Evaluating ischemic exposure in cryonics (S-MIX metric)

One hour without circulation at normal body temperature is not biologically equivalent to one hour after the patient has been cooled near 0°C.

Elapsed time alone therefore gives a poor summary of early cryopreservation conditions.

The Standardized Measure of Ischemic Exposure, or S-MIX, compresses a changing time, temperature and support history into estimated equivalent time at normal body temperature.

That makes cases easier to compare. It does not turn an estimate into a direct measurement of brain damage.

What S-MIX is trying to measure

After circulation stops, oxygen delivery falls while cells continue consuming energy.

Ion gradients fail, cells swell, membranes become permeable and biochemical processes begin damaging tissue. These processes do not all proceed at one universal rate.

Cooling slows many reactions and reduces metabolic demand. This is why the temperature history matters alongside the time since circulatory arrest.

S-MIX estimates the duration of unsupported ischaemia at 37°C that its model treats as equivalent to the recorded exposure.

The output is expressed as time, often hours and minutes. It is not literal elapsed time unless the entire interval occurred at 37°C without metabolic support.

Where the model came from

R. Michael Perry proposed an earlier Measure of Ischemic Exposure in 1996. Related work later explored equivalent normothermic exposure under cooling.

Perry and Aschwin de Wolf published the standardized S-MIX formulation in 2020, followed by an implementation discussion in 2021.

Their paper presents linear-cooling, Newtonian-cooling and metabolic-support variants.

Michael Benjamin later worked with de Wolf to apply S-MIX systematically across historical case records in a 2021 meta-analysis.

Benjamin did not originate the formula. He made an important contribution to its operational use, data reconstruction and comparative case analysis.

The authors describe the measure as an imperfect indicator and explicitly identify assumptions that need refinement.

That intellectual history matters. S-MIX is a practitioner-developed model intended to improve case reporting, not a clinically validated neurological outcome scale.

The core calculation

The published model uses 37°C as its normothermic reference and a Q10 value of 2.

Under that assumption, each 10°C decrease halves the rate assigned to ischaemic exposure.

One hour at 37°C contributes one S-MIX hour. One hour at 27°C contributes 30 minutes, and one hour at 17°C contributes 15 minutes.

At 7°C, the same elapsed hour contributes 7.5 minutes. Near 0°C, it contributes approximately 4.6 minutes under the model.

In compact form, the temperature weight is 2 raised to the power of (T minus 37) divided by 10.

S-MIX integrates that weight over time, then applies any stated metabolic-support weight for each interval.

The arithmetic is exact once the inputs and rules are chosen. Whether those rules accurately represent a particular brain is the uncertain part.

A simple example

Imagine one unsupported hour at 37°C followed by one unsupported hour at 27°C.

The first interval contributes one hour. The second contributes half an hour, producing an S-MIX of 1 hour 30 minutes.

Now imagine two unsupported hours at 0°C. Under Q10 equals 2, they contribute roughly nine minutes of normothermic-equivalent exposure.

The metric therefore preserves an obvious fact that ordinary elapsed time loses: prompt cooling changes the biological significance of later delay.

It does not prove that nine minutes at 37°C and two hours at 0°C create identical tissue changes.

Why the Q10 assumption is only an approximation

Q10 is a convenient way to describe how a biological or chemical rate changes across a 10°C temperature interval.

There is no single Q10 for every damaging process. Energy failure, enzyme activity, membrane disruption, oedema and autolysis can respond differently to temperature.

The relationship can also change across the wide range from body temperature to near 0°C.

S-MIX fixes Q10 at 2 to obtain a transparent common scale. The original authors call this a starting approximation and caution against pressing it too far.

A different Q10 would change the score, especially for long cold intervals.

Published S-MIX values are therefore comparable only when they use the same temperature model and parameter choices.

Measured temperature is better than assumed cooling

The most direct calculation uses frequent temperature readings and numerically sums the weighted contribution of each short interval.

Older or incomplete case records may contain only a starting temperature, ending temperature and duration.

A linear model assumes temperature fell at a constant rate. This is simple, but external cooling rarely behaves that way throughout a case.

A Newtonian model assumes cooling slows as the patient's temperature approaches the surrounding medium. This can be more realistic, but still requires fitted or assumed parameters.

Neither interpolation can recreate fluctuations, pauses or temporary warming that were never recorded.

The hierarchy is clear: measured continuous data, then well-documented segmented data, then inferred curves with explicit uncertainty.

Temperature location matters

A human body does not cool uniformly. Skin, oesophagus, rectum, nasopharynx and deep brain can show different temperatures at the same moment.

The probe location should therefore accompany the S-MIX value.

A surface measurement may respond quickly to ice water while deeper tissue remains warmer. A single unlabelled curve can overstate how rapidly the brain cooled.

Sensor lag, displacement and missing readings add further uncertainty.

Using multiple calibrated probes improves the case record. It does not automatically identify which recorded temperature best predicts every form of neural injury.

How metabolic support enters the score

Mechanical cardiopulmonary support with ventilation can move oxygenated blood after circulatory arrest. It does not reproduce normal cerebral perfusion.

The published implementation assigns a weight of 1 to unsupported ischaemia and 0.5 to periods of ventilated cardiopulmonary support.

It assigns a weight of 0 to oxygenated washout under its stated conditions.

These weights reduce the contribution of those intervals before they are added to the total score.

The 50 percent discount is not derived from direct cerebral oxygen measurements in each patient. The implementation paper explicitly describes it as somewhat arbitrary.

Compression depth, airway quality, interruptions, vascular resistance, blood viscosity and perfusion pressure vary between cases.

A binary label such as "CPS present" therefore contains less information than end-tidal carbon dioxide, pressure, flow and interruption records.

Why a case is divided into segments

The early procedure changes modes several times. No-flow delay, external cooling, cardiopulmonary support, surgery, washout and transport have different conditions.

Segmenting the timeline allows each interval to use its own temperature curve and support weight.

The total S-MIX is the sum of all segment contributions until the chosen endpoint.

A useful report names every segment, gives its start and end time, records temperatures and explains any metabolic discount.

If two analysts draw segment boundaries differently or apply different weights, they can obtain different totals from the same narrative.

Publishing the calculation table makes that disagreement inspectable.

The endpoint is part of the definition

The original formulation estimates exposure before cryogenic cooling, typically ending around 0°C.

That boundary is practical, not a claim that ischaemia becomes metaphysically zero at exactly 0°C.

If the endpoint differs between reports, their totals are not directly comparable.

The measured or estimated time at which the relevant patient temperature reached that endpoint must be stated.

Later cryoprotective perfusion, cooling to dry-ice temperature and deep cooldown introduce different injury mechanisms and measurements.

S-MIX should not be silently extended into those phases without defining a new model.

What S-MIX leaves out

The score generally starts at circulatory arrest. It may miss hypoxia, fever, hypotension or regional ischaemia during the agonal period before arrest.

It does not measure pre-existing brain injury, clotting, oedema, vascular obstruction or the uniformity of subsequent perfusion.

It does not measure cryoprotectant concentration, toxicity, ice formation, dehydration, thermal stress or fracture.

It also does not measure synapses, cell membranes or whether identity-critical neural structure remains interpretable.

Those require complementary evidence from procedural telemetry, CT and consented electron microscopy.

This is why S-MIX belongs inside a multidimensional preservation-quality profile, not at the top of a single leaderboard.

A lower score is better within the model

All else equal, a lower S-MIX means less estimated normothermic-equivalent exposure before cryogenic cooling.

All else is rarely equal.

A case with a low score can still have poor cryoprotective perfusion. A delayed case can retain substantial ultrastructure despite a higher estimated exposure.

The score ranks one modeled dimension. It does not produce a probability of revival or a percentage of preserved neurons.

No validated threshold separates a "successful" from an "unsuccessful" human cryopreservation.

Reporting false precision, such as a minute-level result from estimated temperatures, can make the output look more certain than its inputs.

Why S-MIX is still valuable

An imperfect quantitative model can be much better than adjectives such as "fast," "delayed" or "well cooled."

S-MIX forces a team to reconstruct the timeline, preserve temperature data and state where support was present.

It can reveal which interval dominated a case and where operational improvements would reduce exposure most.

Across many cases, consistently calculated values can show trends in response, cooling and documentation quality.

The metric also makes uncertainty visible. Missing temperature data is not merely an inconvenience for calculation; it is itself a quality problem.

Used this way, S-MIX is less a verdict on a patient than a feedback instrument for the team.

How Tomorrow.bio reports S-MIX

Tomorrow.bio publishes anonymized case reports on its Documents page. The two public 2023 reports reviewed here do not state an S-MIX value.

Later reports do. A 2024 report and a 2025 report each contain a dedicated, segmented S-MIX calculation.

The intervals include periods such as legal death to cardiopulmonary support, support to surgery or perfusion, and perfusion through the next procedural endpoint.

The case-report format places S-MIX beside temperature curves, end-tidal carbon dioxide where available, perfusion pressure, refractive index, CT findings and documented problems.

The reports also disclose missing, reconstructed or limited measurements. S-MIX is therefore presented as one component of case quality, not as a standalone verdict on preservation.

This is consistent with Tomorrow.bio's broader quality-metrics programme: use inspectable case evidence to compare outcomes and improve procedures.

How to read an S-MIX value

First, check the start and endpoint. Then inspect the underlying temperature curve and probe location.

Next, examine which intervals received support discounts and whether the report contains physiological evidence that the support was effective.

Look for missing data, inferred temperatures and temporary warming. Check that the same calculation rules were used before comparing cases.

Finally, read the S-MIX beside perfusion evidence and post-procedure imaging.

A precise-looking number should make the underlying case more inspectable, not replace it.

The honest conclusion

S-MIX answers a narrow and useful question: how much temperature-weighted ischaemic exposure does a model assign before cryogenic cooling?

It cannot answer the question people care about most: how much memory and identity-relevant structure remains recoverable.

Its value comes from standardization, transparency and operational feedback. Its danger comes from treating modeled exposure as measured biological outcome.

Further reading

• Perry and de Wolf's S-MIX formulation and implementation discussion
• Benjamin and de Wolf's 2021 case-metrics analysis
• Ultrastructural effects of normothermic and cold cerebral ischaemia
• How ischaemia affects cryopreservation
• Standby, stabilization and initial cooling
• Technical challenges for high-quality preservation