Outcomes of cranioplasty with customized artificial bone flap made by 3D printing technique in patients with aneurysmal subarachnoid hemorrhage

Article information

J Cerebrovasc Endovasc Neurosurg. 2026;28(1):35-48

Publication date (electronic) : 2025 October 31

doi : https://doi.org/10.7461/jcen.2025.E2025.09.003

Min Geun Gil¹, Sung-Tae Kim^,2, Se Young Pyo¹, Juwhan Lee¹, Jin Lee¹, Won Hee Lee¹, Keun Soo Lee¹, Sung-Chul Jin², Sung Hwa Paeng¹, Moo Seong Kim¹, Young Gyun Jeong¹

¹Department of Neurosurgery, Busan Paik Hospital, Inje University College of Medicine, Busan, Republic of Korea

²Department of Neurosurgery, Haeundae Paik Hospital, Inje University College of Medicine, Busan, Republic of Korea

Correspondence to Sung-Tae Kim Department of Neurosurgery, Haeundae Paik Hospital, Inje University College of Medicine, 875, Haeun-daero, Haeundae-gu, Busan, Republic of Korea Tel +82-51-890-6144 Fax +82-51-898-4244 E-mail kimst015@hanmail.net

Received 2025 September 14; Revised 2025 October 2; Accepted 2025 October 2.

Abstract

INTRODUCTION

Decompressive craniectomy (DC) can serve as a life-saving intervention in patients with aneurysmal subarachnoid hemorrhage (aSAH), particularly those in poor clinical condition, to reduce the high rates of mortality and disability associated with this condition [2,8].

If there is subsequent recovery in the patient’s neurological and medical conditions, cranioplasty may be considered. Cranioplasty restores the skull’s appearance and protects the brain, enhancing neurological recovery by improving cerebral blood flow, cerebrospinal fluid dynamics, and metabolic activity [4,16,23,29].

Autologous bone flaps are the preferred choice for cranioplasty at most institutions due to their cost-effectiveness, biocompatibility with low rejection rates, radiolucency, and the ideal contour provided by using the patient’s own skull [16]. However, complications such as bone resorption and infection have led to high failure rates in cranioplasty [28,38]. Additionally, the 3-D printing technique provides a superior coverage ratio compared to autologous bone in cranioplasty, offering patients greater satisfaction cosmetically [28,38]. Therefore, due to the inherent difficulties and complications of autologous bone cranioplasty(AC), there is increasing interest in bone substitutes and technologically advanced implants to enhance outcomes and aesthetics [20,33].

Recent studies have compared 3-D printed artificial skull bone flaps with autologous bone in cranioplasty without distinguishing the underlying causes of craniectomy [3,28,38]. However, only a few studies have examined cranioplasty outcomes specifically in patients with aSAH. These patients are often older and tend to show cerebral atrophy with enlarged subdural spaces, and many of them require ventriculoperitoneal (VP) shunting [14]. Such conditions increase the risk of complications and make cranioplasty outcomes more difficult to achieve. In this study, we compared the cosmetic and clinical outcomes of cranioplasty using 3-D printed implants and autologous bone in patients with aSAH.

MATERIALS AND METHODS

This retrospective study analyzed the medical records and radiologic images of patients who underwent decompressive craniectomy followed by cranioplasty for aneurysmal subarachnoid hemorrhage (aSAH). All procedures involving human participants were conducted in accordance with the ethical standards of the Institutional Review Board of Busan Paik Hospital, Inje University College of Medicine, and with the principles of the Declaration of Helsinki (IRB No. 2024-04-009).

Cranioplasty style and material

Preoperative CT scans with 1-mm slice thickness were obtained for all patients. DICOM files were processed using Mimics 25.0 software (Materialise, Ann Arbor, MI, USA) to generate three-dimensional cranial models. The contralateral normal side was mirrored to reconstruct the cranial defect, and the surgeon specified the augmentation site during the design stage, with an additional 2–4 mm of convexity incorporated at the temporal muscle attachment. The reconstructed models were refined using Rhinoceros 7 (Robert McNeel & Associates, Seattle, WA, USA) and Magics 24.01 (Materialise NV, Belgium) for design optimization. Finalized patient-specific models were then used for implant fabrication. Titanium implants (Ti-6Al-4V ELI) were produced via Laser Selective Melting using a CM150 3-D printer (CusMedi Co., Korea). PEEK patient-specific implants (PSIs) were manufactured through subtractive milling from radio-opaque PEEK (Polyetheretherketone) blocks (Invibio Ltd., United Kingdom). Fig. 1 presents 3-D printing program images and representative cases of cranioplasty using customized artificial bone flaps fabricated by a 3-D printing technique with titanium and PEEK implants.

Fig. 1.

Cranioplasty using customized artificial bone flaps fabricated by 3-D printing: (A) preoperative 3-D simulation with skull defect and flap design, (B) implantation of a mesh-type titanium flap, and (C) implantation of a solid-type PEEK flap. 3-D, three-dimensional; PEEK, polyetheretherketone.

Measurement of craniectomy lesions and defects

In the case of three-dimensional cranioplasty (3-DC), the defect values and flap surface area were measured using Mimics Research Version 19.0 and 3-matic Research 11.0 made by Materialize. (Leuven, Belgium) After obtaining the craniectomy area and autologous bone flap area using Materialise Mimics and 3-matic, respectively, the subtraced area was defined as the defect area. The following measurements were obtained from postoperative imaging: circumference of the craniectomy, skull defect area, circumference of the bone flap, bone flap area, percentage of skull defect coverage by the bone flap, and residual defect area (mm²) with its mean residual defect size (mm). Fig. 2 illustrates the measurement method.

Fig. 2.

Schematic illustration of craniectomy and bone flap placement. The outer black line indicates the craniectomy margin, used to measure the craniectomy circumference. The inner gray line delineates the boundary of the autologous bone flap, from which the flap circumference and area were calculated. The gray-filled region represents the bone flap, while the blue-hatched area depicts the uncovered portion of the skull defect, used to determine defect area and bone-flap coverage.

RESULTS

This study included data from 50 patients who underwent cranioplasty using either 3-D printed implants or autologous bone, all of whom had previously undergone decompressive craniectomy following aneurysmal subarachnoid hemorrhage (aSAH). Among these patients, four underwent bilateral cranioplasty, resulting in a total of 54 initial cranioplasties.

the 50 patients (mean age 55.3 years, SD 11.6;16 males), 26 received a 3-D printed bone flap (mean age 58.4 years, SD 10.3; 11 males) and 24 received an autologous bone graft (mean age 51.7 years, SD 12.2; 5 males). Aneurysm locations were 14 in the ACA, 10 in the ICA, 25 in the MCA, and 1 in the posterior circulation (3-DC group: 7 ACA, 4 ICA, 15 MCA, and 0 posterior circulation aneurysms). The modified Fisher grade distribution showed that 18 patients were classified as grade III and 32 as grade IV (3-DC group: 7 patients were classified as grade III, 19 as grade IV). On the Hunt and Hess scale, the total cohort included 9 patients at scale III, 31 at scale IV, and 10 at scale V (3-DC group: 3 patients at scale III, 16 at scale IV, 7 at scale V). Regarding treatment methods of ruptured aneurysm, 36 patients underwent microsurgery, while 14 received endovascular treatment (EVT) (surgery: 3-DC group 19 patients; EVT: 3-DC group 7 patients). Out of a total of 50 patients who underwent cranioplasty, 36 patients had a VP shunt prior to the surgery (3-DC group: 18 patients). The mean interval between craniectomy and cranioplasty was 4.8 months (3-DC group: 5.6 months; AC group: 3.8 months). Overall, no statistically significant differences were observed between the groups in terms of sex distribution, aneurysm location, modified Fisher grade, Hunt and Hess scale, aneurysm treatment methods, number of VP shunt patients, and the interval between craniectomy and initial cranioplasty, except age. Preoperatively, the modified Rankin Scale (mRS) scores in the total patient were 4 points for 3 patients and 5 points for 32 patients. After 6 months, the scores improved to 4 points for 20 patients and 5 points for 10 patients. The demographic data and clinical outcome of the 50 patients are presented in Table 1. Overall, neurological function assessed by the mRS was maintained or improved compared with preoperative status.

Table 1.

Demographic data and clinical outcome of the patients underwent cranioplasty after aneurysmal rupture and decompressive craniectomy

A total of 54 hemispheres underwent initial cranioplasty. Among the 28 cases of 3-D cranioplasty, titanium was used in 24 cases and polyetheretherketone (PEEK) in 4 cases. In unilateral cranioplasty cases, 46 hemispheric procedures were performed (3-DC group: 24; AC group: 22). Bilateral cranioplasty was performed in four patients, totaling eight hemispheric procedures—four with autologous bone and four with 3-D printed implants. The average initial operation time was 159.2±61.7 minutes (3-D group, 164.8±66.8 minutes; AC group, 152±58.7 minutes). The average blood loss during initial operations was 151±143.1 ml (3-D group, 154.3±131.9 ml; AC group, 149.1±123.9 ml). There were no statistically significant differences in operation time or blood loss (p=0.432 and p=0.990, respectively). The distribution of hemispheric laterality, material type, operation time, and blood loss is summarized in Table 2.

Table 2.

Operative and morphometric parameter of cranioplasty

Morphometric and statistical analyses were conducted, as described in Table 2. Comparative analysis was based on morphometric parameters, including the surface area and circumference of the skull defect and bone flap, the bone flap coverage ratio, and the residual defect area. The circumference of craniectomy was significantly larger in the 3-D cranioplasty group compared to the autologous bone cranioplasty group (361.6± 8.8 mm vs. 336.6±53.1 mm, p=0.039). Similarly, the skull defect area was greater in the 3-D cranioplasty group (10278.7±2714.4 mm²) than in the AC group (7951.7±3160.8 mm²), with a statistically significant difference (p=0.003). The circumference of the bone flap was also significantly larger in the 3-D group (352.9±39.3 mm vs. 322.3±54.3 mm, p=0.018), as was the bone flap area (9939.9±2646.9 mm² vs. 7423.2±2813.6 mm², p=0.002). Bone flap coverage of the skull defect was higher in the 3-D group (96.7% vs. 93.4%, p=0.044), and the residual defect area was significantly smaller (338.7±274.2 mm² vs. 528.5±331.3 mm², p=0.049).

A total of 24 operation-related issues were observed, with 9 cases in the 3-DC group and 15 in the AC group, with no statistically significant difference (p=0.0994). In the 3-DC group, complications included subgaleal hematoma (n=2), fluid collection (n=4), epidural abscess (n=1), and wound dehiscence (n=2), with no cases of subdural emphysema. Bone resorption, inherent to autologous bone grafts, was not applicable to the 3-DC group. In contrast, the AC group presented with subgaleal hematoma (n=1), fluid collection (n=2), epidural abscess (n=3), subdural emphysema (n=1), and bone resorption (n=8, Fig. 3). Bone flap removal was performed in 1 patient in the 3-DC group and in 5 patients in the AC group (p=0.0948). Of note, most complications occurred in patients with a VP shunt (3-DC group: 9 out of 9; AC group: 14 out of 15). (Table 3)

Fig. 3.

Serial postoperative radiographs and CT scans following cranioplasty with autologous bone after decompressive craniectomy. (A, B) Immediate postoperative images showing a well-positioned bone flap without evidence of resorption. (C, D) Follow-up images obtained 175 days postoperatively demonstrating marked bone flap resorption (white arrows). Both the radiograph (C) and axial CT scan (D) reveal bilateral frontal bone sinking secondary to severe resorption. CT, computed tomography

Table 3.

Operation related issue

As summarized in Table 4, revision surgery was required in six patients. Five (Cases 1–5) initially received autologous bone cranioplasty, and one (Case 6) received a 3-D titanium implant as the primary material. Four patients (Cases 1–4) initially underwent autologous bone cranioplasty but later required replacement with 3-D printed implants due to complications. Case 1 developed an epidural abscess approximately 50 days after autologous bone cranioplasty, necessitating removal of the bone flap and infection control, followed by repeat cranioplasty with a 3-D titanium implant about seven months later (Fig. 4). Case 2 had received the initial autologous bone cranioplasty at an outside institution before undergoing revision at our center (Fig. 5), Case 3 experienced a recurrent epidural abscess, which was successfully treated with a second revision using the same sterilized 3-D titanium implant.

Table 4.

Reoperations for post-cranioplasty complications

Fig. 4.

Radiologic images of Case 1.

(A) MRI image obtained approximately 50 days after autologous bone cranioplasty, revealing an air–fluid level in the right epidural space, suggestive of an epidural abscess. (B) Immediate postoperative CT following removal of the autologous bone flap and irrigation of the abscess cavity. (C) Postoperative CT obtained about seven months after infection control, following cranioplasty with a patient-specific 3-D printed titanium implant. (D) Postoperative skull radiograph (anteroposterior view) 3-D, three-dimensional; CT, computed tomography; MRI, magnetic resonance imaging

Fig. 5.

Representative case of revision cranioplasty using a patient-specific 3-D printed PEEK implant. (A) Postoperative CT after revision cranioplasty with a PEEK implant. The patient developed an epidural abscess 116 days after undergoing autologous bone cranioplasty at another hospital, and subsequently underwent PEEK cranioplasty at our institution. White arrows indicate the PEEK implant, which appears indistinct on CT due to its radiolucent property. (B) Post-revision clinical photograph demonstrating a well-contoured cranial shape without cosmetic deformity. (C) The patient-specific PEEK implant used in this case, assembled by joining two PEEK blocks with mini-plates due to technical limitations. 3-D, three-dimensional; PEEK, polyetheretherketone; CT, computed tomography

Of the remaining patients, Case 4 developed bone resorption with right temporal depression, which was corrected using autologous bone harvested from the contralateral parietal region. Case 5 experienced a complicated fluid collection necessitating removal of the autologous bone flap without subsequent cranioplasty. Case 6, who had previously undergone cranioplasty with a 3-D titanium implant, developed wound dehiscence with titanium plate exposure, which was managed with wound closure in collaboration with the plastic surgery department (Fig. 6). Post-revision outcomes were generally favorable, with no recurrence in Cases 1, 2, 4, 5, and 6.

Fig. 6.

Clinical photograph of Case 6 showing wound dehiscence with exposure of a 3-D titanium plate and mesh, occurring approximately 770 days after the initial cranioplasty. The patient subsequently underwent wound closure in collaboration with the plastic surgery department, with no recurrence thereafter. 3-D, three-dimensional

DISCUSSION

Aneurysmal subarachnoid hemorrhage (aSAH) has high disability and mortality rates, and decompressive craniectomy (DC) can improve survival in patients with poor neurological status [2,8,40]. After DC, once cerebral edema subsides, cranioplasty is performed to protect intracranial structures, promote neurological recovery, and restore cranial contour, thereby improving psychological well-being and social reintegration [5,32,39]. Autologous bone cranioplasty (AC) remains the most widely used method, but it has limitations such as bone storage issues, high infection rates, bone flap resorption (BFR), and cosmetic deformities [24,37]. To overcome these limitations, customized three-dimensional (3-D) printed implants have increasingly been adopted in cranioplasty [7,13].

As demonstrated by morphometric analysis (Table 2), 3-D implants achieved significantly higher defect coverage and smaller residual defects compared with autologous bone grafts, supporting the superiority of 3-DC in anatomical restoration. In DC, the temporal bone is often extensively removed to achieve maximal decompression of the midbrain [19]. However, AC often fails to fully restore this area, leading to residual defects and temporal hollowing. In contrast, 3-D printed implants are customized to replicate the normal cranial contour, providing superior cosmetic restoration [28]. Apart from the precise fit of the implant to the margins of the cranial defect, patients generally assess cosmetic outcomes based on the external cranial contour together with the condition of the overlying soft tissues. Both the DC and the dissection required to separate adhesions between the dura and the temporalis muscle during subsequent cranioplasty can lead to temporalis muscle injury and progressive atrophy [15,41]. In the case of 3-D cranioplasty, the bone flap is custom-made, allowing consideration of factors such as temporalis muscle atrophy and the adequacy of the overlying skin and soft tissues. During the design stage, the surgeon specified the area for augmentation, and an additional 2–4 mm of convexity was incorporated at the temporalis attachment site. This approach allows flexible adjustment of the degree of prominence or concavity, thereby improving aesthetic harmony and demonstrating a clear advantage of customized 3-D printed implants [28,34,41]. In our study, patients in the 3-DC group tended to be older, which may reflect a preference for stable and convenient 3-D implants in elderly patients, where minimizing the risk of reoperation due to wound complications or infection is of greater concern. Also, the mean defect size was larger in the 3-D cranioplasty (3-DC) group than in the AC group. Larger cranial defects and older patients are generally associated with higher complication rates [32]. Nonetheless, complication rates were higher in the AC group than in the 3-DC group.

When complications were compared in detail (Table 3), wound dehiscence and fluid collection were more frequent in patients receiving 3-D printed implants, whereas BFR and epidural abscess occurred more frequently in the AC group. Among the 26 patients who underwent AC cranioplasty, eight developed BFR. Patients diagnosed with BFR not only experienced cosmetic deformities but also a loss of protective function of the skull [25,35]. One patient experienced scalp sinking due to severe BFR, while another required additional bone harvesting from the contralateral parietal bone because of marked resorption. Such complications, including BFR and epidural abscess, are often difficult to manage conservatively and usually require additional surgery, representing severe adverse events. Therefore, these complications occurred predominantly in the AC group and were a major factor contributing to the higher rate of revision procedures in this cohort (Table 4).

BFR is a well-recognized long-term complication following cranioplasty [25,30]. Revascularization and osteoinduction are key determinants of bone flap healing [22]. Histological studies have demonstrated that revascularization and osteoblast infiltration progress from the edges of the implanted bone flap [25,36]. Therefore, close contact between the autologous bone flap and the surrounding bone defect margin is considered a critical factor for graft survival [17,36]. During emergency craniectomy, extensive hemostasis is often required, involving coagulation or ligation of local vessels and meticulous bleeding control before wound closure [1,11]. Additional hemostasis is also performed during subsequent cranioplasty, and such repeated vascular hemostasis may hinder revascularization of the implanted autologous bone flap at the craniectomy site [1]. Also, prolonged cryopreservation of autologous bone increases osteoclast activity, which may exacerbate BFR [6,35]. Additional risk factors for BFR include larger flap size, a longer interval between DC and cranioplasty, and younger patient age [17,25,35,38]. Although younger patients with aSAH often achieve favorable outcomes after DC, they paradoxically show a higher risk of BFR following cranioplasty [8,25]. Moreover, ventriculoperitoneal (VP) shunting has been proposed as another risk factor for BFR [3,26,38]. Shunt-related alterations in intracranial pressure may interfere with firm integration of the bone flap and impair osteogenesis. Furthermore, decreased intracranial pressure reduces the contact between the dura mater and the inner table of the skull, potentially affecting cranial growth through dural signaling mechanisms [3,10,12,35].

Wound dehiscence and titanium exposure observed in the 3-DC group, similar to BFR, may be associated with local perfusion impairment caused by hemostasis performed during both craniectomy and subsequent cranioplasty [18,21,27]. Excessive coagulation can compromise the vascular supply of the scalp flap and weaken its healing capacity [9,27]. In addition, shunt-related negative intracranial pressure may induce scalp depression and adherence of the thin skin flap to the underlying implant, thereby further promoting wound dehiscence and, in some cases, leading to titanium mesh exposure [22,31].

Infection is a frequent complication of cranioplasty (CP). In patients with aSAH, the risk may be increased because they are often elderly and present with severe cerebral atrophy, enlarged subdural spaces, and multiple comorbidities [14]. In our study, three cases of epidural abscess occurred exclusively in the AC group, all requiring revision surgery. In each case, the infected autologous bone was removed and replaced with a 3-D printing implant [33]. One patient underwent two reoperations, during which the previously used 3-D PEEK implant was sterilized by autoclaving and reused, without further complications. Previous studies have associated infection risk with factors such as bone storage method, implant material, surgical interval (DC–CP), and the presence of VP shunts [14,26].

Although Autologous bone cranioplasty (AC) remains a feasible method but is vulnerable to complications such as bone flap resorption, infection, and cosmetic deformities, which may lead to repeated surgeries, higher healthcare costs, and worse patient outcomes. Since most aSAH patients belong to a poor outcome group with mRS scores of 4–5, the use of costly 3D cranioplasty in this population may be controversial in terms of cost-effectiveness. Therefore, patients should be fully informed of the advantages and disadvantages of both approaches. If cost is not a limiting factor, 3-DC may serve as a reasonable alternative in elderly aSAH patients, who frequently require VP shunts, given their potential to reduce complications and improve surgical outcomes.

In interpreting these findings of this study, the status of ventriculoperitoneal (VP) shunting should be taken into account. VP shunt has been reported as an important factor closely associated with postoperative complications following cranioplasty [14,38]. In our cohort, a high proportion of shunt-dependent patients was observed in both groups, with no significant difference between them. This likely reflects the poor baseline condition of patients undergoing decompressive craniectomy, in whom shunt insertion was prioritized before cranioplasty. Therefore, the interaction between VP shunt and cranioplasty outcomes, particularly in patients with aSAH, warrants further investigation. This study has several limitations. It was a single-center retrospective analysis with a small sample size and short follow-up, limiting the evaluation of long-term outcomes. The non-randomized selection of cranioplasty materials may have introduced selection bias, as patient or guardian choice rather than predetermined criteria determined implant use. Unmeasured factors such as patient age, defect size, VP shunt status, or surgical timing could therefore have influenced outcomes. Moreover, cosmetic assessment relied mainly on morphometric data without validated patient-reported outcomes. In addition, variability in patient characteristics and surgical techniques may have affected the results. Furthermore, in the 3D group, titanium implants predominated, with only a few PEEK implants. As these materials differ in biomechanical and complication-related properties [39], this heterogeneity is a study limitation. Larger multicenter prospective studies with longer follow-up are needed to confirm our findings.

CONCLUSIONS

In aSAH patients undergoing cranioplasty after decompressive craniectomy, customized 3-D printed implants achieved significantly better anatomical restoration and showed a trend toward fewer complications compared with autologous bone. While autologous bone remains a feasible option, its risks of resorption and infection often necessitate revision. Therefore, in aSAH patients with multiple risk factors, including advanced age, cerebral atrophy, and a high prevalence of VP shunts, 3-D printed implants may be considered a reasonable alternative.

Article information Continued