AI has demonstrated remarkable progress and received widespread recognition for its usefulness in the medical imaging world. AI technologies -- machine learning in particular -- automate image reading, repetitive tasks and the reporting process, which leads to an increasing efficiency in clinical workflow. The use of AI technologies allows radiologists to see and compare similar cases and learn from other specialists, which increases the overall accuracy diagnoses and enhances the individual radiologist’s expertise. It prioritizes emergency room MR studies so that the most severe cases requiring urgent referral are flagged for quick interpretation by the radiologist. Finally, incorporation of AI tools offers solutions to the growing issue of radiologist shortages.

3D printing technologies have also come along the way over the past decade and have been used in various areas, including medical imaging. The images made by Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) are often used by clinicians to more effectively diagnose patients’ conditions. The image quality produced by these two modalities has improved tremendously in recent years, from two-dimensional (2D) multislice to three-dimensional (3D) imaging sequences. The development of radiological imaging from 2D to 3D aligns with the development in 3D-printing, which can transform the imaging data from virtual to physical models.

3D printing produces a three-dimensional model of anatomical structures, which allow surgeons to define and simulate a surgical plan, including device selection and surgical equipment, before they enter the operation room. These simulations reduce the risk to the patient, according to one study from the Mayo Clinic.

The application of utilizing 3D printing for pre-operative planning at the hospitals, however, is not widely adopted yet, and the segmentation process is one of its impediments. Segmentation is the process of partitioning and categorizing images into body regions with similar properties (e.g. gray level, contrast, texture) to identify regions of interest (e.g. abnormalities). This process is one of the crucial steps to producing a printed 3D image, but it is an extremely time-consuming process. Machine learning could help to provide an automation of the segmentation process. By utilizing AI for the segmentation process, the time takes by the radiologist to do this activity could be reduced or eliminated.

Informative and personalized treatments to patient

Patient education, which is one of the top priorities for most healthcare providers, could be also one of the beneficial results of a marriage between AI and 3D printing. Compared to 2D images, 3D printed models are obviously superior in terms of helping the patient to more readily understand a conversation with the doctor or surgeon. Consequently, patients would better understand their illness, leading to an informed treatment decision and a better-informed consent process. Improved doctor-patient communication, along with the numerous other benefits to radiologists and surgeons, are reasons for hospitals and outpatient clinics to consider the simultaneous use of AI and 3D printing technologies.

Author: Nia Ammann