Hemorrhage Evolution on MRI

Why blood is the most chemically dynamic signal in the brain

A hematoma is not a static blood clot but a slow chemical reactor. From the moment vessel integrity fails, the iron atom at the center of every hemoglobin molecule begins a sequence of redox and conformational transitions, each of which changes the electronic structure of iron and therefore its magnetism. Because MRI signal is exquisitely sensitive to local magnetic fields, the imaging appearance of blood evolves predictably over hours, days, weeks, and years. No other intracranial process rewrites its own signal so dramatically over time, which makes hemorrhage both a diagnostic prize and a trap: the same lesion can be T1-bright today and T1-dark next month. Mastering hemorrhage means mastering the magnetism of a single transition-metal ion and the way biology repackages it. This module builds that understanding from the electron up, then assembles it into the classic temporal staging, the physics of susceptibility-weighted imaging, and the Bayesian discipline required when the signal is ambiguous.

The magnetism of iron: unpaired electrons set the rules

All of hemorrhage signal flows from one fact: the magnetic behavior of a hemoglobin breakdown product is dictated by the number of unpaired electrons on its iron atom. Electron magnetic moments are roughly 658 times larger than proton moments, so even a few unpaired electrons dominate the local field experienced by nearby water protons. A species with no unpaired electrons is diamagnetic and weakly opposes the applied field; a species with unpaired electrons is paramagnetic and concentrates field lines, creating local gradients. The decisive variables are therefore the oxidation state of iron (ferrous, with the iron in the plus-two state, versus ferric, plus-three), whether the heme pocket is oxygenated, and the resulting electron configuration.

Equation 1 makes the hierarchy quantitative. Oxyhemoglobin holds ferrous iron in a low-spin state with all six d-electrons paired, so $n = 0$ and the species is diamagnetic. Deoxyhemoglobin is ferrous but high-spin with four unpaired electrons ($n = 4$, giving an effective moment near 4.9 Bohr magnetons). Methemoglobin is ferric and high-spin with five unpaired electrons ($n = 5$, the maximum for iron and the most strongly paramagnetic species, near 5.9 Bohr magnetons). The single most important conceptual point of this module is that paramagnetism is necessary but not sufficient to predict signal. How that paramagnetism is packaged in space, whether evenly dissolved or sequestered inside intact cells, determines whether it shortens T1, shortens T2, or unleashes the dramatic susceptibility effects that dominate gradient-echo imaging.

Compartmentalization and magnetic susceptibility

Why should it matter whether paramagnetic iron is dissolved uniformly or locked inside red cells? The answer is the geometry of the magnetic field it produces. A uniformly distributed paramagnetic solute shifts the field everywhere by nearly the same amount; spins precess slightly faster but stay in phase with one another, so transverse decay is barely affected. The instant that same iron is sequestered into discrete compartments, intact erythrocytes, the susceptibility difference between the iron-rich interior and the iron-poor plasma creates steep field gradients at every cell boundary. Water molecules diffusing through this corrugated field accumulate random phase, and the transverse signal decays rapidly. This is preferential T2 and T2 shortening with little T1 effect*, the hallmark of intracellular paramagnetic blood products.

Two features of Equation 2 deserve emphasis. First, the relaxation rate scales with the square of both the susceptibility difference and the field strength, so susceptibility effects grow steeply at higher fields and with more strongly paramagnetic species. Second, the angular term shows that an isolated paramagnetic source produces a classic dipolar field pattern, positive along the field and negative perpendicular to it, which is precisely the spatial signature that susceptibility-weighted phase imaging exploits to separate paramagnetic blood from diamagnetic calcium. When red cells eventually lyse and methemoglobin spills into solution, compartmentalization is lost: the field homogenizes, T2 lengthens and brightens, and the now-accessible iron is free to exert its close-range T1-shortening effect on surrounding water. A single chemical event, cell membrane rupture, thus flips the dominant mechanism from susceptibility dephasing to dipolar relaxation.

The five stages: a chemical clock you can read

The temporal evolution of an intraparenchymal hematoma follows a stereotyped chemical sequence that begins at the periphery, where oxygen and metabolic activity are higher, and progresses inward. Each stage is defined by the dominant iron species and its compartmental state, and each has characteristic T1 and T2 signatures. The timing below is approximate and depends on hematoma size, location, oxygen tension, and patient physiology; the chemistry, not the calendar, is the reliable guide.

Hyperacute (first hours): oxyhemoglobin, diamagnetic

Fresh extravasated blood is fully oxygenated. Intracellular oxyhemoglobin is diamagnetic and exerts essentially no relaxation or susceptibility effect, so signal is dominated by the high water content of liquid blood and serum: isointense to mildly hypointense on T1 and bright on T2. A hyperacute hematoma can therefore masquerade as edema or a cystic lesion on conventional sequences. A useful sign is a thin rim of early deoxygenation at the margin producing peripheral T2 hypointensity, and the clinical context, sudden deficit, raised intracranial pressure, almost always accompanies it.

Acute (hours to about 3 days): deoxyhemoglobin, intracellular and paramagnetic

As tissue oxygen is consumed, hemoglobin within intact red cells converts to deoxyhemoglobin, which is paramagnetic ($n = 4$) but has its iron tucked into a hydrophobic pocket that water protons cannot approach, so there is no T1 shortening. The species remains intracellular, however, so its susceptibility creates strong intracompartmental gradients. The result is the classic acute pattern: isointense to slightly hypointense on T1 and markedly hypointense on T2 and T2*. This is why hyperacute and acute hemorrhage are far better detected on gradient-echo and susceptibility imaging than on T1.

Early subacute (about 3 to 7 days): intracellular methemoglobin

Sustained hypoxia and depletion of the red-cell reductase systems allow ferrous iron to oxidize to ferric, forming methemoglobin while the cell membrane is still intact. Methemoglobin is the most strongly paramagnetic species ($n = 5$), and crucially its conformational change opens the heme pocket to water, enabling close-range dipolar relaxation that powerfully shortens T1. The lesion therefore becomes bright on T1, classically beginning at the periphery and marching centrally. Because the methemoglobin is still intracellular, susceptibility dephasing persists and the lesion remains dark on T2 and T2*. The combination of T1-bright and T2-dark is nearly pathognomonic of early subacute blood.

Late subacute (about 1 week to months): extracellular methemoglobin

Red cells eventually lyse, releasing methemoglobin into the extracellular space as a free solution. The iron is still ferric and highly paramagnetic, so the dipolar T1 shortening persists and the lesion stays T1-bright. But the loss of compartmentalization abolishes the susceptibility gradients: the field homogenizes, transverse dephasing collapses, and the lesion becomes bright on T2 as well. Late subacute hematoma is therefore the only stage that is bright on both T1 and T2, the most conspicuous and easily recognized appearance of blood on conventional imaging. A surrounding rim of T2 hypointensity from early hemosiderin may already be forming at the margin.

Chronic (months to years): hemosiderin and ferritin

Macrophages phagocytose the breakdown products and store iron as ferritin and its aggregated, partially crystalline form hemosiderin. These superparamagnetic-to-paramagnetic storage proteins are intensely compartmentalized within cells and are essentially inaccessible to water, so they exert little T1 effect but enormous susceptibility, producing very short T2 and T2* and dramatic blooming. The mature chronic hematoma typically shows a cavity that is variable on T1 and T2 surrounded by a rim of profound T2 and T2-star hypointensity. Because hemosiderin is metabolically stable, this rim can persist for the patient's lifetime, an indelible scar marking where blood once was.

| Stage | Time (approx.) | Iron species and state | T1 | T2 and T2* | | --- | --- | --- | --- | --- | | Hyperacute | First hours | Oxyhemoglobin, intracellular, diamagnetic | Iso to slightly low | Bright (with thin dark rim) | | Acute | Hours to 3 days | Deoxyhemoglobin, intracellular, paramagnetic | Iso to slightly low | Markedly low (blooms on T2*) | | Early subacute | 3 to 7 days | Methemoglobin, intracellular, paramagnetic | Bright (periphery to center) | Low (blooms on T2*) | | Late subacute | 1 week to months | Methemoglobin, extracellular, paramagnetic | Bright | Bright | | Chronic | Months to years | Hemosiderin and ferritin, intracellular | Iso to low cavity, dark rim | Very low rim, marked blooming |

Susceptibility-weighted imaging, blooming, and phase

Because the most diagnostic stages of blood are defined by susceptibility, the sequences most sensitive to hemorrhage are those that let susceptibility act unopposed. A gradient echo never applies a refocusing pulse, so the static field offsets created by paramagnetic iron, the $\gamma \cdot \Delta B_0$ term of transverse decay, are never reversed. The signal void therefore blooms, appearing larger than the true anatomic extent of iron, and grows with echo time and field strength. Susceptibility-weighted imaging (SWI) pushes this further by acquiring a high-resolution, flow-compensated, long-echo gradient echo and then combining the magnitude image with a processed phase image to amplify the contrast of any substance whose susceptibility differs from tissue.

Equation 3 is the engine of SWI and of its quantitative successor. The accumulated phase encodes the sign of the local field offset, and sign is exactly what distinguishes a paramagnetic from a diamagnetic substance: their dipolar fields point in opposite directions. After background field removal and a phase-unwrapping step, a paramagnetic source such as blood and a diamagnetic source such as calcium produce opposite phase, which on a filtered phase map renders them with opposite brightness (the exact polarity depends on whether the vendor uses a left-handed or right-handed convention). Quantitative susceptibility mapping (QSM) inverts the dipole kernel to recover the underlying susceptibility distribution itself, giving a map in which paramagnetic and diamagnetic materials have opposite-signed values regardless of geometry, the most robust way to separate the two.

Microbleeds and the diagnostic payoff of susceptibility

Cerebral microbleeds are small foci of hemosiderin-laden macrophages, the chronic residue of prior microscopic hemorrhage, and they are functionally invisible on routine T1 and T2 yet conspicuous on gradient echo and SWI as punctate blooming foci a few millimeters across. Their distribution carries diagnostic and prognostic weight: a strictly lobar pattern in an older patient favors cerebral amyloid angiopathy, which raises the risk of future lobar hemorrhage and constrains anticoagulation decisions, whereas a deep and infratentorial pattern (basal ganglia, thalamus, pons, cerebellum) points to hypertensive arteriopathy. Counting and mapping microbleeds has therefore become a quantitative biomarker, which is precisely why field strength and sequence standardization matter so much.

Failure modes, mimics, and Bayesian reasoning

The greatest interpretive danger in hemorrhage imaging is mistaking the cause of a signal. Susceptibility blooming is a final common pathway for several physically distinct substances, and conventional T1 brightness has its own roster of impostors. Disciplined interpretation treats every ambiguous finding as a likelihood problem: the prior probability set by clinical context multiplied by the likelihood that each candidate substance would produce the observed multi-sequence pattern.

• Calcium versus blood. Both bloom on gradient echo, but calcium is diamagnetic and blood products are paramagnetic, so they show opposite phase on SWI and opposite-signed values on QSM. CT remains the rapid arbiter: dense calcium has high attenuation, while most blood, except hyperacute, is denser than brain but lower than calcium. When phase and CT disagree, trust the combination over either alone.
• Melanin and melanoma metastasis. Melanin is intrinsically paramagnetic through stable free radicals and shortens T1, so melanoma metastases can be T1-bright without hemorrhage; these lesions also bleed readily, compounding the appearance. Enhancement pattern, multiplicity, and clinical history reweight the differential.
• Proteinaceous and fat-containing material. Concentrated protein (colloid cyst, Rathke cleft cyst, inspissated secretions) and fat (dermoid, lipoma) are both T1-bright but are not dark on T2-star and do not bloom; fat suppresses on fat-saturated sequences and shows chemical-shift artifact, separating it cleanly from methemoglobin.
• Flow and entry effects. Slow or in-plane flow and entry-slice phenomena can mimic intravascular T1 signal or susceptibility, but they localize to vessels, change with sequence and slice position, and lack the persistent blooming of true iron.
• Gadolinium and contrast. Recently administered gadolinium shortens T1 and can mimic subacute blood on a T1 image; the timing of contrast administration and the absence of T2-star correlate resolve it.

The throughline of this module is a chain of inference you can run forward or backward. Iron's oxidation state and oxygenation set the number of unpaired electrons; the number of unpaired electrons sets the paramagnetism; compartmentalization decides whether that paramagnetism shortens T1 by close-range dipolar relaxation or shortens T2 and T2-star by long-range susceptibility dephasing; and the orderly biological progression of these states writes the five-stage clock onto the image. Read the clock with the full multi-sequence toolkit, weight your differentials by clinical prior and by phase polarity, and the most chemically restless lesion in the brain becomes one of the most readable.