Loading field map
For decades, a central question has driven tumor immunology: can the immune system recognize and eliminate cancer, and if so, why does it so often fail? The answers have come not as a single discovery but as a sequence of competing and overlapping frameworks, each reframing the relationship between host immune cells and developing tumors. From the early notion of immune surveillance to the integrated cancer-immunity cycle, these frameworks have shaped both basic research and clinical treatment.
In the 1950s, Sir Macfarlane Burnet and Lewis Thomas proposed the concept of immune surveillance: that the immune system constantly patrols for nascent tumor cells and eliminates them before they form clinical cancers. This framework drew on observations of transplant rejection and the newly recognized role of T cells. It predicted that immunodeficient individuals should have higher cancer rates. However, experimental tests in the 1970s and 1980s—such as the use of nude mice lacking T cells—showed that while viral cancers increased, spontaneous non-viral tumors did not. The surveillance hypothesis could not explain why established tumors grew despite an intact immune system. This crisis set the stage for a more nuanced view.
In 2001, Robert Schreiber and colleagues proposed a new framework: cancer immunoediting. Rather than a simple on/off surveillance, they argued that the immune system both protects the host and sculpts the tumor. The process unfolds in three phases: elimination (the original surveillance concept), equilibrium (a period where immune pressure holds tumor growth in check but selects for resistant variants), and escape (where tumors evade immune destruction). Immunoediting immediately resolved the experimental paradox: nude mice did not show increased spontaneous tumors because the immune system had already, in evolutionary time, shaped the incidence; but in humans, immunoediting could operate over years. This framework also explained why tumors that form in immunocompetent hosts are often more aggressive—they have been edited. Immunoediting remains a cornerstone, but it did not replace earlier concepts entirely; rather, it absorbed them into a dynamic process.
While immunoediting described natural history, two parallel frameworks arose from therapeutic and mechanistic angles. The immune checkpoint paradigm, pioneered by James Allison and Tasuku Honjo in the mid-1990s, focused on molecular brakes—negative regulatory receptors like CTLA-4 and PD-1 that limit T cell activation. Blocking these checkpoints could unleash anti-tumor immune responses, leading to durable clinical responses in melanoma, lung cancer, and other malignancies. This framework was primarily defined by a specific intervention: checkpoint blockade.
Simultaneously, a separate but complementary framework emerged: the tumor microenvironment and immune suppression. This approach emphasized the cellular and chemical context within tumors—the presence of regulatory T cells, myeloid-derived suppressor cells, immunosuppressive cytokines (e.g., IL-10, TGF-β), and metabolic barriers like hypoxia. Unlike checkpoint blockade, which targets a single receptor-ligand interaction, the microenvironment framework views suppression as a multifaceted, multicellular network. These two frameworks often coexist: checkpoint therapy can be frustrated by a suppressive microenvironment, leading to combinatorial approaches. Where the checkpoint paradigm assumes that restoring T cell activation is sufficient, the microenvironment framework argues that the local milieu must also be normalized. This tension remains a central debate.
The early 2000s also saw efforts to quantify immune responses in tumors. In 2006, Jérôme Galon introduced the immunoscore, a standardized method to assess the density and location of T cells (CD3+ and CD8+) in the tumor center and invasive margin. Unlike previous staging criteria that focused solely on tumor characteristics, the immunoscore treated the local immune infiltrate as a prognostic tool. It is a static, diagnostic framework—it tells you the state of the immune response at a single time point, not the dynamic evolution.
Around 2010, the neoantigen paradigm shifted focus to the molecular targets of immune attack. Advances in sequencing revealed that tumors accumulate hundreds of mutations, some of which generate novel peptides (neoantigens) that can be presented on MHC molecules and recognized by T cells. This framework explained why checkpoint therapy works best in tumors with high mutation burden and opened the door to personalized cancer vaccines. It interacts intimately with immunoediting: neoantigens are the substrate for T cell recognition in the elimination phase, and tumors often lose neoantigen expression during escape. It also intersects with the checkpoint paradigm because neoantigen-specific T cells are frequently the ones unleashed by checkpoint blockade.
By 2013, the field had accumulated disparate insights but lacked a unifying conceptual model. Daniel Chen and Ira Mellman proposed the cancer-immunity cycle, a seven-step process that links the generation of tumor antigens to the destruction of cancer cells. The steps include: antigen release, antigen presentation by dendritic cells, T cell priming and activation, T cell trafficking to tumors, infiltration into the tumor, recognition of cancer cells, and killing. The cycle framework did not replace earlier ideas; instead, it organized them into a coherent sequence. For example, checkpoint blockade operates at step three (T cell activation) and step seven (effector function), while the tumor microenvironment modulates steps five through seven. Immunoscore quantifies the outcome of steps four and five. The cancer-immunity cycle has become a dominant teaching tool and design template for combination therapies.
Today, the leading frameworks are no longer isolated; they are used together in research and clinical practice. There is broad agreement that immunity can eliminate tumors (elimination phase, immunoediting), that checkpoints limit responses, and that the tumor microenvironment is suppressive. However, deep disagreements persist. One major debate concerns the dominant mechanism of immune suppression: is it primarily driven by checkpoint receptors, by cellular suppressors like Tregs and MDSCs, or by metabolic constraints? Another debate centers on the relative importance of neoantigens versus shared tumor-associated antigens; the neoantigen paradigm holds that specific mutations are key, but some tumors with low mutation burden still respond to immunotherapy, hinting that other targets exist. A third tension is between the static snapshot of the immunoscore and the dynamic cycle: prognostic tools need predictive power, but the cycle emphasizes temporal sequence and feedback loops. These debates are productive, driving the next generation of integrated models that may eventually merge the strengths of all seven frameworks.
The field today is characterized by pluralism: no single framework has emerged as complete. The cancer-immunity cycle provides a helpful synthesis, but its linearity may obscure the reality of intertwined suppression and activation. The neoantigen paradigm is powerful but may be too narrow. The checkpoint paradigm transformed therapy but left many patients resistant. The future likely lies in frameworks that can capture the dynamic, personalized interplay between tumor genetics, immune editing, microenvironment, and treatment.