RE: MKA placing1 Apr 2026 11:49
An AI interpretation of what the boundaries might be based on known Exatecan Myelosupression.
1) Start with the BOIN target
You first define a target DLT rate (ϕ):
Typical oncology BOIN: ϕ = 0.25–0.30
For exatecan (known myelosuppression):
👉 You might choose ϕ ≈ 0.25 (more conservative)
BOIN then creates escalation/de-escalation boundaries (λe, λd) around this.
2) Translate cytopenias into BOIN-compatible DLT events
You still use CTCAE, but you must collapse hematologic toxicity into a yes/no DLT within cycle 1.
Recommended hematologic DLT definition for BOIN (exatecan-focused)
Neutropenia
Count as DLT if any occur in cycle 1:
Grade 4 neutropenia (ANC <0.5 × 10⁹/L) lasting >5–7 days
Febrile neutropenia
Grade 4 neutropenia with infection/sepsis
👉 Short Grade 4 (<3–5 days) often NOT counted, to avoid over-penalizing transient nadirs.
Thrombocytopenia
DLT if:
Platelets <25 × 10⁹/L (Grade 4)
Platelets <50 × 10⁹/L with bleeding
Platelet transfusion required
Recovery-based hematologic DLT
Critical for BOIN:
Failure to recover by next cycle:
ANC ≥1.5 × 10⁹/L
Platelets ≥100 × 10⁹/L
Within ≤14 days of scheduled dosing
👉 If not → DLT
Anemia (usually secondary)
Grade ≥3 with transfusion or symptoms → optional DLT
Often excluded unless frequent (to avoid noise in BOIN model)
3) What NOT to count as DLT (important in BOIN)
To avoid biasing escalation downward:
Brief Grade 3 neutropenia → NOT DLT
Short-lived Grade 4 neutropenia (<3–5 days) → usually NOT
Asymptomatic thrombocytopenia >25 × 10⁹/L → NOT
Routine transfusions (if expected) → sometimes excluded depending on design
👉 This is crucial: BOIN is sensitive to how you define DLTs.
4) BOIN decision boundaries (example)
Let’s say:
Target ϕ = 0.25
Cohort size = 3
Typical BOIN rules:
0/3 DLTs → escalate
1/3 DLTs → stay
≥2/3 DLTs → de-escalate
For larger cohorts, decisions use:
Escalate if observed DLT rate < λe (~0.20)
De-escalate if > λd (~0.30–0.35)
5) Exatecan-specific nuance
Because cytopenias are:
Predictable
Delayed (nadir-based)
Duration-dependent
You should:
A) Use a slightly longer DLT window
Standard: 28 days
Consider: 28–35 days to fully capture nadir + recovery
B) Carefully define duration thresholds
Prevents overcalling DLTs from transient marrow dips
C) Consider time-to-event BOIN (TITE-BOIN) If using prolonged infusion schedules
Allows partial follow-up weighting before full DLT window completes.
6) Practical summary (what your BOIN boundaries look like)
Your hematologic BOIN DLT definition would effectively be: DLT = severe OR prolonged cytopenia OR clinically complicated cytopenia OR delayed recovery.
Concretely:
Prolonged Grade 4 neutropenia ✅
Febrile neutropenia ✅
Grade 4 thrombocytopenia or bleeding ✅
Failure to recover counts by next cycle ✅
Short, asymptomatic drops ❌
Bottom line
In a BOIN design for exatecan, the “boundary” is not just the lab value—it’s: Whether that lab abnormality cro
Sign In
Follow the stocks
