Chapter 26 — Further Reading

A note before the list. This chapter's subject matter moves faster than any other in the book, and some of it concerns treatments that people are making decisions about right now. Two consequences.

First, prefer primary sources and prefer recent ones. A review from three years ago is a historical document in this field. Where a source below is described generically rather than named, that is deliberate — the category is durable, the specific paper is not, and you should find the current version rather than the one I happened to know about.

Second, be careful with search results. Search terms like "personalized cancer vaccine" return press releases, patient-recruitment sites, and a certain amount of outright predation alongside real science. Sources selling something are not sources.

Tier 1 — Start here

Your national cancer institute's public information pages on cancer vaccines and immunotherapy. Government cancer institutes maintain patient-facing and clinician-facing explanations that are carefully worded, regularly revised, and explicit about what is approved versus investigational. They are the correct first stop for anyone reading this chapter because of a diagnosis rather than curiosity. Look specifically for the distinction they draw between preventive and treatment vaccines.

A current review of neoantigen-based cancer vaccines in a major immunology or oncology review journal. Search for reviews published within the last two years covering neoantigen prediction, vaccine platforms, and clinical trial status. Read the limitations section first — good reviews in this field are unusually candid about prediction accuracy and about how much remains unproven.

Public clinical trial registries. Search for individualized or personalized neoantigen vaccine trials and read the actual registry entries: eligibility criteria, arms, primary endpoint, estimated completion. Ten minutes with a registry entry will teach you more about what is actually known than any amount of press coverage. Note in particular how often the primary endpoint is recurrence-based rather than survival, and what the comparator arm receives.

A standard immunology textbook chapter on antigen processing and presentation. Any current undergraduate or graduate immunology text will cover class I and class II pathways, HLA polymorphism, and costimulation far more thoroughly than §26.2 and §26.3 could. This is the single best investment if you want the mechanism to become intuitive rather than memorized.

Tier 2 — Go deeper

The 2017 first-in-human neoantigen vaccine reports. Two studies published in the same year in a leading journal established feasibility in melanoma — one using synthetic long peptides with a TLR-agonist adjuvant, one using an RNA-based construct, each in fewer than twenty patients. Read them as design documents. Notice what the authors claim and, more instructively, what they carefully do not claim.

The primary publications of the randomized combination trials in melanoma. When you read these, read the confidence intervals and the follow-up duration before the headline estimate, and check whether the analysis you are reading is the pre-specified primary one.

The pancreatic cancer single-arm neoantigen vaccine report. Valuable specifically as an exercise in recognizing a responder-versus-non-responder comparison and understanding why it cannot establish benefit. The authors are clear about this; secondary coverage frequently was not.

Reports of the major therapeutic cancer vaccine failures. The large randomized trials of a cancer-testis antigen in lung cancer and melanoma, a mucin-derived antigen in lung cancer, a telomerase-derived peptide in pancreatic cancer, and a mutant growth factor receptor peptide in glioblastoma are all published and all worth reading. Read at least one in full. The experience of reading a well-conducted trial of a well-motivated therapy that did not work is educational in a way that no summary reproduces.

Literature on synthetic long peptides versus minimal epitopes. The argument that minimal epitopes can be tolerizing, and that long peptides requiring professional processing are preferable, is one of the field's genuine methodological corrections and is well documented.

Reviews of adjuvants for T-cell-directed vaccines. Look for material covering TLR agonists, saponin-containing combinations, and lymph node targeting. The preclinical work on depot-mediated T-cell sequestration is worth finding in the original.

The clinical literature on peptide-based allergen immunotherapy. The phase 2 and phase 3 reports from the cat allergy program, read in sequence, are an unusually clean illustration of promising mid-stage results failing to replicate — and of how large placebo responses in allergy trials complicate everything.

Tier 3 — For the committed

Structural immunology of peptide-MHC complexes. The crystallographic literature on peptide binding grooves, anchor residues, and T-cell receptor engagement is where §26.2's 🧬 The Molecule callout comes from. If Chapter 1's material on side chains appealed to you, this is the payoff.

MHC restriction: the original work. The discovery that T cells recognize antigen only in the context of self MHC — recognized with a Nobel Prize in Physiology or Medicine in 1996 — is worth reading in the original or in a good historical treatment. It reframes what "recognition" means.

Epitope prediction methodology. The literature on binding prediction algorithms, their training data, and their validation is technical but important, particularly the work assessing performance across HLA alleles with sparse reference data. This is where the equity concerns in §26.3 stop being abstract.

Tumor immunoediting, antigen loss, and HLA loss of heterozygosity. The literature on how tumors escape antigen-specific immunity explains why a successful vaccine can still fail, and why clonal target selection matters.

Tumor heterogeneity and clonal evolution. Multi-region sequencing studies established that a single biopsy may not represent a tumor, which is a foundational constraint on step 1 of the pipeline.

Health economics and access literature on individualized therapies. The autologous cellular immunotherapy approved in prostate cancer is a well-documented precedent for what individualized manufacturing costs and what that does to uptake. If neoantigen vaccines succeed clinically, this becomes the next question immediately.

If you only do one thing

Open a clinical trial registry, find one currently recruiting individualized neoantigen cancer vaccine trial, and read its entry from top to bottom.

Not a paper. Not a review. A registry entry.

You will see the eligibility criteria — including the tissue requirements that make step 1 of the pipeline a real filter on who can participate. You will see what both arms receive, which will make the attribution problem concrete rather than theoretical. You will see the primary endpoint, and whether it is recurrence or survival. You will see the estimated completion date, which will tell you how long it will be before anyone knows.

Twenty minutes with that document will inoculate you against every headline about this field for the next several years, and it is the single most direct way to understand what 🔬 means: not a guess, not a hedge, but a specific set of questions that specific people are currently running specific experiments to answer.