DoseMind ingests individual patient PK profiles, tumor biomarkers, and real-time lab values to compute dose adjustments that fit the patient - not the population average.
Each component of DoseMind maps directly to a step in the oncology dose-optimization workflow - from PK sample collection through regulatory submission.
Fits individual patient pharmacokinetic curves using a two-compartment model with MAP Bayesian estimation. Updates AUC and Cmax projections after each observed concentration, with 95% credible intervals surfaced to the clinician.
Connects to HL7 FHIR endpoints to ingest CBC, creatinine clearance, and hepatic function values as they post. Dose recommendations automatically update when organ function values cross predefined thresholds.
Estimates probability of grade 3/4 adverse events per CTCAE v5.0 based on the projected PK exposure. Surfaces risk scores before each cycle so investigators can preemptively reduce or hold dose.
Generates dose rationale documents formatted for IND annual reports, DSMB packet appendices, and clinical study report sections. Output includes the PK model parameters, observed vs predicted concentrations, and decision log.
Handles population-level PK parameter sharing across sites while enforcing site-specific data access controls. Each site's CRC sees only their patients; the central biostatistics team views the aggregated population model.
Every dose recommendation, override, and parameter change is timestamped and attributed to a specific user role. Electronic signature workflows conform to FDA 21 CFR Part 11 and EU Annex 11 requirements.
Import patient demographics, creatinine clearance, BSA, and prior exposure history from the EDC via HL7 FHIR or CSV. Population PK priors are automatically assigned based on the trial protocol.
Lab results post in real time or via manual entry. Each new concentration point triggers a Bayesian update, narrowing the individual PK parameter distribution and improving the AUC forecast.
The engine computes the dose required to land within the target AUC band at the next cycle, adjusting for changes in organ function and cumulative exposure. Confidence interval is displayed alongside the point estimate.
The treating physician reviews the recommendation, toxicity probability, and alternative dose scenarios. Approval or override is recorded with an electronic signature. No dose proceeds without a documented decision.
Replaces Calvert formula calculations with continuous Bayesian dosing that tracks actual clearance throughout treatment. Reduces underdosing in patients with improving renal function.
Provides exposure-response modeling support for Phase I oncology studies running modified BOIN or ROMI designs. Integrates DLT observations with PK data for the dose-escalation decision.
Applies age-corrected allometric scaling and maturation functions to generate weight-band independent dosing recommendations for pediatric patients from neonates through adolescents.
We offer a 30-day pilot with your protocol and a sample patient dataset so the team can validate recommendations before going live.
Schedule a Pilot