Detecting breast cancer early improves patient outcomes and survival rates. However, traditional screening methods like mammograms often come with significant drawbacks.. Patients are forced to wait a long time for results, which can cause anxiety and uncertainty.
Singapore-based company FathomX has recognized the need for innovative solutions in breast cancer detection. This company is using the power of artificial intelligence (AI) to revolutionize mammography screenings, particularly in its home country as well as in Southeast Asia generally.
With a goal to make early diagnosis easier and streamline the screening process, FathomX has the potential to transform breast cancer screenings and detections worldwide. Let’s look closer at its FxMammo product, which uses AI to detect breast cancer in acquired mammography images.
Challenges in Traditional Mammography Screening
Traditional mammography screening methods present significant drawbacks due to their reliance on manual readings, which are susceptible to human error. This approach results in a false positive rate of between 7 and 12 percent, which varies based on the age of the patient. The inaccuracies inherent in manual interpretations mean we need more advanced screening methods to improve diagnostic accuracy.
Additionally, the screening process takes a long time, with many patients experiencing lengthy waiting times before receiving results. Some patients have to wait up to two weeks in standard screening programs. The uncertainty surrounding a patient’s health status during this period can be emotionally taxing. There is an urgent need for technologies that expedite the screening process and provide timely, accurate results.
Artificial Intelligence in Breast Cancer Detection
FathomX’s advanced AI technology has the potential to revolutionize breast cancer detection and diagnosis. Its FxMammo platform uses AI for a more precise and efficient screening option. This platform uses deep learning algorithms and advanced image processing techniques to analyze mammograms with unparalleled accuracy, with significantly fewer false positives and negatives. This innovative approach allows for more accurate diagnoses and accelerates the screening process, which could ultimately improve patient outcomes. Currently, AI does not outperform human professionals, but when radiologists use them to inform their decision-making processes, accuracy improves.
Using AI in breast cancer diagnosis also addresses the pressing issue of waiting times and patient anxiety. FathomX’s AI technology allows for faster interpretation of mammograms and prompt delivery of results. These minimized waiting times and timely diagnoses can help alleviate patient anxiety and increase overall satisfaction with the screening process.
FathomX’s Approach
FathomX’s FxMammo product is a radiological computer-assisted breast cancer detection software that was developed through collaborative research between FathomX and leading healthcare institutions like the National University of Singapore (NUS) and the National University Health System.
FxMammo is a sophisticated AI assistant that allows for easier and more accurate breast cancer detection through augmented mammography screening. The platform draws on deep learning algorithms developed over three years of research. It has been proven to significantly reduce the occurrence of false positives and false negatives in mammogram interpretations.
This technology addresses a critical limitation of manual readings in traditional screening tools, where false positive rates hover around 9 percent. This leads to unnecessary anxiety and additional diagnostic procedures for patients. FxMammo’s ability to accurately identify suspicious cases on mammography images streamlines the screening process and provides clinicians with precise insights, enabling timely interventions for early-stage breast cancer patients.
Collaborative Testing and Development
FathomX has partnered with Hewlett Packard Enterprise (HPE) and other key collaborators to test and develop its AI-assisted mammogram solutions. The company is using HPE’s high-performance computing systems and advanced GPU models to refine its AI algorithms and make them more effective in real-world applications.
The development and testing of FathomX’s platform requires high computational power. To meet this need, the company is using HPE computing platforms like HPE Edgeline EL1000 Converged Edge systems equipped with NVIDIA® T4 graphics processing units (GPUs) and the HPE Apollo 6500 system with NVIDIA A100 GPUs. These high-performance systems are capable of rapidly processing large volumes of medical images used to train AI models and to detect abnormalities, such as microcalcifications in breast tissue.
The collaboration with HPE also includes the creation of a secure production-level data infrastructure to support the training of AI models. This infrastructure lets FathomX use larger batch sizes of images during training, so the AI can learn faster and improve its accuracy in identifying lesions.
For on-site, real-time diagnosis, FathomX uses HPE Edgeline systems. This allows the program’s AI assistant to analyze mammograms directly at clinics or hospitals. This approach eliminates the need for data transfer to central servers, reducing cybersecurity risks and processing delays. Deploying these compact and powerful devices on-site gives healthcare providers the opportunity to obtain diagnostic results quickly and securely.
Prospects and Expansion
Currently, FathomX’s operations are based in Singapore and primarily focused on the Asia Pacific region. However, the company hopes to broaden its reach across this region and potentially into other global markets. This expansion may involve deploying AI-assisted mammogram technologies in diverse healthcare settings. The long-term goal is to provide faster, more accurate breast cancer screening solutions to a broader audience, reduce mortality rates, and improve patient outcomes worldwide.
FathomX is also exploring the application of its AI technologies to other areas of medical imaging. The company hopes to develop AI-assisted diagnostic tools for other types of cancer and medical conditions.