Skip to content
TraceReadRequest a demo

Case study, page T-0046

Found every statement

Unspecific wide QRS and anterolateral infarct pattern

PTB-XL record 12262, a woman aged 89. A public research ECG from Germany, printed on A4, photographed by TraceRead's photo simulator and read end to end. Below is the AI first read, next to the original recording and what the dataset's cardiologists recorded.

0.78

median match with the original recording

0.98 after lining up by up to a tenth of a second

3 of 3

cardiologists' statements that TraceRead AI also flagged

8

extra patterns flagged for the doctor to keep or remove

92

heart rate from the photo, beats a minute

92 on the original recording

The AI first read

What the reading doctor gets

Hover a lead to see how closely the red rebuilt line follows the original recording. Untick the box to see the page without it.

Flattened page of PTB-XL record 12262

Numbers under each lead: how closely the rebuilt line follows the original recording, where 1.00 is a perfect match. Median for this page: 0.78.

Photo usable, hold rule version 2AI first read, for a doctor to check

PTB-XL record 12262, a woman aged 89. A4 paper, 3x4 with one lead II strip, pulses read 4 of 4.

Measures

MeasureFrom the photoOriginal recording
Heart rate (beats a minute)9292
PR interval (ms)151149
QRS width (ms)111108
QT interval (ms)356348
QTc, corrected for heart rate (ms)410401
Heart axis (degrees)-2-3

Patterns flagged, with TraceRead AI's score

  • Sinus rhythm

    the usual rhythm, started by the heart's natural pacemaker

    98

  • Left anterior fascicular block

    a block in one small branch of the left signal pathway, which changes the direction of the electrical signal

    43

  • Minor T wave abnormalities

    small changes in the last part of the beat that are not enough to point to a cause

    42

  • Unspecific T wave changes

    changes in the last part of the beat that do not point to one cause

    12

  • Unspecific ST changes

    small changes in the middle part of the beat that do not point to one cause

    10

  • ST depression

    the middle part of the beat sits lower than usual

    7

  • Low T waves

    the last part of each beat is smaller than usual

    6

  • Unspecific wide QRS

    the signal spreads slowly inside the lower chambers, so the main spike is wide

    3

  • Anterolateral infarct pattern

    a pattern linked to damage to the front and side wall of the heart muscle

    3

  • Left atrial enlargement

    the upper left chamber may be larger or under more strain than usual

    2

  • Q waves present

    a small downward dip at the start of the main spike, which can be normal or can follow damage to the heart muscle

    2

What the dataset's cardiologists recorded for this ECG

  • Unspecific wide QRS
  • Anterolateral infarct pattern
  • Sinus rhythm

Against the cardiologists

TraceRead AI and the dataset's cardiologists, side by side

TraceRead AI flagged 11 patterns. 3 of them are also among the dataset's statements (Sinus rhythm, Unspecific wide QRS and Anterolateral infarct pattern). Not in the dataset's statements: Left anterior fascicular block, Minor T wave abnormalities, Unspecific T wave changes, Unspecific ST changes and 4 more.

Found by both 3

  • Unspecific wide QRS

    The signal spreads slowly inside the lower chambers, so the main spike is wide.

    3

  • Anterolateral infarct pattern

    A pattern linked to damage to the front and side wall of the heart muscle.

    3

  • Sinus rhythm

    The usual rhythm, started by the heart's natural pacemaker.

    98

Recorded by the cardiologists only 0

Nothing: TraceRead AI flagged every statement they recorded.

Flagged by TraceRead AI only 8

  • Left anterior fascicular block

    A block in one small branch of the left signal pathway, which changes the direction of the electrical signal.

    43

  • Minor T wave abnormalities

    Small changes in the last part of the beat that are not enough to point to a cause.

    42

  • Unspecific T wave changes

    Changes in the last part of the beat that do not point to one cause.

    12

  • Unspecific ST changes

    Small changes in the middle part of the beat that do not point to one cause.

    10

  • ST depression

    The middle part of the beat sits lower than usual.

    7

  • Low T waves

    The last part of each beat is smaller than usual.

    6

  • Left atrial enlargement

    The upper left chamber may be larger or under more strain than usual.

    2

  • Q waves present

    A small downward dip at the start of the main spike, which can be normal or can follow damage to the heart muscle.

    2

The doctor's review

What the reading doctor would check here

Worked out from this run's own numbers: where the rebuilt line strays, which measures moved, and which flags to keep or remove.

  1. 1The rebuilt line strays from the original recording in aVR (0.78), V1 (0.73), V4 (0.55), V2 (0.60) and V5 (0.81) and 3 more. Compare those leads with the paper first.
  2. 2Every measure from the photo is close to the same measure on the original recording.
  3. 3TraceRead AI flagged 8 patterns the cardiologists did not record: left anterior fascicular block, minor T wave abnormalities, unspecific T wave changes, unspecific ST changes, ST depression, low T waves, left atrial enlargement and Q waves present. The doctor keeps or removes each one.
  4. 4The lead II piece in the first column and the long lead II strip agree less than they should (correlation 0.70), so part of the trace may be misread.

An AI first read is not the last word.

  • A qualified doctor reviews every AI first read. Nothing reaches a patient until a doctor signs it.
  • Do not rely on TraceRead alone in an emergency.
  • A normal AI first read does not rule out heart disease.
  • For adults, 18 and over.

Stage by stage

How TraceRead read this page

Every step of this run, with what went in and what came out.

Step 1: the photo

On the phone

What went in: one photo of a printed ECG

PTB-XL record 12262, a public research ECG from Germany, was printed as page T-0046 on A4 at 25 mm a second and 10 mm per millivolt. TraceRead's photo simulator then took a photo of it: tilted, lit unevenly, slightly blurred, with sensor noise and JPEG compression, the way a phone camera would.

This record was in PTB-XL's test set, so TraceRead AI never saw it during training. The person was a woman aged 89. Age and sex are all TraceRead keeps.

Photo made by computer of the printed PTB-XL record 12262
Shown at 1600 by 1200 pixels.

Step 2: page finder

TraceRead AI

Find the paper and how it is turned

TraceRead AI looks at the whole photo and marks the four corners of the page, and which way up it is. Fixed code then measures the printed millimetre grid to correct the scale and the tilt, and makes the page flat.

Page found
yes
Turned
0 degrees
Paper
A4
The photo with the four page corners TraceRead AI found
The four corners TraceRead AI found, in red.

Step 3: layout reader

TraceRead AI

Work out where each lead sits

Every row of a printout starts with a calibration pulse, a small square step that is 10 mm tall for 1 millivolt. TraceRead AI finds the pulses, and fixed code matches them against 14 known page layouts to work out where each of the 12 leads sits and how long it runs.

Layout
3x4 with one lead II strip
Calibration pulses read
4 of 4
Height of 1 millivolt
10 mm
Paper speed used
25 mm a second
Layout match
0.961

Warnings this step raised

  • The lead II piece in the first column and the long lead II strip agree less than they should (correlation 0.70), so part of the trace may be misread.

Step 4: trace rebuild

TraceRead AI

Rebuild every lead as a digital signal

TraceRead AI marks which pixels are trace ink. Fixed code then follows the ink of each lead, column by column, and turns it into millivolts, 500 points a second. The rebuilt line is drawn back over the paper so a doctor can see at a glance if a lead was misread.

Because this ECG is public, we can check the rebuilt signal against the original recording. 1.00 would be a perfect match; the second number allows the two to be lined up by up to a tenth of a second.

LeadSeconds shownInk foundMatchLined up
I0 to 2.5100%0.960.97
aVR2.5 to 5100%0.780.97
V15 to 7.5100%0.731.00
V47.5 to 10100%0.550.98
II0 to 2.5100%0.930.95
aVL2.5 to 5100%0.860.99
V25 to 7.5100%0.600.99
V57.5 to 10100%0.810.98
III0 to 2.5100%0.880.99
aVF2.5 to 5100%0.750.99
V35 to 7.5100%0.640.98
V67.5 to 10100%0.970.98
II, rhythm strip0 to 100.770.89

Median match for this page: 0.78.

The flattened page with every rebuilt lead drawn over it in red
The flat page with every rebuilt lead drawn over it in red.

Step 5: coverage map

Fixed code

Mark which seconds the paper shows

A 3x4 printout shows each lead for only 2.5 of the 10 seconds the machine recorded. The coverage map tells the pattern finder exactly which seconds of each lead it has, so a part that is not on the paper is never read as a flat line.

I
aVR
V1
V4
II
aVL
V2
V5
III
aVF
V3
V6
II strip
0 s5 s10 s

Step 6: hold for retake

Fixed code

Decide whether the photo can be read

Before any pattern is read, a fixed hold rule decides whether the rebuild can be trusted. Version 1 looked only at how well the page matched a known layout and how unsure the AI was about the five main groups. Version 2 also checks the calibration, the page shape, how much trace was found, whether the overlapping lead II pieces agree, and clipped or flat parts. If the rebuilt trace looks unreliable, it asks for a new photo instead of giving a first read. Both versions' decisions for this page are shown below.

Version 1Photo usable

The layout match score was 0.961 (a photo is held back when the score is below 0.888) and 4 of 4 calibration pulses were read. The AI's own uncertainty about the five main heart groups was 0.414 (a photo is held back above 0.497). Neither check fired, so the photo is used.

Version 2Photo usable
  • Every check passed, so the rebuilt trace is used.
CheckValueLimitResult
page_foundpassed
gain_standard0.0090.05passed
paper_speed25.042434passed
layout_match0.9610.889passed
page_shape1.0041.7passed
ink_median10.9passed
coverage11passed
lead_ii_agree0.970.49passed
lead_ii_amplitude0.0230.5passed
goldberger0.1151.45passed
edge_clipping00.027passed
flat_run0.0090.34passed
lead_ii_raw0.6970.2passed

Step 7: measures

Fixed code

Measure the heartbeat

Fixed code finds every beat, lines the leads up and measures them together. The same code run on the original recording shows how much the photo changed each number, and a reference ECG program from the PTB-XL+ dataset gives a second opinion.

MeasureFrom the photoOriginal recordingReference program
Heart rate (beats a minute)929292
PR interval (ms)151149152
QRS width (ms)11110896
QT interval (ms)356348348
QTc, corrected for heart rate (ms)410401401
Heart axis (degrees)-2-3-19
PR 164 msQRS 108 msQT 354 ms
The median heartbeat of lead II, from the photo, with the points the PR, QRS and QT measures start and end at.

Step 8: pattern finder

TraceRead AI

Score 71 patterns and flag the ones that pass their line

TraceRead AI reads the 12 rebuilt leads and the coverage map together and scores 71 patterns of rhythm, conduction, thickened heart muscle and ST and T changes. A pattern is flagged when its score passes a line set in advance on other ECGs, never on this one. The lines are set so that patterns are not missed, which is why a pattern can be flagged with a low score, and why some flags are wrong: the doctor decides.

What the dataset's cardiologists recorded for this ECG

  • Unspecific wide QRS
  • Anterolateral infarct pattern
  • Sinus rhythm

TraceRead AI flagged 11 patterns. 3 of them are also among the dataset's statements (Sinus rhythm, Unspecific wide QRS and Anterolateral infarct pattern). Not in the dataset's statements: Left anterior fascicular block, Minor T wave abnormalities, Unspecific T wave changes, Unspecific ST changes and 4 more.

TraceRead AI's score for each pattern, from 0 to 100. The thin line is where that pattern gets flagged.

  • Sinus rhythm

    flagged

    the usual rhythm, started by the heart's natural pacemaker

    98

  • Left anterior fascicular block

    flagged

    a block in one small branch of the left signal pathway, which changes the direction of the electrical signal

    43

  • Minor T wave abnormalities

    flagged

    small changes in the last part of the beat that are not enough to point to a cause

    42

  • Unspecific T wave changes

    flagged

    changes in the last part of the beat that do not point to one cause

    12

  • Unspecific ST changes

    flagged

    small changes in the middle part of the beat that do not point to one cause

    10

  • ST depression

    flagged

    the middle part of the beat sits lower than usual

    7

  • Low T waves

    flagged

    the last part of each beat is smaller than usual

    6

  • Unspecific wide QRS

    flagged

    the signal spreads slowly inside the lower chambers, so the main spike is wide

    3

  • Anterolateral infarct pattern

    flagged

    a pattern linked to damage to the front and side wall of the heart muscle

    3

  • Left atrial enlargement

    flagged

    the upper left chamber may be larger or under more strain than usual

    2

  • Q waves present

    flagged

    a small downward dip at the start of the main spike, which can be normal or can follow damage to the heart muscle

    2

More case studies

Previous, T-0045

Complete right bundle branch block and atrial fibrillation

All 12 cases

Case studies

Next, T-0048

Reduced blood flow pattern, bottom and atrial fibrillation

See TraceRead read a paper ECG

We show you TraceRead reading public research ECGs, step by step, and talk through how your team prints, photographs and reads ECGs today.

Request a demoSee the case studies