The Pill That Could · What Cure Means

What Cure Means, part 4

What this means is that even apparent functional cure, in CML, may not be the same as biological cure in the strictest sense. The cancer cell that started everything — the original Philadelphia-chromosome-positive stem cell from which all the other leukemic cells descended — may still be there, dormant, in some fraction of patients who appear cured. The patient is, in clinical terms, indistinguishable from someone who has never had CML. They feel well; their blood is normal; they don't take a drug; they live a normal life. But the disease, at the deepest level of molecular biology, may be in something like a permanent stalemate rather than a complete eradication.

Whether this distinction matters depends on how you weight different ways of being healthy. If a patient lives a normal life span, with normal blood counts, off all therapy, on the basis of a stalemate in which their immune system is suppressing a small population of leukemic cells they don't know about, are they cured? In the practical sense that matters to them, almost certainly yes. In the strict biological sense that someone like Sawyers might use, perhaps not quite. The semantic gap between these two answers is, in the era of imatinib and its successors, the gap that contains a great many patients.

* * *

One thing the long-term data have made clear is that imatinib does not, in general, fail in the way the early skeptics had feared.

Recall the worry from chapter five: a class of drugs that took years to develop, the skeptics had said, and that might be quickly outpaced by resistance, was not the kind of bet a publicly-traded pharmaceutical company should make. The fear was that imatinib would work briefly, the cancer would evolve around it, and the drug would be obsolete within a few years of approval.

What actually happened was different. Resistance did emerge in some patients, especially those who started in advanced disease, and chapter nine described how the field responded with second-generation inhibitors, then ponatinib, then asciminib. But for the bulk of patients — the patients who started in chronic phase, the patients who were the great majority of CML diagnoses going forward — imatinib has continued to work, in many cases for the lifetime of the patient on the drug, with no need to switch to anything else. The IRIS cohort patients who responded to imatinib in 2001 are, twenty-five years later, mostly the same patients on the same drug. Their disease has not evolved out from under them. The reservoir from which resistance might have emerged has, in their case, remained suppressed.

Why imatinib has worked so durably in chronic-phase CML, when the early models of how cancer evolves under selection pressure suggested it might not, is a question with several partial answers. The number of cells generating new mutations is smaller in chronic phase than in advanced disease. The selective advantage that any single resistance mutation confers is partially offset by the fitness cost of the mutation itself — a mutated kinase is often a less efficient kinase. The deep responses imatinib produces leave only a small reservoir to mutate. The tolerability profile allows patients to remain on consistent dosing for years without the dose interruptions that might create the population dynamics resistance needs to emerge. Some combination of these factors, plus probably some elements not yet fully understood, has made imatinib in chronic-phase CML one of the most durable cancer drugs anyone has ever developed.

This was not what was expected in 2001. It was, in some ways, better than what was expected. The skeptics inside Novartis had been right to worry. They had been wrong about how it would actually go.

* * *

What this all looks like, from the perspective of a CML patient diagnosed in 2026, is a sequence of decisions in which traditional cancer biology has been replaced by something closer to chronic disease management.

The patient is diagnosed, typically incidentally, on routine blood work that finds an elevated white count. A bone marrow biopsy and karyotype confirm the Philadelphia chromosome. The patient is started on a TKI — most commonly imatinib, often one of the second-generation drugs depending on physician preference and patient factors. They take it every day. They get blood drawn every few months for the first year, then less often, with PCR monitoring of BCR-ABL transcript levels. Their disease, in the standard course, achieves complete cytogenetic response within twelve months and major or deep molecular response within two to three years. They stay on the drug.

Side effects, for most patients, are tolerable: muscle cramps, periorbital edema, occasional rashes, low-grade fatigue. Some patients tolerate one drug better than another and switch within the class to find one that suits them. Most patients on imatinib who have achieved deep molecular response after several years remain on the drug indefinitely; some, meeting the criteria, attempt a treatment-free remission and become candidates for life off therapy. They live, in either case, with the underlying expectation that their CML, as it now stands, is not going to be the thing that kills them.

Compare this to the patient sitting across from Druker in 1993, the one the prologue described. The 1993 patient had a few years to live. The 2026 patient has a normal life expectancy. The 1993 patient's options were inadequate. The 2026 patient's options are, by the standards of any other oncology indication, embarrassingly good. The 1993 patient lived in a world where CML was lethal and where the science was right but the drug did not yet exist. The 2026 patient lives in a world where the same molecular fact — the Philadelphia chromosome, the BCR-ABL fusion, the constitutively active kinase — has been turned, by chemistry and clinical trials and several decades of cumulative work, into a piece of clinical management.

Whether the 2026 patient is ’cured' is a question whose answer depends on definitions. By the practical definition that matters to her — will the disease shorten her life, will she suffer from it, will it limit what she can do — the answer is mostly no. By the strict biological definition — are there zero leukemic cells anywhere in her body — the answer is, in most cases, also no. The disease is, in the language the field has settled into using, controlled. It is not the same as cured. It is also not, in any meaningful sense, the disease that killed people in 1993.

* * *

There is a way of thinking about all of this that the field has, slowly, become comfortable with, and that is worth articulating before the chapter closes.

Cancer, on the surgical-oncology model that was still the field's dominant frame in the late twentieth century, is something you cure or fail to cure. You take it out, or you eradicate it with treatment, or you don't, and the person dies. There is a binary built into the way the older oncology talked about its work. Many of the dramatic stories the field tells about itself — about chemotherapy, about transplant, about radiation — are stories in which a patient is either rescued completely or is not rescued at all. The intermediate state, in which the cancer is held in check by an ongoing intervention but not eliminated, did not have an obvious place in the older frame. It was treated, at best, as a temporary holding pattern between cure and failure.

What the imatinib era, with its long-term data and its molecular monitoring and its treatment-free remission attempts, has done is to make that intermediate state legitimate. A disease that is held in check indefinitely, with a daily pill, with normal life expectancy and normal quality of life, is not a failure of treatment. It is a form of treatment that wasn't really possible before, in cancer at any rate, and that has turned out to be enough for an enormous number of patients. It is what most CML patients now experience. It is what most patients with several other targeted-therapy-amenable cancers also now experience. It is, in some respects, the analog of what diabetes treatment looked like fifty years after insulin and sixty years after oral antidiabetics: a chronic disease, well-controlled with daily medication, in patients whose disease is not gone but whose lives are not being abbreviated by it.

This is a different kind of victory than the one the Time cover in 2001 implied. It is a more complicated victory. It is also, for the patients living it, a real one.

* * *

Bud Romine took imatinib for nineteen years before dying of an unrelated cancer in 2017. He never stopped taking it. He could have, by the time he died, met some criteria for attempting treatment-free remission, but he did not try — by his own account, he was content to take the morning pill and not change a thing. His family has said, in interviews after his death, that he understood the drug had given him the years he had with his children and grandchildren and that he saw no reason to risk the loss of any of those years for the sake of finding out whether he could be off the drug.

Some patients in his cohort did try, and stayed off. Some tried and relapsed and went back on. Some chose, like him, not to try. The data the field accumulated on each of these paths — the patients who stopped successfully, the patients who relapsed and restarted, the patients who never tried — became the basis for the guidelines that govern CML practice today.

What none of these data have produced, in CML, is a path that ends with the cancer eliminated as cleanly and as definitively as a tumor cured by surgery. There is, for most patients, always the small reservoir somewhere, the dormant cells, the molecular signal at the limit of detection. The disease has been transformed but, for most patients, not eliminated. What has been transformed is the patient's life, which is, in the end, the thing the field was trying to transform in the first place.

The next chapter is about how this story — a single targeted molecule, a single transformed disease — generalized, and didn't generalize, to the cancers that came after. CML was an unusually clean target. Many cancers are not. The dream the Time cover suggested in 2001 — that targeted therapy would work for cancer the way imatinib worked for CML — turned out to be more complicated when the field tried to apply it to the rest of oncology.

It is the chapter where the book takes seriously what targeted therapy did and did not do across the larger landscape of cancer.

It is also the chapter where the answer is more mixed than the success of the imatinib story might suggest, and where the things that did work — including a story we have so far given less space than it deserves, of a humanized antibody and a kind of breast cancer that, after 2005, became a different disease — deserve their own honest accounting.

All 13 chapters of The Pill That Could · Library · All courses