Liver transplantation (LT) is standard therapy for acute and chronic liver failure, and due to the advancement of immunosuppressive agents and surgical techniques, has evolved rapidly [1]. However, severe organ shortage remains an unresolved issue, and only a limited number of LT candidates can receive liver grafts. To ensure the success of LT, the size of the expected graft should be adequate for the recipient. When the liver graft is too small for the recipient, small-for-size syndrome can occur, which can usually be predicted by calculating the graft–recipient weight ratio (GRWR) [2,3]. On the other hand, when a liver graft is too large to fit inside the abdominal cavity, large-for-size syndrome can occur [2,4]. Although the incidence of large-for-size syndrome is low, fatal outcome can occur due to graft compression, followed by poor oxygen supply, leading to graft dysfunction [2,5]. Hence to prevent large-for-size syndrome, understanding the size of the recipient’s abdominal cavity in relation to graft liver volume is significant. While the risk of large-for-size syndrome is quite low for living donor LT (LDLT), since the donor is also evaluated before transplant, deceased donors are limited in pre-transplant evaluation, and there is always the risk of the graft being too large for recipients who have a small intra-abdominal cavity.

Three-dimensional (3D) printing with its potential as a personalized medical tool has started to be utilized in medicine, especially in maxillofacial and craniofacial surgery [6]. There has also been an effort to use 3D-printing in the field of liver surgery, but mostly focused on understanding the anatomical structures [7,8]. In 2013, the use of 3D-printing in LT was first introduced in the literature of Zein et al. [9], and while it has been continuously developed, it is still too expensive and time-consuming to be practically utilized. To maximize its utility, our team focused on reconstructing the LT recipients’ abdominal cavity to its original size, while minimizing the time for manufacturing. Our initial experience on utilizing our 3D-printing protocol for LT recipients with small abdominal cavity showed successful clinical outcome with short manufacturing time [10]. By the successful application of 3D-printing to prevent large-for-size syndrome in recipients with potential risk, we designed this study to analyze whether the 3D-printed abdominal cavity protocol actually improved our clinical practice, by using propensity score (PS) matching to compare the clinical outcome of patients with 3D-printing to patients without 3D-printing.

Patient selection of 3D-printing of abdominal cavity replica

The study included a total of 760 patients who underwent LT between January 2017 and March 2022. Among these patients, twenty patients had a replica of their abdominal cavity manufactured using a 3D-printed model. The selection of patients for this process was based on the operator’s decision, considering the spatial relationship between the recipient’s liver fossa and the liver graft. Patients with small-sized abdominal cavities, mostly female or pediatric recipients, were chosen. In the cases of LDLT, patients with an expected GRWR exceeding 2%, or a liver graft larger than 1,000 cm³, were selected for 3D-printing of the intra-abdominal cavity.

For patients on the waiting list for a deceased donor match, the following criteria were used to print a 3D model: (1) female recipients who were not 10 cm taller than the allocated male donor, (2) male recipients who were at least 10 cm shorter than the allocated female donor, (3) matches of the same sex where the recipient was at least 10 cm shorter than the donor, and (4) individuals with a small right liver fossa, defined as an anteroposterior (AP) length of ≤ 13 cm, or a lateral space from the inferior vena cava of ≤ 10 cm. Additionally, the transplant surgeon’s judgment played a role in the selection process, especially when the recipient had an abdominal cavity deformity that did not fit the mentioned criteria.

3D modeling of the recipient’s abdominal cavity

Fig. 1 depicts the workflow for creating and applying a 3D-printed model in both living donor and deceased donor LT. The process begins with the recipient’s computed tomography (CT) scan, which is used to outline the inner space of the abdominal cavity where the graft liver will be positioned. Mimics Medical 21.0 (Materialise) is employed for this purpose. The inner surface of the abdominal cavity, excluding the medial two-thirds of the anterior wall, is outlined with a (1 to 3) cm gap between slices. In the case of adult recipients, only the right half of the abdomen is outlined, whereas for pediatric recipients and adults undergoing left LT, both the right and left halves are outlined.

Fig 1. The workflow of manufacturing the 3D-printed abdominal cavity model and its application in both deceased donor liver transplantation and living donor liver transplantation.

Once the outline of the intra-abdominal cavity is marked, the lines are reconstructed into a 3D model using Cinema 4D (Maxon). Subsequently, the 3D model is printed individually using Cubicreator software and Cubicon Single Plus (Cubicon), a fused deposition modeling type 3D-printer. After printing, the pieces are assembled to create the final 3D-printed model. This model assists in the decision-making process of whether or not to utilize a specific liver graft. Fig. 2 depicts the application of the 3D-printed abdominal cavity model during LT decision-making.

Fig 2. The actual application of the 3D-printed model during (A) deceased donor liver transplantation, (B) living donor liver transplantation, and (C) deceased donor liver transplantation of other institution. GRWR, graft–recipient weight ratio.

Data acquisition

Baseline characteristics of the donor and recipient for both the 3D-printed group and the control group were collected. In addition, data of the graft, recipients’ abdominal cavity, and 3D model were collected. Graft data included the type of graft, weight, and GRWR. We measured the AP and lateral length of the right liver fossa, and the AP length of the midline of recipients to roughly estimate the abdominal cavity based on the CT of recipients before LT. Data regarding the 3D-printed model included the amount of materials, cost, and manufacturing time needed for 3D-printing. The clinical courses before and after using the 3D-printed model were also collected. In addition, demographical data and the clinical course of the patients who experienced large-for-size syndrome during the study period were collected.

Statistical analysis

To reduce the impact of selection bias, we performed a 1 : 5 ratio PS matching analysis with the nearest neighbor method between the 3D-printed group and the control group. The control group consisted of patients who underwent LT without using 3D-printing model during January 2017 to March 2022. The PS of each case was calculated using a multivariable logistic regression model by including variables such as sex, age, height, weight, adult or pediatric, LT from living or deceased donor, graft type, graft weight, donor sex, donor age, donor height, and donor weight. After PS matching, 20 cases in the 3D-printed group were matched to 100 cases in the control group.

The outcomes between the 3D-printed group and the control group were compared, such as 30-day complication of recipients, reoperation within 30 days, large-for-size syndrome, graft failure, and death of the recipient. Numerical variables were compared with the Student t test or Mann–Whitney U test, and expressed as the mean ± standard deviation or median (interquartile range [IQR]), respectively. Categorical variables were compared with the chi-square test or Fisher’s exact test. Kaplan–Meier survival curve analysis was performed to estimate the graft survival and overall survival rates after PS matching. Statistical analyses were performed using SPSS 26.0 (IBM Corp.) and R version 4.1.1 (R Foundation for Statistical Computing).

This study was approved by the Institutional Review Board (IRB No. 2022-08-101-001). Informed consent was acquired from the recipients or their parents or legal guardians who were enrolled prospectively, after the approval of the IRB of the Samsung Medical Center. The research was performed in accordance with relevant guidelines/regulations that were in accordance with the Declaration of Helsinki. Neither the recipients or donors were related to procurement from prisoners.

Baseline characteristics before and after PS matching

Among 760 patients who underwent LT between July 2020 and March 2022, 20 patients underwent LT using 3D-printed abdominal cavity models.

Table 1 summarizes the baseline characteristics of donors and recipients between the 3D-printed group and the control group before and after PS matching. Before PS matching, there were higher proportions of female patients (30.3% vs. 75.0%, p < 0.001), pediatric patients (3.6% vs. 30.0%, p < 0.001), re-LT patients (5.6% vs. 20.0%, p = 0.025), and LT from deceased donor (24.2% vs. 55.0%, p = 0.004) in the 3D-printed group, compared to the control group. Also, there were significant differences in age (53.6 ± 12.8 vs. 33.5 ± 25.7 years, p = 0.002), height (163.3 ± 16.2 vs. 135.1 ± 42.7 cm, p = 0.008), weight (66.0 ± 15.5 vs. 43.8 ± 25.7 kg, p = 0.001), body mass index (BMI; 24.3 ± 4.0 vs. 20.7 ± 4.4 kg/m², p = 0.002), etiology (p < 0.001), total bilirubin (9.7 ± 12.5 vs. 23.1 ± 13.6 mg/dL, p < 0.001), MELD/PELD score (19.9 ± 12.5 vs. 30.6 ± 10.7, p < 0.001), graft type (p = 0.027), and GRWR (1.4% ± 0.6% vs. 2.3% ± 0.8%, p < 0.001) between the two groups. After PS matching, 100 patients in the control group, and 20 patients using 3D-printing models were included. After PS matching, there were no significant differences in the baseline characteristics between the two groups.

Table 1 . Comparison of the baseline characteristics of donors and recipients between the 3DP group and the control group before and after PS matching

	Before PS matching			After PS matching
	No 3DP group (n = 740)	3DP group (n = 20)	p-value	No 3DP group (n = 100)	3DP group (n = 20)	p-value	SMD
Sex (male/female)	515/225 (69.6)	5/15 (25.0)	< 0.001	24/76 (24.0)	5/15 (25.0)	> 0.999	0.023
Age (yr)	53.6 ± 12.8	33.5 ± 25.7	0.002	39.2 ± 19.2	33.5 ± 25.7	0.351	0.254
Adult/pediatric	713/27 (96.4)	14/6 (70.0)	< 0.001	79/21 (79.0)	14/6 (70.0)	0.388	0.208
Height (cm)	163.3 ± 16.2	135.1 ± 42.7	0.008	144.8 ± 33.6	135.1 ± 42.7	0.349	0.252
Weight (kg)	66.0 ± 15.5	43.8 ± 25.7	0.001	49.3 ± 20.9	43.8 ± 25.7	0.378	0.234
BMI (kg/m2)	24.3 ± 4.0	20.7 ± 4.4	0.002	21.5 ± 3.7	20.7 ± 4.4	0.448	0.199
Number of LT			0.025			0.485	0.347
First	699 (94.5)	16 (80.0)		86 (86.0)	16 (80.0)
Second	36 (4.9)	4 (20.0)		11 (11.0)	4 (20.0)
Third	5 (0.7)	0 (0.0)		3 (3.0)	0 (0.0)
Etiology (HBV vs. other)			< 0.001			0.19	0.441
HBV	375 (50.7)	1 (5.0)		19 (19.0)	1 (5.0)
HCV	21 (2.8)	1 (5.0)		1 (1.0)	1 (5.0)
Alcohol	181 (24.5)	4 (20.0)		33 (33.0)	4 (20.0)
Others	163 (22.0)	14 (70.0)		47 (47.0)	14 (70.0)
HCC	385 (52.0)	3 (15.0)	0.002	12 (12.0)	3 (15.0)	0.714	0.088
ABO incompatibility	147 (19.9)	0 (0.0)	0.02	12 (12.0)	0 (0.0)	0.214	0.522
Total bilirubin (mg/mL)	9.7 ± 12.5	23.1 ± 13.6	< 0.001	16.8 ± 12.6	23.1 ± 13.6	0.065	0.484
Albumin (g/dL)	3.3 ± 0.6	3.1 ± 0.4	0.109	3.2 ± 0.6	3.1 ± 0.4	0.375	0.193
Creatinine (mg/dL)	1.1 ± 1.1	1.0 ± 1.0	0.63	1.0 ± 0.9	1.0 ± 1.0	0.944	0.019
INR, PT	1.9 ± 1.5	3.8 ± 4.0	0.055	2.6 ± 1.7	3.8 ± 4.0	0.209	0.385
CTP score	8.3 ± 2.6	10.4 ± 2.0	< 0.001	9.6 ± 2.2	10.4 ± 2.0	0.101	0.402
MELD/PELD score	19.9 ± 12.5	30.6 ± 10.7	< 0.001	27.6 ± 13.1	30.6 ± 10.7	0.296	0.242
Donor sex (M/F)	432/308 (58.4)	13/7 (65.0)	0.716	56/44 (56.0)	13/7 (65.0)	0.62	0.185
Donor age (yr)	39.7 ± 15.1	42.1 ± 16.6	0.528	46.4 ± 20.2	42.1 ± 16.6	0.32	0.231
Donor height (cm)	167.8 ± 8.6	167.2 ± 10.5	0.8	166.2 ± 7.5	167.2 ± 10.5	0.714	0.100
Donor weight (kg)	66.8 ± 12.2	68.2 ± 13.6	0.639	67.0 ± 12.1	68.2 ± 13.6	0.715	0.094
Donor BMI (kg/m2)	23.6 ± 3.2	24.4 ± 4.3	0.431	24.2 ± 3.5	24.4 ± 4.3	0.815	0.062
Living/deceased donor	561/179 (75.8)	9/11 (45.0)	0.004	42/58 (42.0)	9/11 (45.0)	> 0.999	0.061
Graft type (whole vs. partial)			0.027			> 0.999	0.060
Whole	165 (22.3)	9 (45.0)		48 (48.0)	9 (45.0)
Right/extended right	540/8 (74.1)	3/1 (20.0)		31/1 (32.0)	3/1 (20.0)
Left/extended left	5/2 (1.0)	1/1 (10.0)		0 (0.0)	1/1 (10.0)
Left lateral	20 (2.7)	5 (25.0)		20 (20.0)	5 (25.0)
Graft weight (g)	866.8 ± 371.6	910.0 ± 576.3	0.743	1,022.5 ± 538.4	910.0 ± 576.3	0.427	0.202
GRWR (%)	1.4 ± 0.6	2.3 ± 0.8	< 0.001	2.2 ± 0.8	2.3 ± 0.8	0.67	0.105
Macrosteatosis (%)	5.2 ± 6.3	3.5 ± 3.0	0.029	5.8 ± 7.6	3.5 ± 3.0	0.03	0.389
Microsteatosis (%)	5.7 ± 7.4	3.9 ± 3.0	0.015	5.4 ± 6.7	3.9 ± 3.0	0.097	0.307
Cold ischemic time (min)	144.0 ± 129.8	241.2 ± 213.7	0.072	220.0 ± 204.6	241.2 ± 213.7	0.701	0.101
Warm ischemic time (min)	40.0 ± 30.9	35.3 ± 11.1	0.112	33.6 ± 19.6	35.3 ± 11.1	0.592	0.111

Values are presented as number (%) or mean ± standard deviation.

SMD, standardized mean difference; PS, propensity score; 3DP, 3-D printed; BMI, body mass index; LT, liver transplantation; HBV, hepatitis B virus; HCV, hepatitis C virus; HCC, hepatocellular carcinoma; INR, international normalized ratio; PT, prothrombin time; CTP, Child-Turcotte-Pugh Score; MELD/PELD, Model for End-Stage Liver Disease/Pediatric End-Stage Liver Disease; GRWR, graft–recipient weight ratio.

Clinical outcomes

Table 2 represents clinical outcomes between the two groups before and after PS matching. Before PS matching, there were no significant differences in 30-day complication (63.6% vs. 50.0%, p = 0.310), reoperation within 30 days (19.6% vs. 20.0%, p > 0.999), large-for-size syndrome (1.8% vs. 0.0%, p > 0.999), graft failure (9.2% vs. 0.0% p = 0.245), and death (12.7% vs. 15.0%, p = 0.726) between the two groups. However, after PS matching, patients without using 3D-printing models experienced a significantly higher rate of graft failure than the 3D-printing group (23.0% vs. 0.0%, p = 0.013).

Table 2 . Clinical outcomes between the 3DP group and the control group before and after PS matching

	Before PS matching			After PS matching
	No 3DP group (n = 740)	3DP group (n = 20)	p-value	No 3DP group (n = 100)	3DP group (n=20)	p-value	SMD
30-day complication	471 (63.6)	10 (50.0)	0.310	64 (64.0)	10 (50.0)	0.356	0.286
C-D class ≥IIIb	176 (23.8)	5 (25.0)	> 0.999	32 (32.0)	5 (25.0)	0.724	0.156
C-D class ≥IV	69 (9.3)	1 (5.0)	> 0.999	18 (18.0)	1 (5.0)	0.193	0.416
Reoperation within 30 days	145 (19.6)	4 (20.0)	> 0.999	27 (27.0)	4 (20.0)	0.709	0.166
Large-for-size syndrome	12 (1.6)	0 (0.0)	> 0.999	7 (7.0)	0 (0.0)	0.599	0.388
Graft failure	68 (9.2)	0 (0.0)	0.245	23 (23.0)	0 (0.0)	0.013	0.751
Death	94 (12.7)	3 (15.0)	0.726	23 (23.0)	3 (15.0)	0.560	0.107

C-D class, Clavien-Dindo Classification; PS, propensity score; 3DP, 3-D printed; SMD, standardized mean difference.

Graft survival and overall survival rate between the 3D-printed group and PS-matched control group were also evaluated using Kaplan–Meier survival curve analysis (Fig. 3). Graft survival of the PS-matched control group was inferior to that of the 3D-printed group (p = 0.034), while the overall survival rate seemed similar between the two groups (p = 0.665).

Fig 3. Graft survival (A) and overall survival (B), between the 3D-printed group and the control group after propensity score matching.

Patient and printing-related data of the 3D-printed group

Ten out of thirteen adult recipients (76.9%) and five out of seven pediatric recipients (71.4%) were female. The median age of adult recipients was 50 years (IQR 25.5–64.0), while that of pediatric recipients was 0.67 year (IQR 0.5−0.8). The median heights of adult and pediatric patients were 158 cm (IQR 163–166) and 73 cm (IQR 60–130), respectively. We also evaluated the size of the abdominal cavity of the recipients by measuring the length of the AP right liver fossa, lateral right liver fossa, and AP midline, according to the CT image of those patients. The mean length of the AP right liver fossa, lateral right liver fossa, and AP midline of adult recipients were 15.50, 8.94, and 9.62 cm, respectively. In comparison, the mean length of the AP right liver fossa, lateral right liver fossa, and AP midline of pediatric recipients were 10.39, 5.67, and 7.15 cm, respectively (Supplementary Table 1).

The mean amount of materials required to manufacture the 3D model was 54.3 g, and the mean cost of the materials was USD $1.25. The mean total time taken to create the 3D abdominal cavity model was 527.0 minutes. Fig. 4 shows the trend of total manufacturing time according to case number. The total manufacturing time included modeling, printing, and assembling time, and as the case was repeated, it became shortened, and recently, it has only taken about 6 to 7 hours.

Fig 4. Total manufacturing time according to (A) only cases (n = 20) included in the study, and (B) all the manufactured cases (n = 29), including non-operated cases.

Baseline characteristics and clinical course of the patients with large-for-size syndrome

A total of 12 patients without using a 3D-printed model experienced large-for-size syndrome after LT during the study period. Table 3 summarizes the baseline characteristics and clinical course of those patients. Among them, there was only one pediatric patient, and besides two adult recipients who received liver grafts from living donors, most of the cases were deceased donor LTs. In ten out of twelve patients, their wound was unable to be closed immediately after the transplantation because of increased intra-abdominal pressure, leading to inadequate perfusion to the graft and bowel ischemia. Two patients required allograft fascia to completely close their wound. Three patients underwent re-transplantation due to graft failure, and seven patients died after experiencing large-for-size syndrome.

Table 3 . Case information of patients with large-for-size syndrome, including the recipient’s and donor’s characteristics and clinical outcome

No.	Recipient					Donor					Outcome
	Sex/age	Height/weight (cm/kg)	Abdominal cavity (mm) (AP of right liver fossa/lateral of right liver fossa/AP of midline)	Cause of liver disease and brief history	MELD/PELD	Sex/age	Height/weight (cm/kg)	Donor type	Graft type	Graft weight (g)/GRWR (%)	Complication associated with large-for-size syndrome	Re-LT	Death
1	M/69	165/54.8	134.90/85.75/83.07	Liver cirrhosis d/t hepatitis B s/p previous abdominal surgery d/t pancreas neuroendocrine tumor	39	M/43	164/60	DD	Whole liver	1,923/2.90	Poor intrahepatic portal flow Unable to close the abdomen	No	Yes
2	M/12 m	88/11.0	87.50/59.12/49.20	Carbamoyl phosphate synthetase deficiency	3	F/34	161/51	DD	Left lateral	376/3.30	Poor intrahepatic portal flow Delayed wound closure with alloderm graft Small bowel ischemia with perforation	Yes	Yes
3	M/51	170/62	183.53/108.74/106.61	Alcoholic liver cirrhosis	28	M/64	157/72	DD	Whole liver	1,895/3.05	Delayed wound closure	No	No
4	F/64	154/63	136.93/82.33/103.66	Toxic hepatitis	40	M/51	176/89.4	DD	Whole liver	1,780/2.83	Cardiac arrest d/t reperfusion syndrome Poor intrahepatic portal flow Delayed wound closure	No	Yes
5	M/49	165/53.3	150.90/88.41/88.06	Acute liver failure d/t hepatitis A	40	M/24	175/75	DD	Extended right lobe	1,263/2.37	IVC compression by large liver graft Delayed wound closure with fascia lata allograft	No	No
6	M/53	178/59.70	166.31/114.09/85.68	HCC-B Acute on chronic liver failure d/t ischemic biliopathy after salvage LDLT	38	M/47	168/79.4	DD	Whole liver	1,803/3.14	Delayed wound closure Poor hepatic arterial flow Luminal narrowing of extrahepatic portal vein	Yes	Yes
7	F/32	158.8/50.8	155.49/119.61/95.28	Alcoholic liver cirrhosis Acute on chronic liver failure d/t diffuse bile duct injury after LDLT	37	M/44	175/69.9	DD	Whole liver	1,523/3.00	Delayed wound closure Primary non-function of large liver graft	No	Yes
8	F/48	169/63.8	127.54/89.76/62.20	Fulminant hepatitis A	39	M/58	171/58.5	DD	Whole liver	1,300/2.04	IVC compression by large liver graft Delayed wound closure	No	No
9	F/62	162/63.5	171.77/76.45/118.15	Liver cirrhosis d/t hepatitis B s/p previous abdominal surgery d/t endometrioid adenocarcinoma	40	F/59	155/55	DD	Whole liver	1,076/1.69	Delayed wound closure Small bowel ischemia with perforation	No	No
10	F/60	160/72.7	145.35/84.47/79.05	Acute liver failure d/t hepatitis B flare	40	M/35	168/84	DD	Whole liver	1,980/2.72	Unable to close the abdomen	No	Yes
11	F/41	158.3/63.8	152.10/76.75/92.37	Autoimmune hepatitis with HCC	15	M/40	174.9/73.75	LD	Right hemi liver	771/1.21	Hepatic artery thrombosis Unable to close the abdomen	Yes	No
12	M/41	167.4/81.8	141.08/109.39/89.93	HCC Atrophy on half of the body d/t 3rd degree burn	6	F/42	159.8/75.35	LD	Extended right lobe	644/0.82	Poor intrahepatic portal flow Unable to close the abdomen	No	Yes

M, male; F, female; AP, anteroposterior; MELD/PELD, Model for End-Stage Liver Disease/Pediatric End-Stage Liver Disease; GRWR, graft–recipient weight ratio; LDLT, living donor liver transplantation; HCC, hepatocellular carcinoma; LD, living donor; DD, deceased donor; d/t, due to; s/p, status post.

Large-for-size syndrome after LT is uncommon, but once it occurs, the consequences can be disastrous [11]. This usually occurs when a recipient with small abdominal cavity receives a large liver graft, in cases such as pediatric LT, or LT from a male donor to a female recipient. Transplanting a large liver graft can lead to increased intra-abdominal pressure of the recipient, followed by graft compression, and increases the risk of vascular complication, including inadequate portal and hepatic arterial flow, as well as the stenosis of hepatic veins. Poor oxygenation and hepatic congestion may lead to graft failure, and perhaps the death of a patient [4,12]. Recently, one study reported that the mortality rate of large-for-size syndrome after adult LT was up to 40% [13].

3D-printing technology has developed rapidly in recent years, and is being applied in various fields, including medicine. There also has been an effort to use 3D-printing for liver surgery, but mostly focused on liver malignancy for better anatomical understanding [8,14-16]. 3D-printing in LT has also been introduced in several studies [9,17,18]. Conventionally, to avoid size discrepancy between the liver graft and the recipient, surgeons measured graft volume to predict the graft-to-recipient weight ratio based on 2D CT images [3]. However, since the shape and thickness are also important, the occurrence of large-for-size syndrome could not be completely avoided, especially in pediatric LT. Wang et al. [18] successfully produced several half-sized 3D-printed models of livers for pediatric LT. These studies showed that the 3D-printed model could help to reduce the risk of large-for-size syndrome. 3D-printing techniques for LT in those studies focused on accurately reconstructing the anatomical structure of the liver graft, and showed that 3D-printing technology might be helpful for precise surgical planning. However, due to long production time and high cost, it is hard to routinely apply their methods. Therefore, no study has yet compared whether 3D-printing technology can lower the complication rates and improve the outcomes of LT. In this study, we managed to make a 3D-printed model that is easier to manufacture and can be applied to actual surgery.

We developed a technique for our 3D-printed model that can be a helpful tool for decision-making for selecting the appropriate liver graft for LT, and showed the new concept of using 3D-printing technology in LT. In our preliminary study, we have already confirmed the potential of our 3D-printed model in LT, since every LT case at risk of large-for-size syndrome was successfully performed using this 3D model [10]. As previously reported in our preliminary study, there were three cases where deceased donor LT was aborted due to large liver size. These cases were decided based on the 3D-printed models. The three patients eventually received liver from other donors. Among them, a 61-year-old female with liver cirrhosis had a chronic right hemi-thorax empyema, leading to contracture of the right liver fossa and extremely small abdominal cavity. After comparison with a 3D-printed abdominal cavity model, the initially selected deceased donor liver graft was rejected, and a LDLT was then planned. The 3D-printed model helped determine that the right liver graft from the living donor was too large, so an extended left liver graft was used, leading to a successful transplant. This is the first systematic study to show that 3D-printing technology can improve the outcome of LT, compared to the traditional management. To minimize selection bias in the study, we used PS-matched analyses between the two groups, and reviewed the clinical outcomes. Even though there was statistically no significant difference in the presence of large-for-size syndrome between the 3D-printed group and control group, none of the patients in the 3D-printed group experienced large-for-size syndrome after LT. Patients who underwent LT using a 3D-printed model showed significantly lower graft failure than the control group after PS matching. After PS matching, the 3D-printed group also showed superior graft survival to the control group.

One point that requires attention is that the control group after PS matching showed lower graft survival, compared to the original patient group. For PS matching, variables such as sex, age, height, and weight were matched. This resulted in selecting a patient group that was predominantly comprised of female, male with small body size, and pediatric recipients. Recipients with this entity have the potential of ending up with poor clinical outcome, compared to ordinary body sized recipients. The reason may be multifactorial, including lower probability of deceased donor matching due to small body size, higher risk of large-for-size syndrome, or inadequate donors for pediatric patients due to size mismatch. Although our 3D-printed model cannot solve the problem of low probability of deceased donor matching, it can help clinicians minimize false decision-making in using a potentially large liver graft. The present study is the result of applying 3D-printing for those purposes.

While we showed superior graft survival, overall survival was not different between the two groups. There were three cases with patient death in the 3D-printed group. One pediatric patient who expired due to asphyxia had comorbid condition, due to combined anomaly. The patient had tracheostomy cannula, which was accidentally displaced while the patient was taken care of by the family members after discharge. One female recipient also expired due to asphyxia due to aspiration after discharge and staying in her house. One female patient expired due to invasive aspergillosis that appeared four days after her re-transplantation using the 3D-printed model. All three cases were successfully transplanted without large-for-size syndrome. Although the 3D-printed model can help the surgeon prevent large-for-size syndrome, the medical condition and other surgical procedures should be handled with the same method as the conventional approach.

Although large-for-size syndrome is rare, surgeons should always take into account the possibility of the graft not fitting into the abdominal cavity. Therefore, understanding of the shape and size of the recipient in relation to the liver graft is essential. The 3D-printed model is not designed to reproduce the abdominal cavity as realistically as possible, but only to focus on the major anatomical structures, such as the diaphragm, parietal peritoneum, kidney, and inferior vena cava. During the study period, the manufacturing time that previously took about 2 days dramatically reduced to 6 to 7 hours. Unlike previous 3D models that must be printed without interruption at once, our model undergoes the procedure of printing the outlines of the recipient’s abdominal cavity, and then assembling the parts of the printed outlines. Therefore, multiple 3D-printers can be used simultaneously to print each part of the model, which greatly reduces printing time. Currently, we managed to reduce the manufacturing time to less than 3 hours by using two 3D-printers to reduce the printing time. This shows that in the future, we can further reduce the manufacturing time.

In addition to the 3D-printing technique, virtual simulation is a method that can help prevent large-for-size syndrome. In cases of deceased donor LT where radiological information, such as CT scans, is unavailable, and the surgery must proceed urgently, it can be challenging to perform a virtual simulation, making the 3D-printing technique highly useful. Regarding virtual simulation, the 3D-printed guide can be beneficial regarding mechanical contact and the 3D approach. Currently, there is no specific simulator that can demonstrate the size comparison that functions between two structures, which are the liver graft and the abdominal cavity. Of course, we demonstrated virtual simulation when 3D-printing was unavailable due to shortage of time or difference in region. However, virtual simulation also requires specific 3D model simulators, which are unfamiliar to most surgeons. Furthermore, we use 3D-printed liver grafts with a thermoplastic material that can demonstrate the elasticity of the liver graft. During virtual simulation, the elasticity of the liver graft, as well as the mechanical interaction between structure, is difficult to demonstrate. Additionally, during the bench procedure, if graft reduction is necessary, the use of a 3D-printed model allows for a more effective and appropriate graft reduction by comparing it with the model, even before the recipient’s total hepatectomy is completed, without the need for direct fitting.

Low production cost is another strength of our 3D-printed model. Compared to other 3D models in previous studies, our 3D-printed model requires only one material known as polylactic acid, which is cheaper than other materials. The fused deposition type 3D-printer that we used for manufacturing the 3D model also has the advantages of cost. The low cost and short manufacturing time allowed our team to apply the 3D-printed model in both elective cases, and in emergency LT. In addition, due to its easy manufacturing procedure, we expect that our 3D-printed model can be easily applied in other institutes.

One of the limitations in this study was distinguishing between the different causes of graft dysfunction, particularly in the context of large-for-size syndrome. While large-for-size syndrome is typically characterized by a rapid drop in blood pressure upon abdominal closure, other forms of graft dysfunction may present in a more subacute, systemic manner. These two phenomena, although related, are not always clearly separable, especially when relying on retrospective data where the available medical records may be limited.

In addition, there may be selection bias due to the small number of cases. To reduce selection bias, we performed PS matched analysis. In the near future, we are planning to conduct a multicenter study of the use of our 3D-printed model.

In conclusions, our novel 3D-printed model has the potential to improve the graft survival and prevent large-for-size syndrome after LT. Moreover, with the advantages of cost-effectiveness and fast manufacturing time, this model could be practically utilized in LT.

Supplementary data related to this article can be found at https://doi.org/10.14701/ahbps.24-153.

The manuscript has been placed in research square as a preprint (https://doi.org/10.21203/rs.3.rs-4157626/v1). This paper was presented as an abstract in Annals of Hepato-Biliary-Pancreatic Surgery in 2023. Ann Hepatobiliary Pancreat Surg 2023;27 Suppl 1:S82.

None.

No potential conflict of interest relevant to this article was reported, in that the protocol for producing the 3D-printed intra-abdominal cavity model is under the process of patent registration in the Republic of Korea.

Conceptualization: SP, JR. Data curation: SP, JR, GSC, JMK, SL, JWJ. Methodology: SP, JR. Supervision: JR. Writing - original draft: All authors. Writing - review & editing: All authors.