Duration and reinvestment risk — a 4.6-month response is a short-duration instrument, and short duration changes everything about how you value it
also called Macaulay duration analogy · bridge-to-transplant as reinvestment
A remission that lasts under five months is worth very little on its own. In finance, a short-lived investment is valued almost entirely by what you can move the money into next, not by how good the investment itself was.
The same thing, for a scientist
Fixed-income analysis separates two things that oncology endpoints routinely conflate: the size of a payment and how long it lasts. An instrument with a short duration returns your capital quickly, and its value is therefore dominated not by its coupon but by reinvestment risk — what you can roll into when it matures. Menin monotherapy in relapsed or refractory KMT2A-rearranged leukemia is a short-duration instrument. The ziftomenib registration cohort reported a median duration of response of 4.6 months [E1]; AUGMENT-101 reported CR+CRh of 22.8% against an overall response rate of 63.2% [E2]; median survival after menin-inhibitor failure is 4.4 months [E3]. Under this frame, CR rate is the coupon and the conversion rate into a curative-intent consolidation is the reinvestment. Interfant-06 supplies the sharpest illustration: in the high-risk group, only 46% of infants with an intention to transplant actually received a transplant, because many had early events [E4]. The reinvestment failed, and the reinvestment was the whole plan.
Why the mapping is more than a metaphor
Duration in finance is a weighted average time to cash flows, and it is useful because it tells you the sensitivity of value to what happens later. Translated: the value of a remission is the integral of survival benefit it generates, and for a short remission that integral is dominated by whether the patient reaches something curative while still in it.
The endpoint the field reports, CR/CRh, is a point estimate of coupon size at a single moment. It carries no duration information at all. Restricted mean survival time, and the proportion reaching consolidation in remission, do.
The numbers, laid out as an instrument would be
| Quantity | Financial counterpart | Value |
|---|---|---|
| CR+CRh rate | coupon | 22.8% (AUGMENT-101 phase 2, n=57) 2 |
| Overall response rate | gross yield including non-par payments | 63.2% 2 |
| Median duration of response | duration | 4.6 months (ziftomenib registration cohort) 1 |
| Survival after failure | recovery value on default | median 4.4 months 3 |
| Proportion reaching transplant | reinvestment rate | 46% of those with intent, in Interfant-06 high risk 4 |
The gap between 63.2% and 22.8% is itself informative under this frame: most responses are not complete, which in instrument terms means most of the payments are partial and the duration on those is likely shorter still.
Where the analogy misleads
Finance assumes cash flows are fungible and that reinvestment is always available at some price. Neither holds here. A transplant is not available at any price to a patient who is too sick, too young, or lacks a donor, and the reinvestment option can be destroyed by the very therapy that created the remission — prior intensive therapy appears to install menin-inhibitor resistance 5, which is the opposite of a market where cash is cash.
The analogy also encourages treating survival months as commensurable with each other, which they are not: a month of remission in an infant carrying decades of anthracycline and neurocognitive risk is not the same quantity as a month in a 69-year-old. Any model built on this frame must say whose months it is counting.
| domain | finance |
|---|---|
| source concept | Macaulay duration, reinvestment risk, and the distinction between coupon and total return |
| endpoint implication | CR/CRh is a coupon measure. The decision-relevant measures are restricted mean survival time and the proportion of treated patients reaching a curative-intent consolidation while in remission, neither of which is a primary endpoint in any menin trial we identified. |
mapping
[
{
"bio": "complete remission rate (CR/CRh)",
"analog": "coupon — the size of the payment, with no information about how long it lasts"
},
{
"bio": "median duration of response, 4.6 months",
"analog": "duration — the weighted average time to the cash flows"
},
{
"bio": "overall response rate exceeding CR rate by 40 points",
"analog": "partial payments that are counted in gross yield but carry shorter duration"
},
{
"bio": "allogeneic transplant or CAR-T consolidation achieved while in remission",
"analog": "reinvestment at maturity — where nearly all the value of a short-duration instrument lives"
},
{
"bio": "patients too unwell, too young or without a donor at the moment remission is achieved",
"analog": "reinvestment risk: the roll is not available when the instrument matures"
},
{
"bio": "prior intensive therapy installing menin-inhibitor resistance before the drug is ever given",
"analog": "an impaired instrument bought at par — see analogy-finance-pre-existing-impairment-diligence"
},
{
"bio": "survival after menin failure, median 4.4 months",
"analog": "recovery value on default"
}
]generated hypotheses
[ "hyp-menin-value-dominated-by-bridge-rate" ]
- If value really is dominated by the bridge rate, then a drug that produces many short partial responses and few complete ones may be worth less than its response rate suggests, and the field's headline numbers systematically overstate benefit.
- Framing therapy as a bridge institutionalises transplant in infants, whose long-term cost — growth, endocrine, neurocognitive, second malignancy — is exactly the burden this whole programme exists to reduce.
- What proportion of KMT2A-rearranged patients responding to a menin inhibitor actually reach transplant or CAR-T in remission? This is reported inconsistently across trials and is arguably the field's most decision-relevant number.
- Does the reinvestment option differ for infants, where transplant carries severe long-term cost, versus adults, where it is more routine? If so the optimal sequencing genuinely differs by age and no trial is designed to detect that.
- Is there a duration threshold below which a remission has no value at all because no consolidation can be organised in time?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the registration-enabling phase 2 of KOMET-001, 92 patients with relapsed or refractory NPM1-mutant AML (median age 69) received ziftomenib 600 mg daily and met the primary endpoint with a CR/CRh rate of 22%, overall response 33%, and a median duration of response of 4.6 months.
Ziftomenib in Relapsed or Refractory NPM1-Mutated AML · Wang ES, Issa GC, Erba HP, et al. · Journal of Clinical Oncology 2025What this evidence cannot show
NPM1-mutant AML in a median-69-year-old population, not KMT2A-rearranged infant ALL. Transferring the duration figure to infants is an assumption, not a finding, and it may be pessimistic.
Evidence that argues against this (2)
- WHERE THE ANALOGY BREAKS. Macaulay and modified duration are defined as weighted average times to receipt of a *known, scheduled* set of cash flows, and the standard treatment is explicit that when the cash flows are not well defined the measure fails: for bonds with embedded options, which 'do not have well-defined internal rates of return', 'yield duration statistics such as Modified and Macaulay Durations do not apply', and effective duration must be used instead. A remission is the embedded-option case in its purest form: there is no schedule, only a stochastic hazard of relapse, and the 'maturity date' is realised rather than contracted. So 'duration' in this entry is being used in the loose sense of 'how long the response lasts', which is just median duration of response — a quantity the field already reports — rather than in the technical sense that would license the reinvestment-risk machinery. The analogy earns the framing (coupon versus reinvestment) but not the arithmetic. AnalystPrep 2019
- The short-duration premise is also a property of monotherapy rather than of the instrument. Adding decitabine/cedazuridine and venetoclax to revumenib produced a median duration of CR/CRh of 10.5 months in relapsed/refractory AML, against the 4.6 months of ziftomenib monotherapy that anchors this entry. If the tenor more than doubles with combination partners, reinvestment risk is not the dominant term. Caveat: cross-trial comparison across different agents, populations and response-duration definitions. Issa GC et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In AUGMENT-101 phase 2, among 57 efficacy-evaluable patients with relapsed or refractory KMT2A-rearranged acute leukemia, CR+CRh was 22.8% (95% CI 12.7-35.8) and overall response 63.2% (95% CI 49.3-75.6), with 15 of 22 assessed having no detectable residual disease; 94 patients were treated with a median age of 37 years, range 1.3-75.
Menin Inhibition With Revumenib for KMT2A-Rearranged Relapsed or Refractory Acute Leukemia (AUGMENT-101) · Issa GC, Aldoss I, Thirman MJ, et al. · Journal of Clinical Oncology 2024What this evidence cannot show
Single-arm, heavily pretreated, adult-dominated. Duration of response is not in the abstract and must be taken from the label or the full text.
Evidence that argues against this (1)
- This is one phase 2 cohort, and the figure most often quoted from it is the least favourable framing of a heavily pretreated adult population: median age 37, no prespecified survival endpoint in the registered protocol, and a gap between 63.2% overall response and 22.8% CR+CRh that consists largely of responses without count recovery. Whether CR+CRh or overall response is the right input depends on what the model is for, and the beat uses both without saying which. Fifteen of 22 assessed with no detectable residual disease is also a subset of a subset: MRD was assessed in fewer than a quarter of treated patients, and who gets assessed is not random. Journal of Clinical Oncology 2025
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 84 adults with relapsed or refractory AML after menin-inhibitor failure, 40% received supportive care only, the CR/CRi rate for non-trial therapies was 19% and overall response 32%, and median overall survival from the start of next therapy was 4.4 months — 15.4 months for those achieving CR/CRi versus 3.4 months otherwise.
Outcomes of relapsed or refractory acute myeloid leukemia after menin inhibition failure · Multicentre retrospective consortium · Blood Advances 2026What this evidence cannot show
Retrospective, adult, and selected for patients well enough to receive next therapy; the 40% receiving supportive care only is the more sobering number and is itself selection-dependent.
Evidence that argues against this (1)
- The 'reinvestment rate' the entry proposes to measure — conversion into curative-intent consolidation — is an exposure that can only be acquired by surviving in remission long enough to receive it. Suissa's analysis shows that defining exposure this way creates immortal time, that misclassifying or excluding it biases the rate ratio in favour of the exposed group in proportion to the length of the immortal period, and that the bias is more pronounced under a decreasing hazard — the hazard shape documented for KMT2A-rearranged infant ALL in the sibling analogy node. The 15.4-versus-3.4-month contrast cited here has this structure. The reinvestment frame is therefore most misleading exactly where the entry finds it most compelling, and any analysis built on it needs a landmark or time-dependent-covariate design. Suissa S 2008
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In Interfant-06, 651 infants had 6-year event-free survival of 46.1% and overall survival 58.2%; in the high-risk group 6-year EFS was 20.9% with an intention to transplant, and only 46% actually received stem-cell transplantation because many had early events.
Outcome of Infants Younger Than 1 Year With Acute Lymphoblastic Leukemia Treated With the Interfant-06 Protocol: Results From an International Phase III Randomized Study · Pieters R, De Lorenzo P, Ancliffe P, et al. · Journal of Clinical Oncology 2019What this evidence cannot show
Pre-dates menin inhibitors and frontline blinatumomab; the 46% conversion rate is from a chemotherapy-only era and is the historical baseline the analogy is built on, not a current figure.
Evidence that argues against this (1)
- Interfant-06 is used here as the sharpest illustration of failed reinvestment. The counter-observation is that the menin-inhibitor class was approved and produced its survival signal with only approximately 30-40% of responders in relapsed/refractory trials proceeding to allogeneic transplantation, and that the transformative result in infant KMT2A-rearranged B-ALL came from adding blinatumomab (2-year DFS 82% versus 49%) rather than from raising the transplant rate. If a low conversion rate were the binding constraint, neither observation should look the way it does. Alati C et al. 2026
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In serially sampled patient-derived xenografts, highly pretreated KMT2A::AFF1 ALL samples were much less sensitive to menin inhibition than earlier samples from the same patients despite no prior menin-inhibitor exposure, with sustained on-target transcriptional effects in the resistant cells.
Upfront menin-inhibitor resistance in multiply pretreated leukemias · Mahdavi L, Bernt KM, et al. · Experimental Hematology 2025What this evidence cannot show
Very small: serial samples from two multiply relapsed patients. Directionally important, statistically thin.
Evidence that argues against this (1)
- The pretreated-PDX observation is bounded by clinical data: relapsed/refractory patients with a median of two prior lines, 52% venetoclax-exposed, reached composite complete remission of 71% on a revumenib-containing triplet with MRD negativity in 80% of responders. Prior therapy does not reliably extinguish the coupon. Issa GC et al. 2026
duration statistics require a defined cash-flow schedule and do not apply when cash flows are uncertain; the model's conversion parameter also carries immortal time bias
menin monotherapy valued as a short-duration instrument
- analogous to →Ziftomenib (KO-539, Komzifti) — once-daily menin inhibitor approved in NPM1-mutant AML
the 4.6-month median duration of response is the duration figure the frame is built on
transplant as the reinvestment vehicle whose availability dominates value
generated by reinvestment risk on a short-duration instrument