A Survey of Medical Device and ICT Links in Medicine

Aspects of Medical Device and ICT Integration (12)

• Instrument interfacing

• Continuous data logging

• Digital medical imaging systems

• Medical device control

• Process automation

• Hospital ICT systems

• Robotics

• Simulation

• CAD Artificial Intelligence Machine learning

• Telemedicine

• Prosthetics

• Smart apps

Instrument Interfacing

Linking instruments to computers so that data can be captured automatically into computer memory

What do we mean by ‘digital’ in instrumentation?

Use of discrete signals and numerical values for measurement processing and data representation

Challenges of Medical Instruments

• Signals from medical instruments can be complex (e.g., ECG/EEG)

• Larger signal complexity leads to larger computer file sizes

Large file sizes cause (4)

• Memory requirements issues

• Long signal processing times

• Long data transfer times

• Problems with email attachments

Solutions in Medical Instrument Signal Processing (2)

• Time Sampling

• Quantisation of the Voltage

Time Sampling (5)

• Continuous (analogue) real-world signals

• Not all values saved, only a sample

• Low sampling frequency reduces file size

• High sampling frequency reduces uncertainty but increases file size

• Too low a sampling frequency causes signal distortion

Quantisation of the Voltage (5)

• Limiting voltage values to save file size

  • (‘quantising’ the measurement range)

• 1 byte of memory saves 256 levels

• # of Quantisation levels = 2ⁿ

  • n is bit-depth of the ADC

• Higher bit-depth improves precision but increases file size

ADC

Analogue-to-Digital Converter

DAC

Digital-to-Analogue Converter

Nyquist Theorem

To avoid output signal distortion the sampling frequency must be at least double the highest frequency in the Fourier Spectrum of the original analogue signal

Calculating Filesizes - Suppose we want to save a 5 minute long signal as a digital file (3)

• Sampling frequency = 100 Hz (100 samples per second)

• Bit depth = 16 (2 bytes per measurement)

Formula: 100 samples per second × 300 seconds × 2 bytes per sample = 60,000 bytes = 60 kB

ADC and DAC in Signal Processing (3)

• ADC:

  • Converts analogue signals to digital

• DAC:

  • Converts digital signals back to analogue for display on output devices (e.g., computer monitor or printer)

  • operation smooths the output by connecting sampled points

Advantages of digital systems (2)

• Easier signal processing (e.g., amplification, smoothing, filtering)

• Processing is programmable and can be automated

Continuous Data Logging (2)

Automatic collection of data at a preprogrammed sampling rate over an extended time period
Fx.: Holter-ECG monitoring; CGM

Holter-ECG monitoring (4)

• Portable device worn by a patient during normal activity

• Electrocardiograph interfaced to a data capture system

• Stores up to 24 hours of ECG data

• Useful for detecting cardiac arrhythmia not visible in a brief ECG at a doctor's office

Continuous Glucose Monitoring (CGM) Components (4)

• Sensor

• Transmitter

• Smartphone app, receiver

• Insulin pump

CGM Sensor

A small piece of material inserted under the skin to measure glucose levels, lasting 7-15 days or months for implantable systems

CGM Transmitter

Wirelessly sends glucose data from the sensor to a viewing device

CGM Smartphone app, receiver, or insulin pump

Displays real-time glucose levels, trends, and history, with some systems compatible with insulin pumps

Continuous Glucose Monitoring (CGM) (3)

• Wearable technology that tracks glucose levels in real-time

• Measures glucose in interstitial fluid just under the skin 24/7

• Primarily used by people with diabetes for managing glucose levels

Digital Medical Imaging Systems

• measure the variation of a physical variable (e.g., density, temperature) across the patient's body

• Not all values can be measured due to large file sizes

• fx.: Spatial Sampling

Spatial Sampling definition

The process of sampling a continuous signal (such as a medical image) at discrete intervals in space, typically measured in pixels per unit of distance (e.g., pixels per cm)

Spatial Sampling (3)

• The higher the spatial sampling frequency (pixels per cm) the more detailed the image

• A higher number of megapixels increases spatial sampling frequency improving image detail

• A higher spatial sampling rate improves image detail but increases file sizes

Medical Imaging Device definition (3)

• Instruments with a measuring function

• The device splits the patient into a 3D set of voxels (like a Rubik's cube)

• Each voxel represents a physical property (e.g., density, temperature)

Tomogram

A 2D plane (slice) of voxels

Linear Attenuation Coefficients in HU (Hounsfield Units) (4)

• measure how much X-ray radiation is absorbed or scattered as it passes through a material

• In CT imaging, these coefficients are used to calculate Hounsfield Units (HU)

  • represents tissue density

• HU values help differentiate between various tissues

Visualizing Physics Patient Voxel Data (photo!)

Medical Device Control (5)

Medical devices are controlled by instructions programmed into a computer:

• Example Pathology Lab

• Neonatal incubator

• CT scanner

Automated instructions ensure precise + consistent imaging → ↑ accuracy +efficiency of the scan

Example Pathology Lab (2)

• take measurements of temperature every 60s

• switch off heater when temperature of sample reaches 40 degrees

Neonatal incubator (2)

• take measurements of temperature every 60s

• maintain temperature within a given range

CT scanner (4)

• move patient by a specific distance - 10mm

• rotate x-ray tube

• collect x-ray tomogram data

• repeat sequence from diaphragm to knees

Automation in the Medical Lab (4)

• BOD (beginning-of-day):

  • automatic start-up and warm-up of instruments and automatic calibration check

• Automatic sorting of different laboratory samples

  • (using colour, bar-code systems, etc)

• Automatic checking of reagent levels and replenishment

• Automatic number of repeat acquisitions on each sample and calculation of average and uncertainty

Levels of Medical Laboratory Automation (2+1)

• Total Laboratory Automation (TLA)

• Modular Laboratory Automation (MLA)

Due to the high cost of TLA, the trend is towards MLA

Total Laboratory Automation (TLA) (3)

• complete automation from specimen input through pre-analytical, analytical phases to final result report generation

• repeat runs are done automatically

• extremely expensive

Modular Laboratory Automation (MLA)

workflow in the laboratory is analyzed to identify 'bottle-necks' and automate those specific parts

Hospital ICT Software Types (2)

• Patient Information Systems (PIS)

• Medical image management

Patient Information Systems (PIS) (4)

manage and store patient-related data in various departments:

• Laboratory Information Systems (LIS) in Pathology

• Radiological Information Systems (RIS) in Radiology

• Oncology Information System (OIS) in Oncology

Medical image management (3)

deals with the storage and handling of medical images like CT, MRI, USI, and digital microscopy

• PACS (Picture Archiving and Communication Systems) for managing medical images

• DICOM (Digital Imaging and Communications in Medicine) file format used for medical images (e.g., .dcm files)

Simple Robotics

Medical Laboratory robots that pipette samples and reagents from containers and transfer them into cuvettes for measurements

More complex Robotics

Surgical robotic systems
fx.: Da Vinci System

Very complex Robotics

Oncology linac X-ray beam radiosurgery systems

CyberKnife® (5)

• part of Robotic ‘Radiosurgery’ Systems

• components:

  • Small linear accelerator producing 6MV beams

  • Robotic arm to aim the linac towards the tumour from various angles

  • X-ray imaging and optical devices to continuously track patient and tumour position in real time for accurate radiation targeting

Simulation (3)

• Practice difficult or dangerous procedures on virtual patients

• Create virtual personalized models for planning procedures like radiotherapy

• Use CT scanner and linac simulator for virtual treatment planning

Type of Simulation (3)

• Breast Treatment Plan

• Cancer Treatment Planning

• Treatment plan for an optic nerve

Breast Treatment Plan (photo)

Cancer Treatment Planning

Treatment plan for an optic nerve

AI (5)

• CAD = Computer Aided Diagnosis/Detection

  • uses AI to assist in diagnosing or detecting diseases based on medical images

• ML = Machine Learning

  • predicting outcomes or aiding in diagnosis

  • algorithms that learn from data and improve over time

Telemedicine (3)

• Use of telecommunication systems to provide medical information and services

• Can range from simple consultations between health professionals to complex telesurgery

• Involves satellite technology, videoconferencing, and remote surgical robot manipulation

Prosthetics

Design, creation, and fitting of artificial limbs or devices to replace missing body parts

Health Care Apps for Smartphones (2)

• These apps are considered medical devices

• Fx.: Welldoc system

  • for managing type 2 diabetes, approved as a medical device