Core Framework & Diagram Future Directions in Tumor Immunology
July 14, 20261 Min Read How Does the Immune System Recognize Cancer Cells?
July 14, 2026We're standing at the entrance of the most exciting decade in tumor immunology history — what comes next
—— From bispecific antibodies to AI-driven neoantigen discovery, from cancer vaccines to 'living drugs': a roadmap for the next decade.
I. Bispecific antibodies: 'forced handshake' — pulling T cells directly next to cancer cells
Bispecific Antibodies (BsAbs) are one of tumor immune engineering's cleverest designs: a single antibody molecule carrying two different binding arms — one end grabs tumor cell surface antigen, the other grabs T cell surface CD3 (TCR complex signal transduction subunit). This literally 'forces T cells next to cancer cells,' bypassing the complex steps normally needed for activation (DC presentation, lymph node activation, antigen-specific TCR binding), forcibly triggering T cell killing programs.
This mechanism solves the fundamental problem of traditional immune checkpoint inhibitors in 'cold tumors': PD-1 inhibitors depend on 'already-present but suppressed T cells existing in the tumor.' If the tumor has essentially no T cell infiltration, releasing the brake doesn't help. Bispecific antibodies bypass this limitation: they don't depend on tumor-specific T cells already entering the tumor but rather, through the 'CD3 binding arm,' recruit any T cells in circulating blood, forcibly guiding them to the tumor site, activating killing in situ.
Approved representatives: Blinatumomab (targeting CD19×CD3) was the first FDA-approved bispecific antibody, achieving approximately thirty to forty percent complete remission rate in relapsed/refractory ALL. In 2023, mosunetuzumab (targeting CD20×CD3) was approved for relapsed/refractory follicular lymphoma, with complete remission rate approximately sixty percent. Multiple bispecific antibodies for solid tumors (like targeting HER2, CEA, EGFR) are in clinical trials, with CD3-mediated T cell recruitment showing preliminary activity in some tumor types.
2. Next-generation CAR-T: the difficult expansion from 'blood cancer specialist' to 'solid tumors'
CAR-T cell therapy in hematological malignancies has achieved historic success — in some relapsed/refractory patients, achieving long-term remission and even functional cure previously thought impossible. But solid tumors remain territory where CAR-T is almost powerless. Solid tumor CAR-T challenges involve three dimensions: target absence (solid tumors lack 'exclusive antigens' as highly specific as CD19, most targets also expressed in normal tissues); physical barriers (dense TME stroma blocking CAR-T cells from physically entering tumor cores); and T cell exhaustion (in TME's chronic antigen exposure and immunosuppressive environment, they quickly exhaust).
Three breakthrough directions: neoantigen-targeting CAR-T (through gene sequencing identifying patient-specific tumor neoantigens, designing CAR-T recognizing that neoantigen-MHC I complex — the highest specificity strategy); 'Armored CAR-T' (additionally transducing genes secreting IL-12, IL-15 and other cytokines into CAR-T, enabling CAR-T to 'self-produce ammunition' after entering TME, maintaining functional durability); and universal 'off-the-shelf' CAR-T (Allogeneic CAR-T, using healthy donor T cells, CRISPR editing deleting TCR and HLA, producing instantly usable 'ready-made' CAR-T — shortening preparation time, reducing costs, especially important for patients with rapidly progressing disease).
3. AI revolution in tumor immunology: from 'data ocean' to 'precise direction'
Artificial intelligence is fundamentally changing every research aspect of tumor immunology. Neoantigen prediction revolution: a patient's tumor genome sequencing generates hundreds to thousands of somatic mutations, of which only a few can produce neoantigens effectively recognized by the immune system. AlphaFold's protein folding prediction (2021), and subsequent models like RoseTTAfold and ESMfold, elevated protein structure prediction precision to experimental level — making neoantigen-MHC binding three-dimensional structure prediction possible, greatly improving neoantigen screening accuracy and efficiency. In personalized tumor vaccine design workflow, AI compresses neoantigen candidate screening time from traditional weeks to hours.
De novo design of TCR and antibodies: traditional antibody and TCR discovery depends on lengthy 'library screening' processes. AI (especially protein language models and generative AI) is fundamentally changing this paradigm: starting from a given target's structure, directly generating TCR/antibody sequences predicted to have high affinity and specificity, compressing discovery cycles from months to days. Immunotherapy response prediction: in 2024, AI pathology analysis based on whole slide images (WSI) was proven able to directly predict PD-1 inhibitor response from HE-stained tumor sections, with precision exceeding most traditional molecular markers.
4. Cancer prevention vaccines: 'preventing it before cancer happens'
All current cancer vaccines (including personalized neoantigen vaccines) are 'therapeutic vaccines' — helping the immune system recognize and combat cancer after it has already occurred. But one more exciting direction is 'preventive cancer vaccines' — before cancer occurs, letting the immune system pre-learn to recognize early carcinogenic features, immediately eliminating them when cancerous signals appear.
KRAS G12D prevention vaccine: KRAS is one of the most common oncogene mutations in human cancer, with KRAS G12D appearing frequently in pancreatic cancer (approximately forty-five percent), colorectal cancer (approximately twelve percent), and lung adenocarcinoma (approximately four percent). This mutation produces a neoantigen peptide 'exclusive' to cancer cells — normal cells don't carry this sequence. The US National Cancer Institute (NCI) has launched a preventive KRAS G12D vaccine Phase I trial targeting high-risk populations (those with KRAS G12D mutation precancerous lesions, like intraductal papillary mucinous neoplasm IPMN), aiming to pre-activate CTLs against KRAS G12D before cancer transformation completes, placing them in 'immune alert' status to immediately clear cells with this mutation once they begin abnormally proliferating.
Lynch syndrome preventive vaccine: the ongoing MACE clinical trial evaluates a preventive vaccine against 'shared' Lynch syndrome-related neoantigens, with hopes of significantly reducing colorectal cancer incidence in this extremely high-risk population (lifetime colorectal cancer risk up to seventy to eighty percent).
5. Toward 'functional cure': reasonable expectations for 2035
'Functional cure' — not necessarily having absolutely no cancer cells in the body, but: after the patient stops treatment, the immune system, relying on its own surveillance and memory, continuously controls residual cancer cells at a 'symptom-free equilibrium' level; patients can live, work, and reproduce normally without substantial cancer-related restrictions. This is a reasonable expression of the 'big goal' that tumor immunology has the best chance of achieving before 2035.
Which cancers are most likely to achieve functional cure before 2035? High TMB tumors (melanoma, highly MSI colorectal cancer, some lung cancers) + early diagnosis + personalized neoantigen vaccines + PD-1 combination: some patients have already achieved over five years of relapse-free remission; as follow-up time extends, the proportion achieving 'functional cure' is expected to be further confirmed. Relapsed/refractory hematological malignancies (B-ALL, DLBCL, MM) + CAR-T or bispecific antibodies: some patients have achieved ten-plus years of durable remission.
Which challenges remain hardest? Low TMB solid tumors (pancreatic cancer, prostate cancer, glioblastoma) — lacking sufficient neoantigens and with extremely immunosuppressive microenvironments, even combining multiple strategies current response rates remain extremely low. This may require 'creating immunogenicity from scratch' approaches — oncolytic viruses, STING agonists, or direct gene editing to introduce artificial neoantigens in tumors.
6. A final page for every reader: understanding is the prerequisite for action
This is the final article of the entire 'Tumor Immunology Foundations' section (Articles 66 to 80). In these fifteen articles, we traveled from 'what is tumor immunology' (Article 66), through 'how the immune system finds cancer cells' (67), 'how cancer cells evade' (68), 'what the tumor microenvironment is' (69), 'tumor antigens' (70), 'immune surveillance and immunoediting' (71–72), 'why cancer relapses and metastasizes' (73–74), 'how inflammation promotes cancer' (75–76), 'how tumors manipulate immunity' (77–79), to 'future directions' (80).
This knowledge isn't to make you a tumor immunologist. It's to help you understand one thing: your immune system is your most important anti-cancer force, and its health is significantly influenced by your daily choices. Regular exercise, adequate sleep, managing chronic inflammation, reducing chronic stress, not smoking, maintaining healthy body weight, standardized screening — these aren't generic 'health advice' but specific actions proven again and again throughout these fifteen articles, with mechanisms and evidence, to maintain immune surveillance efficiency. Understanding why these things matter is what converts 'knowing' into 'genuinely doing.' That's what these fifteen articles ultimately want to convey.
Frequently Asked Questions