How precise is excision and reconstruction using 3D printing technology for total sacrectomy: accuracy validation in 9 consecutive cases

Abstract

Background

With 3D printing technology, we can now use preoperative imaging for precise surgical plan. We can also use patient-specific surgical jig to improve the accuracy of osteotomy and 3D-printed custom-made endoprostheses combined with a screw-rod system to restore lumbosacral stability. The aim of this study was to evaluate the accuracy of 3D printing technology for precise osteotomy during total sacrectomy.

Methods

Nine patients with primary malignant tumors of the sacrum who underwent total sacrectomy at our center were enrolled. Osteotomy was planned based on preoperative imaging (CT, MRI). Generally, an additional 8–10 mm margin beyond the tumor was determined by the fusion of MR and CT images. Patient-specific surgical jigs and 3D-printed sacral endoprostheses were then designed based on the planned osteotomy planes. Pre- and postoperative 3D models of the lumbosacral and pelvic regions were constructed using the fiducial registration model of 3D slicer software 5.1.0. Postoperative CT scans were compared with the planned osteomy planes based on preoperative CT scans, in order to evaluate the accuracy of the osteotomy and endoprosthetic reconstruction. For each patient, four levels of osteotomy planes were chosen, including the upper edge of the sacroiliac (SI) joint, the S1 and S2 foramen levels, and the caudal edge of the SI joint, for analyzing position and angular deviations between the preoperative plan and actual osteotomy along with the endoprosthesis position.

Results

Pathological diagnoses included four cases of osteosarcoma, four cases of chordoma, and one case of Ewing sarcoma. All osteotomies in nine patients achieved R0 resection, as verified pathologically. An average angular deviation of 4.27° (interquartile range[IQR] 4.15) and an osteotomy position deviation of 4.00 mm (IQR 2.90) were observed. The mean angular deviations of the four levels were 3.50° (IQR 6.02), 3.86° (IQR 2.55), 4.81° (IQR 4.37), and 4.92° (IQR 3.27). The mean position deviations at the four levels were 3.15 mm (IQR 3.54), 3.55 mm (IQR 1.37), 4.26 mm (IQR 2.61), and 4.86 mm (IQR 3.93). No significant difference was found among the angular and position deviations at different levels. However, the proportions of individuals with position deviations > 2 mm and > 5 mm were significantly greater at the caudal end of the SI joint than at the upper end. All position deviations were within 8 mm. The average follow-up duration was 24.4 months. At the last follow-up, three patients experienced local recurrence, and one patient died of disease. All endoprostheses were in place without significant displacement. The mean Musculoskeletal Tumor Society scoring system (MSTS93) and MUD scores (function and sensation of lower limbs (M), urination and uriesthesia (U), and defecation and rectal sensation (D)) were 19.4 (16 to 24) and 16.3 (12 to 24), respectively.

Conclusion

Notably, 3D-printed patient-specific surgical jigs exhibit high accuracy of osteotomy and lead to optimal surgical margin and reconstruction in total sacrectomy. Effective and reliable reconstruction can be achieved with a custom-made 3D-printed endoprosthesis. The application of 3D printing technology using patient-specific surgical jigs and the custom-made 3D-printed implants exhibited high surgical accuracy in total sacrectomy, as evidenced by accuracy validation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41205-025-00295-6.

Background

Surgical resection is the main treatment option for primary sacral tumors. Compared to patients with tumors originating from the extremities, patients with primary sacral tumors have notably poorer prognoses and higher local recurrence rates [1–3]. Preoperative chemotherapy is crucial in the treatment of bone sarcomas since it can reduce tumor size, thus improving surgical and patient outcomes [4]. However, the primary types of sacral tumors, chordoma and chondrosarcoma, are insensitive to chemotherapy [5–7]. Owing to the concealed symptoms of sacral tumors and the large space of the pelvic region, sacral tumors tend to be large at the time of diagnosis [7]. R0 resection with a wide margin is crucial for the survival and prevention of local recurrence in patients with sacral tumors [8, 9]. However, the proximity of vital structures, including visceral organs, major vessels and nerves, increases the difficulty and complexity of surgical excision [8]. In cases involving the sacroiliac (SI) joints with pronounced curvature, precise control of the position and direction of the osteotomy is more challenging than in the extremities. A Gigli saw is an efficient tool for osteotomy but its position and direction control is poor [10, 11]. During total sacrectomy, wide resection with excessive removal of normal bone tissue leads to a large residual bone defect and impedes reconstruction and postoperative function. Therefore, precise osteotomy and successful reconstruction during total sacrectomy are highly important for improving survival and functional outcomes in patients with sacral tumors.

To preserve as much unaffected tissue as possible while ensuring R0 resection, computer-assisted surgery (CAS) and the patient-specific instrument (PSI) were introduced to the field of orthopedic oncologic surgery. Studies have shown that both CAS and PSI could improve surgical accuracy [8, 12]. With CAS, surgeons can view the clinical scenario in 3D and in real time, thus decreasing the risk of inaccurate resection [13]. However, during CAS, the scope of resection cannot be accurately defined due to limitations associated with navigating the saw. The entry point may be identified, but the exit point may be uncertain without adequate assessment [14]. Based on preoperative images, a PSI can be manufactured using 3D printing technology and used in surgery to guide resection. The combined use of a PSI and an oscillating saw allows the precise creation of entry and exit points for total sacrectomy since there are notable limitations of using only a PSI or only an oscillating saw to ensure adequate resection. Using a patch-type PSI, Fragnaud et al. reported a 3 mm deviation of osteotomy on average (1 to 8 mm) during pelvic tumor resection in cadavers [15]. Siegel et al. reported that osteotomy of the SI joint, which is a crucial step during total sacrectomy, may be more precise with a PSI than with CAS [16]. Moreover, Bosma et al. reported that surgical accuracy was higher with the PSI than with CAS and that the duration of osteotomy with a PSI was shorter than that during CAS in a cadaveric study [17]. 3D printing technology has been adopted to repair bone defects following total sacrectomy. Custom-made 3D-printed total sacral endoprostheses can be used to reconstruct the lumbosacral, lumboiliac and pelvic posterior rings in one step [18]. Surgical accuracy during total sacrectomy for sacral malignancy is vital not only for surgical margin control but also for 3D-printed endoprosthetic reconstruction, which requires adequate surface match. Based on the preoperative CT and MRI scans, 3D virtual models of the structures of the lumbosacral and pelvic rings were generated using the fiducial registration model of the 3D slicer software 5.1.0 (https://www.slicer.org/) [19]. The planned surgical margin was designed based on the theoretical surgical margin plus the systematic error of the PSI. Osteosarcoma patients with a margin distance greater than 4 mm showed no local recurrence [20], and a 4–6 mm deviation was observed in bone tumor resection using a PSI [16, 21, 22]. Therefore, an 8–10 mm margin beyond the tumor was planned. The locations of the bilateral and upper osteotomy planes are determined by the tumor location and the location of the L5 lower endplate. After determining the osteotomy planes, patient-specific osteotomy guides and 3D-printed custom-made sacral endoprostheses were designed accordingly.

In this study, we present the accuracy validation of total sacrectomy and reconstruction with patient-specific surgical jigs and 3D-printed sacral endoprostheses. The accuracy of the surgical procedures was analyzed among multiple levels. To our knowledge, this is the largest study to reveal the use of patient-specific surgical jigs and 3D-printed sacral endoprostheses during total sacrectomy and the first study to validate the accuracy of this procedure.

Methods

Nine patients with primary sacral malignant tumors who underwent total sacrectomy using patient-specific surgical jigs and 3D-printed sacral endoprostheses between March 2019 and January 2024 were recruited for this study. The average age of the patients was 43 years (15 to 66). There were four cases of osteosarcoma, four cases of chordoma, and one case of Ewing sarcoma. A patient-specific surgical jig was adopted to improve surgical accuracy and achieve R0 resection. A 3D-printed sacral endoprosthesis was designed to match the bone defects that remained after the osteotomy, which was performed with the assistance of a patient-specific surgical jig. All operations were performed by the senior author (TJ). Informed consent was obtained from all patients, and the study was approved by the Medical Ethical Committees at the authors’ institution.

Design of patient-specific surgical jig and 3D-printed sacral endoprosthesis

Bilateral iliac osteotomy lateral to the SI joint was planned based on CT and MRI scans. The PSI included three major elements: (1) a matching surface, which could fit onto the bone surface of the spinal iliaca posterior superior (SIPS); (2) fixation holes, which were made to allow passage of 2.5 mm Kirschner wires to fix the jig onto the bone surface. The sites of fixation were on the lateral side of the osteotomy for a safe surgical margin. (3) osteotomy slots, which improved the accuracy of osteotomy by physically restricting the oscillating saws to a specific position and orientation. Additionally, the depth of the slots on preoperative CT and MRI served as a reference for depth control during osteotomy with an oscillating saw. Additionally, an endoprosthesis placement jig was designed to achieve accurate placement during endoprosthetic fixation.

A 3D-printed sacral endoprosthesis was designed to match the bone defect based on the osteotomy planes. Generally, the design was similar to that of the endoprosthesis reported previously [11]. The prosthesis restored the structure from the lumbar spine to the bilateral ilium with a closed pelvic posterior ring in one step, and the metallic porous surface facilitated bone ingrowth at the bone-implant interface, including the inferior endplate of the L5 vertebra and bilateral iliac osteotomy site, thus allowing reconstruction of the lumbosacral joint and bilateral SI joints. Usually, at least two screws (diameter, 6.5–7 mm) are used to fix the endoprosthesis. The combination of a rod-screw system and an endoprosthesis provided adequate stability at the bone-implant interface for bone ingrowth. For all the patients, bilateral L5 screws were connected to the sacral endoprosthesis, and L3-4 screws were connected to the iliac pedicle screws, which were placed most caudal to the iliac osteotomy just superior to the ischiatic notch.

Surgical procedures

Preoperative selective artery embolization and intraoperative aortic balloon occlusion were performed to reduce blood loss during total sacrectomy [23]. An inverted Y incision was made, and dissection and exposure were performed as reported previously [24]. Following exposure of the posterior sacrum, the surgical jig was placed based on the surface profile of the bilateral PIS, and there was also a column in the middle of the jig that contacted the spinous process of S1. Then, K-wires were used to fix the jig. An oscillating saw was used for osteotomy in the appropriate direction and to an adequate depth with a jig. Following thorough osteotomy, L5/S1 discectomy was performed to ensure adequate patient stability. Then, the sacral nerve and vessels were ligated directly. Then, 3D-printed sacral endoprostheses were used for reconstruction. The bilateral iliac surface was first fixed with reference to the endoprosthetic placement jig. Then, screw fixation was performed at L5. Two iliac screws were placed distal to the endoprosthesis, and four rods were used to connect the L5 implant with the L3-4 ilium. Crosslinking was used to enhance the traction of the posterior ring.

Follow-up

All patients were followed up regularly every three months for the first two years postoperatively and every six to twelve months thereafter. At each follow-up visit, X-ray and CT scans (if necessary) were performed to evaluate the surgical region, and the implant, chest CT scan and bone scan were regularly performed to rule out metastasis. Functional outcomes were evaluated using the Musculoskeletal Tumor Society scoring system (MSTS93) [25] and MUD (function and sensation of lower limbs (M), urination and uriesthesia (U), and defecation and rectal sensation (D)) [26] scores.

Surgical accuracy validation

The surgical accuracy was validated based on the comparison between the postoperative CT scans and the planned osteotomy planes based on preoperative CT scans. Post- and preoperative 3D models of the lumbosacral and pelvic regions were constructed using the fiducial registration model of 3D slicer software 5.1.0 (https://www.slicer.org/) [19]. Measurements of the 3D models were used for surgical accuracy validation. The deviation between the planned osteotomy planes and the achieved osteotomy planes was assessed according to three types of differences: (1) if the achieved osteotomy plane was parallel to the planned osteotomy plane, the angular deviation was recorded as 0°, and the position deviation was recorded as the distance between the two planes (Fig. 1A); (2) if the achieved osteotomy plane crossed the planned osteotomy plane, the angular deviation was recorded based on the angle between the two planes, and the position deviation was recorded as “NA” (not applicable) (Fig. 1B); (3) if the achieved osteotomy plane was not parallel to and did not cross the planned osteotomy plane, the angular deviation was recorded as the angle between the extension lines of the two planes, and the position deviation was recorded as the mean value of the longest and shortest distance between the two planes (Fig. 1C).

Fig. 1.

Surgical accuracy validation protocols. (A) The angular deviation was recorded as 0° if the achieved osteotomy plane was parallel to the planned osteotomy plane, and the position deviation was recorded as the distance between the two planes, as shown by the red dotted line. (B) The position deviation was recorded as “NA” if the achieved osteotomy plane crossed the planned osteotomy plane, and the angular deviation was recorded based on the angle between the two planes. (C) The angular deviation was recorded as the angle between the extension lines of the two planes if the achieved osteotomy plane was not parallel to and did not cross the planned osteotomy plane, and the position deviation was recorded as the mean value of the longest and shortest distance between the two planes. The red dotted line represents the distance, and the red arrow represents the angle. NA, not applicable

The position and angular deviations were analyzed at four different levels for each patient (Fig. 2): the upper edge of the SI joint (Level 1), the S1 neural foramen (Level 2), the S2 neural foramen (Level 3), and the caudal edge of the SI joint (Level 4). At each level, bilateral osteotomy deviations were analyzed.

Fig. 2.

Schematic diagram of the four different levels for accuracy validation. Levels 1–4 represented the upper edge of the SI joint, the S1 neural foramen, the S2 neural foramen, and the caudal edge of the SI joint, respectively. At each level, the bilateral position and angular deviations were calculated

Statistical analysis

Statistical analyses were conducted via GraphPad Prism 8.0. Student’s t test was utilized for normally distributed continuous data, and the Mann–Whitney test was utilized for nonnormally distributed continuous data. Categorical variables were assessed using the X² test or Fisher’s exact test. A p value < 0.05 was considered to indicate statistical significance.

Results

Demographics of the patients

The present study included nine patients (6 males and 3 females) with primary malignant sacral tumors who underwent total sacrectomy surgery (Table 1). The mean age of the cohort was 43 years (15 to 66). The pathological diagnoses included four cases of osteosarcoma, four cases of chordoma, and one case of Ewing sarcoma. The average tumor size was 13.1 cm (8 to 17). Based on the level and extent of the lesion, a posterior approach was used in 8 patients, and a combined anteroposterior approach was used in one patient. The mean operation time was 545 min (range 330–675 min), and the mean estimated intraoperative blood loss was 3211 ml (range 2000–4300 ml). Wide resection with R0 resection was achieved in all nine patients. Perioperative complications occurred in three patients—two with a deep infection and one with wound dehiscence. Deep infections were treated with DAIR procedures (debridement, antibiotics, and implant retention). No major implant failure was observed, and screw breakage occurred in one patient at 8 months after surgery.

Table 1.

Characteristics of nine patients who underwent patient-specific instrument resection and custom-made 3D-printed endoprosthesis reconstruction

Case	Sex	Age (yr)	Pathological diagnosis	Tumor size (cm)	Follow-up (month)	Status*	R0 resection	Blood loss (ml)	Perioperative complications and treatment	Implant failure	Critical structures involved	Postoperative hospital stay (day)	MSTS 93/ MUD scores*
1	Male	45	Chordoma	12	62	NED	Yes	3600	Wound dehiscence, debridement	None	None	18	24(80%)/24
2	Male	15	Ewing’s sarcoma	15	37	DOD	Yes	3600	None	None	None	12	16(53%)/14
3	Male	53	Osteosarcoma	17	26	NED	Yes	2500	None	None	None	23	18(60%)/20
4	Male	22	Osteosarcoma	16	20	DOOD	Yes	4000	None	None	None	22	21(70%)/20
5	Male	44	Osteosarcoma	8	12	AWD	Yes	4200	Deep infection, DAIR procedures	None	None	42	20(66%)/12
6	Female	54	Chordoma	8	14	NED	Yes	2000	None	None	None	18	22(73%)/14
7	Male	66	Chordoma	10	18	NED	Yes	4300	None	None	None	15	19(63%)/16
8	Female	64	Chordoma	16	14	NED	Yes	2200	Deep infection, DAIR procedures	None	None	19	18(60%)/13
9	Female	24	Osteosarcoma	16	17	AWD	Yes	2500	None	None	None	13	17(56%)/14
Mean	\	43	\	13.1	24.4	\	\	3211.1	\	\	\	20.2	19.4(65%)/16.3

*NED, no evidence of disease; DOD, dead of disease; AWD, alive with disease; DOOD, dead of other diseases; * MSTS 93 and MUD scores at follow-up. MSTS, Musculoskeletal Tumor Society scoring system, MUD, function and sensation of lower limbs (M), urination and uriesthesia (U), and defecation and rectal sensation (D)

Illustrative case

A 66-year-old male developed pain in the buttocks for 2 years and got dysuria and defecation difficulties for 1 year. An MRI scan revealed a lesion in the sacrum (Fig. 3A), and a sacral chordoma was confirmed by biopsy. The extent of bone and soft tissue resection was determined via 3D virtual models, and the positions of the bilateral osteotomy planes were determined via preoperative CT scans (Fig. 3B). A virtual 3D model of the lumbosacral and pelvic region was constructed using 3D slicer software (Fig. 3C), which aided in the design of patient-specific surgical jigs, endoprosthesis placement gaskets and custom-made 3D-printed endoprostheses. The surgical jig had 2.5 mm holes suitable for Kirschner wires, 2.0 mm osteotomy slots to restrict the oscillating saws during osteotomy (Fig. 3D). The patient received a total en bloc sacrectomy through the posterior approach, with the use of a patient-specific surgical jig for osteotomy. Custom-made 3D-printed endoprostheses were used for bone reconstruction, while LARS artificial ligaments were used for soft tissue reconstruction (Fig. 3E). The surgery was performed successfully, and the tumor was dissected en bloc (Fig. 3F). The volume of bleeding was 2000 ml. No perioperative complications were observed. Accuracy was validated by comparing the actual osteotomy planes and the planned osteotomy planes using pre- and postoperative CT scans (Fig. 3G).

Fig. 3.

A 66-year-old male patient with a sacral chordoma. (A) Preoperative MRI scan. (B) 3D schematic of the bone and soft tissue to be resected and planned bilateral osteotomy sites based on preoperative CT. Generally, an additional 8 mm margin beyond the tumor range was added. (C) Preoperative design of patient-specific surgical jigs, endoprosthesis placement gaskets (red arrows) and custom-made endoprostheses in 3D virtual models. (D) 3D-printed PSI and custom-made endoprosthesis. The endoprosthesis had porous surfaces to promote bone ingrowth. Placement of the PSI during osteotomy. (E) The reconstruction of bone structures using the custom-made endoprosthesis and reconstruction of soft tissue structures using the LARS artificial ligament. Radiograph of the surgical region two weeks after surgery. (F) Resected tumor sample and its radiograph. (G) Accuracy validation by comparing the actual reconstructed endoprosthesis and osteotomy planes (green) versus the planned endoprosthesis and osteotomy planes (yellow) using pre- and postoperative CT images

Surgical accuracy

The surgical accuracy was validated by comparing the planned and acutal osteotomy planes in 3D virtual models (Fig. 4A). The angular and position deviations were analyzed at four different levels for each patient: the upper edge of the SI joint (Level 1), the S1 neural foramen (Level 2), the S2 neural foramen (Level 3), and the bottom edge of the SI joint (Level 4). At each level, bilateral deviations were analyzed (Fig. 4B-E).

Fig. 4.

Validation of surgical accuracy using 3D printing technology following total sacrectomy. (A) Post- (green) and preoperative (yellow) 3D models of the lumbosacral and pelvic regions constructed by the fiducial registration model of 3D slicer software. (B-E) The angular and position deviations of the osteotomy were evaluated by comparing postoperative CT scans with the planned osteomy planes based on preoperative CT scans. (B), (C), (D) and (E) represent accuracy validation at Level 1, Level 2, Level 3 and Level 4, respectively

An average angular deviation of 4.27° (interquartile range[IQR] 4.15) and an osteotomy position deviation of 4.00 mm (IQR 2.90) were observed (Figure. 5 A). The deviations at the four different levels of each case are shown in Table 2. The mean angular deviations at Levels 1–4 were 3.50° (IQR 6.02), 3.86° (IQR 2.55), 4.81° (IQR 4.37), and 4.92° (IQR 3.27), respectively (Figure. 5B). The mean position deviation at the four levels were 3.15 mm (IQR 3.54), 3.55 mm (IQR 1.37), 4.26 mm (IQR 2.61), and 4.86 mm (IQR 3.93) (Figure. 5 C). Overall, the angular and position deviations were not significantly different among the four different levels (Figure. 5B&C). The angular and position deviations of left and right sides of osteotomy at the four different levels exhibited no significant difference (Fig. 5D&E). Since a 2-mm deviation may turn a wide margin into an intralesional margin [15], we used 2-mm as the cut-off value to evaluate the surgical accuracy. A total of 40.0%, 11.1%, 16.7%, 0% of the position deviations were less than 2 mm at Level 1, Level 2, Level 3, and Level 4, respectively (Figure. 5 F). We also used 5-mm as the cut-off value (Figure. 5G) because 5 mm was a common deviation in bone tumor resection usind a PSI [16, 21, 22]. Significant differences were found between Level 1 and Level 2 and between Level 3 and Level 4 using 2 mm and 5 mm as cut-off values (Figure. 5 F&G). All position deviations were below 8 mm, which made an additional 8 mm margin beyond the tumor become a safe design of planned osteotomy planes. Although position deviations at the four levels were similar statistically, significantly greater proportions of position deviations greater than 2 mm and 5 mm were observed at the caudal end of the SI joint than at the upper end.

Fig. 5.

Angular and position deviations at the four different levels. (A) Angular and position deviations of the cohort. (B) Angular deviations at the four different levels. (C) Position deviations at the four different levels. (D) Angular deviations of the left and right side at the four different levels. (E) Position deviations of the left and right side at the four different levels. (F) Percentage of position deviations > 2 at the four different levels. (G) Percentage of position deviations > 5 at the four different levels. Level 1–4 represented the upper edge of the SI joint, S1 neural foramen, S2 neural foramen, the bottom edge of the SI joint, respectively. Multiple t-test, Fisher’s exact test and Chi-Squared test were used for the statistical analysis. ns, not significant. *, p-value < 0.05. **, p-value < 0.01. ***, p-value < 0.001. ****, p-value < 0.0001

Table 2.

Evaluation of surgical accuracy at different levels

Level	Left		Right		All
	Angular	Position	Angular	Position	Angular	Position
1	3.42° (6.60)	3.06 mm (1.36)	3.58° (4.30)	3.21 mm (3.84)	3.50° (6.02)	3.15 mm (3.54)
2	4.49° (3.40)	3.18 mm (2.95)	3.23° (1.50)	3.84 mm (0.79)	3.86° (2.55)	3.55 mm (1.37)
3	3.94° (1.60)	4.09 mm (2.63)	5.66° (4.50)	4.43 mm (2.50)	4.81° (4.37)	4.26 mm (2.61)
4	4.66° (3.10)	5.16 mm (3.56)	5.18° (2.20)	4.49 mm (3.99)	4.92° (3.27)	4.86 mm (3.93)

The deviations were given as the mean and IQR

Outcomes and prognosis

After a mean follow-up of 24.4 months (12–62 months), three patients experienced local recurrence (two cases of osteosarcoma and one case of Ewing’s sarcoma), and one patient experienced distal metastasis. The mean overall survival (OS) and mean recurrence-free survival (RFS) times were 20.1 months and 22.2 months, respectively. Up to the last follow-up, one patient died of tumor progression (local recurrence and distal metastasis), and one patient died because of viral myocarditis. Two patients with local recurrence alone were both alive and living with the disease. The mean MSTS93 and MUD [26] scores were 19.4 (16 to 24) and 16.3 (12 to 24).

Discussion

Total en bloc sacrectomy is a major therapeutic strategy for malignant sacral tumors, especially chordoma and chondrosarcoma, which are resistant to chemotherapy or immunotherapy [5–7]. R0 resection with a wide surgical margin has been indicated as the most important factor for survival and local recurrence in patients with sacral chordoma [9], making the achievement of R0 resection one of the primary objectives of total sacrectomy. However, the complex anatomy of the pelvic region, limited surgical area, adjacent neurovascular structures, and curved SI joint make surgical resection of sacral tumors difficult compared with tumors in the extremities. Precise total sacrectomy is crucial because osteotomy with adequate but not redundant margins can not only ensure R0 resection but also preserve enough bone tissue and neurovascular structures to ensure postoperative recovery of function [9].

Owing to the increased use of 3D printing technology, PSIs have been applied in bone tumor resection, including sacral tumor resection, to improve surgical accuracy and achieve better outcomes [21, 27]. Müller et al. reported a combined error of the osteotomy ranging from 0.74 ± 0.96 mm to 3.60 ± 2.46 mm in 11 patients who underwent bone tumor resection using a PSI [21]. Evrard et al. performed 31 bone tumor resections with a PSI and revealed that R0 resection was achieved in all patients and that the deviations between the obtained margins and planned margins were within the range of -5 to + 5 mm [27]. However, few studies have focused on total sacrectomy using a PSI, and to our knowledge, no study has evaluated the surgical accuracy of total sacrectomy with a PSI. Therefore, we conducted this study to assess the surgical accuracy of osteotomy with a PSI during total sacrectomy.

PSI is suitable for osteotomy, and an oscillating saw is suitable for osteotomy in the preset orientation of the PSI. Therefore, using both a PSI and an oscillating saw can ensure a certain entry point and exit point for osteotomy during total sacrectomy. Furthermore, custom-made 3D-printed sacral endoprostheses are capable of reconstructing bone defects precisely following total sacrectomy since both the endoprosthesis and PSI are made according to the osteotomy planes planned preoperatively. The match between the residual bone interface and the endoprosthesis is primarily dependent on the surgical accuracy of the osteotomy. To assess the surgical accuracy of osteotomy using a PSI, we calculated the angular and position deviations at four different levels (the upper edge of the SI joint, the S1 neural foramen, the S2 neural foramen, and the bottom edge of the SI joint). In general, the mean angular and position deviations were 4.27° (IQR 4.15) and 4.00 mm (IQR 2.90), respectively, with 16.7%, 73.8% and 100% of position deviations ≤ 2 mm, ≤ 5 mm and ≤ 8 mm, respectively. We planned the osteotomy planes to be 8–10 mm beyond the tumor range preoperatively. Compared with 2 mm and 5 mm, 8 mm beyond the tumor range was safer because, in some cases, the position deviations may exceed 5 mm, and all deviations were within 8 mm. Although the values of position deviations among the four different levels were statistically similar, the deviations increased gradually during SI joint cuts in total sacrectomy when 2 mm and 5 mm were used as cutoff values. The proportion of position deviations greater than 2 mm and 5 mm at the exit point was greater than that at the entry point. These results demonstrated that even though the PSI had osteotomy slots that could limit the direction of osteotomy and the oscillating saw had advantages over the Gigli saw when cutting bones with curved surfaces, the accuracy and deviation of the entry point still increased during osteotomy. Previous studies have revealed the use of PSIs during bone tumor resection in the extremities and pelvic region (Table 3). K.C. Wong et al. reported a deviation ranging from 1.3 to 4.0 mm when using a PSI for pelvic tumor resection [22]. Daniel A Müller et al. performed osteotomy using a PSI in the pelvic region and long bone and scapula tumor resection and achieved an accuracy ranging from 0.74 ± 0.96 mm to 3.60 ± 2.46 mm [21]. Matthew A. Siegel BS et al. compared the surgical accuracy of PSI and CAS in an ex vivo idealized sawbone model and indicated that surgical accuracy was higher with the PSI than with CAS [16]. To our knowledge, this is the first study to investigate the surgical accuracy of using a PSI during total sacrectomy in vivo.

Table 3.

Literature review of the surgical accuracy achieved in tumor surgery

Authors	Year	Number of cases	Type of study	Site	Method	Accuracy/Deviation
K. C. Wong et al. [22]	2015	1	In vivo	Pelvic region	PSI	1.3–4.0 mm
Matthew A. Siegel BS et al. [16]	2020	22	Ex vivo	Sacropelvic region	Freehand	67.0%≤5 mm 25.8%≤2 mm
					Navigation	71.1%≤5 mm 32.5%≤2 mm
					PSI	85.6%≤5 mm 47.5%≤2 mm
Qing Zhang et al. [14]	2020	10	In vivo	Pelvic region	Navigation	95%CI: [-3.95, -3.27]
		16	In vivo	Long bones	Navigation	95%CI: [-2.69, -2.34]
Daniel A Müller et al. [21]	2020	11	In vivo	Pelvic region, long bones and scapula	PSI	Range from 0.74 ± 0.96 mm to 3.60 ± 2.46 mm
Current Study	2024	9	In vivo	Sacral	PSI	4.27° (IQR 4.15) 4.00 mm (IQR 2.90)

PSI, patient-specific instrument

The results of the present study should be interpreted carefully owing to the following limitations. First, it should be noted that the study design was retrospective and monocentric, which may cause selection bias. Second, the postoperative 3D models were approximately manually aligned with the preoperative 3D models. Third, the small sample size and short follow-up period may have resulted in some of the conclusions being inaccurate. Last but not least, the present study did not include patients who underwent total sacrectomy via the freehand technique as a control group. Future studies should include a larger sample size with a longer follow-up period and include a group of patients who underwent total sacrectomy via the freehand technique to investigate whether the use of a PSI could actually decrease the risk of local recurrence and improve patient prognosis.

Conclusions

The combined use of the PSI and custom-made 3D-printed endoprosthesis during osteotomy in total sacrectomy exhibited high surgical accuracy. The mean angular and position deviations of the PSI during osteotomy were 4.27° (IQR 4.15) and 4.00 mm (IQR 2.90), and the deviations increased during the osteotomy process. An additional 8 mm margin beyond the tumor range may be a safe osteotomy plane for preoperative design.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

T.J. design of the work and review the manuscript; Q.Y. S. perform the data analysis and write the manuscript; J.Z. Z. analyze the surgical accuracy; H.J. L., R.F. W. and S.Y. H. collected the data; W. G. interpretation of data; X.D. T. supervision.

Funding

This work was supported by the National Natural Science Foundation of China (81872180), Beijing Natural Science Foundation (L234063) and Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0525804).

Data availability

Data of the current study is available from the corresponding author on reasonable requests.

Declarations

Ethics approval and consent to participate

Informed consent was obtained from all patients, and the study was approved by the Medical Ethical Committees at the authors’ institution (Peking University People’s Hospital).

Consent for publication

All authors have agreed to publish this manuscript.

Competing interests

The authors declare no competing interests.