top of page

Meet Our Software Engineer - Jasser Dhaouadi

In our "Meet the Team" series, our knowledgeable teammates discuss the major trends, impact, and technology of artificial intelligence in the diagnostic space. Today we speak with Jasser Dhaouadi, our Software Engineer.

You have a very interesting background, you’ve worked and lived all over the world, what inspired you to join the Bot Image team in Omaha, NE?

Bot Image team is leading the way to a bright future where the early detection of prostate cancer is possible. Bot Image’s unique perspective has inspired me to serve humanity in a better way. I am very impressed by the commitment of the team to change the world and revolutionize the healthcare industry by providing hope to the people who are living in the darkness of prostate cancer.

What initially attracted you to the field of healthcare AI?

We are living in a computerized era, where AI is the new oil of the century. As an engineer with a big focus on Data Science and Machine Learning, I have been implementing a wide variety of AI techniques and I am fully aware of the AI capabilities. One of AI's biggest potential benefits is to save lives and improve the ability of healthcare professionals to better monitor and detect diseases at more treatable stages. The potential of healthcare AI is limitless and it can go beyond our imagination.

Since joining the team you have made some significant improvements to the algorithm, how do you identify areas for improvement?

I have developed an intelligent pipeline that automates the different phases the software goes through before generating a color map revealing the likelihood of prostate cancer on top of an MRI image. Thus, improving the usability of ProstatID by making the processing steps seamless to the user.

In addition, I have introduced the concept of microservices to the software to reduce the sophistication of ProstatID and avoid the tight coupling of its internal components. Besides, I have developed a distributed architecture for the software increasing its availability and boosting its resilience to errors.

Furthermore, I have designed several deep learning models and implemented innovative neural nets architectures enhancing the capabilities of ProstatID by improving its sensitivity and specificity. I have performed an explanatory data analysis over the generated MRI features in order to identify critical patterns. The results are promising.

What do you think is the most innovative aspect of the ProstatID algorithm?

ProstatID enhances the performance of healthcare professionals and makes them understand better the needs of their patients. The software enables doctors to oversee early-stage prostate cancer while reducing the need for unnecessary biopsies. With that understanding, ProstatID supports clinical decision-making with appropriate cancer predictions and it puts physicians in control of health and well-being of their patients.

The ProstatID algorithm use interprets multi-parametric and bi-parametric MRI using random forest with instance weight and MR segmentation by deep learning with holistically-nested networks, can you explain what that means to those of us with a less technical background?

ProstatID is an AI software that takes a specific set of MRI modalities, more specifically T2w, ADC and DWI, as input and generates a probability map with colors highlighting the location and the severity of prostate cancer. Before the inference step, the software performs a variety of data preprocessing techniques such as input quality check, co-registration and normalization of MRI volumes, prostate segmentation and feature computation. ProstatID implements the random forest as a learning model. The random forest was trained on a set of well-studied cases and learnt on its own the critical patterns leading to a cancer within a prostate. During the inference step, the trained model is fed by a set of image features characterizing the input MRI modalities and predicts the existence of cancer within an MRI image.

When designing the product delivery system, how did you ensure the product platform would be user-friendly?

The centralized architecture of ProstatID makes it easy for the user to get the benefits of the software services. The user is not required to install a single piece of software or worry about compatibility issues. The software is available to a list of authenticated users through the cloud. Bot Image team is charged to establish a secure tunnel of communication between the user PACS and the local PACS of ProstatID. The main design principle of ProstatID is to make everything simple to the user. In fact, the internal sophistications of the software are seamless and not exposed to the user. Just a simple click is needed to send a medical study to ProstatID. Furthermore, the software informs the user, in real-time, via email about the different processing stages of the medical study. At the end, the user should expect to receive on his/her PACS a probability map with meaningful colors revealing the likelihood of cancer and a summary report that includes a 3D visualization of any detected tumor within the prostate.

Who do you think will benefit the most from diagnostic AI tools like ProstatID?

One of the major benefits of diagnostic AI tools is to help people stay healthy and sound. The early detection of a cancerous disease can save lives and the high accuracy of AI tools can avoid causing harm to patients by reducing the need for unnecessary biopsies.

AI can help healthcare professionals understand better the needs of their patients, take a more suitable approach for disease management, better coordinate care strategies and help patients to better comply with their long-term treatment plans.

You are very passionate about the transformative power of AI, what do you see as the biggest impact of AI in the diagnostic healthcare space?

The biggest impact of diagnostic AI is improving the healthcare space. Using appropriately the tremendous amount of health data trains AI to learn patterns and trace conclusions. AI supports clinical decisions, suggests actions prioritizing administrative tasks and reduces medical costs. AI puts humans in a better control of health.

Bot Image has several upcoming products in the pipeline, what are you most excited to work on next?

Being considered the leading cause of cancer deaths worldwide, lung cancer is a challenging type of cancer that I feel excited to work on. Symptoms of lung cancer typically occur when the disease is advanced. Hence, AI would be a great approach to catch signs during its early stages.

What are your predictions for the future of healthcare artificial intelligence?

AI can review facts, store information, adapt to changes and learn patterns exponentially faster than any human. Using pattern recognition and predictive analytics, AI might outperform the best clinicians and researchers all around the world. Besides, the combination of machine learning and neuroscience may give birth to a powerful general-purpose learning algorithms that exceed the capabilities of the human brain and solve accurately real-world healthcare problems.

370 views0 comments
bottom of page