Real-World Complication Profiles of Fully and Semi-Implanted Intrathecal Drug Delivery Systems for Cancer Pain: A Large Single-Center Chinese Cohort

Abstract

Introduction

Intrathecal drug delivery systems (IDDS) are used for selected patients with refractory cancer pain. Fully implanted and semi-implanted systems differ in implantation pathway, maintenance requirements, external component exposure, and tissue-device interaction. This study aimed to describe real-world complication profiles of these two IDDS models in a large single-center Chinese cohort.

Methods

This retrospective registry-based cohort study included 409 patients with advanced cancer who underwent initial IDDS implantation at Sichuan Cancer Hospital between 1 January 2023, and 31 December 2025. Complications were classified into level I categories and level II subtypes. Each level II complication subtype was counted once per affected patient. Frequencies were reported as counts, patient-level proportions using the device group as denominator, and event-level proportions using recorded complication events as denominator. Because detailed baseline characteristics, standardized follow-up durations, mortality timing, event timing, and paired efficacy outcomes were not systematically available, analyses were descriptive and hypothesis-generating.

Results

Among 409 patients, 38 received fully implanted systems and 371 received semi-implanted systems. A total of 101 complications were recorded: 10 in the fully implanted group and 91 in the semi-implanted group. In the semi-implanted group, postdural puncture headache occurred in 38 patients (10.2% of 371; 41.8% of 91 events), cerebrospinal fluid leak in 22 patients (5.9%; 24.2%), and catheter blockage in 17 patients (4.6%; 18.7%). In the fully implanted group, skin erosion occurred in four patients (10.5% of 38; 40.0% of 10 events), and inadequate analgesia occurred in two patients (5.3%; 20.0%). Findings in the fully implanted group were based on single-digit subtype counts and should not be interpreted as stable comparative risk estimates.

Conclusions

Fully implanted and semi-implanted IDDS showed different recorded complication profiles in this real-world Chinese cohort. Semi-implanted systems were characterized mainly by puncture-related and catheter-pathway events, whereas fully implanted systems showed tissue-integrity and treatment-adaptation events among the small number of recorded complications. These findings support individualized device selection but cannot establish device superiority or time-dependent risk because baseline, follow-up, mortality, event-timing, and efficacy data were incomplete.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40122-026-00862-7.

Plain Language Summary

Intrathecal drug delivery systems can be used when cancer pain remains difficult to control despite standard pain treatment. These systems deliver pain medicine directly into the fluid around the spinal cord. This study reviewed 409 patients with advanced cancer who received an intrathecal drug delivery system at Sichuan Cancer Hospital between January 2023 and December 2025. The study did not show that one device model was safer than the other. Instead, it found that the two models had different recorded complication patterns. Semi-implanted systems were mainly associated with postdural puncture headache, cerebrospinal fluid leak, and catheter blockage. Fully implanted systems had fewer recorded complications, but skin erosion and inadequate analgesia were more prominent among those events. Because some important information, such as baseline patient characteristics, follow-up duration, death timing, complication timing, and pain-score changes, was not available, these findings should be viewed as descriptive. Device choice should consider expected survival, nutrition, soft-tissue condition, follow-up access, cost, and caregiver support.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40122-026-00862-7.

Key Summary Points

Why carry out this study?

IDDS (intrathecal drug delivery systems) are used as salvage interventions for refractory cancer pain, but the clinical burden of different device models is not fully captured by crude overall complication counts.

Evidence from large Chinese real-world cohorts remains limited, particularly for distinguishing fully implanted and semi-implanted systems by recorded complication profile.

This study asked whether the two device models show different descriptive complication profiles rather than whether one model is globally safer or more effective.

What was learned from the study?

In 371 semi-implanted cases, postdural puncture headache (PDPH) was recorded in 38 patients (10.2%) and cerebrospinal fluid (CSF) leak in 22 patients (5.9%); these same events represented 41.8% and 24.2% of recorded semi-implanted complications.

In 38 fully implanted cases, skin erosion was recorded in four patients (10.5%) and inadequate analgesia in two patients (5.3%); these estimates are based on sparse counts and are descriptive only.

Device selection should account for expected survival, nutritional and soft-tissue status, follow-up accessibility, financial constraints, and home-care capacity, while future prospective studies should collect baseline, follow-up, event-timing, mortality, and efficacy data.

Introduction

Refractory cancer pain remains one of the most difficult clinical problems in patients with advanced malignancies [1]. Although many patients achieve meaningful relief with standardized, opioid-centered cancer pain management, a clinically important subgroup continues to experience inadequate analgesia or dose-limiting adverse effects during systemic opioid escalation. These adverse effects, including sedation, nausea, vomiting, and other toxicities, often prevent further dose titration [2, 3]. Intrathecal drug delivery systems (IDDS) have, therefore, emerged as an important salvage strategy for selected patients with refractory cancer pain. By delivering analgesic agents directly into the subarachnoid space, IDDS can achieve effective pain control at substantially lower doses while reducing systemic drug exposure [4]. Recent systematic reviews and meta-analyses further support the role of IDDS in improving pain control, reducing systemic opioid burden, and enhancing quality of life in appropriately selected patients [5, 6].

In clinical practice, IDDS are implemented primarily as fully implanted or semi-implanted systems. These two configurations differ in implantation technique, maintenance requirements, drug-adjustment logistics, exposure to external components, and long-term tissue-device interactions. Current expert guidance recognizes procedure-related and device-related complications as important considerations in IDDS use. However, the existing evidence has often emphasized overall complication rates or isolated adverse events, such as postdural puncture headache (PDPH), cerebrospinal fluid (CSF) leak, infection, or pump-related malfunction [7, 8]. This approach is clinically incomplete because it may suggest that one device type is simply safer than the other. In routine practice, however, fully implanted and semi-implanted systems may generate different patterns of clinical burden.

Evidence from Chinese patients with advanced cancer remains limited, despite the increasing use of IDDS in high-volume oncology centers. A large real-world single-center cohort from China therefore provides an opportunity to characterize how complications are recorded under routine multidisciplinary care. This perspective is clinically relevant because device selection in cancer pain depends not only on technical performance but also on expected survival, nutritional reserve, soft-tissue coverage, family caregiving capacity, access to follow-up, and economic constraints.

Accordingly, this study analyzed a retrospective registry-based cohort of 409 patients who underwent initial IDDS implantation at a tertiary cancer center in western China. Rather than determining whether fully implanted or semi-implanted IDDS is globally superior, we aimed to map level I and level II complication profiles and identify the dominant recorded clinical challenges associated with each intervention model. This profile-based framework may provide a more clinically actionable basis for individualized device selection and help inform the design of future prospective controlled studies.

Methods

Study Setting and Patient Population

This study was conducted at Sichuan Cancer Hospital & Institute (Sichuan Provincial Cancer Prevention and Control Center, and the Affiliated Tumor Hospital of the University of Electronic Science and Technology of China), a large Class-3A specialized cancer hospital in western China integrating cancer prevention, diagnosis, treatment, rehabilitation, research, and education. The cohort reflects routine clinical practice in a high-volume tertiary oncology setting.

Study Design and Reporting Guideline

This retrospective, registry-based observational cohort study included patients who underwent initial IDDS implantation between 1 January 2023, and 31 December 2025. The manuscript was prepared in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations [9]. Because the registry was not originally designed as an analytic comparative database, detailed patient-level baseline characteristics, standardized longitudinal follow-up data, mortality timing, exact complication timing, and paired efficacy outcomes were not systematically available. We therefore did not construct a baseline characteristics table from incomplete data and did not perform adjusted comparative analyses.

Indications and Preimplantation Procedures

IDDS implantation was considered for patients with refractory cancer pain after multidisciplinary clinical assessment. Before implantation, candidates underwent cognitive screening to confirm alertness and the ability to understand and cooperate with the treatment protocol. A routine screening test to verify analgesic responsiveness was performed before formal device implantation.

Device Selection and Perioperative Management

The surgical approach was determined according to pain location, anatomical considerations, and multidisciplinary consensus. Device type was selected in routine clinical practice after discussion among the treatment team, patient, and caregivers. The decision was individualized rather than governed by a formal algorithm. Factors considered included pain distribution, anticipated need for dose adjustment, expected survival, soft-tissue and nutritional condition as assessed clinically, patient preference, financial circumstances, home-care feasibility, caregiver capacity, expected maintenance burden, and access to follow-up. Prophylactic antibiotic therapy administered 0.5 h before surgery was uniform across cases. All implantations were performed by the same experienced surgical team.

Postimplantation Management

After implantation, device flow rate was set according to preoperative drug titration results. Patients were instructed to return 15–30 days postoperatively for dose adjustment and clinical assessment. At follow-up visits, analgesic efficacy, side effects, and device parameters were evaluated and adjusted according to changes in clinical condition. These follow-up visits reflected routine clinical care and were not conducted according to a prospective research schedule.

Complication Classification, Definitions, Outcome Documentation, and Counting Rule

Complications were organized into a two-level profile structure. Level I categories included surgery-related, device-related, drug- or treatment-adaptation-related, and other tissue-integrity complications. Level II subtypes included PDPH, CSF leak, infection, catheter-related injury, catheter blockage, pocket seroma, port abnormality, inadequate analgesia, and skin erosion. The mapping of level I categories to level II subtypes and the operational definitions used for this registry-based analysis are shown in Table 1. For each level II subtype, each affected patient was counted only once. Repeated records of the same subtype in the same patient were not counted as additional subtype events. However, because a patient could theoretically experience more than one different subtype, subtype-specific patient percentages should not be summed to estimate the overall proportion of patients with any complication. The analysis reports both patient-level proportions for each subtype and event-level proportions within the recorded complication profile.

Table 1.

Two-level complication classification, subtype mapping, and operational definitions used for the revised registry-based analysis

Level I category	Level II subtype	Operational definition used for this registry-based analysis	Clinical interpretation
Surgery-related	Postdural puncture headache (PDPH)	Orthostatic or postprocedure headache documented after dural puncture or catheter implantation and managed clinically as postdural puncture headache	Procedure/puncture-related event
Surgery-related	Cerebrospinal fluid (CSF) leak	Clear fluid leakage from the puncture site, wound, or catheter pathway, or clinician-documented CSF leakage after implantation	Procedure- or tract-related leakage event
Surgery-related	Infection	Clinically documented local or systemic infection related to the surgical site, catheter pathway, pocket, or device requiring antimicrobial therapy and/or clinical intervention	Potentially serious infectious event
Device-related	Catheter-related injury	Clinician-documented tissue, neural, or structural injury related to catheter placement, manipulation, or device pathway	Potentially device-modifying event
Device-related	Catheter blockage	Impaired drug delivery, flushing, aspiration, or function attributed to catheter obstruction and requiring clinical assessment or management	Mechanical device-pathway event
Device-related	Pocket seroma	Fluid collection at the pump, port, or device-pocket site documented clinically or by imaging and managed conservatively or procedurally	Local pocket event
Device-related	Port abnormality	Abnormal port access, displacement, leakage, malfunction, or other port-related problems requiring clinical management	Mechanical access event
Drug- or treatment-adaptation-related	Inadequate analgesia	Insufficient analgesia requires unscheduled reassessment, dose escalation, regimen adjustment, or evaluation after excluding obvious mechanical failure	Treatment-adaptation signal; not analyzed as an efficacy end point
Other tissue-integrity	Skin erosion	Skin breakdown, thinning, incision-site erosion, or device-adjacent tissue compromise with concern for progression or device exposure	Soft-tissue integrity event

Definitions are operational and registry-based. They were added to improve transparency; they do not represent a validated IDDS-specific severity grading system

Severity and Unavailable Outcomes

A validated IDDS-specific severity grading system was not available in the registry, and retrospective common terminology criteria for adverse events (CTCAE)-grade reconstruction was not reliable. As an alternative, complications that required device revision, procedural intervention, unscheduled clinical management, treatment escalation, or raised concern for device exposure were identified descriptively as clinically important or device-modifying events where applicable. Deaths during the study period were not analyzed as device-related outcomes because detailed mortality timing and cause-of-death data were not systematically recorded.

Statistical Analysis

Analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Data visualization was conducted using Python Matplotlib (version 3.5.1) and Seaborn (version 0.11.2). Categorical complication profiles were summarized as counts and proportions. For each level II subtype, the device-group denominator was used to calculate patient-level affected proportions, and the total number of recorded complications in each device group was used to calculate event-level profile proportions. Exploratory relative-risk estimates were moved to the Supplementary Material because sparse cells and zero events generated very wide confidence intervals.

Ethical Approval

The study protocol was approved by the Ethics Committee of Sichuan Cancer Hospital (ethical approval number: SCCHEC-03–2024-071). The study was performed in accordance with the Declaration of Helsinki of 1964 and its later amendments. The requirement for informed consent was waived by the ethics committee because this was a retrospective study using anonymized registry data.

Results

Cohort Denominators and Overall Profile

The cohort comprised 409 patients with refractory cancer pain who underwent initial IDDS implantation. Of these, 38 received fully implanted systems and 371 received semi-implanted systems. Overall, 101 complications were documented: 10 in the fully implanted group and 91 in the semi-implanted group. Because the fully implanted group contributed only ten recorded complications, each additional event would shift an event-level proportion in that group by ten percentage points. Therefore, subtype proportions in the fully implanted group should be read as descriptive profile signals rather than stable estimates.

Denominator Clarification

Table 2 presents the same complication data using two denominators: patient-level proportions within each device group and event-level proportions within each recorded complication profile. For each level II subtype, each affected patient was counted once.

Table 2.

Complication counts, patient-level proportions, and event-level profile proportions by intrathecal drug delivery system modality

Level	Category or subtype	Fully implanted (n = 38; 10 recorded events)	Semi-implanted (n = 371; 91 recorded events)	Interpretive note
Level I	Surgery-related	1 (2.6%; 10.0%)	64 (17.3%; 70.3%)	Aggregated procedure-related events
Level I	Device-related	3 (7.9%; 30.0%)	27 (7.3%; 29.7%)	Aggregated mechanical/device-pathway events
Level I	Drug- or treatment-adaptation-related	2 (5.3%; 20.0%)	0 (0.0%; 0.0%)	Treatment adaptation, not formal efficacy
Level I	Other tissue-integrity	4 (10.5%; 40.0%)	0 (0.0%; 0.0%)	Soft-tissue integrity events
Level II	PDPH	0 (0.0%; 0.0%)	38 (10.2%; 41.8%)	Dominant semi-implanted puncture-related subtype
Level II	CSF leak	1 (2.6%; 10.0%)	22 (5.9%; 24.2%)	Reported with patient and event denominators
Level II	Infection	0 (0.0%; 0.0%)	4 (1.1%; 4.4%)	Clinically documented infection
Level II	Catheter-related injury	1 (2.6%; 10.0%)	1 (0.3%; 1.1%)	Sparse event counts
Level II	Catheter blockage	0 (0.0%; 0.0%)	17 (4.6%; 18.7%)	Mechanical catheter-pathway subtype
Level II	Pocket seroma	1 (2.6%; 10.0%)	7 (1.9%; 7.7%)	Local pocket event
Level II	Port abnormality	1 (2.6%; 10.0%)	2 (0.5%; 2.2%)	Mechanical access event
Level II	Inadequate analgesia	2 (5.3%; 20.0%)	0 (0.0%; 0.0%)	Treatment-adaptation signal
Level II	Skin erosion	4 (10.5%; 40.0%)	0 (0.0%; 0.0%)	Soft-tissue integrity event; sparse fully implanted count

Values are n (% of device group; % of recorded events within the same device group). For each level II subtype, each affected patient was counted only once. Level I rows aggregate recorded events and should not be interpreted as mutually exclusive patient-level categories when a patient had more than one subtype

Level I Complication Profiles

In the semi-implanted group, level I complications were predominantly surgery-related, representing 64 of 91 recorded events (70.3%) and corresponding to 17.3% of the 371 patients in this group. Device-related complications accounted for 27 events (29.7%; 7.3% of patients), while no drug- or treatment-adaptation-related, or other tissue-integrity complications were documented.

Fully Implanted Level I Profile

In the fully implanted group, the level I profile was more heterogeneous but based on sparse counts. Other tissue-integrity complications accounted for four of 10 events (40.0%; 10.5% of patients), device-related complications for three events (30.0%; 7.9% of patients), drug- or treatment-adaptation-related complications for two events (20.0%; 5.3% of patients), and surgery-related complications for one event (10.0%; 2.6% of patients).

Level II Complication Profiles

At the level II subtype level, the semi-implanted group was characterized primarily by puncture-related and catheter-pathway events. PDPH was recorded in 38 patients (10.2% of 371; 41.8% of 91 events), CSF leak in 22 patients (5.9%; 24.2%), and catheter blockage in 17 patients (4.6%; 18.7%). Less frequent events included pocket seroma in seven patients (1.9%; 7.7%), infection in four patients (1.1%; 4.4%), port abnormality in two patients (0.5%; 2.2%), and catheter-related injury in one patient (0.3%; 1.1%).

Fully Implanted Level II Profile

In the fully implanted group, skin erosion was recorded in four patients (10.5% of 38; 40.0% of 10 events), and inadequate analgesia was recorded in two patients (5.3%; 20.0%). CSF leak, catheter-related injury, pocket seroma, and port abnormality were each recorded in one patient (2.6% of patients and 10.0% of events for each subtype). No PDPH, infection, or catheter blockage events were recorded in this group. Because all level II findings in the fully implanted group involved single-digit counts, these data cannot support precise comparative risk estimates.

Visual Profile Summary

Figures 1, 2 and 3 were retained as the primary figures because they display raw-count patterns and denominator-dependent profile structure more transparently than sparse relative-risk estimates. The exploratory relative-risk figure was moved to the Supplementary Material as Supplementary Fig. S1. Exploratory relative-risk visualization for level II complication subtypes. This figure is retained only as a descriptive supplementary visualization because sparse cells and zero-event categories generated very wide confidence intervals. It should not be interpreted as evidence of causal effect, device superiority, or time-dependent risk. The main visual findings were interpreted as descriptive clustering of recorded complication types rather than as evidence of device superiority, causal effect, or time-dependent risk.

Fig. 1.

Level I real-world complication profiles by IDDS modality. The figure summarizes level I recorded event profiles. Patient-level denominators and event-level denominators are provided in Table 2

Fig. 2.

Level II real-world complication profiles by IDDS modality. The figure shows the distribution of recorded level II subtypes. Sparse counts in the fully implanted group should be interpreted descriptively. CSF, cerebrospinal fluid; PDPH, postdural puncture headache

Fig. 3.

Heatmap of level II complication event counts by IDDS modality. The heatmap visualizes descriptive profile separation across recorded subtypes; it should not be interpreted as a causal or time-to-event analysis. CSF, cerebrospinal fluid; PDPH, postdural puncture headache

Discussion

In this retrospective registry-based cohort of 409 Chinese patients with refractory cancer pain, fully implanted and semi-implanted IDDS showed different recorded complication profiles. The central finding is descriptive: semi-implanted systems were characterized mainly by puncture-related and catheter-pathway events, whereas fully implanted systems showed tissue-integrity and treatment-adaptation events among the small number of recorded complications.

PDPH represented 41.8% of recorded semi-implanted complications, but the patient-level proportion was 38 of 371 patients (10.2%). Similarly, CSF leak represented 24.2% of recorded semi-implanted complications, but the patient-level proportion was 22 of 371 patients (5.9%).

Semi-implanted systems contain externalized or partially externalized components, making catheter-site care, connector stability, and home maintenance central to their safe use. In this cohort, PDPH, CSF leak, and catheter blockage were the dominant recorded subtypes in the semi-implanted pathway. These events are clinically plausible consequences of dural puncture, catheter passage, tunneling-related tissue trauma, external catheter movement, traction on connecting components, and persistent CSF egress along the puncture or catheter tract [10, 11].

Fully implanted systems presented a different vulnerability profile, but the data was sparse. Skin erosion was recorded in 4 of 38 patients and represented four of ten recorded complications. Fully implanted pumps are larger implanted devices, and many patients with advanced malignancy have cachexia, hypoalbuminemia, reduced subcutaneous fat, or impaired wound healing. Under these conditions, chronic pressure from the pump pocket may plausibly contribute to local ischemia, delayed wound repair, and skin compromise [12]. Nevertheless, because the fully implanted group included only 38 patients and ten recorded complications, the observed profile cannot be treated as a stable estimate of expected complication distribution.

In this study, inadequate analgesia was defined as a treatment-adaptation event requiring unscheduled reassessment or regimen adjustment rather than as a formal efficacy end point. Advanced cancer pain can evolve rapidly as tumor burden, neuropathic invasion, skeletal involvement, inflammation, and treatment-related injury change over time. Fully implanted systems depend on scheduled refills, parameter adjustment, and specialized access procedures, which may delay dose adaptation compared with systems allowing more direct external management [13–15]. However, because paired pre- and postimplantation pain scores, systemic opioid consumption, functional status, and quality-of-life outcomes were not systematically collected, this study cannot evaluate analgesic benefit or the risk–benefit balance of either device model.

Systematic reviews by Duarte et al. and Perruchoud et al. support the analgesic role of IDDS in cancer pain and identify complications such as infection, CSF leak, device malfunction, and local tissue events as clinically important barriers to durable benefit [5, 16]. A recent Chinese study also examined risk factors for postoperative CSF leak after IDDS and external pump implantation in patients with cancer [8]. In the present cohort, the semi-implanted patient-level proportions for PDPH and CSF leak were 10.2% and 5.9%, respectively. Direct classification as higher or lower than pooled international estimates remain inappropriate because prior studies differ in device models, definitions, follow-up windows, reporting denominators, and whether complications were recorded per patient or per event. The main contribution of the present analysis is, therefore, not a pooled-rate comparison but transparent profile mapping in a large Chinese real-world cohort.

Clinically, device selection should not be reduced to initial cost or a crude overall complication count. For patients with limited expected survival, financial constraints, reliable caregiver support, and feasible catheter-site care, a semi-implanted approach may remain practical, but clinicians should anticipate puncture-related and catheter-pathway events. For patients with longer expected survival and reliable access to specialized follow-up, fully implanted systems may reduce external maintenance burden, but soft-tissue coverage, nutritional reserve, and wound-healing capacity require careful assessment. Thus, the two approaches are best understood as different intervention ecosystems rather than as competing devices on a single safety scale.

Several limitations should be acknowledged. First, this was a retrospective registry-based descriptive study, not a prospectively designed comparative effectiveness study. Detailed patient-level baseline characteristics, including age, sex, cancer type and stage, performance status, body mass index, albumin level, nutritional status, prior opioid exposure, social support, and caregiver capacity, were not systematically available. We, therefore, could not construct a reliable baseline characteristics table, evaluate baseline comparability, or perform propensity-score matching or multivariable regression. Second, the fully implanted group was small, comprising 38 patients and only ten recorded complications; subtype-level percentages in this group are unstable and should be interpreted as descriptive signals only. Third, individual follow-up durations, mortality timing, and complication event timing were unavailable. The apparent distinction between puncture-related and tissue-integrity profiles may partly reflect differential observation windows and surveillance rather than intrinsic device biology. Fourth, paired efficacy outcomes, including pain scores, opioid consumption, functional status, and quality-of-life measures, were not systematically collected. Consequently, the present study describes recorded complications but cannot define the overall risk–benefit balance of either IDDS model. Fifth, because this was a single-center study conducted in a high-volume Chinese cancer center, generalizability to other health systems, device models, reimbursement environments, and home-care structures should be interpreted with caution. Methodologically, these unavailable variables increase susceptibility to selection bias, missing-data bias, and exposure-time-related distortion in observational registry analyses [17, 18].

Conclusions

This large single-center Chinese cohort suggests that fully implanted and semi-implanted IDDS for refractory cancer pain show different recorded complication profiles rather than a simple high-versus-low risk hierarchy. Semi-implanted systems were characterized mainly by puncture-related and catheter-pathway events, whereas fully implanted systems showed tissue-integrity and treatment-adaptation events among sparse recorded counts. These findings support individualized device selection based on expected survival, nutritional and soft-tissue status, follow-up accessibility, financial constraints, and home-care capacity. They also provide a hypothesis-generating foundation for future prospective studies with standardized baseline variables, follow-up duration, mortality data, event timing, and efficacy outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization: AZ; methodology: SL; data curation: ZZ; formal analysis: ZN; investigation: HP; writing—original draft: ZN; writing—review and editing: HW; supervision: AZ and PF. The authors critically reviewed, verified, and approved all revised content and took full responsibility for the integrity and accuracy of the manuscript. All named authors meet the International Committee of Medical Journal Editors criteria for authorship, take responsibility for the integrity of the work, and have approved the final version for submission. No authorship changes were made during revision.

Funding

No funding or sponsorship was received for this study or publication of this article. The Rapid Service Fee was funded by the authors.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request, subject to institutional privacy and ethical restrictions.

Declarations

Conflict of Interest

Zikun Ning, Zheng Zou, Shunxin Liu, Hui Pan, Huaiming Wang, Pengjiu Feng, and Aimin Zhang have nothing to disclose.

Ethical Approval

The study protocol was approved by the Ethics Committee of Sichuan Cancer Hospital (ethical approval number: SCCHEC-03–2024-071). The study was performed in accordance with the Declaration of Helsinki of 1964 and its later amendments. The requirement for informed consent was waived by the ethics committee because this was a retrospective study using anonymized registry data.