Worth a Pound of Cure? Emerging Strategies and Challenges in Cancer Immunoprevention

Cancer is the second leading cause of mortality worldwide, accounting for nearly 10 million deaths in 2020 alone (1). The global burden of cancer is rapidly growing with projections estimating 27.5 million new cases and 16.3 million deaths per year by 2040 (2). In light of these projections, there is a critical need to develop prevention strategies that can reduce cancer-associated mortality, morbidity, and financial costs. Immunoprevention—modulation of the immune system to prevent and intercept cancer—has emerged as a promising approach to achieve this goal in recent years (3–6). The renewed Cancer Moonshot initiative highlights immunoprevention as a key priority, emphasizing the identification of novel targets and early interventions such as vaccines for high-risk individuals (7).

Cancer prevention is a multitiered concept that includes primary, secondary, and tertiary interventions (8, 9). Primary prevention refers to risk-reducing measures implemented prior to the onset of carcinogenesis. Such interventions include lifestyle modifications to decrease exposure to known carcinogenic factors (e.g., smoking cessation, sun safety), as well as vaccination against virally mediated cancers (10–13). Secondary prevention refers to risk-reducing measures implemented after the initial onset of carcinogenesis. Such interventions include screening, early detection, and early treatment (e.g., removal of precancerous lesions) that aim to prevent progression to advanced disease (14–16). Cancer interception is an emerging form of secondary prevention that involves strategies targeting the progression of precancerous lesions to invasive carcinoma after the first carcinogenic insult has occurred, including vaccination against nonviral cancers (Fig. 1; refs. 4, 17, 18). Tertiary prevention refers to risk-reducing measures implemented after an established cancer diagnosis. These interventions seek to improve quality of life and long-term survivorship by minimizing the risk of recurrent disease (e.g., adjuvant endocrine therapy in breast cancer) and secondary malignancies (19, 20).

Successful development of prophylactic vaccines against two oncogenic viruses—hepatitis B virus (HBV) and human papillomavirus (HPV)—has established immunoprevention as a feasible strategy for viral-mediated cancers (6, 13, 21–32). Since the first HBV vaccine became commercially available in 1981, universal vaccination programs have dramatically reduced chronic HBV infections and the resulting incidence of hepatocellular carcinoma worldwide (33–36). More recently, the advent of multivalent vaccines against HPV is predicted to significantly decrease the global incidence of cervical and other HPV-related cancers (13, 37–39). In addition, public health measures to screen for and eradicate Heliobacter pylori infection have led to a decrease in rates of gastric cancer (40). However, given that <20% of the global cancer burden is infectious in etiology, there is an urgent need to develop novel approaches for the prevention of cancers that are not driven by microbial pathogens (4, 6, 41).

In this article, we provide a comprehensive overview of emerging strategies in cancer immunoprevention. Our focus lies on reviewing molecular targets and novel immune-based approaches for early interception (i.e., primary and secondary prevention) of nonviral cancers (Table 1). Furthermore, we discuss translational challenges and future avenues of investigation in the field of cancer immunoprevention.

Mutant Kirsten rat sarcoma virus (KRAS) is an oncogenic driver found in approximately 20% of all cancers. Notably, ∼90% of pancreatic ductal adenocarcinomas (PDAC) as well as the majority of associated precancerous lesions, including pancreatic intraepithelial neoplasia (PanIN) and intraductal papillary mucinous neoplasms (IPMN), express mutant KRAS (42, 43). Although previously deemed an “undruggable” target, recent advances have led to the development of mutant KRAS-specific therapies, such as the KRAS G12C inhibitors sotorasib and adagrasib (44–47). Pan-KRAS inhibitors are also under clinical development and have shown initial anticancer activity (48, 49). Furthermore, in the realm of immunotherapy, adoptive T-cell therapy targeting the KRAS G12D driver mutation has been reported to achieve durable tumor regression in metastatic PDAC (50).

Mutant KRAS is an ideal interception target in PDAC given its widespread prevalence in precursor lesions and the large window of opportunity available for intervention in the precancerous setting. Indeed, clonal lineage data suggest that the development of established PDAC from the first driver event takes approximately 11 years (51). In this context, there is growing interest to advance immune-based interception strategies for patients who are at high risk of developing PDAC due to an underlying genetic predisposition and/or presence of high-risk pancreatic lesion. As an estimate, a fully effective vaccine targeting KRAS G12D alone—a driver mutation found in ∼40% of all PDAC tumors—could prevent more than 30,000 deaths per year in the United States (4, 52).

In a proof-of-concept study, Keenan and colleagues demonstrated that a Listeria-based vaccine targeting KRAS G12D combined with regulatory T cell (Treg)-depleting agents can slow the progression of early PanINs to overt PDAC in Kras G12D/+; Trp53 R172H/+; P48-Cre (KPC) mice and improve survival (53). Informed by this rationale, an ongoing pilot trial is studying a pooled synthetic long peptide (SLP) mutant KRAS vaccine in individuals who are at high risk of developing PDAC based on family history and/or germline mutational status (NCT05013216). This vaccine consists of poly-ICLC adjuvant and 21-mer SLPs targeting six KRAS mutation subtypes (G12D, G12V, G12C, G12A, G12R, G13D) that are commonly found in PDAC. The co-primary endpoints of this trial are safety and immunogenicity, as measured by the change in mutant KRAS-specific T-cell density using serial IFNγ enzyme-linked immunosorbent spot (ELISpot) assays.

In addition to PDAC, immunoprevention strategies targeting mutant KRAS may be applied to other malignancies, including colorectal cancer and lung cancer. For example, Pan and colleagues studied the use of a multi-peptide mutant KRAS vaccine in a doxycycline-inducible KRAS G12D murine model of lung adenocarcinoma (54). Compared with their non-vaccinated counterparts, mice who were vaccinated prior to doxycycline induction demonstrated significant reductions in both the number and volume of lung surface tumors. Given the compelling preclinical rationale, further studies are needed to develop and optimize clinical-grade strategies for the interception of KRAS-driven tumorigenesis.

Mucin-1 (MUC1) is a heterodimeric transmembrane glycoprotein that is aberrantly expressed in multiple solid tumor malignancies, including breast, ovarian, pancreatic, lung, and colorectal cancers (55, 56). In 2009, MUC1 was ranked as the second most promising target for the development of a cancer vaccine based on a tumor antigen prioritization project initiated by the NCI (57). Several approaches to develop a MUC1 vaccine have been attempted since that time, including the use of dendritic cell (DC), viral vector, DNA-based, and peptide-based platforms. To date, the majority of studies testing clinical-grade MUC1 vaccines have occurred in the therapeutic setting for patients with established cancers (58–67).

As an immunopreventive strategy, MUC1 vaccines have been studied in individuals at high risk of developing colorectal cancer. In a single-arm open-label phase I/II trial, a peptide-based MUC1 vaccine was studied in 39 evaluable patients with advanced colorectal adenomas (NCT00773097). The vaccine tested in this trial consists of a 100-mer MUC1 peptide admixed with poly-ICLC adjuvant. The primary endpoint of immunogenicity was measured by the anti-MUC1 IgG response. A twofold or greater increase in anti-MUC1 IgG at week 12 compared with pre-vaccination baseline was considered a positive immune response and observed in 17/39 (43.6%) patients. Interestingly, flow cytometry analysis of baseline peripheral blood samples revealed a significantly higher proportion of myeloid-derived suppressor cells (MDSC) among the 22/39 (56.4%) patients classified as immune nonresponders compared with responders (68).

The same peptide-based MUC1 vaccine with poly-ICLC adjuvant was also studied in a randomized, double-blind, placebo-controlled phase II trial to evaluate efficacy outcomes in patients with advanced colorectal adenomas (NCT02134925). Following booster vaccination, a twofold or greater increase in anti-MUC1 IgG at week 55 compared with pre-vaccination baseline was considered a positive immune response and observed in 11/51 (21%) of MUC1 vaccine recipients compared with 0/44 (0%) of placebo recipients. Despite vaccine immunogenicity, there was no statistically significant difference noted in the adenoma recurrence rate between the placebo (66.0%) and MUC1 (56.3%) groups in an intention-to-treat analysis. Of note, among the subgroup of immune responders, there was a 38% reduction in absolute risk of adenoma recurrence noted in 3/11 (27.3%) patients compared with placebo that approached statistical significance (P = 0.08; ref. 69).

MUC1 vaccines are also being studied as an immunopreventive strategy in patients at high risk of developing lung cancer. In an ongoing pilot study, a peptide-based MUC1 vaccine with poly-ICLC adjuvant is being tested in current and former smokers with a 30 pack-year or greater smoking history (NCT03300817). The expected sample size of this trial is 40 evaluable subjects with co-primary endpoints of vaccine safety and immunogenicity. Similar to prior studies, immunogenicity is defined as a twofold or greater increase in anti-MUC1 IgG at week 12 compared with pre-vaccination baseline. Secondary outcome analyses are planned to evaluate differences in immunogenicity between current versus former smokers, pre-vaccination baseline levels of circulating MDSCs and inflammation markers (hsCRP, IL16), and the potential association between chronic obstructive pulmonary disease status and immunogenicity.

HER2/neu overexpression is a driver event found in approximately 25% to 30% of invasive breast cancers (70). HER2/neu overexpression has also been identified in the precancerous lesions that precede invasive breast cancer, including ductal carcinoma in situ (DCIS). Although the prognostic significance of HER2/neu positivity in DCIS is evolving, initial reports have suggested an association with increased aggressiveness and higher rates of local disease recurrence (71, 72). HER2/neu-directed vaccines for breast cancer have been developed using multiple approaches, including DC, tumor cell, carbohydrate antigen, DNA-based, and peptide-based platforms. These vaccines have been tested in clinical trials spanning the entire spectrum of breast cancer—from interception of DCIS to the treatment of metastatic disease (73–79).

One of the earliest vaccines for cancer interception was developed using an antigen-pulsed type 1 polarized dendritic cell (DC1) platform and tested as a neoadjuvant strategy in a pilot trial of patients undergoing surgical resection of HER2/neu-expressing DCIS (NCT00107211). A total of 27 evaluable patients underwent weekly intranodal vaccination for four weeks prior to surgery. The primary endpoints were safety and feasibility, while secondary endpoints included immune response and clinical response. The feasibility rate was 100%, defined as the percentage of patients who completed all four vaccine doses on schedule. The immune response rate was 88%, defined as a twofold or greater postvaccination increase in HER2/neu-specific IFNγ-secreting CD4⁺ T cells measured by ELISpot (80). There were 5/27 (18.5%) patients who achieved a complete response with no residual DCIS at the time of surgery. Furthermore, 13/22 (59.1%) patients exhibited a >20% decrease in HER2/neu-expressing cells after vaccination (81). In a subsequent trial, the DC1-based vaccine was administered by intralesional, intranodal, or both intralesional/intranodal injections to 54 patients with HER2/neu-expressing DCIS or early invasive breast cancer (NCT02061332). The results of this study demonstrated that there was no statistically significant difference in immune or clinical responses based on the route of vaccine delivery (82).

Another immunopreventive strategy tested in patients with HER2/neu-expressing DCIS is the nelipepimut-S (NPS) vaccine plus GM-CSF. This vaccine consists of a 9-mer peptide from the extracellular domain of the HER2 receptor that serves as an immunodominant CTL epitope with high binding affinity for HLA-A2 and HLA-A3 molecules (83). In a randomized two-arm phase II study, patients received two doses of either NPS + GM-CSF (n = 9) or GM-CSF alone (n = 4) at 2-week intervals prior to surgical resection. The primary endpoint of immune response was determined by the percentage change in NPS-specific CD8⁺ CTLs following vaccination compared with baseline, as measured by a flow cytometry-based dextramer assay. There was an 11-fold increase in NPS-specific CD8⁺ CTLs noted at 1 month post-surgery in the NPS + GM-CSF arm compared with a 2.25-fold increase in the GM-CSF only arm (84). In addition to NPS, H2NVAC is a multi-peptide vaccine targeting four HER2/neu-derived epitopes that is currently being studied in patients with HER2/neu-expressing DCIS in an ongoing phase IB trial (NCT04144023).

A number of emerging immunopreventive strategies across cancer types seek to simultaneously target multiple antigens instead of a single-antigen approach. For example, the PVX-410 vaccine has been studied with or without lenalidomide in patients with smoldering multiple myeloma (SMM) as a potential interception strategy to halt progression to active multiple myeloma. The PVX-410 vaccine consists of four 9-mer peptides derived from three multiple myeloma–associated antigens: X-box binding protein 1 (XBP1), syndecan-1 (CD138), and CS1. In an initial phase I/IIA trial, encouraging safety and immunogenicity data were observed with a 95% immune response rate among all patients immunized with the PVX-410 vaccine (NCT01718899). Modest clinical responses were observed in the PVX-410 vaccine plus lenalidomide cohort (n = 12) with 5 patients achieving clinical response, 6 patients achieving stable disease, and 1 patient progressing on study (85). A subsequent phase IB trial is designed to study the PVX-410 vaccine and histone deacetylase inhibitor citarinostat with or without lenalidomide in patients with SMM (NCT02886065).

Other immunopreventive strategies based on targeting multiple antigens are being studied in high-risk subjects with inherited cancer susceptibility syndromes. In an ongoing phase IB trial, INO-5401 with electroporation is being investigated with or without IL12 DNA plasmid in germline BRCA1 and BRCA2 mutation carriers (NCT04367675). INO-5401 is a DNA-based vaccine that targets Wilms tumor gene-1 (WT1), prostate-specific membrane antigen (PSMA), and human telomerase reverse transcriptase (hTERT; ref. 86). Furthermore, in a planned phase IIB trial, the efficacy of the Tri-Ad5 vaccine combined with IL15 superagonist N-803 will be tested in individuals with Lynch syndrome (NCT05419011). Tri-Ad5 is a novel adenovirus-based vaccine that targets MUC1, carcinoembryonic antigen (CEA), and brachyury as a set of diverse antigens expressed in a broad range of human carcinomas (87).

Neoantigen-specific vaccines are another compelling approach that has been studied in the precancerous setting. For instance, the safety and feasibility of a preventive neoantigen-loaded DC vaccine was established in a phase I/II trial of 20 patients with Lynch syndrome (NCT01885702). In this study, 15 of 20 vaccinated subjects demonstrated functional neoantigen-specific or CEA-specific T-cell responses in cultures of skin test biopsies (88). Moreover, the Nous-209 vaccine is currently being studied in another phase I/II trial of 45 patients with Lynch syndrome (NCT05078866). Nous-209 is an adenovirus-based vaccine that consists of 209 shared frameshift neoantigens enriched in microsatellite instability-high tumors (89–91). In addition to hereditary cancer syndromes, an ongoing phase I trial is investigating the safety and feasibility of developing personalized neoantigen vaccines as an interception strategy in patients with SMM to prevent progression to active multiple myeloma (NCT03631043).

Although vaccine-based strategies have prevailed in cancer immunoprevention, recent interest has grown in the use of immune checkpoint inhibitors and other non-vaccine approaches in the precancerous setting. In a randomized, placebo-controlled phase II trial (n = 81), pembrolizumab is being studied in patients found to have high-risk indeterminate pulmonary nodules without established lung cancer (NCT03634241). Similarly, an ongoing phase II trial (n = 42) is testing the efficacy of nivolumab to reverse endobronchial dysplasia in high-risk current or former smokers (NCT03347838). In addition to programmed cell death protein 1 (PD-1) blockade, the anti-IL1β antibody canakinumab is being tested in a phase II trial for patients with high-risk indeterminate pulmonary nodules (NCT04789681). Finally, outside of lung cancer, pembrolizumab was studied as an interception strategy in a pilot trial of 13 evaluable patients with intermediate/high-risk SMM (NCT02603887). 11 patients maintained stable disease, 1 patient progressed to active multiple myeloma, and 1 patient achieved a stringent complete response in this study. Interestingly, multiparametric flow cytometry of CD8⁺ T cells obtained from the baseline bone marrow biopsy of the exceptional responder demonstrated high PD-1 expression and low levels of exhaustion markers (TIM-3, LAG-3, BTLA) compared with nonresponders (92). Overall, the benefits of using immune checkpoint blockade in the prevention setting need to be weighed carefully against the risks of causing immune-related adverse events in healthy, disease-free populations.

Unique immunologic requirements to develop a successful immunoprevention vaccines should be considered. First, in contrast to prevention vaccines for virally driven cancers which primarily evoke an antibody response to neutralize viral components, vaccines targeting non-virally driven cancers must evoke a robust and durable, high-quality cytotoxic T-cell response particularly when targeting intracellular precancerous antigens. Second, vaccines must also be able to generate memory T cells to maintain durability of antitumor response as well as T-cell diversity to target the evolving genetic landscape. Thirdly, vaccines may also need to safely co-deliver other immunomodulatory signals to fully halt the progression from premalignancy to malignancy as we will discuss below. Finally, timing of interception must also be considered. Growing evidence supports cancer immunoediting as the earliest premalignant lesion attracts high quality T cells in the “elimination phase” prior to T-cell tolerization and exhaustion in the equilibrium phase (93). Without agents to reeducate these T cells, tumor escape occurs as increasing number of immunosuppressive signals allow for tumor evasion. Vaccination particularly during the elimination phase to reeducate and reinvigorate T cells provides an opportunity to tip the balance in favor of elimination. To generate an effective vaccine for cancer interception, we will therefore consider the optimal antigen target, vaccine delivery platform that allows for versatility to co-deliver antigen with immunomodulatory signals, as well as optimal timing and refinement of patient selection for cancer vaccine interception as discussed below.

The selection of the optimal target antigen is critical for the successful design of a preventive cancer vaccine. Several groups have attempted to target tumor-associated antigens (TAA), which are unmutated proteins that are highly expressed by cancer cells and found at low levels in nonmalignant cells (94, 95). TAAs can be subdivided into several categories including oncofetal, differentiation, and overexpressed antigens. Oncofetal antigens are found in cancer cells as well as nonmalignant cells in embryonic/fetal tissue (e.g., CEA, AFP, 5T4; ref. 96). Differentiation antigens are found in nonmalignant cells derived from the same lineage as the primary tumor (e.g., CD20, EpCAM, gp100). Overexpressed antigens are normal molecules that are upregulated in cancer cells (e.g., HER2/neu, PSA, EGFR). Although they are widely represented in many cancers, TAAs are self-molecules that may elicit limited T-cell responses due to central and peripheral tolerance mechanisms (97, 98). Moreover, given the presence of TAAs in nonmalignant tissues, the risks associated with serious on-target, off-tumor toxicities must be carefully considered—for example, as previously reported with HER2-specific chimeric antigen receptor T-cell therapies (99).

Compared with TAAs, tumor-specific antigens (TSA) are ideal targets for a prophylactic vaccine due to enhanced immunogenicity and decreased risk of off-target adverse events. TSAs are mutant proteins uniquely expressed by cancer cells that can be further subdivided into shared versus personalized neoantigens. Shared or “public” neoantigens encoding either driver oncogenes (e.g., KRAS, BRAF, PIK3CA) or tumor suppressor genes (TP53) common to subsets of multiple patients are particularly attractive given that they are more likely to result from recurrent hotspot mutations and are ubiquitously expressed (100, 101). Personalized or “private” neoantigens are TSAs specific to an individual patient and result from passenger mutations that are less likely to be recurrent (102, 103). Recent advances have been made in the development of computational pipelines to predict patient-specific neoantigens for cancer immunotherapy (104, 105). Although a personalized strategy is likely not feasible in the precancerous setting, implementation of an “off-the-shelf” paradigm using shared neoantigen vaccines is an attractive approach for the prevention of non-virally driven cancers. However, prior to widespread clinical use, the high costs and complexity related to manufacturing and delivery of neoantigen-specific vaccines will need to be thoughtfully addressed.

The selection of the ideal vaccine platform for cancer immunoprevention remains an open question that requires further study. Multiple delivery platforms have been attempted in the development of preventive cancer vaccines, but high-quality data from head-to-head trials are limited. Cancer vaccine platforms can be classified into five broad categories: cell-based, viral vectors, peptide, DNA, and mRNA (106, 107). The advantages and disadvantages of different vaccine platforms are summarized in Table 2.

With the advent of mRNA vaccines during the COVID-19 pandemic, there is renewed interest in optimizing cancer vaccine delivery using mRNA-based platforms due to their rapid scalability and superior versatility (108). Although prophylactic vaccines against infectious pathogens leverage B cell–mediated humoral immunity, an effective vaccine for cancer interception must induce a potent and diverse repertoire of CD8⁺ CTLs (109, 110). Indeed, CD8⁺ CTL-mediated adaptive immunity is essential to target the evolving antigen landscape of precancerous lesions as they undergo immunoediting during malignant transformation (18, 111, 112). Furthermore, CD4⁺ T cells have recently emerged as key players in orchestrating antitumor immunity through secretion of effector cytokines, support of CD8⁺ CTL priming, and direct cytotoxicity (113, 114). Given the unique ability to target multiple antigens for stimulation of both CD4⁺ and CD8⁺ T-cell responses, the development of mRNA-based vaccines for cancer immunoprevention is a promising area of ongoing investigation (115).

Another key consideration in the delivery of preventive cancer vaccines is the addition of co-stimulatory molecules to overcome immunosuppressive signaling and optimize efficacy. Due to their versatility, mRNA-based vaccine platforms may allow for co-encoding of immunomodulatory signals. While the precancerous microenvironment (PCE) exhibits less immune evasion, preclinical evidence suggests that coadministration of immunomodulatory agents with cancer vaccines enhances antitumor efficacy (116–121). Specific approaches to consider in the precancerous setting include co-encoding of cytokines/chemokines (e.g., GM-CSF, IL12, IL15) and immunostimulatory molecules (e.g., Toll-like receptor agonists, CD40L, CD70, CD137, CCR7, OX40). Moreover, combinatorial strategies of mRNA vaccines with immune checkpoint blockade are being studied in the adjuvant setting to prevent post-resection recurrences of established cancers (NCT03359239, NCT04117087, NCT03897881). Of note, preliminary findings from the recent phase IIB KEYNOTE-942 trial indicate improved relapse-free survival with the use of a mRNA-based neoantigen vaccine plus pembrolizumab compared with pembrolizumab alone in resected stage III/IV melanoma patients (n = 157; ref. 122).

The selection of the appropriate patient population is another major factor influencing the development of a successful immunopreventive strategy. Historically, clinical trials in this realm have included patients deemed at an increased risk of malignancy based on family history, germline mutation status, the presence of precancerous lesions, and/or risk-enhancing lifestyle factors (e.g., tobacco exposure). Although such elements play an important role in determining inclusion eligibility, patient selection in immunoprevention trials could be refined further by incorporating predictive biomarkers of immune response. At present, there are no validated immune-based biomarkers in the precancerous setting. Compared with established cancer, the discovery of novel biomarkers in premalignancy has been limited by the absence of readily available tissue samples. Ongoing efforts are employing state-of-the-art techniques such as single-cell RNA/T-cell receptor (TCR) sequencing and mass cytometry (cytometry by time of flight) to identify candidate parameters of T-cell response to immunotherapy in peripheral blood (123–125). In the future, next-generation immune profiling may lead to the validation of predictive biomarkers that can guide accrual of the patients who are most likely to experience clinical benefit from prophylactic cancer vaccines and other immunopreventive approaches.

Finally, given the wide window of opportunity for intercepting premalignancy, determining the optimal timing of immune-based interventions in high-risk patients is a crucial consideration. Notably, the PCE is thought to be less immunosuppressive and more receptive to T cell–mediated cytotoxicity than the tumor microenvironment. Prior studies have shown that infiltrating T-cell immunophenotypes become more suppressive as pancreatic precursors progress towards invasive PDAC (126, 127). Moreover, Hernandez and colleagues conducted multiplex immunofluorescence analysis of a cohort of IPMN tissue samples and noted that high-grade lesions exhibited diminished immune surveillance compared with low-grade lesions (128). Similar phenomena are noted in other solid tumor malignancies, including the progression of DCIS to invasive breast cancer as well as cervical intraepithelial neoplasia to invasive cervical cancer (129, 130). In light of these findings, we posit that vaccine-based interception should be implemented during the earliest stages of premalignancy in high-risk individuals (Fig. 2). This approach is expected to maximize antitumor efficacy by circumventing mechanisms of immune evasion that arise later during the development of advanced precancerous lesions and invasive carcinomas.

Given advances in early detection technologies, the identification of high-risk patients with early precancerous lesions for inclusion in immunoprevention trials is readily feasible (15, 131). While the scientific rationale to intervene at the earliest stages of pre-carcinogenesis is compelling, clinical regulatory standards do not favor upfront testing of first-in-human cancer vaccines in healthy, disease-free subjects. One potential strategy to address this issue is to demonstrate the initial safety of an experimental vaccine as an adjuvant treatment in patients with resected cancer prior to use in the precancerous setting. For instance, the preliminary safety of a first-in-human mutant KRAS vaccine was established in a pilot study of patients with resected PDAC (NCT04117087) before vaccinating high-risk individuals with pancreatic cysts in a subsequent phase I trial (NCT05013216; ref. 132).

Despite significant therapeutic advances, the burdens of cancer-associated mortality, morbidity, and financial toxicity continue to multiply across the globe. To address these burdens, current efforts in cancer immunoprevention seek to use vaccines and other immunomodulatory agents to intercept premalignancy at its earliest stages. These efforts leverage advances in the understanding of premalignant biology and cancer immunoediting, highlighting the transition from immune surveillance to immune escape during malignant transformation. Future avenues of investigation related to preventive cancer vaccines include optimizing antigen selection and delivery platforms, identifying predictive biomarkers of immune response, and establishing the ideal window of opportunity for cancer interception. If successful, the development of clinical-grade preventive vaccines for non-virally driven cancers will have paradigm-shifting public health implications for millions worldwide.

E. Vilar reports grants and personal fees from Janssen Research and Development; personal fees from Recursion Pharma, Guardant Health; and personal fees from Tornado/Cambrian outside the submitted work. A. Maitra reports other support from Cosmos Wisdom Biotechnology, Freenome, Tezcat Biotechnology; and other support from Thrive Earlier Detection, an Exact Sciences Company, outside the submitted work. N. Zaidi reports personal fees from Genentech; other support from Adventris; and nonfinancial support from Bristol-Myers Squibb outside the submitted work; in addition, N. Zaidi has a patent for Neoantigen Vaccines for Cancer Prevention pending. No disclosures were reported by the other author.