Category Archives for "AI Software"

Artificial Intelligence in Clinical Radiology

Artificial Intelligence in Clinical Radiology

One of the largest issues in veterinary radiology today is an incredibly high caseload. Radiologists in North America consult on 2.5 million cases per year, and that number is projected to more than double within the next three years. With fewer educational opportunities available in radiology, how can this vital specialization keep up with the demand?

This week on the Veterinary Innovation Podcast, Shawn and Ivan speak with Dr. Seth Wallack, the founder and CEO of Vetology, about how artificial intelligence can improve the workflow of clinical radiologists, whether we’re too late in adopting it, and how the best veterinarians are those who are most eager to learn.

Topics Covered

  • The Shrinking Number of Radiologists and How AI Can Fill Those Gaps
  • The Adoption of New Technologies in Clinical Radiology
  • Higher Amounts of Specialization in Veterinary Medicine

Innovations in veterinary radiology: using AI to improve patient care

Innovations in veterinary radiology: using AI to improve patient care

The implementation of artificial intelligence in the field of veterinary radiology allows doctors to receive crucial diagnostic information almost immediately.

The world of medicine is constantly changing, and the veterinary medical field is no different. Technology is advancing rapidly, and the rule is to adapt to these changes or succumb to the consequences. The use of artificial intelligence (AI) in veterinary radiology is a relatively new area. Leaders in this field are ethically responsible for providing correct product knowledge to the veterinary community, and should follow the principles of transparency, honesty, and integrity. At the same time, veterinary professionals eager to use this new technology must understand that the field is ever-evolving. As such, offered products will be in different development stages. Good Machine Learning Practices (GMLP) should be adhered to and documented.1

AI in veterinary medicine – potential areas of impact

Implementing an artificial intelligence strategy is a must for veterinary practices moving into the future. Access to specialists is not always available to hospitals or clients, and AI offers an attractive solution. However, experts believe that depending solely on AI can be detrimental. A strategic combination of both human competency and AI technology is important to drive the best care.

Some of the products related to AI in veterinary practice have already been in the marketplace for years. These are expected to gain immense popularity in the near future. The market for global wearable devices for the remote monitoring of pet health and activity is forecast to reach over $8 billion by year 2025.

When it comes to the human health market, similar products have gained in popularity. These are known for measuring simple parameters such as movement, heart rate, and body temperature. In addition, these devices keep tabs on food intake. They will make recommendations based on appropriate behavioral responses. For instance, these devices will inform diabetic patients about requirements for glucose or insulin.

New devices with the capability of measuring other important parameters will soon be available. In the veterinary field, cattle can be fitted with movement sensors to identify the onset of estrus. Similar technology will no doubt become available for other species, including pigs. At present, these sensors are quite expensive, especially for routine usage. Special efforts are being made to produce low-cost versions of the technology that would make them more accessible to farming operations.

This process will emulate the rapid decrease in the cost of genome analysis, resulting in possible developments of custom-made medicines in animals and human patients. Apart from identifying individuals with genotypes that may make them more or less vulnerable to the effects of specific drugs, professionals will also carry out genomic analyses on microbiome samples from the skin, gut, and other sites to evaluate which disease-causing organisms may exist in the patient’s body. There have been speculations on whether these approaches will ultimately lead to a reduction in the demand for antibiotics.

Advancements in monitoring technologies will similarly lead to challenges in processing and applying the available data. It is believed that doctors will have access to 200 times more data than the human mind can process. Consequently, another priority is the development of artificial intelligence software that analyzes this mass of information and draws precise conclusions about its meaning.

Google and Apple have special teams working on these issues. Significant development has already been seen; a team has produced diagnostic software that can easily identify human patients with early indications of diabetes. It is based on constant measurements of heart rate variation. This technology and others will change the way clinical practice operates.

Current applications

1. Medical imaging processing and assessment

The use of AI in this area includes quick yet precise and sensitive interpretations of radiographs, MRI images, CT scans, ultrasound images, and cytology assessments. At present, all aspects of AI are progressing exponentially, from computer processing power, speed, and affordability, to the development of machine vision reference directories. Hence, it is expected that most standard tasks involving clinical interpretation in veterinary practice will be allotted to AI. This will help veterinarians obtain quick, accurate, and detailed reports as well as consistency in interpretation, a factor that currently depends on the experience and skill level of the individual practitioner.

2. First-line primary consultations

This is another area where AI shows enormous potential. For human patients, smart kiosks are available that cut down on wait times, a major source of dissatisfaction among patients and stress for physicians. However, a combination of AI and detailed AR (augmented reality) instructions ensures a consistent, accurate, and detailed collection of crucial patient history and physical examination data. This is provided and collected by the patients themselves while they are guided through the entire procedure. After collecting the relevant data (usually completed in under 15 minutes), the doctor is sent a detailed patient work-up that includes a proper breakdown of predicted illnesses and treatment options. The medical practitioner then conducts a video consultation in order to confirm the authenticity of the AI-collected information, verify the diagnosis, and approve or alter the treatment plan. AI will also help maintain detailed and accurate healthcare records, and will automatically follow up with patients within a few days of the consultation.

Special efforts are being made to roll out veterinary-related versions of this system. Experts in the field believe that this technology will significantly augment the experience of clients and their pets. Implementation of this technology will also improve the professional lives of veterinarians, who face similar issues of stress and overload as their human physician counterparts.

Using artificial intelligence software provides an attractive option for all doctors, allowing them to receive crucial diagnostic information almost immediately.

Author disclosure statement: Eric Goldman is President of Vetology AI, a company that designs and delivers service innovation for the veterinary industry. He has a financial interest in Vetology Innovations LLC. For more information visit vetology.ai.

_______________________________

1American Association of Veterinary Radiologists guidelines, submission and review process for Veterinary Radiology Artificial Intelligence (AI) AAVR GMLP SaMD Product Certification — American Association of Veterinary Radiologists. Available at: http://www.aavr.org/index.php?option=com_content&view=article&id=99&Itemid=263.


Vetology In the News

Vetology In the News

The American Medical Veterinary Association interviewed Dr. Seth Wallack, Founder of Vetology and Eric Goldman, President of Vetology. We discussed with the AVMA the challenges of not having enough radiologists to meet the demand. This is not just a problem for our teleradiology practice, but a global problem that impacts general practice managers, specialty and emergency hospitals. Turn-around times for obtaining a radiology over-read by a boarded radiologist can take upwards of 8 hours to 5 days. A 2018 study predicted that, by the end of 2022, 66 percent of the North American veterinary teleradiology caseload will not be met.

Vetology recognized the power of AI technology to automatically read images and interpret results. In 2017, Vetology developed software to use visual object resolution to help diagnose diseases. The techniques were first used internally to support our own radiologists to be more efficient and meet demand. In 2018, the capabilities were expanded and we were the first to market to offer an augmented interpretation of radiographs using artificial intelligence and machine learning. This allowed for a complete radiograph interpretation in 5 minutes or less and on par to that of a human veterinary radiologist.

The AVMA identified that artificial intelligence has the ability to speed up radiology interpretations and provides a substantial benefit to both the patient health and the DVM. Faster, more reliable results mean faster diagnoses, better treatment, and healthier pets.

To see more details read the full article the AVMA

Artificial intelligence & veterinary medicine

Can machine learning live up to expectations?

Radiology Technician

How Intelligent Is Artificial Intelligence?

How Intelligent Is Artificial Intelligence?

AI For Human And Veterinary Radiology

Radiology Artificial Intelligence, commonly referred to as AI, is in full development, and the FDA is actively testing AI in human radiology. In veterinary radiology we’re not far behind and are quickly catching up. Veterinary medicine AI product development’s greatest strength is also its greatest weakness. That is, oversight, or more specifically, the lack of oversight by a governing body.

In human radiology, AI products require FDA oversight and approval prior to coming to market (see the American Association of Veterinary Radiologists GMLP write-up here to learn more). Some veterinary products do fall under FDA oversight BUT veterinary radiology AI isn’t one of them. This means a veterinary radiology AI company can develop quickly but also have no formal obligation to demonstrate their product actually works.

If a company doesn’t provide clinical test results, it is up to you, the veterinarian, to determine the worthiness of the product. A true 'caveat emptor.'


Assessment Is Critical

So how should a veterinarian assess a veterinary radiology AI product? The same way the entire medical community evaluates any diagnostic test for a condition, by measuring clinical performance.

The two standard measures of clinical performance are SENSITIVITY and SPECIFICITY. To better understand how these measures can assist us to asses clinical performance, we will briefly revisit a couple of formulas from that old favorite - statistics class.


Sensitivity And Specificity

SENSITIVITY is the probability that a test will identify a patient who HAS a condition (true positive). It is calculated by the following formula:

Sensitivity = True Positives / (True Positives + False Negatives)


SPECIFICITY is the probability a test will correctly identify a patient who DOES NOT have a condition (true negative). It is calculated by the following formula:

Specificity = True Negatives / (True Negatives + False Positives)


The Confusion Matrix

These two standard measures of clinical performance lead us to the four outcomes possible for each patient:

True Positive, False Positive, True Negative, False Negative


The four outcomes are typically reported in a 2 x 2 table called a confusion matrix showing the total numbers of true and false positives and negatives. A generic example is shown here next.

A Typical Confusion Matrix

Name of Condition or Disease

Total Number of Cases Measured

% Sensitivity

% Specificity


Radiologist Positive

Radiologist Negative

AI Positive

# of cases

# of cases

AI Negative

# of cases

# of cases


Vetology's AI Testing

Vetology’s AI testing evaluates AI results against veterinary radiologist reports as a reference standard. The results and confusion matrix tables are displayed just below on this page.


The Truth Is In The Confusion Matrix

To help you evaluate a product’s performance, always ask an AI vendor for their confusion matrix tables.

Also, be aware that an AI product MUST be 100% autonomous to have a valid result. If a human intervenes during any part of the result creation, it’s not artificial intelligence, it’s human intelligence.

Next we show the confusion matrix for several diseases among 75 random cases:

Cardiomegaly

47 Cases

90% Sensitivity

76% Specificity


Radiologist Positive

Radiologist Negative

AI Positive

17

19

AI Negative

3

8

Heart Failure

39 Cases

100% Sensitivity

89% Specificity


Radiologist Positive

Radiologist Negative

AI Positive

3

4

AI Negative

0

32

Dynamic Airway Pattern

30 Cases

100% Sensitivity

85% Specificity


Radiologist Positive

Radiologist Negative

AI Positive

3

4

AI Negative

0

23

Dynamic Airway Collapse

30 Cases

67% Sensitivity

93% Specificity


Radiologist Positive

Radiologist Negative

AI Positive

2

2

AI Negative

1

25

Buyers Beware

As we said earlier in this article, AI brings to the forefront of your purchase decision-making, the phrase 'caveat emptor,' or buyer beware. Make sure you review the provider's confusion matrix tables (if they have them at all), and make sure their AI is fully autonomous, else you'll just be buying expensive human intelligence.

At Vetology, we assertively and proactively test ourselves and continually train and improve the AI for everyone's benefit. We are as transparent as possible. We have the data and are willing to publish it.

If you have any questions about our veterinary radiology software or services, we encourage you to reach out to us via email here.

AI Machine Learning

Veterinary Radiology Artificial Intelligence Software Made Simple: A Step-By-Step Guide

Veterinary Radiology Artificial Intelligence Software Made Simple: A Step-By-Step Guide

A Brief History

Need a brief history of AI software in general here.


Teleradiology

Need a history of the current state of veterinary teleradiology here.


Image Match

Need to discuss image matching; its strengths and drawbacks.


AI In Medicine Is Growing Very Fast



AI For Veterinary Radiology

This is where we would pivot to the specific application of AI and Machine Learning applied to veterinary radiology


Sensitivity

Need to explain what sensitivity means.


Specificity

Need to explain what specificity means.


Confusion Matrix

Need to explain our confusion matrix and enough to give confidence.


Vetology's AI Guardian Software

Need to give overview of our software.


Intended Use

This is where we would provide intended use


Post-Market Monitoring Plan

What are our post-market ongoing monitoring, reporting and corrective plans? This should consider latest data and new data sources.