Why the Calvert Formula Misses AUC Targets in 38% of Patients
The Cockcroft-Gault assumption baked into Calvert's original derivation was calibrated on a 1970s dataset. Here is what that means for your 2025 trial...
BOIN vs 3+3: A Practitioner's View After Running Both
Neither design is universally better. The choice depends on your DLT window, expected accrual rate, and sponsor's appetite for MTD precision over speed...
Sparse Sampling for PK Estimation: Where MAP Bayesian Works and Where It Fails
Two-sample Bayesian estimation is practical in Phase II. But three common mistakes in sampling time selection undermine the MAP fit before the lab even runs...
The DLT Window Problem Nobody Talks About in Myelosuppressive Regimens
Standard 28-day DLT windows were designed for targeted therapies. In regimens with delayed hematologic nadir, you may be making escalation decisions before the worst toxicity appears...
Busulfan TDM in Pediatric Transplant Conditioning: What the Published Nomograms Get Wrong
Three academic medical centers published busulfan dosing nomograms in the past decade. They disagree by up to 19% at the same body weight. The reason is maturation function parameterization...
Anatomy of a PK/PD Toxicity Model for Myelosuppression
Friberg's myelosuppression model has been applied to 40+ cytotoxic drugs. This is a walkthrough of the transit compartment structure, what the proliferation rate constant means biologically, and how cycle-to-cycle nadir prediction degrades...
Training CRCs on TDM Software Without the PK Background
Not every CRC has taken a pharmacokinetics course. That is fine. Here are the four concepts they need to understand to operate a TDM platform safely - and the two they do not need to know...
Connecting a PK Dosing System to Medidata Rave: What FHIR R4 Actually Supports
The FHIR R4 spec covers the resources you need - Patient, Observation, MedicationAdministration. Getting Rave to emit them in a form a downstream system can parse is a different problem...
21 CFR Part 11 Compliance for Dosing Software: The 7 Requirements That Get Missed
Most clinical software vendors claim 21 CFR Part 11 compliance. Fewer than half implement all seven requirements correctly. The audit trail sequencing requirement alone disqualifies several common platforms...
What Clinical Teams Need to Know About the Population PK Model Running Their Dosing Software
Your TDM system is applying a population PK model that was built in NONMEM by a modeler who may no longer be at your institution. Here is how to interrogate that model before you trust it with patients...
AUC-Based Dosing vs Flat BSA Dosing: What the Evidence Actually Shows for Carboplatin
The evidence base for AUC-based carboplatin dosing is larger than for almost any other cytotoxic TDM application. The three trials worth reading - and why two of them are routinely misinterpreted...
Designing the PK Sampling Scheme for a Phase I Oncology Study: 5 Decisions You Need to Make Early
PK sampling design is often finalized two weeks before first-in-human dosing. By then, three of the five critical decisions have already been constrained by choices made in protocol writing...