The Pill That Could · The Generalization Problem

The Generalization Problem, part 2

ALK-rearranged lung cancer, identified in 2007, accounts for a smaller fraction of lung adenocarcinomas — perhaps three to five percent — but is driven by a fusion gene that is, structurally, more like BCR-ABL than like the EGFR mutations: a chromosomal rearrangement creating a fusion protein with constitutively active tyrosine kinase activity. The first ALK inhibitor, crizotinib, was approved in 2011, with subsequent generations — alectinib, lorlatinib — producing progressively deeper and more durable responses. Patients with ALK-rearranged lung cancer, on the most current drugs, can have median survival measured in many years rather than months. This is, again, a cancer that turned out to look molecularly more like CML than like generic lung cancer.

BRAF-mutant melanoma, identified in 2002 in work led by Mike Stratton's group at the Sanger Institute, is driven in about half of melanomas by a single recurrent mutation — V600E, a substitution of valine by glutamate at position 600 of the BRAF kinase. BRAF inhibitors, beginning with vemurafenib (approved 2011), produced rapid and dramatic responses in BRAF-mutant melanoma, although the responses were less durable than the kinase-inhibitor responses in CML or ALK lung cancer. The combination of a BRAF inhibitor with a MEK inhibitor (a drug aimed at the next kinase down in the same signaling pathway) extended the responses, and dual BRAF/MEK therapy is, in 2026, standard of care for advanced BRAF-mutant melanoma. Five-year survival rates that were essentially zero in the early 2000s are now in the thirties.

These cancers, and a growing list of others — PARP inhibitors in BRCA-mutant ovarian and breast cancer, FLT3 inhibitors in FLT3-mutant AML, IDH inhibitors in IDH-mutant cancers, KRAS-G12C inhibitors in lung cancers carrying that specific mutation — collectively form the part of the targeted-therapy story that lived up to the 2001 hype. They are not all imatinibs. The responses are sometimes shorter, the drugs are sometimes harder to live with, the cancers sometimes evolve out from under the drugs faster than CML did. But each represents a real clinical benefit, in a defined patient subset identified by a molecular test, of a kind that did not exist before targeted therapy as an idea took hold.

* * *

There is one targeted therapy whose success story has been, in some ways, undertold relative to its impact, and it is worth pausing on that one in particular before this chapter moves to the cancers where the paradigm did not work as well.

Herceptin in metastatic breast cancer, approved in 1998 and described in chapter eight, was the first major non-CML success of the targeted paradigm. It extended survival in HER2-amplified metastatic disease by a meaningful margin and established the small-molecule-vs-antibody dichotomy that has shaped much of oncology since. But Herceptin's bigger story, in some ways, came after the 2001 Time cover, in a different setting.

The setting was adjuvant therapy — treatment given right after surgery for localized breast cancer, in patients whose tumors had been completely removed but whose disease was at meaningful risk of recurrence elsewhere in the body. The standard of care, before 2005, was adjuvant chemotherapy and (for hormone-receptor-positive disease) tamoxifen or an aromatase inhibitor. Patients with HER2-positive disease did less well on the standard regimens than patients with HER2-negative disease; the HER2 amplification, as Slamon's 1987 paper had originally established, predicted worse outcomes.

Beginning around 2000, several large randomized trials launched to ask whether adding Herceptin to adjuvant chemotherapy in HER2-positive early breast cancer would reduce recurrence and extend survival. The trials were enormous — the HERA trial enrolled five thousand patients, the joint analysis of two North American trials (NSABP B-31 and NCCTG N9831) covered another three thousand. They took years to mature. Results began appearing in 2005.

What the trials showed was, in the language of breast oncology, a paradigm shift. Adding one year of Herceptin to adjuvant chemotherapy reduced the risk of recurrence by about half. The reduction translated, on follow-up, into substantial improvements in overall survival — absolute differences of about ten percentage points at ten years between the chemotherapy-alone and chemotherapy-plus-Herceptin arms. Patients with HER2-positive early breast cancer, who had previously been the breast cancer subset with the worst prognosis, were now, with adjuvant trastuzumab, doing as well as or better than HER2-negative patients. The prognostic stratification that had defined breast cancer for nearly twenty years — HER2 amplification as a marker of more aggressive disease — had been not just neutralized but, in some respects, inverted.

The breast cancer community absorbed the result quickly. Adjuvant Herceptin was, by 2006, standard of care for any HER2-positive breast cancer with sufficient risk of recurrence. The treatment is, in 2026, given to roughly fifty thousand women a year in the United States and many times that worldwide. The number of HER2-positive breast cancer recurrences prevented, and the number of breast cancer deaths prevented, runs into the hundreds of thousands cumulatively over the past two decades. It is, by any measure of total clinical impact, one of the largest contributions of targeted therapy to oncology so far.

The story did not end with Herceptin. A second-generation HER2 inhibitor, pertuzumab, was approved in 2012 and is now used in combination with trastuzumab for both metastatic and high-risk early disease. An antibody-drug conjugate — trastuzumab emtansine, sold as Kadcyla — was approved in 2013, attaching a cytotoxic chemotherapy molecule to the trastuzumab antibody so that the antibody delivers the chemotherapy specifically to HER2-positive cells. A more recent antibody-drug conjugate, trastuzumab deruxtecan (Enhertu), approved in 2019, has produced even larger response rates, and has, in a particularly striking 2022 result, been shown to benefit not just HER2-amplified tumors but tumors with merely modest HER2 expression — a much larger fraction of breast cancers — effectively expanding the HER2-targeted franchise from a quarter of breast cancers to perhaps two-thirds of them.

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