Building the dosing infrastructure that oncology trials have needed for a decade.
In 2022, Dr. Lena Hartmann was running PK modeling for a carboplatin re-dosing study at Dana-Farber. The protocol called for individualized AUC targeting using Calvert's formula. What she found was that three different clinical pharmacologists at the same institution were applying different creatinine clearance estimation methods, resulting in dose recommendations that varied by as much as 22% for the same patient.
None of the three answers was wrong by the letter of the protocol. All three were defensible. But two of them put the patient outside the target AUC band. When she looked at the published literature on Calvert-derived dosing in relapsed ovarian cancer, the average AUC target attainment rate in TDM-naive sites was 61.4%.
She built a spreadsheet. Then a MATLAB script. Then, in late 2023, she co-founded DoseMind with two colleagues from the clinical pharmacology team at HMS to replace the spreadsheet with something a clinical research coordinator could actually use.
Active trial protocols on the platform
Patient-cycles processed to date
Academic medical center partners
Founded in Boston, MA
DoseMind shows its work. Every dose recommendation includes the PK curve fit, the AUC projection with confidence interval, and the assumption log. Clinicians are not asked to trust a black box. They are given the information needed to exercise judgment.
Every feature is grounded in published PK/PD science. We do not ship a model parameter without a published reference. When the literature conflicts, we show the conflict and let the investigator configure the appropriate prior for their protocol population.
A dose recommendation that cannot be submitted to a regulatory authority is not useful. DoseMind is designed so that every decision is documented to the standard required by FDA 21 CFR Part 11 and EU Annex 11. The audit trail is not an afterthought - it is the core workflow.
A Northeastern academic medical center running a dose-escalation study in AML used DoseMind to replace manual Bayesian calculations performed in an R script by a single biostatistician. PK update time dropped from 4 hours to under 3 minutes per patient cycle.
4h to 3min per cycle update
A three-site carboplatin re-dosing study standardized dose calculations across sites using a shared DoseMind protocol configuration. Site-to-site variance in cycle-1 AUC attainment fell from 31% to 8.4%, and the DSMB required no unplanned safety reviews during the first 12 months.
Site variance reduced from 31% to 8.4%
A pediatric sarcoma study using busulfan conditioning required weight-band dosing with maturation-function adjustments for patients aged 3-16. DoseMind eliminated the per-patient Excel calculation that had previously required 45 minutes of pharmacist review time before each conditioning cycle.
45-minute review replaced by real-time calculation
We work directly with clinical pharmacologists, principal investigators, and CROs to configure DoseMind for your protocol before the first patient is enrolled.
Get in Touch