutical partner and longer-term outcomes are delivered through academic funding.

Early engagement with regulators is also vital to chart the path to regulatory approval. Similarly, ethics boards and institutional review boards need to gain familiarity with PPSs, particularly when approving experimental interventions in high-risk populations, as these may be somewhat different from healthy volunteer studies. In very high–risk and rare conditions, evidence from a PPS could be used to provide a regulatory approval for an intervention in that indication, particularly as these communities have the greatest need for rapid approvals. It is important to include relevant health economic or quality of life questionnaires into the PPS if this is planned.

Developing future talent

PPSs lack a clear medical specialization. They fall outside the remit of general medicine but, by targeting a cancer precursor state, are not yet considered oncology studies either. Although geneticists can conduct PPSs in genetically high-risk individuals, they have limited access to those at high risk of sporadic cancers. Equally, the requirement to biologically match individuals to interventions, assess safety/tolerability, and conduct longitudinal sample collection places PPSs outside the ownership of public health. In most centers with experience of conducting PPSs, studies tend to be conducted by clinical pharmacologists, oncologists, or subspecialists who care for high-risk individuals (e.g., gastroenterologists, pulmonologists, gynecologists, etc.) with clinical trial experience in primary as well as secondary care teams and draw on multiple other disciplines, including trialists, geneticists, basic biologists, epidemiologists, translational scientists, radiologists, oncologists, surgeons, ethicists, biostatisticians, patient and public groups, and regulatory bodies. Involvement of clinical teams within specific cancer pathways (particularly those involved in screening) and who care for individuals to be targeted for cancer prevention is critical for trial success and the subsequent incorporation of successful strategies into routine practice.

It is therefore essential that people from these backgrounds are all included in prevention conferences and that the subject is integrated into teaching modules as part of undergraduate and postgraduate courses to increase knowledge and generate interest in prevention as a whole and PPS specifically. The onus is on the current generation of researchers in this field to provide education, support, and mentorship to future generations of trialists interested in conducting PPSs. Networks can help achieve this, such as the NCI’s Cancer Prevention Clinical Trials Network, covering multiple institutions.

Conclusions/Perspective

Precision prevention presages an era of cancer research in which tumorigenesis biology and precancer-targeted drug discovery will be interwoven, fields of research that are currently in their infancy. PPSs complement traditional population-based research but are rooted in an enhanced understanding of cancer risk, tumorigenesis biology, and biologically matched interventions. They can evaluate interventions in participants at high cancer risk who do not yet have a cancer diagnosis and, through regular biological sampling, include concomitant exploration of carcinogenesis or risk intervariation and its impact on response.

Although PPSs offer a potentially more time-efficient way to evaluate cancer preventive agents, they require a coordinated and multidisciplinary approach. This makes them inherently more complex to set up, and hence platform or master protocol study designs offer an efficient use of resources. Whereas academic institutions have access to multiple professions and networks, resource limitations can be alleviated by partnering with commercial companies possibly using a hybrid model with early and late endpoints collected separately. Learning from ongoing and future studies will help address the many questions that remain unanswered, including the point along the precancer trajectory that interventions should be given, the minimum duration and intensity of interventions to achieve clinical impact, identifying noninvasive methods (e.g., liquid biopsies) as surrogate markers of precancer, and defining acceptable risk/benefit in high cancer risk versus more general populations. The establishment of national and international networks will be essential to share learning and motivate and train future generations of PPS trialists. Ultimately, PPSs will not only benefit those at the highest risk of cancer but could unravel the molecular process of cancer formation and identify interception opportunities that can be disseminated to the general populations.

Authors’ Disclosures

S.P. Blagden reports other support from Prenostics and personal fees from Oxford Drug Design, Theolytics, UCB, and Oxford Hormone Clinic outside the submitted work; in addition, S.P. Blagden has a patent for WO2021064361A1 issued to RNA Guardian/Prenostics and a patent for WO2025093876 pending to RNA Guardian/Prenostics and institutional support by GlaxoSmithKline toward precancer biology research. Funding by Cancer Research UK for LungVax clinical trial. K. Brown reports grants from the Leicester Experimental Cancer Medicine Center, the National Institute for Health and Care Research Leicester Biomedical Research Center, and Cancer Research UK during the conduct of the study, as well as personal fees from the Scientific Advisory Board of the European Institute of Oncology, Milan, Italy, and the World Cancer Research Fund International Regular Grant Program Panel outside the submitted work. K. Brown reports being a member of the Scientific Advisory Board for the Cancer Research UK City of London Centre from August 2021, a member of the American Association for Cancer Research Cancer Prevention Working Group Steering Committee from 2023 to 2026, and a member of the Award Nominating Subcommittee from 2024. No disclosures were reported by the other authors.