Educational tool. Paste/drop any MRI image to sample intensity per sequence. Distinguish Ca²⁺ vs venous blood on SWI phase using handedness-aware reference.
📖 How to Use This Tool
The primary workflow is to characterize an unknown substance by its signal pattern across multiple sequences. Follow the steps below, then study the worked examples.
Step-by-step workflow
⚠ 0. Anonymize FIRST (always do this before anything else)
Press F6 in PACS to remove patient name, MRN, DOB, accession number, and all identifying overlay text from the display BEFORE taking any screenshot. Verify visually that the screen no longer shows PHI before capturing. This is non-negotiable for HIPAA compliance and patient privacy. If your institution uses a different anonymization key, use that — F6 is the default for most PACS (Epic Radiant, GE Centricity, Sectra, Visage, etc.). Only F6-anonymized images should be loaded into this educational tool.
1. Load the anonymized image
After confirming F6 anonymization, drag, click-to-browse, or paste (⌘V/Ctrl-V) the screenshot from PACS into the dashed drop zone on the Signal → Substance tab.
2. (Optional) Capture tissue anchors
The CSF, GM, WM, and FAT anchor circles show what normal tissue looks like for each sequence. They serve as visual comparators — your lesion circle next to them tells you instantly whether the lesion is brighter, equal to, or darker than each normal tissue.
3. Sample the lesion on each sequence
Click 🔬 pick next to a sequence (e.g., T1), then click the lesion in the loaded image. The slider snaps to the sampled greyscale value, and the lesion circle repaints. Or drag the yellow thumb directly along the black→white strip if you don't have an image loaded.
4. Repeat for every available sequence
T1, T2, SWI magnitude, SWI phase, DWI, ADC. Skip any sequence you don't have — the DDx still works with partial input.
5. Set scanner handedness
Choose Right-handed (GE / Philips) or Left-handed (Siemens / Canon) so the SWI phase slider's polarity is interpreted correctly. This is critical for blood-vs-calcium calls.
6. Anchor the phase voxel to a known structure if possible
"Matches known vein" / "Matches known calcium" / "Opposite polarity to vein" produces the most decisive paramagnetic-vs-diamagnetic call. Use a cortical vein in the same image as your phase reference standard.
7. Read the ranked differential diagnosis
The Differential diagnosis panel ranks entities by similarity to your sampled fingerprint, with one-line rationales for each.
8. Sanity check
Glance down the lesion-circle column vs the anchor circles. If a sample looks visually wrong, re-sample. Consider whether the lesion has multiple components (e.g., hemorrhagic mucin) and characterize the dominant region.
9. Cross-check with the forward tab
Switch to Protein → Predicted Signal, drag the protein slider to your suspected concentration, and confirm the predicted lesion circles match what you actually sampled. This validates your DDx call.
Pattern shortcut table
| What you see | Likely substance |
|---|
| T1 dark, T2 very bright, free diffusion | Simple fluid / arachnoid cyst / CSF |
| T1 bright, T2 dark | Colloid cyst / Rathke cleft cyst / mucin / intracellular metHb |
| T1 bright, T2 bright | Late-subacute hematoma (extracellular metHb) / fat (drops on fat-sat) / proteinaceous fluid 9–12 g/dL |
| T1 iso/dark, T2 markedly dark, SWI dark with phase paramagnetic | Acute hemorrhage (deoxyHb) or hemosiderin |
| T1 iso/dark, T2 markedly dark, SWI dark with phase diamagnetic | Calcification |
| T1 dark, T2 bright, restricted diffusion | Epidermoid / pyogenic abscess / hypercellular tumor / acute infarct |
| T1 dark/iso, T2 dark, no SWI signal change, restricted diffusion | Hypercellular tumor (lymphoma, medulloblastoma) |
Worked examples
Example 1 — Colloid cyst
A 42-year-old presents with positional headaches. MRI shows a rounded hyperdense lesion at the foramen of Monro with mild lateral ventricular enlargement.
| Step | Action | Sample value |
|---|
| 1 | Capture T1 anchors | CSF≈28, GM≈115, WM≈175, FAT≈245 |
| 2 | Sample lesion on T1 | ≈225 (brighter than WM, near-fat) |
| 3 | Sample lesion on T2 | ≈75 (darker than WM — protein signature) |
| 4 | Sample SWI mag + phase, set RH | mag≈170, phase≈130 (no blooming, no polarity) |
| 5 | Sample DWI + ADC | DWI≈95, ADC≈165 (no restriction) |
| 6 | Read DDx | Colloid cyst ~92%, Rathke ~88%, Mucinous ~83% |
| 7 | Cross-check forward tab at 14 g/dL | Predicted circles match sampled within ±5 greyscale |
What you learned: The "T1 bright, T2 dark" pair is the protein/colloid fingerprint. SWI excludes blood and calcium; DWI/ADC excludes epidermoid and abscess. Forward-tab cross-check at ~14 g/dL confirms internal consistency.
Example 2 — DeoxyHb dural venous sinus thrombosis
A patient presents with headache and papilledema. The right transverse sinus appears unusually dark on T2 — flow void or acute thrombus?
| Step | Action | Sample value / observation |
|---|
| 1 | Forward-tab sanity check | Set DeoxyHb stage; confirm circles freeze regardless of protein slider position |
| 2 | Sample sinus on T1 | ≈110 (iso to GM — ambiguous, flow void looks the same) |
| 3 | Sample sinus on T2 | ≈28 (markedly dark — still ambiguous with flow void) |
| 4 | Decisive sequence — SWI mag + phase | mag≈25 (marked blooming), phase≈45 paramagnetic, matches cortical vein anchor — flow void shows neutral phase ~128 with no blooming |
| 5 | Set "Phase reference target → Matches known vein" | Locks the paramagnetic call |
| 6 | Sample DWI + ADC | DWI≈150, ADC≈80 — mild restriction, flow voids do not restrict |
| 7 | Read DDx | Acute hemorrhage (deoxyHb) ~88%, Venous structure ~75%, Hemosiderin ~70% |
What you learned: SWI magnitude blooming + paramagnetic phase + matching to a known vein anchor are the differentiators between flow void and acute deoxyHb thrombus. Flow voids are dark on T1/T2 magnitude but show no blooming and neutral phase. Add Admixture mode only for genuine mixed protein + blood lesions.
Example 3 — Hemosiderin vs calcification
A patient with prior surgical resection AND tuberous sclerosis. Two foci appear dark on T2: one in the surgical bed (suspected hemosiderin), one in a cortical tuber (suspected calcification). They look identical on T1, T2, and SWI magnitude — how do you tell them apart?
| Sequence | Hemosiderin | Calcification |
|---|
| T1 | iso/dark (~95) | iso/dark (~100) |
| T2 | markedly dark (~25) | markedly dark (~30) |
| SWI magnitude | very dark, blooms (~20) | very dark, may bloom (~30) |
| SWI phase RH (GE/Philips) | paramagnetic — DARK (~50) | diamagnetic — BRIGHT (~210) |
| SWI phase LH (Siemens/Canon) | paramagnetic — BRIGHT (~210) | diamagnetic — DARK (~50) |
| DWI | mildly dark (~110) | mildly dark (~115) |
| ADC | normal (~140) | normal (~140) |
Step-by-step:
- Set scanner handedness FIRST. This determines whether paramagnetic substances render dark (RH/GE/Philips) or bright (LH/Siemens/Canon) on phase.
- Sample T1, T2, SWI magnitude on both lesions — they will look essentially identical.
- Sample SWI phase on both lesions, AND on a known cortical vein in the same image. The vein is paramagnetic (deoxyHb in venous blood) and serves as your reference anchor.
- Hemosiderin's phase will MATCH the vein. Calcification's phase will be OPPOSITE the vein.
- Set "Phase reference target → Matches known vein" for the surgical-bed focus, and "Opposite polarity to vein" for the tuber focus.
- Sample DWI + ADC — both will show no significant restriction, ruling out hypercellular tumor and acute infarct.
- Read DDx — hemosiderin should rank highest for one focus and calcification for the other.
What you learned: SWI phase polarity referenced against a known vein is the single decisive sequence for distinguishing hemosiderin from calcification. The polarity rule INVERTS between RH and LH scanner conventions, so handedness must be set first or you will get the diagnosis exactly backwards. T1, T2, and SWI magnitude alone cannot make this call — they are identical for both substances.
Take-home points
- Protein-bright fingerprint: T1↑ + T2↓ at protein concentrations ≈10–17 g/dL identifies colloid/mucin/Rathke.
- SWI is the blood/calcium exclusion sequence: magnitude blooming flags both; phase polarity (referenced to a known vein) discriminates them — but only after you set scanner handedness.
- DWI/ADC is the cellularity discriminator: low ADC + high DWI = true restriction (epidermoid, abscess, hypercellular tumor, acute infarct); always check ADC, not DWI alone.
- Hematoma signal is independent of bulk protein concentration: the protein slider in this tool is decoupled when a hemorrhage stage is selected, because hemoglobin chemistry (not water-protein interactions) dominates relaxation in pure blood.
- SWI phase polarity rule (RH conventions): blood = dark, calcium = bright. INVERTED on LH scanners. Always anchor against a cortical vein in the same image.
Model Notes
Hematoma signal: protein decoupling
In this tool, when a hemorrhage stage is selected, the predicted T1/T2/SWI/DWI signal is computed from a fixed canonical fingerprint per stage rather than from the protein slider. This reflects the physical reality that a pure hematoma's relaxation is dominated by hemoglobin's iron oxidation state and intra/extracellular location, not by bulk protein concentration. Whole blood does contain ~7 g/dL protein, but coupling that to the slider would produce misleading sensitivity to a parameter that does not actually drive hematoma signal. For mixed proteinaceous-and-blood lesions (e.g., hemorrhagic mucinous tumor), characterize the dominant component first.
Protein → T1/T2 model
Below ~9 g/dL, protein has subtle T1/T2 effects. From ~9–17 g/dL, T1 progressively shortens (lesion becomes T1-bright) and T2 also shortens (lesion becomes T2-dark) — the colloid/mucin/Rathke fingerprint. Above ~20 g/dL the lesion approaches solid behavior with marked T1 and T2 shortening (signal-void appearance on T2).
SWI handedness
Right-handed scanners (GE/Philips) display paramagnetic substances (deoxyHb, hemosiderin) as dark on phase, and diamagnetic substances (calcification) as bright. Left-handed scanners (Siemens/Canon) invert this convention. Anchoring against a known cortical vein in the same image neutralizes the convention question.