Healthcare
13 columns
200 rows
8 formats

Clinical Notes Image Matching

This dataset provides a comprehensive mapping between de-identified clinical notes and their corresponding diagnostic images, enabling advanced research in multi-modal AI for healthcare. Each entry includes rich metadata for both text and imaging, supporting tasks such as automated diagnosis, cross-modal retrieval, and explainable AI in clinical settings.

last updated 12 Jul 2025
by GoMask

The brief that made it

Training and benchmarking multi-modal AI models for automated diagnosis

Sample rows

preview · 8 of 200 rows · all 13 columns

note_idstring	image_typestring	match_confidencefloat	synthetic_flagboolean	modalitystring	image_idstring	patient_idstring	note_datedate	image_file_pathstring	image_datedate	body_partstring	diagnosisstring	note_textstring
N0001	X-ray	0.98	true	Siemens Ysio Max	IMG0001	P10001	2019-03-12	/images/xray/P10001_20190312.png	2019-03-12	Chest	Normal study	Chest X-ray shows clear lung fields and no acute pathology.
N0002	MRI	0.94	false	Siemens Avanto	IMG0002	P10002	2022-07-09	/mri/P10002_brain_20220709.nii	2022-07-09	Brain	Multiple sclerosis	MRI brain demonstrates multiple periventricular lesions suggestive of MS.
N0003	CT	0.89	true	Toshiba Aquilion One	IMG0003	P10003	2020-05-21	/ct/P10003_chest_20200521.dcm	2020-05-21	Chest	Viral pneumonia	CT scan of chest reveals bilateral ground-glass opacities, likely viral pneumonia.
N0004	Ultrasound	0.82	true	GE Logiq E9	IMG0004	P10004	2021-02-17	/ultrasound/P10004_abdomen_20210217.dcm	2021-02-17	Abdomen	Mild hepatomegaly	Ultrasound of abdomen shows mild hepatomegaly without focal lesions.
N0005	PET	0.71	true	Siemens Biograph mCT	IMG0005	P10005	2023-11-01	pet/P10005_20231101.img	2023-11-01	Chest	Lymphoma suspicion	PET scan reveals increased uptake in mediastinal nodes. Lymphoma cannot be excluded.
N0006	MRI	0.91	false	GE Discovery MR750	IMG0006	P10006	2021-06-05	/mri/spine/P10006_lumbar_20210605.nii	2021-06-05	Spine	Lumbar disc herniation	MRI spine shows disc bulge at L4-L5 with mild nerve root impingement.
N0007	X-ray	0.95	true	Canon Radrex	IMG0007	P10007	2018-12-26	/images/xray/P10007_20181226.jpg	2018-12-26	Knee	Mild osteoarthritis	X-ray of left knee shows early signs of osteoarthritis.
N0008	CT	0.93	true	Siemens Somatom Force	IMG0008	P10008	2020-01-13	/ct/P10008_head_20200113.dcm	2020-01-13	Head	Normal study	CT head shows no acute intracranial abnormality.
N0009	Ultrasound	0.79	true	Philips EPIQ 7	IMG0009	P10009	2022-03-23	/us/P10009_pelvis_20220323.dcm	2022-03-23	Pelvis	Complex ovarian cyst	Ultrasound pelvis demonstrates complex ovarian cyst, recommend follow-up.
N0010	X-ray	0.97	true	Philips DigitalDiagnost	IMG0010	P10010	2024-01-15	/images/xray/P10010_20240115.png	2024-01-15	Chest	Mild pneumonia	X-ray chest demonstrates left lower lobe infiltrates, consistent with mild pneumonia.
N0011	X-ray	0.96	true	Canon CXDI	IMG0011	P10007	2019-01-09	/images/xray/P10007_20190109.jpg	2019-01-09	Knee	Progressive osteoarthritis	X-ray knee follow-up shows progression of osteoarthritis.
N0012	MRI	0.92	false	Philips Ingenia	IMG0012	P10011	2023-04-01	/mri/P10011_head_20230401.nii	2023-04-01	Head	Chronic infarcts, atrophy	MRI head reveals chronic infarcts and mild cerebral atrophy.
N0013	CT	0.85	true	Hitachi Supria	IMG0013	P10012	2021-08-13	/ct/P10012_abdomen_20210813.dcm	2021-08-13	Abdomen	Normal study	CT abdomen shows no evidence of acute appendicitis.
N0014	MRI	0.89	true	Siemens Avanto	IMG0014	P10013	2022-11-18	/mri/P10013_head_20221118.nii	2022-11-18	Head/Neck	Carotid artery stenosis	MRI head indicates carotid artery stenosis on the left side.
N0015	Ultrasound	0.75	true	Samsung HS70A	IMG0015	P10014	2018-05-30	/ultrasound/P10014_liver_20180530.dcm	2018-05-30	Liver	Normal study	Ultrasound liver is normal in size and echotexture.
N0016	Other	0	true	blank	IMG0016	P10015	1900-01-01	/misc/P10015_19000101.file	1900-01-01	blank	blank	blank
N0017	PET	0.65	true	GE Discovery IQ	IMG0017	P10016	2021-09-21	pet/P10016_20210921.img	2021-09-21	Whole Body	Normal PET	PET scan shows physiologic distribution, no abnormal uptake.
N0018	X-ray	0.91	true	Shimadzu RADspeed	IMG0018	P10017	2020-10-14	/xray/P10017_leg_20201014.jpg	2020-10-14	Leg	Femoral fracture healing	X-ray of right leg shows healing femoral fracture.
N0019	CT	0.82	true	Canon Aquilion Prime	IMG0019	P10018	2019-04-17	/ct/P10018_pelvis_20190417.dcm	2019-04-17	Pelvis	Normal study	CT pelvis demonstrates no evidence of acute fracture.
N0020	Ultrasound	0.86	true	Philips EPIQ 7	IMG0020	P10019	2024-04-10	/ultrasound/P10019_leg_20240410.dcm	2024-04-10	Leg	No DVT	Ultrasound right leg negative for deep vein thrombosis.

What the 200 rows show

from the 200-row sample

PET (image type) stands out: mean match_confidence is 0.51, against 0.83 for the rest.

85%synthetic_flag = true
0.85median match_confidence
12body parts

Mean match_confidence by image_type200 rows

match_confidence200 rows, in bands of 0.1

Median 0.85, from 0.0 to 0.99.

modality192 rows with a value · 8 left blank

GE Discovery MR75014
Siemens Ysio Max13
GE Logiq E913
Samsung HS70A13
Philips Ingenia12
Siemens Somatom Force11
Canon CXDI11
GE Discovery IQ11
Siemens Avanto10
Siemens Biograph mCT10

13 columns by typefrom the column list below

string 9
float 1
date 2
boolean 1

Columns

13 columns in four groups

blueprint · 13 columns
column	type	description	example
Text 9 columns
`note_id`	string	Unique identifier for the clinical note entry.unique	N0001
`image_id`	string	Unique identifier for the diagnostic image associated with the note.	IMG0001
`patient_id`	string	De-identified unique identifier for the patient.	P10001
`note_text`	string	De-identified clinical note or report text describing the case.	Chest X-ray shows clear l…
`image_type`	string	Type of diagnostic image (e.g., X-ray, MRI, CT, Ultrasound).6 values	X-ray
`image_file_path`	string	File path or URI to the diagnostic image file.	pet/P10005_20231101.img
`modality`	string	Imaging modality or equipment used (e.g., Siemens Avanto, GE Discovery).optional	Siemens Ysio Max
`body_part`	string	Body part or anatomical region imaged (e.g., Chest, Head, Abdomen).12 parts · optional	Chest
`diagnosis`	string	Primary diagnosis or finding described in the note or image.optional	Normal study
Numbers 1 column
`match_confidence`	float	Confidence score (0-1) indicating the reliability of the note-image match.0 to 1 · optional	0.98
Dates and times 2 columns
`note_date`	date	Date when the clinical note was recorded.	2019-03-12
`image_date`	date	Date when the diagnostic image was captured.	2019-03-12
True or false 1 column
`synthetic_flag`	boolean	Indicates whether the entry is synthetic (true) or real (false).	true

Use it for

A healthcare dashboard
The synthetic_flag rate, match_confidence by image_type and a breakdown of modality. Excel, Power BI or Tableau.
Why do the 25 PET rows have a mean match_confidence of 0.51?
A root-cause class exercise
Hand out the rows and one question. The answer is in the data, not in the brief.
Notes200
N00010.98X-ray
N00020.94MRI
N00030.89CT
A software demo
Believable notes with image_id, patient_id and note_text to fill a screen in front of a buyer.

This dataset200 rows13 columns

Yours1,000 rows13 columns

blueprint · clinical-notes-image-matching

Behind this dataset

Same schema. As many rows as you need.

These 200 rows came out of a blueprint — 13 columns with generation rules behind each one. Open it in Data Factory to retune a column, add your own, wire in foreign keys, and run it at the size you actually need.

Rules it was built with

Each patient visit includes one clinical note and one image
Diagnosis codes align with provided imaging findings
Randomize patient demographics for privacy
Imaging modality varies (X-ray, MRI, CT)
Clinical notes contain both structured and unstructured text

Rows1,000 rows10,000 rows100,000 rows1,000,000 rows

Open the blueprint in Data Factory

1 credit per row. Preview 20 rows free. Buy rows for the full run.

Exports: CSV, JSON, JSONL, Parquet, SQL, Excel, TSV, XML
Licence: yours to use, including commercially
API slug: clinical-notes-image-matching

What should your data show?

No signup. No card.

Sample rows

What the 200 rows show

Columns

Use it for

A healthcare dashboard

A root-cause class exercise

A software demo

Make it yours.

Same schema. As many rows as you need.

More in Healthcare

What should your data show?