Robotic‐Assisted Partial Revision of Total Knee Arthroplasty: First Case Report and Literature Review

ABSTRACT

Background

Revision total knee arthroplasty (rTKA) poses substantial technical challenges. Although robotic assistance has demonstrated efficacy in primary TKA, its application in partial revision procedures remains poorly understood. This study presents a case of robotic‐assisted partial rTKA—specifically, isolated tibial component revision using the retained femoral component as a registration reference—and provides a comprehensive review of the current literature.

Case Presentation

A 66‐year‐old male presented with progressive knee pain 8 years after left TKA. Imaging revealed a well‐fixed femoral component but isolated aseptic loosening with posterior tilting of the tibial component. Given the stability of the femoral component, a robotic‐assisted partial revision targeting only the tibial component was performed. The robotic system was registered intraoperatively using the well‐fixed retained femoral component as the primary reference, enabling accurate spatial orientation despite the altered tibial anatomy. A critical prerequisite for this procedure was implant compatibility between the retained Triathlon femoral component and the planned Triathlon TS revision system. After removal of the loose tibial tray and debridement, robotic‐arm guidance was used for precise tibial osteotomy with a 10 mm augment. Postoperative imaging confirmed optimal component positioning. The patient achieved 0°–130° range of motion within 2 weeks, with favorable short‐term clinical and radiological outcomes at 1 year follow‐up.

Conclusion

This case demonstrates the technical feasibility of robotic‐assisted partial TKA revision, with favorable short‐term clinical and radiological outcomes at 1 year follow‐up. The technique enabled precise osteotomy, balanced gap restoration, and preservation of the well‐fixed femoral component. While these preliminary results are encouraging, larger studies with longer follow‐up are needed to establish comparative effectiveness and generalizability.

Robot‐assisted partial TKA enables precise osteotomy, balanced gap restoration, and excellent clinical and radiological outcomes. As robotic platforms continue to advance and clinical evidence accumulates, this technology demonstrates significant promise for improving the outcomes of complex partial knee revision procedures.

1. Introduction

Total knee arthroplasty (TKA) is a highly successful and increasingly common surgical intervention for end‐stage knee osteoarthritis, offering significant pain relief and functional improvement to millions of patients globally [1]. However, with the rising number of primary TKA procedures and an aging population, the demand for revision total knee arthroplasty (rTKA) is projected to rise substantially in the coming decades [2]. Revision surgery poses considerably greater technical challenges than primary TKA. Factors such as extensive bone loss, ligamentous instability, altered anatomy, and the need for accurate component positioning and alignment within a compromised surgical field all contribute to increased procedural complexity and the risk of suboptimal outcomes [3]. Although conventional manual techniques are well established, they rely heavily on surgeon experience and intraoperative decision‐making, making it difficult to achieve consistently precise implant placement and optimal soft‐tissue balance in revision cases [4].

The advent of robotic‐assisted surgery has revolutionized orthopedic procedures, particularly in primary TKA [5]. Systems such as the Mako robotic‐arm‐assisted platform (Stryker Corporation) enable enhanced preoperative planning using detailed 3D imaging, provide real‐time intraoperative feedback, and allow precise bone preparation with haptic boundaries [6]. This technology has demonstrated potential advantages in improving implant placement accuracy, restoring limb alignment, optimizing soft tissue balance, and minimizing unnecessary bone resection in primary TKA [7]. These advantages suggest that robotic assistance may play a pivotal role in addressing the inherent complexities of rTKA, where precision and preservation of remaining bone stock are essential.

While robotic assistance is increasingly being adopted in primary TKA, its application in revision procedures—particularly partial revisions involving the targeted replacement of only failed component(s)—remains poorly documented. Partial revisions, when clinically feasible, offer potential advantages, including reduced surgical morbidity, preservation of well‐fixed components, minimized bone loss, and possibly faster postoperative recovery compared with complete component exchange. However, accurate assessment of component stability, meticulous planning of the optimal revision strategy, and precise bone preparation adjacent to retained stable components present substantial technical challenges.

To the best of our knowledge, this is the first case report describing a patient who underwent robotic‐assisted partial TKA (prTKA) involving isolated tibial component revision with intraoperative registration performed using the retained, well‐fixed femoral component. We describe preoperative planning using the Mako system's 3D CT‐based modeling, an intraoperative procedure that leveraged robotic precision for targeted bone preparation and accurate implant positioning relative to retained components, and the early postoperative outcomes. Importantly, successful execution of this partial revision approach requires explicit verification of implant compatibility between the retained component and the planned revision system—a consideration that has significant implications for patient selection and surgical planning. Additionally, this novel case is supplemented by a comprehensive literature review synthesizing current evidence on robotic assistance in rTKA, examining its potential benefits, technical considerations, limitations, and future directions. Our objective is to contribute initial clinical insights and provide a critical evaluation of the evolving role of this advanced technology in optimizing outcomes in the complex setting of pre‐TKA.

2. Case Report

2.1. General Information

A 66‐year‐old male patient was admitted with a chief complaint of persistent, nonrest pain in the left knee for 2 years following a TKA performed 8 years earlier. The patient's medical history revealed that he had been diagnosed with bilateral knee osteoarthritis (more severe in the left knee) 8 years prior, for which he subsequently underwent left TKA using a Triathlon prosthesis (Stryker, Kalamazoo, MI, USA). The postoperative recovery was uneventful. Two years prior, without any identifiable trigger, the patient developed persistent left knee pain that worsened with physical activity. He received conservative treatments, including acupuncture and hot compress therapy, which provided temporary symptomatic relief. Subsequently, he returned to the hospital for follow‐up evaluation, during which radiological and clinical assessments suggested loosening of the left knee prosthesis, and surgical intervention was recommended.

Additionally, the patient had undergone surgery for left knee joint effusion 40 years earlier. His past medical history included hypertension for more than 20 years, with a peak blood pressure of 170/115 mmHg. He was currently taking a combination tablet of valsartan and hydrochlorothiazide once daily. The patient had also been diagnosed with diabetes mellitus over 20 years ago and was being treated with oral pioglitazone and metformin. More than 10 years earlier, he experienced a cerebral infarction, which was managed conservatively, and he was subsequently maintained on oral traditional Chinese patent medicine. He also underwent bilateral tonsillectomy more than a decade ago.

On physical examination, a “cross”‐type surgical scar was observed anteriorly over the left knee. No erythema or swelling was observed, though mild tenderness was present. The active range of motion (AROM) was 0°–120°, and the passive range of motion (PROM) was 0°–130°. Medial and lateral stress tests of the left knee joint were positive (Figure 1). Laboratory tests revealed normal complete blood counts, a C‐reactive protein level of 0.27 mg/L (reference range: ≤ 0.8 mg/L), an erythrocyte sedimentation rate of 3.00 mm/h (reference range: 0–15 mm/h), an antistreptolysin O titer of < 25 IU/mL (reference range: < 125 IU/mL), and a rheumatoid factor < 20 IU/mL (reference range: ≤ 20 IU/mL). Preoperative radiographic and computed tomographic evaluations demonstrated a marked posterior tilt of the tibial plateau prosthesis, with suspected periprosthetic bone resorption and prosthetic loosening (Figure 2). A sectional three‐phase bone scan revealed no signs of infection. Based on these findings, we diagnosed aseptic loosening of the tibial plateau prosthesis after left TKA.

FIGURE 1.

(A–D) Preoperative general view of the patient's thigh.

FIGURE 2.

(A–C) Preoperative X‐ray data (D) preoperative CT data.

After a multidisciplinary discussion, we decided to perform pre‐TKA using the Mako robotic system (MAKO Surgical Corp., Stryker, Fort Lauderdale, Florida, USA) in conjunction with a Stryker Triathlon TS knee revision implant (Stryker, Kalamazoo, MI, USA).

2.2. Intraoperative Information

Under anesthesia, the left knee was dissected layer by layer, and 15 mL of bright yellow, viscous synovial fluid was aspirated via needle puncture for routine analysis, metagenomic next‐generation sequencing (mNGS) for pathogens, and bacterial culture (Figure 3A,B). The joint was accessed via the medial parapatellar approach, and all hyperplastic scar tissue surrounding the joint was thoroughly excised (Figure 3C,D). The intraoperatively obtained frozen sections of the knee joint synovial membrane revealed papillary hyperplasia, synovial cell proliferation, and a paucity of neutrophils. Combined with the patient's clinical history, these findings were consistent with a synovial histiocytic reaction secondary to prosthetic loosening.

FIGURE 3.

(A) Surgical incision (B) extraction of joint fluid (C and D) the medial para‐patellar approach enters the joint to completely remove the hypertrophic scar tissue around the joint.

Following exposure of the total knee joint, instability was noted due to a significant posterior tilt of the tibial metal tray, although no gross loosening was observed. The femoral component remained stable without evidence of loosening (Figure 4A). The polyethylene insert was then removed (Figure 4B). Next, two 3.2‐mm‐diameter screws were inserted into the medial tibia, approximately 12 cm below the tibial tubercle, and into the medial distal femur, approximately 12 cm proximal to the articular surface. The screws were secured using bicortical fixation and connected to the tracker (Figure 4C,D).

FIGURE 4.

(A and B) Polyethylene gasket was removed (C and D) two 3.2‐mm diameter screws were inserted and bicortical fixation was performed.

Following tracker placement, registration was performed using the MAKO system's standard protocol. For the femoral component, 40 registration points were acquired. For the tibial side, prior to component removal, 40 points were acquired on the tibial tray and exposed bone surfaces. Registration accuracy was quantified by the system's proprietary algorithm, which calculates the root‐mean‐square error between acquired points and the preoperative 3D model. After initial registration, the surgeon used a calibrated probe to verify accuracy by checking 6 additional points on each component, including the tip of the femoral pegs and the periphery of the tibial baseplate. The system displayed a “verification accuracy” of 0.3 mm on the femoral side and 0.4 mm on the tibial side, indicating excellent correspondence between the registered anatomy and the preoperative plan. Preoperative CT imaging was performed using a standardized protocol (120 kV, 200 mA, 0.625 mm slice thickness) with metal artifact reduction software (MARS) enabled to minimize implant‐related artifacts and optimize 3D model reconstruction accuracy.

According to the preoperative plan, registration was performed for both the femoral condyle and tibial plateau prostheses, achieving an accuracy of 0.3 mm on the femoral side and 0.4 mm on the tibial side (Figure 5A,B). The knee exhibited a 7° flexion deformity and 2° varus deformity. Under varus and valgus stress, the medial and lateral joint gaps measured 27 and 28 mm, respectively, at full extension. During flexion, the medial gap measured 24 mm, while the lateral gap remained 28 mm (Figure 5C).

FIGURE 5.

(A and B) Conponent registration (C) the knee parameters after registration.

The interface between the tibial tray and bone cement was carefully dissected using a thin bone curette, followed by removal of the tibial tray and thorough debridement of residual bone cement and fibrous scar tissue from the tibial plateau (Figure 6). Based on preoperative planning and intraoperative assessment of bone defects, robotic‐assisted tibial plateau osteotomy was performed using a 10‐mm augment (Figure 7A,B).

FIGURE 6.

(A and B) The tibial component was removed.

FIGURE 7.

(A and B) The robotic‐assisted tibial plateau osteotomy was performed with a 10 mm thickness augment. (C) The interface of the robot during tibial osteotomy.

The standard polyethylene insert was replaced with a restrictive polyethylene insert. Manual varus and valgus stress testing confirmed well‐balanced medial and lateral gaps throughout flexion and extension. The robotic interface displayed gaps of 32 mm medially and 31 mm laterally at extension, and 30 mm medially and 31 mm laterally at flexion (Figure 8A,B).

FIGURE 8.

(A) The standard polyethylene insert trial was replaced by a restrictive polyethylene insert trial. (B) The robot interface shows the medial and lateral gaps.

Patellar tracking was satisfactory, and the patellar bone was prepared for replacement. All samples were removed, and the joint cavity was thoroughly irrigated and soaked in hydrogen peroxide and iodophor for disinfection. Bone cement was prepared, and the stem (50 × 12 mm), tibial tray (No. 4), and augments (10 mm medial and 10 mm lateral) were assembled and implanted (Figure 9A). The 13‐mm restrictive polyethylene insert and its corresponding reinforcing post were then installed, followed by placement of the patellar prosthesis (Figure 9B). After the bone cement had fully hardened, all excess cement was carefully removed, and the stability of the knee joint was confirmed. The surgical procedure was completed with layered closure of the knee joint tissues. Intraoperative blood loss was 20 mL.

FIGURE 9.

(A and B) The tibial side component was fitted and a polyethylene spacer fitted.

2.3. Postoperative Information and Follow‐Up

Postoperative radiographs demonstrated optimal prosthesis positioning, with satisfactory lower limb alignment and patellar tracking (Figure 10A–D). Postoperative mNGS and bacterial cultures were negative for all target organisms, and no abnormalities were detected in the synovial fluid.

FIGURE 10.

(A–D) Postoperative X ray.

During postoperative follow‐up, the patient achieved independent ambulation 1 week after surgery, with an active range of motion (ROM) of 0°–120° (Figure 11A–C). By 2 weeks postoperatively, active ROM had improved to 0°–130° (Figure 11D–F).

FIGURE 11.

(A–C) The gross observation of the affected limb was followed up at 1 week after operation. (D–F) The gross observation of the affected limb was followed up at 2 weeks after operation.

The Knee Society Score (KSS), Hospital for Special Surgery Knee Score (HSS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), satisfaction score, quality of life score, and Short Form‐12 scores (SF‐12) at 3 and 6 months postoperatively are shown in Table 1. At the 6‐month follow‐up, no abnormalities were observed, and the patient demonstrated good recovery (Figure 12). Radiographic evaluation at 3 months (Figure 13), 6 months (Figure 14), and 1 year (Figure 15) confirmed that the prosthesis remained stable and in good condition.

TABLE 1.

Various scores at 3 and 6 months after surgery.

Time	KSS (clinical scores)	KSS (functional scores)	HSS	WOMAC	Satisfaction scores	Quality of life scores	SF‐12 (PCS)	SF‐12 (MCS)
Before surgery	50	20	74	35	/	/	30	44
3 months after surgery	84	70	80	7	20	Good improvement	37	62
6 months after surgery	96	80	95	6	20	Great improvement	55	61
1 year after surgery	100	100	98.25	2	20	Great improvement	68	72

Note: Satisfaction scores consist of four items: pain relief, household ability, improvement in leisure activity ability, and overall satisfaction. The options are: very satisfied (5 points), relatively satisfied (4 points), moderately satisfied (3 points), relatively dissatisfied (2 points), and very dissatisfied (1 point).

FIGURE 12.

(A–C) The gross observation of the affected limb was followed up at 6 months after operation.

FIGURE 13.

(A–C) X‐ray of the affected limb at 3 months after surgery.

FIGURE 14.

(A–C) X‐ray of the affected limb at 6 months after surgery.

FIGURE 15.

(A–C) X‐ray of the affected limb at 1 year after surgery.

3. Discussion

This study is the first to report the application of the Mako robotic‐assisted system in prTKA, specifically for replacing only the loosened tibial prosthesis. Fortunately, the patient had previously undergone primary surgery with a Stryker implant, which provided a valuable opportunity for a successful revision. Compared with traditional total component revision, this partial approach significantly reduces surgical trauma, preserves stable components, and accelerates postoperative rehabilitation. However, the technical challenges primarily involve accurately assessing the stability of the retained components (in this case, the femoral component was very stable without loosening), carefully planning the mechanical environment to ensure compatibility between the new and existing prostheses—including osteotomy depth, gap balance, mechanical alignment, and rotational alignment—and executing the procedure with millimeter‐level precision near the retained components. By leveraging 3D‐CT‐based preoperative planning and real‐time intraoperative navigation with the Mako robotic system, we successfully addressed these challenges and demonstrated the feasibility and potential of robotic‐assisted technology in partial revision TKA procedures.

3.1. Current Literature on Robotic‐Assisted Primary and Revision TKA

The current literature primarily focuses on robotic‐assisted primary TKA, which has demonstrated superior outcomes compared with conventional surgical techniques in several key aspects, including improved prosthesis alignment accuracy (with errors of less than 1°), higher rates of alignment restoration (exceeding 95% compliance), reduced soft tissue trauma, and enhanced early‐ to mid‐term functional recovery—such as greater postoperative pain relief, accelerated functional rehabilitation, and shorter hospital stays [8, 9, 10, 11, 12, 13, 14]. Long‐term clinical and radiological outcomes are generally comparable to those of conventional TKA [8]. In the field of revision arthroplasty, robotic systems have not yet been licensed or approved by the U.S. Food and Drug Administration (FDA). Since the first report of robotic‐assisted revision knee arthroplasty (UKA → TKA) was published in 2020 [9], only a few studies have described the use of robots in total component revision [9, 10, 11, 12, 13, 14, 15, 16, 17] (Table 2). The reported advantages include improved accuracy of bone cuts with greater bone preservation, precise implant positioning and mechanical alignment despite the absence of typical anatomic landmarks, and enhanced survival rates of revision prostheses. In 2021, Steelman et al. first reported a case of rTKA using the MAKO robotic system for aseptic loosening of the tibial component, achieving satisfactory postoperative alignment and favorable clinical outcomes [10]. In 2023, Danoff et al. used the MAKO robot for revision surgery in a case of periprosthetic joint infection following TKA, demonstrating its excellent ability to restore knee joint function [18]. In 2024, Cochrane et al. used the CORI robotic system to perform revision surgery in 115 TKA patients with aseptic and septic failures, achieving favorable outcomes at the 90‐day follow‐up [15]. In contrast to these full‐revision approaches, our case represents a distinct technical paradigm: isolated tibial revision with the femoral component retained as the registration reference. This distinction is critical, as partial revision preserves bone stock, reduces surgical morbidity, and maintains the well‐fixed femoral component's native alignment. Table 2 provides a comprehensive comparison of published robotic revision cases, highlighting the unique position of our partial‐revision approach within this evolving landscape.

TABLE 2.

Case report evaluating revision knee arthroplasty.

Author, Year	Sample size	Mean age	Type of surgery	Technology	Cause for revision
Kalavrytinos D., 2020 [9]	1	87	rUKA	Mako	Failure of UKA
Yun A.G., 2020 [12]	17	70	rTKA	Robotic‐assisted conversion technique	Osteoarthritis progression/Aseptic loosening/Infection
Tuecking L.R., 2021 [13]	20	62.4	rTKA	NAVIO	Not Report
MacAskill M., 2022 [14]	2	66.5	rTKA	Mako	Persistent pain and subjective instability
Steelman K, 2021 [10]	1	68	rTKA	Mako	Aseptic loosening
Ngim H.J., 2023 [11]	19	69.7	rTKA/rUKA	Mako	Aseptic loosening/Polyethylene wear/Unbalanced knee with restricted range of movement
Cochrane N.H., 2024 [15]	115	65	rTKA	CORI	Instability/Aseptic loosening
Magruder M.L., 2024 [16]	41	72	rTKA	Mako	Osteoarthritis progression/Aseptic loosening/Unspecified pain/Polyethylene wear/Prosthetic joint infection
Andriollo L., 2025 [17]	35	69.7	rTKA	ROSA	UKA failure/Osteoarthritis progression

3.2. Technical Breakthrough and Core Advantages of Robotic Assistance in prTKA

The case in our study represents the first reported application of robotic technology in pre‐TKA, marking a unique technical breakthrough in which prosthesis and bone registration were first performed, using the fixed femoral prosthesis as the reference coordinate system. Based on this registration, the robotic system calculated the alignment, joint line, and positions for bone resection and implant placement. Satisfactory alignment and gap balance were achieved with ease. The ROM reached 0°–130° within 2 weeks postoperatively, with no complications and excellent early outcomes. At 6 months postoperatively, the clinical outcomes remained favorable, aligning with existing research trends, further supporting the potential of robotic assistance as a valuable tool for addressing the technical complexities of TKA revision procedures. The success of this case can be attributed to the following core advantages of robotic technology in prTKA: (1) Enhanced ability for complex anatomical reconstruction. In this study, a three‐dimensional model of the tibial bone defect was reconstructed based on CT data. Using the retained femoral component as the reference, the robot accurately reestablished the tibial mechanical alignment—postoperative radiographs confirmed satisfactory lower limb alignment—and avoided alignment deviation caused by the loss of anatomical landmarks in traditional techniques. (2) Accurate bone defect management. Bone defects in TKA remain a major challenge that directly affects prosthesis stability. Traditional surgery relies heavily on the surgeon's experience to estimate the extent of bone resection, which can lead to excessive bone resection or inadequate bone preservation. The MAKO system can accurately quantify the extent of bone defects through 3D‐CT‐based preoperative planning. When combined with intraoperative confirmation and mapping of the bone structure after prosthesis removal and debridement, this approach maximizes bone preservation, thereby providing biologically stable support for the revision prosthesis. In addition, bone resection can be performed with high precision according to the preoperative plan using the robotic arm. (3) Dynamic soft tissue balance. Patients undergoing rTKA often present with ligamentous laxity or contractures. Traditional techniques rely on the surgeon's tactile feedback to evaluate gap balance and make iterative “trial‐and‐error” adjustments, which may result in a high margin of error. The robotic system can continuously measure medial and lateral gaps in both flexion and extension throughout the procedure, dynamically guide bone resections, assist in selecting the appropriate polyethylene insert, and monitor real‐time gap changes. For instance, in this case, the difference between the medial and lateral gaps was controlled within 1 mm in both extension and flexion positions.

An important technical consideration in robotic partial revision is the potential alteration of the bony reference surface after removal of the loose tibial component and cement. In this case, the robotic guidance relied primarily on the retained femoral component as the stable reference for overall limb alignment and gap balancing. After tibial tray removal and debridement, the team performed a verification check by re‐acquiring several points on the femoral component to confirm that no tracker displacement had occurred during the debridement process. The system confirmed maintained registration accuracy (0.3 mm). We established the following contingency thresholds: if post‐debridement assessment revealed (1) bone loss exceeding 15 mm requiring structural graft, (2) compromised tibial metaphyseal shell preventing stable implant fixation, or (3) femoral component loosening discovered intraoperatively, the plan would be abandoned in favor of conventional full revision.

A critical prerequisite for this procedure was confirming geometric and mechanical compatibility between the retained Triathlon femoral component (Stryker, primary TKA) and the planned Triathlon TS revision tibial system (Stryker). Implant compatibility verification items included femoral component size, articular surface geometry, locking mechanism, bearing thickness range, stem options and augment availability. Also, we propose the following criteria for considering robotic‐assisted partial revision: Mandatory: well‐fixed, stable retained component (no radiographic or intraoperative evidence of loosening); Implant compatibility between retained component and planned revision system (same manufacturer, verified articular compatibility). Relative: aseptic loosening of isolated component; bone defects amenable to augmentation (not requiring structural allograft or metaphyseal cones). Contraindication: active infection (unless planned 2‐stage with interval spacer); mismatched implant brands without verified compatibility.

3.3. Challenges and Future Directions

However, this technology still faces some challenges and limitations. First, patient selection remains critical. Currently, this technique is most appropriate for cases with aseptic loosening and well‐fixed retained components, as in this case. In patients with periprosthetic joint infection, complete revision surgery for all component removal remains necessary. Also, we recognize that a dedicated postoperative CT protocol with standardized long‐leg alignment assessment would provide more robust accuracy data. This represents a limitation of the current study. We have planned a comprehensive CT evaluation at the 1‐year follow‐up to formally assess component position and alignment compared to the preoperative plan. Moreover, since most current robotic platforms are closed systems—and revision surgeries often occur several years or even decades after the initial procedure—it is necessary to ensure that revision prostheses are compatible both with the robotic system and with the original prosthetic components. This compatibility is a prerequisite for performing partial revision surgeries using robotic assistance. Second, the requirements for registration accuracy are extremely high; therefore, precise segmentation of bone and implant structures is essential. Incorporating metal artifact reduction software can further improve imaging accuracy. Furthermore, there remains a learning curve associated with robotic‐assisted revision surgery, and the technology remains expensive. Therefore, multicenter studies are needed to evaluate long‐term cost‐effectiveness, particularly regarding their potential to reduce revision rates and improve prosthesis longevity. Looking ahead, the development of dedicated software modules tailored for revision surgery is required, and formal regulatory approval for robotic assistance in knee revision procedures by the FDA and other regulatory authorities is anticipated.

4. Conclusion

This study confirmed, for the first time, the feasibility, safety, and accuracy of robotic‐assisted partial knee revision, establishing a new technical paradigm characterized by complex anatomical reconstruction, accurate bone defect management, dynamic gap balancing, and co‐positioning of the retained prosthesis. With continued optimization of robotic platforms and the accumulation of clinical evidence, this technology is expected to become an important option for complex revision surgeries, ultimately improving patient prognosis and reducing the overall medical burden.

Author Contributions

Xinguang Wang: writing – original draft, formal analysis, data curation, conceptualization. Kangpeng Li: methodology, software, data curation, investigation, writing – original draft. Ziyang Dong: formal analysis, data curation, writing – review and editing. Zhencan Han: formal analysis, data curation, writing – review and editing. Yipu Zhang: formal analysis, data curation, writing – review and editing. Yang Li: formal analysis, data curation, writing – review and editing. Zijian Li: conceptualization, formal analysis, data curation, supervision, project administration, writing – review and editing. Hua Tian: conceptualization, writing – original draft, writing – review and editing, formal analysis, data curation.

Funding

This study was founded by Beijing Municipal Commission of Science and Technology, Pharmaceutical Innovation varieties and platform cultivation project (Z241100009024023).

Ethics Statement

This study was approved by the Ethics Committee of Peking University Third Hospital (IRB00006761‐M2023429).

Consent

All data and images presented in this manuscript have been anonymized to protect the identity of the individuals involved. For the cases where detailed clinical information is provided, written informed consent for publication was obtained from the patients or their guardians.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.