Treatment of Unstable Sacral Fractures with Robotically‐aided Minimally Invasive Triangular Fixation

Abstract

Objective

The treatment of unstable sacral fractures is huge challenge to surgeons. Robotically‐aided minimally invasive triangular fixation (RoboTFX) is the most advanced technique up to now. This study is to evaluate the clinical outcomes of unstable sacral fractures treated with RoboTFX.

Methods

From March 2017 to October 2021, 48 consecutive patients with unstable sacral fractures were included in the study. All patients received surgical treatment with triangular fixation (TFX). Patients were divided into four groups according to the number of fractures (uni‐ or bilateral) and surgical method employed (RoboTFX or traditional open TFX). Between these four groups, clinical data on operation time, intraoperative bleeding, intraoperative fluoroscopy time, infection rate, fracture healing rates, insertion accuracy, Majeed pelvic outcome score, Mears' criterion, and Gibbons score were compared. Quantitative data were expressed as mean ± standard deviation and compared using Student's t‐test. Categorical variable were compared using the Pearson's χ² test.

Results

Comparing unilateral RoboTFX versus open TFX, neither fracture healing rate, infection rate, Majeed pelvic outcome score, Mears' radiological evaluation criterion, nor Gibbons score of the two groups were statistically significantly different (p > 0.05). However, operation time, intraoperative bleeding, intraoperative fluoroscopy time, and insertion accuracy in the RoboTFX group were all significantly better than those of the traditional open group (p < 0.05). Likewise, operation time, intraoperative bleeding, intraoperative fluoroscopy time, and accuracy of fixation insertion of the bilateral RoboTFX group were significantly better than those of the bilateral open group (p < 0.05). Meanwhile infection rate, fracture healing rate, Majeed score, Mears' criterion, and Gibbons score of two groups were not significantly different (p > 0.05).

Conclusion

RoboTFX has the advantages of less operation time, less intraoperative bleeding and fluoroscopy, more accurate fixation insertion, and a higher healing rate compared to traditional open methods in the treatment of unstable sacral fractures. However, RoboTFX requires a few critical considerations, and the indications of its operation should be strictly evaluated.

Male, 35 Y, injured by falling. X‐ray (A) and CT reconstruction (B) showed AO C‐1‐3 type pelvic fracture, and the sacral fracture was Dennis type II and unilateral unstable. The patient accepted robotically‐aided minimally invasive triangular fixation (RoboTFX) operation 5 days after injury (C, D). The intraoperative fluoroscopy showed that the fracture obtained anatomic reduction and stable fixation (E), and the postoperative picture (F) showed the incisions was minimal.

Introduction

Sacral fractures account for 10%–45% of all pelvic fractures. Between 17% and 30% of sacral fractures are unstable. These fractures occur vertically in one or both sides of the sacrum and are often associated with horizontal, vertical and rotational deformity. These types of fractures invariably result from high‐energy injuries and have severe complications. ¹ , ² Currently, most surgeons advocate for surgical treatment in treating unstable sacral fractures rather than conservative management, with the main goals of such an operation being: to obtain anatomical reduction and stable fixation so that the fracture healing rate can be increased; to reconstruct the normal weight‐bearing load and the early ambulation can be ensured; to repair nerve damage caused by progressive fracture deformity and to alleviate the pain caused by instability of the lumbosacral junction. ¹ , ² , ³ , ⁴

Triangular fixation (TFX) is a classical type of internal fixation for the treatment of unstable sacral fractures. TFX consists of unilateral or bilateral lumbopelvic fixations and iliosacral screws, and has been widely used for its superior usefulness as a biomechanical model, fixed strength, and ability to improve neurological deficiency. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ TFX can not only simultaneously reduce and fix sacral fractures, but can also formulate a tri‐dimensional interlocking structure to fix the fracture both directly and indirectly, there by obtaining maximum sacral stability.

However, TFX also has a number of notable disadvantages: longer operating time; more intraoperative bleeding; a high rate of infection resulting from excessive exposure; and the risk of iatrogenic vessel and nerve injury during implantation of the iliosacral screws. In addition, excessive intraoperative radiation exposure can cause harm to both patients and surgeons.

Recently, the application of orthopedic robots can overcome these shortcomings effectively. ¹² , ¹³ , ¹⁴ The robotic system can assist surgeons in planning the trajectory, position, and length of the inserted screws by importing intraoperative C‐arm images. This ensures less invasive and more accurate screw placement, leading to less bleeding and intraoperative fluoroscopy, less iatrogenic exposure damage and vessel or nerve injury, and better prognosis. ¹³ , ¹⁴ However, the sample size of the existing report on this procedure is small, diminishing the quality of evidence therefrom.

To address this shortcoming, we retrospectively analyzed data of unstable sacral fracture patients who, from March 2016 to October 2020, underwent RoboTFX surgery. We compared these clinical data with the traditional open operation procedures. Our objectives in undertaking this study were to: (i) evaluate the clinical outcomes of treating unstable sacral fractures with RoboTFX; (ii) discuss the advantages of this technique; and (iii) describe the key points of surgery.

Methods

Inclusion and Exclusion Criteria

Inclusion criteria: (i) unstable sacral fracture treated with TFX; (ii) time from injury to operation <3 weeks; and (iii) continuous follow‐up >12 months with clinical data. Exclusion criteria: (i) open sacral fracture; (ii) associated severe internal medical disease; (iii) sacral fractures associated with severe nerve injuries requiring extra open nerve decompression; and (iv) associated L5 vertebral fracture, severe lumbosacral conjunction, or intervertebral disc injury requiring extra operation.

Patient Data

From March 2017 to October 2021, 48 consecutive cases of unstable sacral fractures were included in the study. Subjects were divided into two groups according to the number of sacral fracture sides—unilateral and bilateral group. Then these two groups were classified to four subgroups according to the different operations (RoboTFX or traditional open TEX) —Uni‐Robo, Uni‐Open, Bil‐Robo and Bil‐Open group.

Every group contained 12 patients, whose medical records were retrospectively reviewed. The study was carried out in accordance with the Declaration of Helsinki. Institutional review board approval was obtained (Tianjin Hospital, 2023/YLS063).

Groups

The Uni‐Robo group was Unilateral sacral fracture and RoboTFX operation. The basic information of the patients is shown in the Table 1. All sacral fractures were associated with anterior pelvic ring injuries, which included 10 pubic rami fractures and two pubic synthesis dissociation. All patients had associated injuries, which included one head injury, four thoracic injuries, two abdominal injuries, two urinary system injuries, and six other part fractures besides the pelvis.

TABLE 1.

Comparison of patients' characteristics between Uni‐Robo and Uni‐Open groups.

Characteristics	Uni‐Robo group (n = 12)	Uni‐Open group (n = 12)	t/χ2	p value
Gender n (%)
Male	10 (83.3%)	11 (91.7%)	χ2 = 0.381	1.000
Female	2 (16.7%)	1 (8.3%)
Age (years)	36.3 ± 1.2 (21–63)	38.2 ± 1.6 (21–68)	t = 3.291	0.003
Injury mechanism n (%)
Falling	8 (66.7%)	8 (66.7%)
Traffic accident	4 (33.3%)	4 (33.3%)
AO classification n (%)
C1‐3	10 (83.3%)	11 (91.7%)	χ2 = 0.381	1.000
C2	2 (16.7%)	1 (8.3%)
Dennis classification n (%)
I	4 (33.3%)	3 (25%)	χ2 = 0.202	1.000
II	8 (83.3%)	9 (75%)
Sacral nerve injuries n (%)	2 (16.7%)	3 (25%)	χ2 = 0.253	1.000
Associated injuries n (%)	12 (100%)	12 (100%)

The Uni‐Open group was unilateral sacral fracture and traditional open TFX. The basic information of the patient is shown in the Table 1. All sacral fractures were associated with anterior pelvic ring injuries, which included 11pubic rami fractures and 1pubic synthesis dissociation. All patients had associated injuries, which included 1 head injury, 4 thoracic injuries, 2 abdominal injuries, 1 urinary system injury, and 5 other part fractures besides the pelvis.

The Bil‐Robo group was bilateral sacral fracture and RoboTFX operation. The basic information of the patients is shown in the Table 2. All sacral fractures were associated with anterior pelvic ring injuries, which included 10 pubic rami fractures and two anterior wall fracture of the acetabulum. All patients had associated injuries, which included four head injuries, eight thoracic injuries, one urinary system injury, and 11 other part fractures besides the pelvis.

TABLE 2.

Comparison of patient characteristics between Bil‐Robo and Bil‐Open groups.

Characteristics	Bil‐Robo group (n = 12)	Bil‐Open group (n = 12)	t/χ2	p value
Gender n (%)
Male	9 (75%)	9 (75%)
Female	3 (25%)	3 (25%)
Age (years)	35.9 ± 1.5	38.4 ± 1.2	t = 4.508	0.000
Injury mechanism n (%)
Falling	8 (66.7%)	8 (66.7%)
Traffic accident	4 (33.3%)	4 (33.3%)
Morphological classification n (%)
U	9 (75%)	10 (83.3%)	χ2 = 1.114	1.000
H	2 (16.7%)	2 (16.7%)
Y	1 (8.3%)	0
ROY‐Camille classification n (%)
II	10 (83.3%)	11 (91.7%)	χ2 = 0.381	1.000
III	2 (16.7%)	1 (8.3%)
Sacral nerve injuries n (%)	5 (41.7%)	6 (50%)	χ2 = 0.168	1.000
Associated injuries n (%)	12 (100%)	12 (100%)

The Bil‐Open group was bilateral sacral fracture and traditional open TFX. The basic information of the patients is shown in the Table 2. All sacral fractures were associated with anterior pelvic ring injuries, which included 10 pubic rami fractures and two pubic synthesis dissociations. All patients had associated injuries, which included three head injuries, eight thoracic injuries, two abdominal injuries, one urinary system injury, and nine other part fractures besides the pelvis.

Preoperative Treatment and Plan

In the emergency room, the injury severity score (ISS) was evaluated for all patients, with advanced trauma life support (ATLS) protocol taking precedence over definitive surgery. Thirty patients were managed based on damage control orthopedics’ (DCO) principles and transferred to the intensive care unit. Definitive operations of sacral fracture were performed until the physiological condition of patients was stable. All patients accepted skeletal traction, with the associated injuries were treated by related specialist surgeons.

The standard preoperative imaging examination plan included vessel ultrasonic examination to diagnose deep vein thrombosis (DVT), as well as anterior–posterior, inlet, and outlet X‐ray and 3D CT reconstruction of the pelvis to assess the displacement and degree of deformity of the sacral fracture. We created a radiological simulation of normal morphology for every injured sacrum in order to measure the anatomic data of the sacral vestibule and judge preoperatively whether the iliosacral screws were suitable for implantation. Magnetic resonance of sacral nerves (MRN) was conducted regularly to inform the surgeons the nature and severity of sacral neural injuries.

Surgical Procedure

All procedures were performed and the related clinical data were recorded by the same group of experienced orthopedic surgeons. Patients were administered general anesthesia and placed prone on a radiolucent table to ensure that pelvic anteroposterior, inlet, outlet, and lateral views could be feasibly obtained. Intravenous antibiotics were administered within 30 min before skin incisions were made.

Operation of Uni‐Robo Group

The navigation tracker of the orthopedic robot (TINAVI Medical Technologies, Beijing, China, Fig. 1) was percutaneously fixed on L3 spinous process. After L5intraoperative CT images were obtained using C‐arm (Siemens, Munich, Germany), the data were transmitted to the robotic planning system. Based on preoperative planning combined with L5 vertebral anatomic feature‐mapping, the length, angulation, and direction of unilateral pedicle screw insertion were designed, with screw placement simulated on the images of screen.

Fig. 1.

The combination of TINAVI orthopedic robotic system.

After navigation planning was thus established, the robotic arm, covered in a sterile, plastic dressing, began to move to the proper position following guidance of the preplanned trajectory. A guide sleeve was placed onto the L5 bone surface via a 3 cm incision. After the trajectory was recalibrated, a guide pin was then inserted into the pedicle through the sleeve. A cannulated universal pedicle screw, 6 mm in diameter (Kanghui Medical Instruments, Suzhou, China), was inserted along the pin, after which the guide pin was removed.

Following insertion of the screw, pelvic images of the lateral, iliac tangential (inlet‐obturator) and teardrop (outlet‐obturator) position were obtained through intraoperative fluoroscopy. These data were then transmitted to the robotic planning system. After the next round of navigation planning, the robotic arm moved to the next assigned position. A 5 cm incision was made on the posterior superior iliac spine (PSIS), through which a 7 mm diameter and 10 cm deep ipsilateral universal iliac screw (Kanghui Medical Instruments, China) was inserted.

At this point, the direction of the screw from the PSIS to the anterior inferior iliac spine (AIIS) and between the medial and lateral lamina of the iliac wing was verified. In most cases, we resected part of the PSIS to avoid skin irritation caused by protruding parts of screw. A pre‐contoured rod, 6.5 mm in diameter (Kanghui Medical Instruments, China), was then inserted through two incisions and connected to the pedicle and iliac screw. Reduction of vertical displacement was achieved through the distraction of the lumbopelvic device, followed by all connectors being screwed in place. A Schantz pin was screwed into the lateral part of sacrum to reduce the horizontal or rotational residual displacement of the pelvic ring; which is called the “joystick” technique. ¹⁵

Once the quality of the reduction was confirmed by intraoperative fluoroscopy of anterior‐posterior (AP), inlet, outlet, and lateral views, the related data of reduced sacrum were transmitted to the robotic planning system, and the navigation planning of the iliosacral screw was designed. The robotic arm then moved to the assigned position and made a 1 cm incision. A 7.3 mm diameter, fully threaded and cannulated screw (Kanghui Medical Instruments, China) was inserted, serving as the iliosacral screw running from the ilium to the vertebral center of S1 and achieving direct fixation of the sacral fracture (Fig. 2). Intraoperative anteroposterior fluoroscopy of L5/S1 was necessary to ensure the balance of both sides of the intervertebral space. If the space of the fracture's side was wider than the contralateral one, the connector was loosened, and the lumbopelvic fixation was adjusted and re‐screwed until both intervertebral spaces were equal.

Fig. 2.

Male, 35 years, injuried by falling. X‐ray (A) and CT reconstruction (B) showed AO C‐1‐3 type pelvic fracture, and the sacral fracture was Dennis type II and unilateral unstable. The patient accepted ROBOTFX operation 5 days after injury (C–H). The intraoperative fluoroscopy showed that the fracture obtained anatomic reduction and stable fixation (I–K), and the postoperative picture (L) showed the incisions were minimal.

Operation of Uni‐Open Group

Aposterior‐median incision was made from the L4 spinous process to S3.The unilateral erector spinae was dissociated, to expose the transverse and superior articular processes of ipsilateral L5. Then the pedicle screw was inserted through the unilateral pedicle of L5.

An iliac screw was inserted 10 cm deep from the PSIS to the AIIS between the medial and lateral lamina of the iliac wing with the monitor of intraoperative fluoroscopy. Approximately 2–3 cm of bone around the iliac screw ends was removed to avoid any prominence from the screw head. The pre‐contoured rods were then placed, connecting the pedicle and iliac screws. Reduction of vertical displacement was achieved through distraction of the lumbopelvic device, followed by all connectors being screwed in place.

The rotational deformity and horizontal displacement were corrected with reduction clamps or an extra Schantz screw as a joystick. Once the quality of the reduction was confirmed by intraoperative fluoroscopy of AP, inlet, outlet and lateral views, aniliosacral screw was inserted into S1. During this procedure, consecutive fluoroscopy of lateral, AP, inlet, and outlet views were monitored to ensure the correct insertion of screws.

Operation of Bil‐Robo Group

Lumbopelvic fixations of both sides were inserted as described for procedure of Uni‐Robo group. The sequence of the screws and rods' insertion could change flexibly. Then the reduction of vertical displacement was achieved through distraction of the bilateral lumbopelvic devices. After all connectors were screwed in place, the transversal connection instruments were fixed subfascially through two parallel incisions in L5 or PSIS. The joystick technique was applied to correct residual displacement of the fracture, and two S1 iliosacral screws were inserted with the guidance of the robot when the intraoperative fluoroscopy showed the reduction of sacral fracture was satisfied (Fig. 3).

Fig. 3.

Male, 45 years, injuried by falling. X‐ray (A) and CT reconstruction (B, C) showed ROY‐Camille type III pelvic fracture, and the sacral fracture was U type and bilateral unstable. The patient accepted ROBOTFX operation 6 days after injury (D, E). The intraoperative fluoroscopy showed that the fracture obtained anatomic reduction and stable fixation (TFX and absorbable screws of pubic ramus fractures, F–H), and the postoperative picture (I) showed the scars of incision was less than the one of open operation (J).

Operation of Bil‐Open Group

Lumbopelvic fixations of both sides were inserted as described for procedure of Uni‐Open group. Procedure of reducing of the fracture and assembling the transversal connection instruments were as same as procedure of Bil‐Robo group in the open incision. Then two iliosacral screws were inserted with guidance by C‐arm manually when the reduction was satisfied.

All patients accepted irrigation and vacuum drainage instruments before their wound was closed. Operations for all anterior pelvic ring injuries were performed simultaneously, for which plates were applied in 38 cases and screws were applied in 10. In three patients with type C‐2 fractures, the contralateral sides that had no unstable sacral fracture were treated with iliosacral screws.

Postoperative Treatment

All patients underwent the same postoperative management, with intravenously administered antibiotics continued for 72 h following surgery. The drainage instrument was removed 48–72 h postoperatively according to the amount of drainage fluid. Low molecular weight heparin was applied regularly as thromboprophylaxis during the immobilization period. Rehabilitation was planned and performed by a physical therapist as soon as the physical and wound condition of each patient permitted. Full weight bearing was permitted 8–16 weeks postoperatively, depending on the type of fracture and associated injuries.

Follow‐up and Clinical Outcome Evaluation

The follow‐up schedule was 4 weeks, 12 weeks, and every 3 months postoperatively. Physical examinations and imaging exams were undertaken to evaluate the accuracy of screw insertion, degree of fracture healing, whether or not fixation had failed, neurological impairment, and improvement in daily activity functionality. The clinical outcome for fractures was evaluated with Majeed pelvic outcome score. ¹⁶ Mears' radiological evaluation criterion was applied to evaluate the reduction quality of fractures, ¹⁷ and Gibbons' score was used to assess neurological deficiency and improvement. ¹⁸

Statistical Analysis

All data were processed using SPSS 20.0 statistical software (SPSS Inc., Chicago, IL, USA). Quantitative data (operation time, bleeding, fluoroscopy time, Majeed score) were expressed as mean ± standard deviation (SD) and compared using Student's t‐test. Categorical variable (infection rate, healing rate, accuracy rate of fixation insertion, satisfied rate of reduction and Gibbons score change) were compared using the Pearson's χ² test. The p‐value was set at <0.05 for significance.

Results

Patient Information

There were no statistically significant differences in patients demographic data, such as gender, age, injury mechanism, fracture type, and sacral nerve injuries, between the two unilateral groups (Group Uni‐Robo and Uni‐Open, Table 1) and the two bilateral groups (Group Bil‐Robo and Bil‐Open, Table 2).

Comparison of Basic Conditions of Operation

All operations were undertaken between 4 and 15 days (average 6.4 ± 1.4 days) following primary injuries. All patients were followed up on continuously, with an average follow‐up time of 21.2 ± 3.2 months (range, 12–54 months). The mean surgical time was 100.3 ± 14.5 (55–190) min in the Uni‐Robo group, 202 ± 18.5 (90–320) min in the Uni‐Open group, 130.3 ± 19.5 (75–249) min in the Bil‐Robo group, and 245 ± 19.8 (90–380) min in the Bil‐Open group. The mean amount of intraoperative bleeding was 180 ± 17.4 (50–350) mL in the Uni‐Robo group, 550 ± 15.2 (350–1500) mL in the Uni‐Open group, 350 ± 19.4 (90–650) mL in the Bil‐Robo group, and 850 ± 17.8 (400–2200) mL in the Bil‐Open group. The mean intraoperative fluoroscopy time of the four groups was as follows: 23.3 ± 4.5 (15–35) s in the Uni‐Robo group, 90 ± 7.7 (30–240) s in the Uni‐Open group, 30.3 ± 3.8 (15–45) s in the Bil‐Robo group, and 110 ± 7.2 (50–280) s in the Bil‐Open group. No patients experienced loss of reduction or failure of fixation.

Comparison of Surgical Reduction

According to Mears' criterion, there were seven anatomic, four satisfied, and one unsatisfied reduction of fractures in the Uni‐Robo, Uni‐Open and Bil‐Robo groups, and six anatomic, five satisfied, and one unsatisfied reduction in the Bil‐Open group. The overall rate of satisfied reduction across all four groups was 91.7%. In the Uni‐Robo and Uni‐Open groups, patients of each group received a total of 36 screws. All screws were placed accurately in Uni‐Robo group, whereas eight screws in Uni‐Open group perforated the cortical edge of the bone, for an accuracy rate of 77.8%. In the Bil‐Robo and Bil‐Open groups, patients of each group received a total of 72 screws. All screws were placed accurately in group 3, whereas 14 screws in group 4 perforated the cortical edge of the bone, for an accuracy rate of 80.6%. In the Uni‐Robo, Uni‐Open and Bil‐Robo groups, all fractures healed properly, whereas two fractures in the Bil‐Open group were non‐union, for a healing rate of 83.3%.

Majeed scores for the last follow‐up for each of the four groups were as follows: Uni‐Robo group, 86.2 ± 3.4; Uni‐Open group, 84.2 ± 2.7; Bil‐Robo group, 82.5 ± 3.7; and Bil‐Open group, 81.4 ± 3.2. Changes in sacral nerve injury in the four groups were as follows: Uni‐Robo group, two patients showed improvement (one change from Gibbons grade II to I, and the other from grade III to II); Uni‐Open group, three patients showed improvement (two from Gibbons grade II to I, one from grade III to II); Bil‐Robo group, four of five patients who were associated with neurological injury showed improvement (two of three patients changed from Gibbons grade II to I, and one from grade III to II); and Bil‐Open group, five of six patients who were associated with neurological injury showed improvement (two of three patients changed from Gibbons grade II to I, one changed from grade III to I, and two changed from grade III to II).

There were no occurrences of wound infection in the Uni‐Robo group, whereas there were two deep and one superficial postoperative infections in the Uni‐Open group, giving an overall infection rate of 25%. One patient in the Bil‐Robo group had a superficial infection (8.3%). Two deep and two superficial postoperative infections occurred in the Bil‐Open group (33.3%).

For the unilateral fracture groups, operation time, intraoperative bleeding, intraoperative fluoroscopy time, and accuracy rate of fixation insertion were significantly better in the robot group than the traditional open group (p < 0.05). Healing rate, infection rate, Majeed pelvic outcome score, Mears' criterion, and Gibbons score change between the two methods were not significantly different (p > 0.05) (Table 3).

TABLE 3.

Comparison of clinical data between four groups.

Groups	Operation time (m)	Intraoperative bleeding (mL)	Intraoperative fluoroscopy time (s)	Infection rate n (%)	Healing rate of fracture n (%)	Accuracy rate of fixation insertion n (%)	Majeed score	Satisfied rate of reduction n (%)	Gibbons score change
Unilateral
Uni‐Robo	100.3 ± 14.5	180 ± 17.4	23.3 ± 4.5	0	100%	100%	86.2 ± 3.4	91.7%	2
Uni‐Open	202 ± 18.5	550 ± 15.2	90 ± 7.7	25%	100%	77.8%	84.2 ± 2.7	91.7%	3
t/χ2/Z	t = 14.99	t = 100.46	t = 32.13	χ2 = 3.43		χ2 = 9.00	t = 1.60		z = 0.45
p value	<0.0001	<0.0001	<0.0001	0.22		<0.0001	0.35		0.76
Bilateral
Bil‐Robo	130.3 ± 19.5	350 ± 19.4	30.3 ± 3.8	8.3%	100%	100%	82.5 ± 3.7	91.7%	4
Bil‐Open	245 ± 19.8	850 ± 17.8	110 ± 7.2	33.3%	83.3%	80.6%	81.4 ± 3.2	91.7%	5
t/χ2/Z	t = 9.85	t = 23.44	t = 29.62	χ2 = 2.27	χ2 = 2.18	χ2 = 7.88	t = 1.23		z = 0.37
p value	<0.0001	<0.0001	<0.0001	0.32	0.48	<0.0001	0.31		0.91

For the bilateral fracture groups, operation time, intraoperative bleeding, intraoperative fluoroscopy time, and accuracy rate of fixation insertion were all significantly better in the robot group than the open group (p < 0.05). Meanwhile infection rate, fracture healing rate, Majeed pelvic outcome score, Mears' radiological evaluation criterion, and Gibbons score of two groups were not significantly different (p > 0.05) (Table 3).

Discussion

Unstable sacral fracture is the most serious type of orthopedic trauma and the prognosis of treatment was uncertain previously, mainly because the internal fixations did not satisfy the needs of surgical requirement. In our study, the related clinical data showed that RoboTFX can achieve the best ideal goal of treatment currently.

Triangular Fixation

Unstable sacral fractures have special pathological and imaging characteristics. The injury energy loads throughout the spine‐sacrum axis, and further shears the sacrum or lumbosacral region in a state of excessive flexion or extension, resulting in fractures of the weak parts of sacrum. The fractures are completely unstable, and, thus, three‐dimensional displacement and deformities are frequent. The goals of surgery on these fractures include pelvic ring reconstruction, lumbopelvic stability restoration, fracture displacement correction, and improvement of neurological damage. Lumbopelvic fixation and iliosacral screws are currently popular fixation methods. However, both have their limitations and deficiencies. ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ Lumbopelvic fixation cannot fix the sacral fracture directly, which leads to a relatively high rate of non‐union because of the excessive motion of the sacrum and malunion caused by re‐displacement of the fracture. Meanwhile, iliosacral screws cannot simultaneously ensure reduction and fixation. As a result, the quality of regular intraoperative reduction is worse than that of lumbopelvic fixation, especially for vertical displacement, which is the most important factor affecting the prognosis.

TFX is more advanced than two of these techniques. This technique combines indirect lumbopelvic fixation and direct screw fixation of sacral fractures, which can restore the normal load of the weight‐bearing axis, and form a tri‐dimensional interlocking structure that provides the healing fracture with maximum stability. Furthermore, the technique avoids placing shear or rotational force on the injured area of the sacrum, thus not only preventing progressive nerve injury and re‐displacement of the fracture, but also creating a stable base to improve the rate of fracture union and re‐enable ambulation earlier. Compared with other methods, TFX can obtain better quality of reduction, better chance of fracture healing, and improved prognosis. ⁵ , ²¹

However, TFX also comes with several significant disadvantages. The excessive exposure required by the operation results in longer operation time, more intraoperative bleeding, and higher rates of wound‐related complications such as infection and wound dehiscence. Our research showed the rate of woundhealing disturbances, which are as high as 33% in bilateral open TFX group patients, were higher than those other surgeries, and the report showed the infection rate was 26%. ²⁵ Furthermore, the healing of fracture may be affected because of iatrogenic soft tissue damage.

Advantages of RoboTFX

1. Compared to open fixation, RoboTFX can shorten operation time, decrease intraoperative bleeding, and reduce infection ¹³ , ¹⁵ , ¹⁶ —results our own study has confirmed. Even the comparison of infection rate showed no significant difference, the data that none of two RoboTFX group patients had infection can proved our opinion indirectly.

2. In our study, the healing rate of bilateral sacral fractures in the RoboTFX group was higher than the open group, which indicated the minimally invasive surgery can maximize the fracture healing. This is most likely due to the reduced soft tissue damage incurred by the former's minimal invasiveness. The reason of the statistical difference between RoboTFX and open groups was not significant as perhaps the sample size was not large enough.

3. The process of translating intraoperative two‐dimensional fluoroscopy images to a comprehension of the three‐dimensional anatomical structure of the sacrum and the skillful insertion of iliosacral screws therein is complicated and challenging for orthopedic surgeons. In order to gather more information, doctors often conduct more fluoroscopic imaging, which can lead to the excessive radiological damage to both patients and surgeons. Furthermore, serious iatrogenic vessel or neural injury can result from inaccurate manual screw insertion. However, screw insertion with the aid of a robot can decrease radiological exposure time effectively and ensure higher insertion accuracy to the traditional method.

Indications and Surgical Experiences of RoboTFX

Although RoboTFX has several advantages, not all unstable sacral fractures are suitable for this procedure. Detailed planning and preoperative evaluation must be undertaken, and indications and contraindications must be strictly controlled. We concluded the indications were as following:

1. The lumbar or sacral nerves should not be entrapped by the fragments of the fracture, as determined by the preoperative MRN exam. If entrapment is present, sacral laminectomy and neural decompression should be performed with open incision before attempting the vertical distraction by lumbopelvic fixation in order to avoid exacerbating nerve injury caused by more serious entrapment. ²⁶ , ²⁷ , ²⁸

2. Upper sacral dysplasia can occur in 30%–50% of all sacrum, ²⁹ which leads to a more narrow and oblique screw pathway in S1 than normal, making screw placement difficult and significantly increasing the risk of iatrogenic vessel and nerve injury. For this reason, the evaluation of sacral morphonology by CT is important, as only patients who have a safe pathway for iliosacral screws are suitable for TFX.

3. Melly et al. ³⁰ reported that the screw pathway of S1 is decreased by 40% if the residual fracture displacement is 10 mm or more. This is true even with minimal displacement of sacral fracture, which is hard to avoid during surgery, the safe path of the iliosacral screw can be narrowed greatly and the difficulty and danger of screw implantation increases sharply. Therefore, patients who undergo RoboTFX must display closed reduction to ensure safe sacral screw insertion. If this is not the case, open reduction must be undertaken instead.

4. If patients present with lumbosacral conjunction injury, only those with the relatively stable (Isler types I, IIA, and IIB ³¹ ) and unnecessary for open operation of fusion are suitable for RoboTFX.

Skill and Experience of Surgery

Lower lumbar pain was the most common complication of TFX insertion, especially in unilateral fractures. This is caused by unequal distracting forces of fracture reduction and unbalanced loads of bilateral lumbosacral space. ³² After the iliosacral screws are inserted, flouroscopy of L5/S1 should be taken intraoperatively to evaluate the balance of both intervertebral spaces. If the spaces of both sides are not equally wide, the connector should be loosened, then the lumbopelvic fixation should be adjusted and re‐screwed until both sides are even.

Fully threaded cannulated screws should be used as iliosacral screws in order to avoid extra compression of fracture lines, which may lead to extrusion of the sacral nerve and iatrogenic injury.

We removed the bone with a small area (2 cm²) around the PSIS and deepened the iliac screw ends in order to avoid associated complications resulting from prominence of the fixation, such as chronic pain and wound dehiscence.

Limitations

The sample size of this study is relatively small, the follow‐up time is not long, and patients who required extra neural decompression were excluded, which may lead to deviation in the results and affect the validity of the conclusions.

Conclusion

RoboTFX had the advantages of reduced operating time, less intraoperative bleeding, less fluoroscopy, more accurate insertion, low infection rate and higher rate of fracture healing compared to traditional open methods in the treatment of unstable sacral fractures. Our study confirms previous efforts and supports RoboTFX as an effective and advanced choice. However, the indications of this operation should be strictly controlled.

Author Contributions

Wei Tian: writing and revising the study. Xiao‐Man Dong and Feng‐Shuang Jia: collecting and analyzing the data. Jian Jia: designing the study and supervising the process.

Conflicts of Interest

The authors declare that they have no conflict of interest.

Ethics Statement

This research was approved by the ethic committee of Tianjin hospital.