Skip to content
Spero Science Innovation — Brazil

Decision support

Clinical data, imaging and genetics in a single read

Success case: knee osteoarthritis

Spero AI brings the three sources together and shows the reasoning behind each result. Knee osteoarthritis is the case already built. The same method can be trained on data from patients, hospitals and companies, for areas such as chronic diseases, lifestyle medicine, orthopedics, cardiology, oncology and more.

Talk to the team
Spero AI illustration: a stylized brain drawn as a neural network that brings together clinical data, imaging and genetics
Spero AIOne reading · three sources
Clinical data
Guided questionnaire
Imaging
Radiograph
Genetics
Risk score

Spero Science Innovation — Brazil

Spero AI is the healthcare artificial intelligence platform of Spero Science Innovation, a Brazilian company. This is the platform website; you can also visit our corporate website. About Spero Science Innovation

Assessment performed with the support of artificial intelligence. It does not replace a medical consultation, diagnosis or report.

Already in production

Clinical data
93%

sensitivity of the clinical model for surgical referral, from a guided questionnaire

83% specificity

Imaging
94.9%

within one grade on X-ray grading

5,175 knees in a test set never used for training

Imaging
68.6%vs 60%

exact grade accuracy: our model versus the best specialist radiologist

Model: 5,175 test knees · Specialist: Thomas et al., Radiology: Artificial Intelligence, 2020, 50 OAI radiographs

The imaging dataset behind the numbers

18,043

radiographs

4,494

participants

What we assess

Three fronts, one read

Osteoarthritis is not explained by a single source. Spero AI reads all three and brings the result together.

in production

Clinical data

Guided questionnaire on pain, function and history, with a factor-by-factor explanation.

in production

Imaging assessment

Today, radiography: Kellgren-Lawrence grading with a map of the regions that mattered.

coming soon

Genetics

Risk score from variants associated with disease progression.

How the AI reads a knee

Three steps, all visible to the physician and the patient.

  1. 1

    Receives the exam

    Standard image or DICOM radiograph, plus the patient's clinical questionnaire.

  2. 2

    Grades and explains

    Distribution across the five osteoarthritis grades (Kellgren-Lawrence) and a heat map of the regions that mattered.

  3. 3

    Hands over the decision

    An editable description of the findings, for the physician's assessment.

Spero AI X-ray result: right knee, KL 4, with heatmap and grade distribution
Output of the Spero AI X-ray model.

From innovation to impact

Secure, effective solutions built for the regulatory path.

  1. 1

    Algorithm development

  2. 2

    Data collection and validation

  3. 3

    Proof of concept

  4. 4

    MVP / product on the market

Who it is for

Hospitals, payers and companies

Service provision: patient portfolio assessment for decision-making, integrated into the care workflow.

Talk to the team

Patients

A self-knowledge tool: answer the questionnaire and/or upload your X-ray and bring the read to your doctor.

Open the store

Physicians and clinics

Clinical data and image intake by QR code or phone, with real-time results on the physician's screen. Physicians get free assessments to try the platform.

Start now

The model learned from more than 300 hours of reading.

Let's talk

Tell us your context.

Talk to the team