Cervical cancer is the fourth most common cancer found in women around the world, per the WHO. To help improve education surrounding cervical cancer detection, Rice and MD Anderson faculty developed the Low-Cost, Universal Cervical Cancer Instructional Apparatus (LUCIA), a portable model that includes a wooden pelvic frame, a vaginal canal, cervical model holders, and the cervix models. Since the model allows for interchangeable cervix models, the LUCIA model supports 20 distinct cervix models, including normal, precancerous, and cancerous cervixes. 

For our midterm project, we (Victoria, Morgan, Shreyas, Archit) fabricated two sets of the LUCIA Nabothian cyst models, one set produced through 3D printing and the other set through a polyurethane cast from a prepared silicone negative mold. 

First, we downloaded the SLDPRT files for the individual models and converted them to STL using Autodesk Fusion360; we also created an STL file of the positive mold of all four cervix models together in one box for creating the silicone negative mold later. We mated each of the four cervixes was mated to the bottom of the box to allow preparation for the silicone pour.

We used the Bambu slicer to prepare the STL files for the Bambu 3D printers. We first used a layer height of 0.1, which resulted in noticeable layer lines near the rounded top of each cervix model. We then tried to improve the smoothness by performing a second print a layer height of 0.08 (extra-fine), the smallest possible setting for our Bambu machine. The total print time for this print was one hour and fifty minutes for the four models, which included tree supports.

We then repeated the 0.08 layer height for the positive mold, which took four hours and nine minutes.

We then created a BJB Platinum Based Silicone negative mold to make polyurethane models of the cervixes. The volume of the positive mold was 248 mL (determined using water), so we prepared ~124 mL samples of the Part A and Part B silicone formulae, first measuring the weight of 124 mL Part A, then measuring the same weight of Part B.

After spraying the inside of the box with mold release agent, we mixed parts A and B in a cup. Silicone is self-healing and takes 24 hours to cure, so we were able to work carefully and mix enough silicone to fill the entire box.

We poured the silicone slowly into one corner of the box to minimize air pockets that would lead to poor structure in the mold. Afterward, we repeatedly tapped the container to further remove any air bubbles. We allowed our silicone mold to cure for 48 hours.

Once the mold was cured, we made the polyurethane molds. We used Smooth Cast 300. We used water to estimate the negative volume of the silicone mold to be approximately 60 mL. Before pouring, the silicone mold was sprayed with a mold release to facilitate easier removal, but this was likely unnecessary due to the hydrophobicity of both materials.

Smooth Cast 300 has a pot life of 3 minutes, so we had to combine the mixture quickly (for about 50 seconds) and pour it into each of the negative sections. We used 40 mL of parts A and B, totaling 80 mL, in case we needed more than estimated (since polyurethane is not self-curing and has a short pot life). It took 10 minutes for the models to cure.

The lettering for 1C and 2D in this process was not perfect; we believed it was likely an issue with the silicone mold, but redid our polyurethane casts to make sure we didn’t have trapped air bubbles reducing the fidelity of the model. However, both iterations yielded similar results. 

Finally, we processed the 3D-printed and polyurethane models using 400-grit sandpaper and paint. We used 80-grit sandpaper to level the bottoms of the models.

We spray-painted the 1C models a candy pink, then made a custom mix of red and pink spray paint for 1D and 2C and applied it with a sponge brush. This initially showed a glossy finish, but revealed some white spots after drying that we covered using another coat of acrylic paint. We finished the detailing of the cysts of all eight models with a combination of pink, white, and red acrylic paints to match our reference pictures. After drying, we applied a gloss spray.

The cost analysis for this project is below:

3D Printing:

Cost Type	Cost	Price	Source	Quantity	Total
Materials	PLA Matte	$22.99/kg	bambulab.com	0.03099 g	$0.71
	Spray Paint	$5.98	homedepot.com	2 bottles	$11.96
	Gloss Clear Coat	$5.98	homedepot.com	1 bottle	$5.98
	Acrylic Paint	$0.69	michaels.com	3 bottles	$2.07
Labor	3D printing technician	$25/hour	ziprecruiter.com	1.824 hours	$45.60
Overhead	Bambu Lab X1C Depreciation	$1.506/day*	bambulab.com raise3d.com	0.228 days	$0.18
Design	Iterations	$22.99/kg	bambulab.com,	2 iterations	$1.42
Misc.	Assorted Sandpaper	$7.99	amazon.com	1 pack	$7.99
	Glue Stick	$1.39	target.com	1 pack	$1.39
	Paint Brushes	$5.99	michaels.com	1 pack	$5.99

TOTAL: $83.29

* Depreciation = (Asset Cost – Salvage Value)/Useful Life

Cost of Bambu X1C Printer: $1449

Salvage Value: $0 (minimum)

Lifespan: 5 years

Depreciation after use= ($1449-$0)/1825 days * 0.076 days * 3 iterations

Depreciation after use= $0.18

Molding and Casting:

Cost Type	Cost	Price	Source	Quantity	Total
Materials	BJB Platinum Based Silicone	$240/2 gal	amazon.com	0.0554 gal	$6.65
	Smooth-Cast 300	$24.71/2 pints	amazon.com	0.12 pint	$1.48
	Silicone Mold Release Spray	$9.29	amazon.com	1 bottle	$9.29
	Spray Paint	$5.98	homedepot.com	2 bottle	$11.96
	Gloss Clear Coat	$5.98	homedepot.com	1 bottle	$5.98
	Acrylic Paint	$0.69	michaels.com	3 bottles	$2.07
Labor	Mold Technician	$20/hr	ziprecruiter.com	1 hr	$20
Overhead	Scale Calibration	$125/year	americanscaleus.com	Within the span of 1 yr	$125
Design	Iterations	$1.48/set of 4	–	0.5 additional	$0.74
Misc.	Assorted Sandpaper	$7.99	amazon.com	1 pack	$7.99
	Disposable Measuring Cups	$14/50 cups	amazon.com	5	$1.40
	Popsicle Sticks	$5.99/100 pcs	amazon.com	3	$0.18

TOTAL: $192.56

To make 50 models, the same procedure could be employed for both 3D printing and molding/casting. It may be somewhat time-consuming but would probably not warrant a revamp of the entire manufacturing process. However, for further scaling (i.e., making 100 models), increasing the number of printers running simultaneously and increasing the number of molds would be useful –– this may increase materials costs (silicone, PLA) and overhead (electricity, asset depreciation), but would cut production time.

To manufacture 1000 or more of these models, it would be worth considering mass production techniques, including industrial-grade 3D printers and injection molding/extrusion molding. This would lead to significant overhead costs (manufacturing space, tooling). 

Overall, we successfully produced 3D-printed as well as casted models of the LUCIA Nabothian cysts. For high-fidelity demonstration models, 3D printers are likely the gold standard due to their exact match to the design. However, for low-resource areas with limited access to 3D printers, it is easier to transport a silicone mold and create polyurethane models from it at a high frequency.

References 

1. World Health Organization. Cervical Cancer. World Health Organization. Published 2024. https://www.who.int/news-room/fact-sheets/detail/cervical-cancer
2. Parra S, Oden M, Schmeler K, Richards-Kortum R. Low-Cost Instructional Apparatus to Improve Training for Cervical Cancer Screening and Prevention. Obstetrics & Gynecology. 2019;133(3):559-567. doi:https://doi.org/10.1097/aog.0000000000003140
3. What is the difference between a positive mold and a negative mold in thermoforming? compression mould. Accessed March 12, 2025. https://www.zjmdc.com/IndustryNews/thermoforming-mold-positive-negative.html.