A filament 3D printing approach for CT-compatible bone tissues replication.

Standard phantoms used in medical imaging often fail to replicate the complex X-ray properties of human tissues. While prior research has shown that 3D printing can produce anatomically accurate structures, few studies have focused on matching radiological characteristics of bone tissue. This gap motivated the development of a method that integrates material mixing with medical imaging data. Existing approaches typically use single materials or lack precise control over radiodensity. No prior work had resolved how to dynamically mix filaments during printing to match bone Hounsfield units. The challenge lies in translating CT data into printable material compositions. Current phantoms may lack the necessary attenuation properties for accurate CT calibration. This paper addresses the need for a reproducible, CT-compatible phantom fabrication method.

The goal was to create a 3D printing methodology that can replicate the X-ray attenuation of human bone tissue. This approach aims to produce anthropomorphic phantoms suitable for CT imaging applications. The specific problem addressed is the lack of phantoms that accurately mimic both the anatomical and radiological properties of bone. The motivation stems from the need for reliable calibration tools in medical imaging. The method seeks to bridge the gap between anatomical accuracy and radiological equivalence. By integrating CT data directly into the printing process, the study aims to improve phantom realism. The focus is on hip bone structures, which are critical in abdominal CT imaging. The study aims to demonstrate that real-time filament mixing can achieve this goal.

The study employed a dual-filament extrusion setup to mix PLA and Stonefil during printing. A custom software linked CT scans directly to the 3D printing process. Three hip bone phantoms were printed under varying conditions. The printing scenarios included different filament ratios and extrusion rates. The materials used were polylactic acid and a calcium-based composite. The phantoms were based on patient-derived CT scan data. Histogram comparisons were conducted to assess radiological equivalence. The methodology focused on matching Hounsfield units of cancellous and cortical bone.

The phantoms achieved Hounsfield units between 700 and 800, matching human bone. A 30% Stonefil and 70% PLA mix produced realistic cancellous bone. Cortical bone was replicated using 100% Stonefil with a specific extrusion rate. The extrusion rate per voxel was 0.0375 for cancellous and 0.04 for cortical bone. The flow rates were 93.75% and 100%, respectively, for each tissue type. Histogram comparisons confirmed the phantoms matched patient data closely. The results suggest that dynamic filament mixing during printing is feasible. The method successfully translated CT data into radiologically accurate phantoms.

The study demonstrated that real-time filament mixing during 3D printing can replicate bone radiodensity. The method successfully matched Hounsfield units of human hip bone tissue. The use of 30% Stonefil and 70% PLA for cancellous bone proved effective. Cortical bone was best replicated with 100% Stonefil and a specific extrusion rate. The approach allows for direct translation of CT data into printable structures. The results suggest potential for CT-compatible abdominal phantom development. The methodology provides a reproducible way to create anatomically and radiologically accurate phantoms. The authors propose that this technique could improve phantom realism in medical imaging.