Designing a Synthetic 3D-Printed Knee Cartilage: FEA Model, Micro-Structure and Mechanical Characteristics

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper presented the design, FEM analysis and experimental tests of a synthetic 3D-printed knee cartilage. Overall, the paper is well written and has solid contribution. However, several issues should still be addressed to further improve the quality of this paper. Below are some comments for the authors to consider:

1. In section 2.2, the authors only described the mesh size of the 2D triangular mesh. The size of the 3D tetrahedral elements was not provided. Please also provide the 3D mesh size.

2. In section 2.4, the authors mentioned that two kinds of synthetic materials were designed for 3D-printing: the soft and medium cartilage. In Table 2 of section 2.5, the Young's modulus of the two printed materials were listed as 2.43MPa and 7.24MPa, respectively. How were these two values determined? Are they calculated or determined by experiments? Please clarify.

3. In line 187 to 188, the authors mentioned that a force of 1150 N was applied perpendicularly to the upper surface of the femur. However, in Figure 3, the left subfigure showed that red force (1150 N) was not applied not on the top but on the side surface of the femur. Please make this point clear to avoid any misunderstandings.

4. In line 179, the authors mentioned that the connections between femur-cartilage and cartilage-tibia were both defined as bonded in the FEM simulation. However, in the real situation, the cartilage and the tibia are in contact, but are not connected to each other. Why didn't the authors define the boundary condition between the cartilage and tibia as two surfaces with contact?

5. The literature study part could be further extended. Currently, there are also other studies using finite element methods to analyze the performance of 3D-printed medical devices with contact surfaces (see the references below). From this point of view, the authors are also recommended to mention those work in the Introduction section as related work. Below are some related references:

"FEM-Based Mechanics Modeling of Bio-Inspired Compliant Mechanisms for Medical Applications". https://doi.org/10.1109/TMRB.2020.3011291

"Design and analysis of a compliant 3D printed energy harvester housing for knee implants". https://doi.org/10.1016/j.medengphy.2020.12.008

Author Response

• Added in line 122 - 124 → In the Control model, the average 3D element mesh size measured 0.81 mm, whereas in the Degenerative model it was 0.79 mm.
• TABLE 2 → caption changed: “Material properties used in the simulation.” >>> “Material properties used in the simulation; Printed Cartilage modules have been calculated as stated in Section 2.6.”

LINE 244 - 246 → “The Young's modulus of each polymeric blend was computed, as average, from the stress-strain charts in the linear region.” + “This has been done by following the corresponding ASTM guidelines for the calculation of elastic parameters.”

• LINE 221 → change the reference to point out the model on the LEFT.

LINE 232-234 → change the reference to point out the model on the RIGHT.

• The Bonded option on Ansys is a type of contact behavior: in particular, the cartilage is defined as the contact body and the tibia as the target body. Defining this contact as Bonded simplifies the analysis, reduces computational complexity, and allows for easier application of loads and boundary conditions. Although minor movements and sliding may occur between the two surfaces, significant movements or sliding are not observed. This assumption is commonly adopted in literature and provides a reasonable approximation of the knee joint behaviour under certain loading conditions. Therefore, it is reasonable to extend this assumption to include the contact between the tibia and femoral cartilage as Bonded as well. While this may not perfectly capture the full range of motion and forces at play in the knee joint, it is a rational simplification given the limitations of computational modeling and the choice to represent the knee joint in full extended position.
• Several references were added, including the first one suggested. Following the entire list:

Sun, Yilun, et al. "Fem-based mechanics modeling of bio-inspired compliant mechanisms for medical applications." IEEE Transactions on Medical Robotics and Bionics 2.3 (2020): 364-373.      

 MacDonald, Eric, and Ryan Wicker. "Multiprocess 3D printing for increasing component                 functionality." Science 353.6307 (2016): aaf2093

Forni, Riccardo, et al. "Heart surgery: septal defect." Handbook of Surgical Planning and 3D Printing.             Academic Press, 2023. 143-169.

Link citation: https://www.stratasys.com/en/3d-printers/printer-catalog/polyjet/j850-digital-anatomy/

Mandrycky, Christian, et al. "3D bioprinting for engineering complex tissues." Biotechnology               advances 34.4 (2016): 422-434.

Vijayavenkataraman, Sanjairaj, et al. "3D bioprinting of tissues and organs for regenerative               medicine." Advanced drug delivery reviews 132 (2018): 296-332

Wu, Yang, et al. "Three-dimensional bioprinting of articular cartilage: a systematic   review." Cartilage 12.1 (2021): 76-92.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dolino et al. used 3D printing to reproduce in-silico the  Knee Cartilage joint. The authors claim the proposed method produces a joint that can be used as a patient simulator that can be clinically used. Some issues that need to be addressed before it can be considered for publication.

 The review lacks recent experimental findings: Ibarra recently measured the pressure distribution near the knee limb and Chaparro measured forces at different knee angles and Eschweiler modeled Cartilage behavior.

Did you perform a mesh convergence analysis?

From table 2 is seen the 3D printed cartilage behaves as linear elastic. It had been shown that cartilage behavior can be described by biphasic poroelastic. The interstitial fluid goes through the porosity giving a nonlinear behavior. So, the proposed material choice will give linear but probably less comfortable behavior as it cannot dissipate energy by viscous damping.

 

Minor issues

Please add details of the experimental equipment and software used for the testing in section 2.3

Please enhance Fig 4 font.

 

***

Ibarra et al. Interface Pressure System to Compare the Functional Performance of Prosthetic Sockets during the Gait in People with Trans-Tibial Amputation. Sensors, 20, 24 7043, 2020 http://dx.doi.org/10.3390/s20247043

Eschweiler J, et al. The Biomechanics of Cartilage-An Overview. Life (Basel). 2021 Apr 1;11(4):302. doi: 10.3390/life11040302

 Chaparro et al. Tibiofemoral contact properties for different flexion angles on injured ACL knee joint. Revista UIS Ingenierías, vol. 17, (2), 2018

 

https://doi.org/10.18273/revuin.v17n2-2018024

Author Response

• ESCHWEILER → Added as a reference in LINE 429.
• ANSWER → To evaluate the quality of the meshes obtained, several parameters were computed such as the aspect ratio, which is the ratio between the longest and shortest sides of an element, and the Jacobian, a metric used to measure the deviation of an element from its ideal shape. The evaluation of these parameters indicated that the obtained meshes had an acceptable level of quality.
• As stated in LINES 190-201, while describing this as linear elastic is a simplification (but still considered to be reliable [*],[**]), it leads to a more direct implementation of a parallel computational approach. Furthermore, this material choice served as a first insight into the capability of PolyJet materials to emulate articular cartilage mechanical behavior. Thus, using a less accurate constitutive model was considered acceptable for this study, and a finer tuning of material properties is of high consideration for future studies.

[*] Trad, Zahra, et al. FEM analysis of the human knee joint: a review. Berlin: Springer International Publishing, 2018

[**] Donzelli, Peter S., et al. "Contact analysis of biphasic transversely isotropic cartilage layers and correlations with tissue failure." Journal of Biomechanics 32.10 (1999): 1037-1047.

• LINES 134-136 → “Collagen-like fibers were realized and integrated into ASTM D695 and D638 standardized samples (Figure 1) on the open-source 3D computer graphics software Blender.
• The font of the figure was updated.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Comments for author File: Comments.pdf

Author Response

Revisions

1. Designing a Synthetic 3D-Printed Knee Cartilage: FEA model, Micro-Structure and Mechanical
2. The abstract was modified adding more numerical results
3. Gender, weight and age were added in the material and methods, in the clinical data section.

Left leg, Age: 26 and 68; Height 1,81 and 1,85; Weight 73 and 94; BMI 22.28 and 27.47.

4. The polymers specification was added in Material and Methods, section 2.4
5. New references were added in the manuscript

Sun, Yilun, et al. "Fem-based mechanics modeling of bio-inspired compliant mechanisms for medical applications." IEEE Transactions on Medical Robotics and Bionics 2.3 (2020): 364-373.       

MacDonald, Eric, and Ryan Wicker. "Multiprocess 3D printing for increasing component                 functionality." Science 353.6307 (2016): aaf2093

Forni, Riccardo, et al. "Heart surgery: septal defect." Handbook of Surgical Planning and 3D Printing.             Academic Press, 2023. 143-169.

Link citation: https://www.stratasys.com/en/3d-printers/printer-catalog/polyjet/j850-digital-anatomy/

Mandrycky, Christian, et al. "3D bioprinting for engineering complex tissues." Biotechnology               advances 34.4 (2016): 422-434.

Vijayavenkataraman, Sanjairaj, et al. "3D bioprinting of tissues and organs for regenerative               medicine." Advanced drug delivery reviews 132 (2018): 296-332

Wu, Yang, et al. "Three-dimensional bioprinting of articular cartilage: a systematic                 review." Cartilage 12.1 (2021): 76-92.

6. The introduction was modified according to the input in the review and with more citations, especially about the 3D printing part. Look at the attachment for the new version of the manuscript.
7. Materials and Methods: more information about printing process, polymers used and printing properties have been updated in the material and method section and with a reference.
8. Results and conclusion:

The two sections were updated according to the input. Following a reply concerning the study of cartilage in the wet state.

We acknowledge that our 3D printed cartilage may not fully replicate the complex biochemical composition and water content of actual cartilage tissue. Nevertheless, our research demonstrates the potential of DAP technology to mimic the mechanical properties of knee cartilage, offering avenues for further advancements in the field of bioprinting and tissue engineering.

Regarding the “limited” part of cartilage under compression in figure 5: the work investigates a specific condition, as stated in the introduction, related to a single stride with an extended leg. In this configuration, the cartilage area under compression is reported. The other parts work in other conditions, like flexed knees, that are out of the scope of the work.

We took subjects from a database that was the output of the EU project mentioned in the study. There are papers cited in the methodology that report the segmentation step and also the pictures of the scanned knee (Ciliberti et. Al. Abounnet et. al.).

The numbers related to the micro-architecture are reported in Material & Methods, section 2.3.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

In this study, focuses on developing patient-specific synthetic cartilage using Digital Anatomy polymers to mimic cartilage morphology and mechanics. It employs Finite Element Analysis (FEA) to validate the setup, revealing the impact of osteoarthritis on cartilage stress-strain distributions. Results demonstrate that 3D-printed polymers closely mimic cartilage properties, highlighting potential benefits for knee joint analysis. However, limitations in representing native cartilage complexity and the need for broader anatomical considerations emerge as areas for further improvement. The content could benefit and be more sentient with detailed discussion. The following issues must be addressed to enhance the quality of the article:    

1.      How does osteoarthritis impact cartilage morphology and mechanical behavior within the knee joint?

2.      Discuss the limitations of creating patient-specific synthetic cartilage using Digital Anatomy polymers compared to traditional manufacturing methods for studying knee joint biomechanics.

3.      How does Finite Element Analysis validate the patient-specific measurement setup in mimicking cartilage stress-strain distributions? Elaborate on the role of morphology in these stress-strain distributions.

4.      Compare the elastic properties of 3D-printed polymers with biological cartilage samples. Discuss the discrepancies observed and the implications for designing synthetic cartilage. Refer to the following:

a.       3D printing of bone tissue engineering scaffolds

5.      Analyse the comparative outcomes between computational simulation (FEA) and mechanical testing of knee joint cartilages. How do these results align or differ, particularly in degenerative morphologies, and what implications do they have for understanding knee mechanics in osteoarthritis?

6.      How do the elastic properties of polymers correlate with the layers of native cartilage? Discuss the observed relationships, including differences and similarities in mechanical behavior.

7.      How many tests/specimens are conducted for consistency on each sample? The ASTM D695 and D638 standards are associated with the testing of rigid plastics; how is it associated with cartilage due to soft matter nature? For reference, look at the following paper:

 

a.       Experimental and numerical investigation on 2.5-dimensional nature-inspired infill structures under out-plane quasi-static loading.

Author Response

Revision – Round 1 – Correction

 

• The impact of OA on the morphology and on the mechanical behavior of knee cartilage is addressed during the introduction. (Rows 33-45)
• The pros and limitations of the DAP compared to the traditional manufacturing process are reported in the introduction section. (Rows 62-78 )
• The role of morphology is described in the results sections. Particularly, the FEA validation is reported in section 3.1 (From row 274).
• The comparison of the elastic properties of synthetic polymers and biological knee cartilage are reported in Discussion 4.1 (Rows 333-337). Section 3.2 discusses the mechanical behavior of synthetic cartilage materials (Rows 288-290). 
• The comparison of FEA and synthetic testing is reported in Discussion 4.2 (Rows 340-347).
• The relationship between mechanical properties and layer depth in synthetic and native cartilage is discussed in Discussion section 4.3. (Row 390-394 and 416-417 plus Tables 3 and 4)
• Since a standardization for the mechanical test of soft plastic is lacking, we started from the well-know references for rigid, reinforced and non-reinforced plastic. We prototyped 10 specimens per group, they were measured in different positions as the normative requires and the reported variability was mainly related to the caliber resolution.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have revised the manuscript according to my comments. Therefore, I recommend to publish this paper in this journal.

Author Response

Thank you!

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have suffciently improved the manuscript. I recommend it for publication in Applied Sciences.

Author Response

Thank you!

Reviewer 3 Report

Comments and Suggestions for Authors

No comments.

 

Author Response

Thank you!

Reviewer 4 Report

Comments and Suggestions for Authors

The comments are not appropriately addressed and still need some improvement. 

1.      Formatting and spelling corrections

·         The term osteoarthritis (OA) (Lines -2,33,328), with abbreviation should be mentioned at first time instead osteoarthritis (Lines -337,428,431,436,445) or only OA (Lines -35,37,44,46,52,94,96,185,276,366,378,383), which is correct one, Maintain the uniformity throughout manuscript.

·         check Spelling of “osteoarthritic” (277), kneew joint(126), synthetic(370), biomechanics(371), desease(373).

·         Overall kindly check the grammar and maintain manuscript without spellings mistakes.

2.      Texted (lines 166-169) in the manuscript “Two different composites were created using these materials. The first composite, called Soft Cartilage, consisted of 70% flexible material (Agilus30), 15% gel polymer (GelMatrix), and 15% stiff polymer (BoneMatrix). The second composite, named Medium Cartilage, had a composition of 85% flexible material (Agilus30) and 15% stiff polymer (BoneMatrix).” and texted (lines 187-189) in the manuscript “Five different materials were chosen during the simulation: Cartilage Literature, Bone Literature, Printed Cartilage Soft, Printed Cartilage Medium and Printed Bone.”

a.       What is the technical procedure of mixing of materials include elastomers, conducting polymers, ionically conducting polymers, and carbon nanotubes are used for fabrication of tensile specimens.

b.      Were there any challenges are presented while mixing?  was that defect free mixing?

c.       What is the effect of the anisotropic nature of cartilage with synthetic cartilage and native cartilage?

d.      How to incorporate the element type and material properties with anisotropic properties for Numerical simulation (FEA)

e.       Was there mesh convergence study conducted.

f.        What is the role of degenerative models and control models with Superficial Deep and Middle composition with 1% and 5% volume contents of fibers.

 

3.      Texted (lines 166-169) in the manuscript “A compressive force of 1150 N was applied perpendicularly to the upper surface of the femur, which matches the force acting during a gait cycle for a full extension position”,

a.       is the force (1500N) being constant which is acting on femur surface? What is the direction of loading and how the load is distribution through the knee joint?

b.      As mentioned reference 31, the loads are applied with combination of   compression, anterior, anterior–posterior loads, valgus torque and frictional coefficient, but in the current work, why the authors are considered only compression loading

c.       will it be different from person to person?

d.      what is the targeting age group of people are considered for this work or is there any special case?

 

4.      Texted (lines 240-246) in the manuscript “Specifically, ten cylindrical samples were printed to meet the dimensional requirements outlined by the ASTM D695 compression protocol, five using the Soft material and five using the Medium material. Additionally, ten dog-bone specimens were printed for the ASTM D638 tensile test, with an equal split of five for each material type. The Young’s modulus of each polymeric blend was computed, as average, from the stress-strain charts in the linear region. This has been done by following the corresponding ASTM guidelines for the calculation of elastic parameters.” and the authors are replied “Since a standardization for the mechanical test of soft plastic is lacking, we started from the well-known references for rigid, reinforced, and non-reinforced plastic.”

a.       If there are no standardization techniques for the characterization of soft plastic, then why the authors are considered the Type IV dimensions of ASTM D638(available in the manuscript-figure-1) and dimension of ASTM D695(not available) standards. Because the specimen dimensions will be varied based on the behaviour of material, for example in the case of metals, ASTM E8(most of the metal samples) and ASTM A370 (different steel grades).  In case of polymers, ASTM D3039(polycarbonate composite materials), ASTM D412(elastomers), ASTM D651(insulating materials) ASTM D882(Thin plastic sheets), etc., kindly justify?

b.      Will the blended polymers produce linear region in the stress-strain charts?

 

c.       Why the compression cylindrical samples, testing illustrations and results are not shown in the manuscript?

Comments on the Quality of English Language

Grammer is fine.

Author Response

1. After leaving the term “osteoarthritis (OA)” only at first time in the abstract and introduction (lines 2 and 33) , we changed the term “osteoarthritis” with “OA” throughout the text. Grammar was checked and spelling was corrected.
2. Several points were addressed:
3. The authors believe there is a misunderstanding because the chemical composition of the available polymers is unknown and proprietary to Stratasys, which decided not to disseminate the datasheet. The J850 DAP automatically mixes cartridges based on the user input. Reference [13] recall Stratasys website and the user manual guide for more information about the functioning of the machine.
4. No challenges arose during the mixing process. The materials were created and printed precisely as intended using the software specifically tailored for the printer we used (J850 Digital Anatomy Printer by Stratasys). Any specific test addressing the actual composition of the blend was performed. A qualitative assessment was done using these technologies for more than five years, where materials were mixed (i.e., colors), and the results were exactly as expected. Further studies will focus on the chemical composition of the polymers to provide alternatives not limited to the closed system used in this study.
5. Effects of anisotropic behavior of native cartilage tissue are cited in lines 26-32 and lines 391-394. 

Fiber-reinforced specimens were designed to give synthetic cartilage an anisotropic behavior instead of the orthotropic one obtained in the first two parts of this study (Soft and Medium). Before considering the implementation of differently oriented fibers inside a single model, we tested stand-alone orientations within the specimens. Further studies will compare the anisotropic nature of native cartilage to synthetic cartilage.

1. Lines 190-191 and 207-208 report that linear elastic isotropic properties were assigned to the materials and lines 190-210 justify that choice. Therefore, no anisotropic properties were included in the computational simulation. Further investigation may include a more accurate constitutive model integrating anisotropic behavior.
2. To evaluate the quality of the meshes obtained, several parameters were computed such as the aspect ratio, which is the ratio between the longest and shortest sides of an element, and the Jacobian, a metric used to measure the deviation of an element from its ideal shape. The evaluation of these parameters indicated that the obtained meshes had an acceptable level of quality.
3. Fiber-reinforced materials were tested using only ASTM-standardized specimens. Therefore, no differentiation was made between “control” and “degenerative” conditions in these cases. Furthermore, the effect of degeneration on cartilage’s composition (volume of fibers and their orientation) is still an open question mark. For this reason, we did not merge the two analyses. 

 

3. Several points were addressed:
4. The force was not constant but ranged from 0 N to 1150 N. The load was applied along the z-axis, directed downward from the upper surface of the femur to the lower surface of the tibia, perpendicular to the joint surface (lines 217-221, Figure 3). The femur was used as an impeller to transmit the load on the cartilage, thus replicating the patient-specific load distribution.
5. In our  model, we did not refer to reference [31] for setting the boundary conditions and applied loads. Lines 217-221 explain that we constrained flexion-extension and varus-valgus rotations for the femur to examine the knee joint in full extension. This was done by restricting translation solely along the load application axis (the z-axis), thus setting rotations to zero along the x- y-  and z-axes. The tibia was fixed at the lower surface, providing stable support. Additionally, we applied a perpendicular compressive force to the upper surface of the femur. This resulted in a purely compressive loading, corresponding to a fully extended leg touching the ground. No anterior-posterior loads, valgus torque, or frictional coefficient were introduced. 
6. It is well-known that the force experienced by the knee joint during a gait cycle in a fully extended position varies among individuals. The mechanical performance of biological cartilage tissue and the loads experienced are influenced by a lot of factors, including age, weight, presence of OA, etc. (lines 326-329 , [39]). To align our findings with those of other researchers, we opted to apply a compressive load of 1150 N. This value corresponds to the maximum force in the gait cycle in full extension obtained by Pena et al. [39], Trad et al. [26], Sathasivam and Walker [36], and many other authors, as mentioned in lines 219-221. 
7. The age of the two patients considered in this work is 26 and 68 (lines 101-102). We chose one healthy subject and one elderly with clinically evaluated cartilage degeneration to establish correlations between material properties and morphological changes due to OA. It’s important to consider this study as a preliminary exploration into assessing the potential and feasibility of the Digital Anatomy Printer technology in emulating the mechanical behavior of knee cartilage. Moving forward, we aim to extend this methodology to a wider range of patients, expanding the cohort for further analysis. 

 

4. Several points were addressed:
5. Our digital polymers do not fit into any category (elastomer, polycarbonate, or insulating material) since the chemical composition is proprietary and hidden to the public. We are aware of these ASTM standards, but since we do not have access to the composition of our basic polymers, we need to base our decision just on the output of the printing process: a non-rigid plastic with a viscoelastic behavior.

Reported from the ASTM D638 “6.1.2 Nonrigid Plastics—The test specimen shall conform to the dimensions shown in Fig. 1. The Type IV specimen shall be used for testing nonrigid plastics with a thickness of 4 mm (0.16 in.) or less.” In addition, the ASTM D638 is for the tensile properties of plastic and is NOT just for rigid plastic. 

For the compression specimens, an adherent standard is not clearly stated. Since the ASTM D695 refers to reinforced plastic where our specimens with fibers may fall, we opted for that test also for the non-reinforced samples to achieve comparability at least within our results. In conclusion, a correct answer is hard to be found since the standardization for these tests is unclear. They all came with a different trade-off that required a decision. To the best of our knowledge, this should be the more suitable way to achieve elastic property comparability. In addition, biological specimens from cartilage are usually tested in a confined setup, immersed in fluid media. In the next future we are planning to perform these kind of tests, but we need a basic knowledge concerning the “dry” behavior of our materials.

1. The blend polymers that are not shown in the paper (Soft and Medium) showed a non-linear behavior during compression test and a linear behavior during tensile test. The Medium non-linear behavior was characterized by a linear region until 0.35 mm/mm strain and then an exponential stress increase with strain. The Soft material showed a quasi-linear behavior where the disalignement from linear trend may be due to the viscosity of the polymer. 

For the blends reported in the manuscript that showed a J-shaped stress strain chart we computed the elastic module within the first 20% of deformation where the approximation with elastic behavior was more accurate.

The images of compression cylindrical samples were omitted due to their poor quality. Conversely, the presence of fibers, with their varying orientations and densities, was more clearly visible in the tensile specimens due to their smaller thickness, providing higher contrast against a black background. The results from the testing of cylindrical samples are presented numerically in Table 2 for the materials without fibers, and in Table 3 and 4 for materials with fibers. The Young’s moduli provided in the tables are calculated as explained above. Additionally, Figure 8 (right) reports the stress-strain curves of the customized cylindrical specimens with fibers. Including more graphs or images would have consumed excessive space, so we opted to include only the results we deemed most significant.