Where Rooty, Sam, and the crew work through the week’s evidence.

The door was already open when Debby knocked. Sam waved her in.

She sat down without her phone. That was the first thing Rooty noticed. Debby always had her phone.

“You okay?” Sam asked.

“Maria has CML,” Debby said. “My friend Maria. From church. She got diagnosed 3 weeks ago. She didn’t tell anyone for a while, but she told me yesterday.”

Nobody said anything for a moment.

Domino, the cute little black and white papichon puppy, padded across the room and put his head in Debby’s lap.

“She is starting Gleevec,” Debby said. “Her oncologist said her prognosis is good. She might be on it for the rest of her life, but her prognosis is good.” She paused. “I looked it up. I read everything I could find.”

“And?” Sam said.

“And I found the same things you have been telling me about. The IRIS trial. The 76%. The DSMB stopping it early.” She looked at her hands. “I did not read it to find the flaw this time. I read it to find out if my friend is going to be okay.”

She looked up.

“I think she is going to be okay.”

“She might be,” Sam said carefully. “The data support optimism. That is different from a guarantee, which her oncologist will have told her. But the data are genuinely good.”

“She asked me what I thought about it. She knows I have been skeptical. About the system. She wanted to know: is this real? Is the drug real? Is the trial real? Should I actually trust this?”

From across the room, Iris, a young girl with inquisitive, analyzing eyes, asked, “What did you say?”

She had been quiet the whole time, watching Debby, Domino now back by Iris’s side, head now resting on Iris’s foot.

“I didn’t know what to say,” Debby said. “I know what the Skeptic’s Toolkit says to do. I could have walked her through the questions. But she does not want to evaluate a claim. She is scared, and she needs to start treatment, and she needs someone who knows something to tell her whether to trust it.”

“That is not actually what she needs,” Sam said.

Debby looked at him.

“What she needs is someone who knows how to listen to what she is scared about specifically, and then respond to that. Not the general claim. Her specific fear.” He paused. “The Toolkit is for evaluating claims. But Maria is not a claim. She is a person with a specific situation and a specific fear. That requires a different set of moves.”

“What moves?”

Sam wrote five letters on a piece of paper and pushed it across the table.

P. A. U. S. E. Protocol

P: Pause before you respond. Not because you do not have an answer. Because she needs to finish the sentence first, and you need to actually hear it.

A: Ask about what she is experiencing. Not the claim. Her. What is Maria actually afraid of? Side effects? That the drug will not work? That the diagnosis means something worse is coming? You do not know yet. You have to ask.

U: Understand the concern before you address it. If she is afraid of side effects, respond to that. If she is afraid the evidence is not real, respond to that. Those are different conversations.

“P and A are the easy part,” Sam said. “You pause. You ask. You actually listen to what she is scared about, not what you think she should be scared about. Most conversations never get that far because people skip both steps.”

“And U?” Debby asked.

Sam looked at the paper.

“U is where it gets hard. You have to understand what it would actually take to change her mind. Not what she says it would take. What it would actually take. Because if no evidence could ever be enough, that is the conversation you are really in.”

Debby was quiet for a moment.

“For Maria,” she said, “her oncologist. She trusts him. If he looked her in the eye and told her the data are real, she would believe it.”

“That is something specific,” Sam said. “That is where you start.”

He set the pen down and looked at her.

“Now apply it to yourself.”

Debby looked up.

“About the system,” Sam said. “About Gleevec. About all of it. What would change your mind?”

A long silence.

“I do not know,” she said finally.

“That is an honest—” Sam stuttered out.

“That is not what I meant.” Said Debby. Her voice had changed, her face had hardened. Her hands clenching.

“I mean I don’t know if anything would. I’ve done the reading. I’ve sat here for 6 weeks and gone through the questions. And I still wake up every morning knowing that the same system that approved this drug in 71 days is the same one that let thalidomide sit on shelves while women lost their babies.”

Without missing a beat, “Thalidomide was not approved in the United States,” Rooty said, breaking his silence. He too had been sitting trying to analyze and understand how best to support Sam and Debby.

Debby retorted, “You know what I mean.”

Rooty paused. One beat. Then: “What specifically are you saying? That the FDA approved something it should not have? That it moved too fast? What is the actual claim?”

“She held the line after enough pressure. After enough time. After Europe already had the data written on children.”

“Kelsey received that application in September 1960,” Rooty said. “She started requesting more evidence immediately. The European market withdrawal did not happen until November 1961. More than a year later. She was already holding the line before any of those children were born. She didn’t need Europe’s data to say no. The safety data wasn’t sufficient and she said so.”

“The US was not zero,” Rooty continued, calmly sticking to the facts. “Investigational samples reached about 20,000 patients here. There were cases. But the European-scale disaster did not happen. And what followed—the Kefauver-Harris Amendment—made efficacy requirements and informed consent mandatory for the first time. The system found the gap. It closed it.”

“That is—”

“You are doing it right now.”

Debby’s voice had gone flat.

“You know what you just did? I said one thing and you both got that look and started explaining. That is what you do. That is what people like you always do.”

Sam set down his pen.

“People like us?” he asked quietly, as Iris and Domino became more and more unsettled seeing the conflict brew.

“Scientists. Doctors. People with enough credentials that you get to decide what counts as evidence and what does not.” She stood up. “Maria asked me what I thought. I am going to sit with her. I am going to tell her I think her oncologist knows more than I do. I am not going to hand her a checklist or ask her to debate her own fear.”

She picked up her bag. “Not everything is a study problem.”

Iris had gone very still, eyes wide, taking everything in. Domino did not move.

The door closed behind Debby. Not hard. Just closed.

Near the doorway, someone had not moved. She had been there for a while, quiet and still. No one had introduced her. She had not asked to be.

Sam looked at the paper.

“P. A. U.”

He had not gotten to the S.

Iris looked at Sam. Then at the paper.

“You were trying to help her see,” she said. “But she needed someone to see her.”

Send this to someone who has been reading to find the flaw. They know what Debby felt in that room. The question Sam asked is the one waiting for them too.

Send this to someone who fact-checks and argues with people they’ve never met online about how safe a medicine or supplement is. They know what it feels like to be Sam and Rooty in that room, and might even follow the PAUSE protocol.

We’re in this Room together!

The deeper dive: trial methodology, regulatory mechanics, and what Gleevec changed about medicine permanently.

From BCR-ABL to approval, the Gleevec story covered roughly 40 years: the identification of the Philadelphia chromosome in 1960, the characterization of BCR-ABL in the 1980s, the development of imatinib in the 1990s, Phase I in 1998, Phase II and the IRIS trial from 2000, and FDA approval on May 10, 2001.

That timeline matters because the 71-day approval figure is sometimes used to argue the system moved too fast. The counterargument is not that 71 days is normal. It is that 71 days was the time taken to review a data package assembled across 40 years of research, 13 years of clinical development, and 1,106 patients.

A Data Safety Monitoring Board is constitutionally independent from the trial. Members are not investigators. They are not employees of the sponsor. They do not share in commercial outcomes. Their mandate, pre-specified in the trial protocol before a single patient enrolls, is to evaluate unblinded interim data against pre-specified stopping rules.

Those stopping rules are set in advance. If the efficacy signal crosses the pre-specified threshold before the full trial completes, the DSMB can recommend stopping for efficacy. If the safety signal crosses its threshold, they can recommend stopping for harm. If the futility boundary is crossed, meaning the trial has become statistically implausible to achieve its primary endpoint, they can recommend stopping for futility.

The IRIS DSMB stopped the trial for efficacy at the first interim analysis. The stopping boundary had been crossed. The recommendation was not a judgment call. It was the application of a rule the investigators, the sponsor, and the FDA had all agreed to before the trial started.

This is worth understanding because it eliminates a particular argument about IRIS: that Novartis pressured the DSMB to stop early in order to accelerate approval. The DSMB’s independence is structural. They see unblinded data that no one at the company sees. Their recommendation is governed by pre-agreed statistical thresholds. If the boundary was crossed, the stopping was correct. If it was not crossed, the stopping would not have been recommended.

The 71-day approval used a specific regulatory mechanism. Accelerated Approval allows the FDA to approve drugs based on a surrogate endpoint, a measurable outcome that is reasonably likely to predict clinical benefit, rather than waiting for overall survival data that may take years to accumulate.

Complete cytogenetic response, the absence of Philadelphia chromosome-positive cells in bone marrow, was the surrogate endpoint for imatinib. It was chosen because prior data established that patients who achieved complete cytogenetic response on interferon had significantly better long-term outcomes than those who did not. The endpoint was not invented. It was validated.

Accelerated Approval also requires post-marketing confirmatory trials: evidence that the surrogate endpoint actually translates to clinical benefit in practice. For imatinib, the IRIS long-term follow-up data at 5, 8, and 10 years provided that confirmation. The 10-year overall survival data, published in 2017, showed that patients who achieved a major molecular response had survival rates comparable to age-matched general population controls, meaning the drug had, for most patients, converted CML from a fatal disease to a chronic one.

The Gleevec story established a template that has since been applied hundreds of times: identify a molecular driver of a specific cancer, develop a compound that targets that driver specifically, validate it in a biomarker-selected population, and apply the results to patients who carry that specific mutation.

The BCR-ABL test, which confirms the Philadelphia chromosome translocation, became the model for what the field now calls a companion diagnostic: a test that is approved alongside the drug and is required to identify which patients should receive it. This was the first major example of precision oncology at commercial scale.

The same logic has since been applied to EGFR-mutant non-small cell lung cancer, HER2-positive breast cancer, BRAF-mutant melanoma, and hundreds of other biomarker-drug pairs. Each went through the same process. What changed was the science making it possible to identify the right patient population before the trial started.

For healthcare professionals and researchers: The IRIS trial remains a methodological benchmark for biomarker-selected oncology trials. The pre-specified statistical analysis plan, the DSMB independence, and the surrogate endpoint validation model are all worth examining against current trial design standards. The IRIS 10-year follow-up (Hochhaus et al., NEJM 2017) confirmed that major molecular response predicts long-term outcomes. Treatment-free remission protocols, emerging since approximately 2010, are a direct extension of the IRIS long-term data.

For patient advocates and policy audiences: The 71-day approval demonstrates what is possible when evidence is clear and the regulatory infrastructure is designed to respond to it. The DSMB mechanism, the pre-specified stopping rules, and the advisory committee vote are structural features designed specifically to protect patients, including from premature approval. They all applied here, and the approval was correct. The access gap that followed—imatinib costs that remain unaffordable in many countries without insurance or generic substitution—is a separate and real problem that the trial result does not resolve.

For patients and general readers: If you or someone you know has CML, the 76.2 percent complete cytogenetic response rate from the IRIS trial is not a ceiling. Long-term follow-up shows that many patients who respond well go on to achieve deep molecular responses and, in some cases, are able to discontinue therapy and remain in treatment-free remission. Discontinuation protocols are not standard practice for all patients, but they exist and are being refined. Talk to an oncologist who specializes in CML.

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.

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: Accelerated Approval. 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, et al. Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells. Nat Med. 1996;2(5):561-566.

Gleevec did not end the story. For most patients, it held. But for a subset, the drug stopped working. A single amino acid substitution in the BCR-ABL kinase domain, the T315I mutation, rendered the binding site structurally incompatible with imatinib. The oncology community called it the gatekeeper mutation. The pipeline called it the next problem to solve.

The response was a second generation of kinase inhibitors, designed specifically to overcome what imatinib could not. Then a third, designed specifically to target T315I itself. How a resistance mutation becomes the design target for the next drug class is Issue 008.

In the Root Room, Sam got to U. He did not get to S or E. Debby walked out before the conversation could finish.

Someone new arrives next week. She has been listening from the doorway for a while. She has something Sam does not. She has a way to lead the PAUSE Protocol to increase its odds of success.

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