A surgeon treats an early esophageal cancer with photodynamic therapy. Two days after the patient receives the drug, a fiber-optic delivers red light through an endoscope, the illuminated tissue dies, and the lesion clears. A colleague proposes the same approach for a bulky tumor several centimeters beneath the skin. That plan fails before it begins — and not because the photosensitizer is worse or the cancer is more aggressive. The problem is light.

Red light at six-thirty nanometers penetrates tissue only a few millimeters before being scattered and absorbed to uselessness. And even where light does reach, photodynamic therapy has a second hidden dependency — it needs molecular oxygen, and the hypoxic core of a bulky tumor may have too little. Two physical constraints. Every approved indication, every clinical success, every endoscopic engineering decision is an answer to the same two facts.

Photodynamic therapy kills cells with a strict three-component chain. The patient receives a photosensitizer — a drug that absorbs light at a specific wavelength — which accumulates preferentially in tumor tissue. Light of the matching wavelength is applied. The excited photosensitizer transfers its energy to molecular oxygen, generating singlet oxygen and other reactive oxygen species that damage membranes, proteins, and DNA. The photosensitizer alone is inert in the dark. Light alone does nothing without the drug. Both are useless without oxygen — because singlet oxygen is made from molecular oxygen, not released by the drug and not produced by the light.

Because the cytotoxic agent is singlet oxygen derived from molecular oxygen, PDT efficacy depends on local oxygen concentration. Tumors are frequently hypoxic in their interiors — their vasculature is chaotic, perfusion is uneven, and rapidly proliferating cells consume oxygen faster than disorganized vessels can supply. PDT worsens this as it runs. Intense illumination consumes oxygen faster than perfusion can replenish it, and the illuminated region can become locally depleted mid-treatment. A protocol that works well at the start may become less effective as it continues, because the oxygen supply present at the beginning has been used up.

Photothermal therapy uses nanoparticles that absorb light and convert the energy to heat, thermally ablating tumor cells. Gold-based nanoparticles — nanorods, nanoshells, nanocages — are the most studied because their plasmon resonance can be tuned to absorb in the near-infrared. The mechanism is simpler than PDT in one important respect: PTT does not require oxygen. Heat kills regardless of oxygen tension. This removes PDT’s hypoxia problem — a real advantage for oxygen-depleted tumor cores. But PTT does not remove the penetration problem. Near-infrared light penetrates somewhat deeper than visible red, but the difference is millimeters, not orders of magnitude.

The clinical record of photothermal therapy is instructive. AuroLase, the gold nanoshell product from Nanospectra, reached clinical trials for prostate cancer. It did not achieve broad regulatory approval. The mechanism works in cell culture and animal models. Translation to patients has been harder, for reasons that appear to involve both delivery and light penetration — the same combination that limits most nanoparticle cancer therapies. PTT has produced real clinical results in accessible settings, but the claim that it solves the depth problem of PDT is not supported by the available evidence.

Photodynamic therapy is a bounded tool, not a broken one. Within its envelope it is selective, repeatable, and carries none of the cumulative marrow or organ toxicity of chemotherapy. Superficial skin cancers are ideal — the tumor is millimeters thick, the light source is placed directly on the skin, oxygen is abundant. PDT here is standard care and genuinely preferred over surgery for field-treatment of widespread lesions. Early esophageal and endobronchial cancers are ideal because the light is delivered endoscopically to the luminal surface. Bladder cancer treated by instillation, with light delivered cystoscopically, follows the same logic. The device engineering in every case is entirely about access.

Photoimmunotherapy extends selectivity further. Cetuximab — an antibody that targets EGFR on many head and neck cancers — is conjugated to the near-infrared dye IR700. The conjugate, approved in Japan, binds cancer cells through the antibody and is activated by NIR light. Even within the illuminated zone, only cells expressing EGFR carry the activated photosensitizer. Fluorescence-guided surgery uses the same principle diagnostically. Patients with glioblastoma drink five-ALA before surgery. Tumor cells convert it to fluorescent protoporphyrin IX, which makes tumor tissue glow pink under blue-violet illumination — the surgeon sees the margin in real time. The imaging use succeeds where curative PDT of glioblastoma would fail: it requires only enough fluorescence to see the surface of the resection cavity.

Here is the chapter’s central claim. Light-activated cancer therapies — both photodynamic and photothermal — are bounded by tissue optics and oxygen availability. That is why they remain treatments for accessible lesions rather than general cancer therapies. The finding that would force revision: a deliverable technology that routinely overcame the penetration ceiling for deep solid tumors without invasive light placement. An efficiently X-ray-activated photosensitizer, or an upconverting nanoparticle system producing durable responses in centimeters-deep lesions in controlled trials. Or photothermal therapy crossing from limited clinical success to multiple broad regulatory approvals with demonstrated survival benefit. Neither has happened. The physics has not changed.