Read this issue’s Web Companion: The Root Room and the Informed

In Issue 006, we left off with Phase I just beginning. Brian Druker had negotiated a small supply of STI571 from a reluctant Novartis after six years of the company declining to develop the compound. Fifty-four patients with CML who had failed interferon-alpha enrolled. Fifty-three of 54 responded. Their white blood cell counts normalized at the very first dose level tested.

Phase I had not exactly found what it was looking for. It was designed to find the maximum tolerated dose, the ceiling above which a drug becomes more dangerous than useful. What it found instead was a drug producing complete responses well below any ceiling anyone expected to reach.

Phase II was next. The question changed: not "is it safe?" but "does it actually work, in more patients, against a harder comparison?"

That question took 3 more years and 1,106 patients across 16 countries to answer.

In 2008, a 60-year-old man sat in a doctor's office in Los Angeles and was told he had cancer. His first public statement about what he felt in that moment: "I thought I might be dead in a few months."

He is 78 years old today. He is still here specifically because of a trial that stopped early. Not because it failed. Because it worked too well to keep going.

His name is Kareem Abdul-Jabbar. The drug that kept him alive was approved by the FDA in 71 days. Not 71 months. Not 71 weeks. 71 days.

Before we get to what that number means, and what it does not mean, we need to understand the trial that made it possible.

The International Randomized Study of Interferon versus STI571. The field called it IRIS. It enrolled 1,106 patients diagnosed with CML in chronic phase who had not yet received treatment beyond hydroxyurea. The design was a head-to-head comparison: imatinib at 400 milligrams once daily versus interferon-alpha plus low-dose cytarabine, which was the best available treatment at the time.

Interferon-alpha worked. In about 10 to 15% of patients. At the cost of daily injections and flu-like symptoms severe enough that many patients reduced their doses or stopped therapy entirely. Bone marrow transplant, the only treatment that offered a chance of cure, required a matched donor and carried a treatment-related mortality rate of 15 to 30% in older patients.

That was what CML treatment looked like in 2000. IRIS launched in June of that year. The first scheduled look at the data was 18 months in.

Here is what they found.

In 3 out of 4 patients on Gleevec, the cancer cells in bone marrow disappeared completely. On the best available treatment at the time, that happened in 1 out of 7.

Progression to blast crisis, the event that turns CML from a manageable condition into a rapidly fatal one, was three times lower in the imatinib arm.

Before Gleevec, the median survival for CML was 3 to 5 years. After, most patients live for decades on a daily pill. Not a death sentence. A chronic condition.

That shift happened inside a single trial.

Every major clinical trial has a Data Safety Monitoring Board, a DSMB. This is an independent group of scientists and statisticians who are not employed by the company running the trial, are not the investigators, and have no financial stake in the outcome. Their job is to review unblinded data at pre-specified intervals and make one of three decisions: keep going, stop for harm, or stop because the evidence is so clear it would be wrong to continue.

That third option is called stopping for efficacy. It is rarer than the others.

The IRIS DSMB met at the 18-month interim analysis. The pre-specified stopping rule for efficacy had been crossed. Continuing to randomize newly diagnosed CML patients to interferon, when imatinib was producing complete responses in 76% of patients and interferon was producing them in 14%, was no longer ethically defensible.

The trial was stopped. All patients on interferon were offered the chance to cross over to imatinib.

The system did not stop the trial because someone applied pressure. The system stopped it because the system was built to stop it when the evidence was that clear. That distinction matters. It means the trial result was not a product of enthusiasm or a company's interest in a headline. It was a pre-specified rule, applied by an independent board, to unblinded data.

The FDA designated imatinib for Priority Review. The review team ran independent statistical analyses. They compared the pre-specified analysis plans against what Novartis submitted, a check designed to detect post-hoc modifications. They found no discrepancies of concern.

On May 10, 2001, the Oncologic Drugs Advisory Committee convened. Nine hours of presentation. Independent FDA analysis. Patient testimony. Drug safety review.

Every member voted to recommend approval. The FDA issued its approval the same day. From NDA submission to approval: 71 days.

When a drug is approved in 71 days, the phrase most often applied is: unprecedented speed. That framing invites a follow-up question worth answering precisely: speed compared to what, and at what cost to rigor?

The 71 days was the time from NDA submission to approval. It did not include the 2 years of Phase I and Phase II clinical trials. It did not include the 4 years of the IRIS trial itself. It did not include the 5 years of preclinical work, or the 13 years from the characterization of BCR-ABL to the IND filing. From a clinical trial project perspective where I'm coming from presently, it did not include the daily meetings with team members keeping projects moving, or the weekly visits to trial sites to monitor the data, or the likely weekly meetings between the sponsor and CRO discussing how best to maintain data quality and patient safety, or grueling data cleaning and data cuts preparing for analyses. Or the countless hours regulatory teams poured over country regulations to ensure the study was operating correctly.

The clinical trial framework that governs all of this requires pre-specified protocols, independent monitoring, adverse event reporting within regulatory timelines, data integrity auditing, and informed consent for every participant before a single dose. None of that was compressed. All of it had happened, across 1,106 patients in 16 countries, before the 71-day clock started.

What the 71 days represents is not a shortcut through evidence. It is the regulatory machinery moving at maximum speed because the evidence was, in the history of cancer drug development, almost uniquely clear. Large effect size. Understood mechanism. Pre-specified analysis matching submitted analysis. Safety signals characterized and manageable. When those conditions are all present simultaneously, the review is faster. Not because the standard dropped. Because the data supported rapid decision-making.

The 71-day approval was not the system moving fast despite insufficient evidence. It was the system moving fast because the evidence was sufficient.

The Skeptic's Toolkit is 10 questions Root to Rx uses to evaluate health claims without a science degree. It debuted in Issue 003b. Every issue since has run at least part of the story through it. This issue, the evidence answers four of them directly. Here is what those questions looked like in practice against the IRIS trial and the FDA's 71-day review.

IRIS measured cytogenetic response. The primary endpoint was pre-specified before the trial started. The submitted analyses matched the pre-specified plan. The FDA independently verified the match. There were no discrepancies. The DSMB agreed to move through the way the study was designed to.

Novartis funded IRIS. This is relevant context, not disqualifying evidence. The pre-specified analysis requirement, the independent DSMB, and the FDA's independent verification are structural safeguards designed to make funder bias detectable. The FDA found none.

76.2% complete cytogenetic response versus 14.5% on interferon. Progression to blast crisis three times lower. The DSMB stopped the trial at 18 months because continuing randomization was not ethically defensible.

The 71-day approval is sometimes framed as evidence of regulatory weakness. The data show the opposite. It was a response to evidence that met the pre-specified criteria for efficacy in a rigorously conducted trial.

This week in the Web Companion, Debby gets news that changes how she reads all of it.

A friend of hers has CML.

She comes to Sam the Skeptic not with a claim to examine. She comes with a person she loves and a question the Skeptic's Toolkit does not fully answer: how do I talk to her? How do I be useful to someone who needs the system to work, when I have spent two years arguing it does not?

That conversation, and what Sam does with it, is in the Web Companion this week. We've seen the crew use the Skeptic's Toolkit, and now, the PAUSE framework debuts.

Read this issue’s Web Companion: The Root Room and the Informed - Skeptical if it’s worth the read? Go give it a shot, and put it through your own test.

Issue 008 follows Gleevec after approval into territory the IRIS trial might not have predicted.

For most patients, the drug kept working. But for a subset, it stopped. A mutation called T315I, a single amino acid substitution in the BCR-ABL kinase domain, made the drug structurally unable to bind. The clinical oncology community called it the gatekeeper mutation.

The response was a second generation of kinase inhibitors, dasatinib and nilotinib, designed specifically to overcome the resistance. And then a third generation, ponatinib, designed specifically to target T315I itself.

In 2008, a man named Kareem Abdul-Jabbar received a CML diagnosis. His oncologist had access to 7 years of post-approval data, second-generation alternatives, molecular monitoring tools, and a treatment algorithm that did not exist in 2001.

He is still here. That is Issue 008.

1. O'Brien SG, et al. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase CML. NEJM. 2003;348(11):994-1004. (IRIS trial primary results)

2. Druker BJ, et al. Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia. NEJM. 2006;355(23):2408-2417.

3. Deininger M, et al. IRIS 8-year follow-up. Blood. 2009;114(22):462.

4. Hochhaus A, et al. Long-term outcomes of imatinib treatment for CML. NEJM. 2017;376(10):917-927.

5. FDA. Gleevec Approval Letter and Summary Basis of Approval. May 10, 2001. accessdata.fda.gov

6. FDA. Guidance for Industry: Expedited Programs for Serious Conditions. 2014. fda.gov

7. American Cancer Society. Survival Rates for Chronic Myeloid Leukemia. cancer.org

8. Kareem Abdul-Jabbar. Associated Press interview on CML diagnosis and treatment. 2019.

9. Druker BJ, Talpaz M, Resta DJ, et al. Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myelogenous leukemia. N Engl J Med. 2001;344(14):1031-1037. (Phase I primary results; 53 of 54 chronic-phase patients achieved complete hematologic response)

10. Kantarjian H, Sawyers C, Hochhaus A, et al. Hematologic and cytogenetic responses to imatinib mesylate in chronic myelogenous leukemia. N Engl J Med. 2002;346(9):645-652. (Phase II results in interferon-refractory patients; basis of the NDA and 2001 approval)

11. Faderl S, Talpaz M, Estrov Z, et al. The biology of chronic myeloid leukemia. N Engl J Med. 1999;341(3):164-172. (CML natural history and pre-imatinib prognosis; supports 3-5 year median survival figure)

12. Bower H, Björkholm M, Dickman PW, et al. Life expectancy of patients with chronic myeloid leukemia approaches the life expectancy of the general population. J Clin Oncol. 2016;34(24):2851-2857. (Supports the "80%+ 10-year survival" and "chronic condition" framing post-Gleevec)

13. DeMets DL, Pocock SJ. The critical role of independent data monitoring committees. J Biopharm Stat. 2004;14(2):355-371. (DSMB structure and stopping-for-efficacy rules; explainer reference for the IRIS board section)

For educational purposes only. Nothing in this newsletter is medical advice. Talk to your doctor before making any health decision.

Instagram: instagram.com/roottorx | Facebook Lab: facebook.com/groups/theroottorxlab | Subscribe: RootToRx.com