Abstract
Background
The submental island flap (SIF) is a versatile reconstructive option for oral cavity defects, but concerns regarding recurrences in level I lymph node basins have persisted in the context of neoadjuvant immunochemotherapy (NICT). This study evaluated the clinical outcomes and the incidence of loco-regional recurrences following SIF reconstruction in patients with oral squamous cell carcinoma (OSCC) treated with or without NICT.
Methods
This retrospective cohort study included 279 patients who underwent SIF reconstruction for OSCC between 2018 and 2024, divided into NICT (n = 135) and upfront surgery (US, n = 144) groups. Propensity score matching (1:1) yielded 85 patients per group with balanced baseline characteristics. Primary outcome was loco-regional recurrence (LRR) analyzed using competing risks regression. Secondary outcomes included overall survival (OS), recurrence patterns, and flap-related complications.
Results
After matching, the NICT group demonstrated significantly less LRR (cumulative incidence 11.8% vs. 30.6%; Gray’s test, p = 0.004) and OS (3-year OS 91.8% vs. 85.7%; log-rank p = 0.041). On multivariable analysis, NICT remained an independent protective factor for LRR (adjusted SHR = 0.38; 95% CI: 0.20–0.72; p = 0.003) and OS (adjusted HR = 0.51; 95% CI: 0.27–0.96; p = 0.038). Critically, no recurrences occurred within level I or along the flap subcutaneous tract in either group. Pathological analysis confirmed complete level I lymph node clearance (median yield 7 nodes) with only two NICT patients having positive level I nodes. Flap complication rates were comparable between groups, with no total flap losses.
Conclusions
In this cohort, SIF reconstruction combined with meticulous level I lymph node clearance was feasible and not associated with observed level I or flap tract recurrences. NICT appears to be associated with significantly improved survival without compromising flap outcomes.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12903-026-08668-x.
Keywords: Submental island flap, Neoadjuvant immunochemotherapy, Oral squamous cell carcinoma, Oncologic safety, Propensity score matching
Introduction
Oral squamous cell carcinoma (OSCC) remains a significant global health burden, with over 350,000 new cases diagnosed annually worldwide [1]. Despite advances in surgical techniques and multimodal treatment approaches, the prognosis for patients with locally advanced disease remains suboptimal, with five-year survival rates stagnating at approximately 50–60% over the past decades [2]. The management of OSCC necessitates a delicate balance between achieving adequate oncologic resection and preserving function and aesthetics, particularly when reconstruction of complex oral cavity defects is required [3]. Among the available reconstructive options, the submental island flap (SIF) has emerged as a versatile and reliable technique for defects of the oral cavity, oropharynx, and facial skin [4]. Harvested from the submental region based on the submental artery, this fasciocutaneous flap offers several advantages, including proximity to the recipient site, ease of harvest, favorable color and texture match for intraoral reconstruction, and minimal donor site morbidity [5].
However, the adoption of the SIF has been historically tempered by legitimate oncologic safety concerns. The flap’s vascular pedicle courses through level I of the neck, and the very lymph node basin constitutes the first echelon of lymphatic drainage for most oral cavity subsites [6]. Critics have raised theoretical concerns that harvesting the flap might compromise the completeness of the level I neck dissection or, worse, inadvertently transplant lymph node tissue containing occult metastases into the reconstructive site, potentially seeding tumor cells along the flap tract or into the oral cavity [7]. These concerns have prompted many surgeons to restrict SIF utilization to patients with clinically node-negative necks, thereby limiting its application in a substantial proportion of OSCC patients who present with cervical metastases.
Parallel to these developments in reconstructive surgery, the treatment paradigm for locally advanced OSCC has undergone a transformative shift with the advent of neoadjuvant immunochemotherapy (NICT) [8]. Immune checkpoint inhibitors targeting the PD-1 pathway have demonstrated remarkable efficacy in recurrent or metastatic head and neck squamous cell carcinoma, leading to their investigation in the neoadjuvant setting [9]. The KEYNOTE-689 trial [10] has reported major pathologic response rates exceeding 50%, and the compelling oncologic benefits have established NICT as an increasingly attractive strategy for downstaging advanced tumors, facilitating surgical resection, and improving long-term outcomes.
The convergence of expanding role of NICT and persistent questions surrounding SIF oncologic safety creates both opportunities and uncertainties. On one hand, tumor downstaging and nodal clearance achieved with NICT could theoretically mitigate the risk of occult level I metastases, potentially expanding the indications for SIF reconstruction. On the other hand, the inflammatory and fibrotic changes induced by immunotherapy in both tumors and lymph nodes might alter tissue planes, potentially complicating flap harvest or obscuring pathologic assessment [11]. Furthermore, no study to date has systematically evaluated whether SIF reconstruction in patients who have received NICT is associated with an increased risk of level I recurrences or iatrogenic tumor seeding, this critical knowledge gap must be addressed before this reconstructive approach can be safely embraced in the NICT era.
In this context, we conducted a retrospective cohort study to evaluate the impact of NICT on patients undergoing SIF reconstruction, specifically monitoring for level I recurrences.
Methods
Study design and setting
The retrospective study was conducted at The First Affiliated Hospital of Zhengzhou University, and was approved by the Institutional Review Board of The First Affiliated Hospital of Zhengzhou University. The study adhered to the Declaration of Helsinki and followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for reporting cohort studies (Supplementary file 1).
Patient selection and cohort definition
We conducted a retrospective review of the medical records of all consecutive patients who underwent surgical resection for OSCC with simultaneous reconstruction using a pedicled SIF between January 2018 and December 2024. Because NICT protocols were introduced at our institution in 2020, patients receiving NICT were treated exclusively between 2020 and 2024, whereas the upfront surgery (US) cohort includes patients treated across the entire study period (2018–2024).
Inclusion criteria
Histologically confirmed diagnosis of squamous cell carcinoma of the oral cavity (including oral tongue, floor of mouth, buccal mucosa, gingiva, hard palate, and retromolar trigone).
Treatment with primary tumor resection and neck dissection (at least levels I-III) with immediate reconstruction using a pedicled SIF.
Availability of complete medical records, including detailed operative notes, pathology reports, and postoperative follow-up data.
Minimum postoperative follow-up of at least 12 months for patients without a documented recurrence or death.
Exclusion criteria
History of prior radiotherapy to the head and neck region.
History of prior neck dissection.
Presence of synchronous or metachronous second primary malignancies at the time of diagnosis.
Clinically or radiologically evident distant metastasis at presentation.
Patients lost to follow-up immediately after discharge without any postoperative oncologic surveillance data.
Incomplete pathological data, specifically regarding level I lymph node yield or status.
Based on their preoperative treatment, eligible patients were assigned to one of two cohorts: NICT Group consisted of patients who received at least two cycles of NICT prior to surgery. Upfront Surgery (US) Group included patients who underwent primary surgery without any neoadjuvant treatment.
Neoadjuvant immunochemotherapy regimen and response evaluation
For patients in the NICT group, the specific NICT regimen, including the choice of agent (sintilimab, toripalimab, or camrelizumab at 200 mg every three weeks combined with cisplatin based chemotherapy at 75 mg/m² every three weeks) and the planned number of cycles, was determined for each patient by our institutional Multidisciplinary Tumor Board based on tumor characteristics, patient comorbidities, and contemporary clinical guidelines.
Clinical and radiographic response to NICT was assessed based on the iRECIST guideline [12]. Pre-treatment imaging (contrast-enhanced CT, MRI, or PET-CT) was compared with restaging imaging performed after the completion of neoadjuvant therapy. The response was categorized as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). For subsequent subgroup analyses, patients were dichotomized into responders (CR/PR) and non-responders (SD/PD).
Surgical technique and pathological assessment
Following a planned washout period of 3–6 weeks after the last NICT dose, a consistent surgical team performed curative-intent resection of the primary tumor in all patients. A comprehensive neck dissection, which routinely included levels I, II, and III at least, was performed in all cases. The SIF was harvested using a standardized antegrade approach. A critical component of the procedure was the oncologic clearance of level I lymph nodes. To achieve this, the flap pedicle (anterior belly of the digastric muscle, submental artery, and vein) underwent meticulous skeletonization. By aggressively dissecting the pedicle, this technique effectively shifts the reconstructive method from a traditional SIF toward a perforator-based flap. All fibrofatty and lymphatic tissue superficial and deep to the pedicle, as well as tissue surrounding the submandibular gland, was removed as part of the level I neck dissection specimen and sent separately for pathological analysis. This skeletonized technique ensured that flap harvest did not compromise the completeness of the oncologic neck dissection and reduced the risk of inadvertent lymph node transfer.
Pathological evaluation was performed by dedicated head and neck pathologists who were aware of the patients’ treatment history as per routine clinical practice. The following data were extracted from pathology reports:
Primary Tumor: Pathological T-stage (according to the AJCC 8th edition), tumor grade, and final surgical margin status (clear [≥ 5 mm], close [< 5 mm], or positive [tumor at the inked margin]).
Lymph Nodes: Pathological N-stage, total number of lymph nodes harvested from the neck dissection specimen, number of positive lymph nodes, and the presence of extranodal extension (ENE).
Level I Analysis: Specific data on level I lymph nodes were collected, including the number of nodes harvested and the number of positive nodes in level I. In the NICT group, a descriptive analysis of treatment-related pathological changes in level I lymph nodes (e.g., fibrosis, necrosis, giant cell reaction, foamy histiocyte infiltration) was performed.
Outcome variables
The primary outcome was loco-regional recurrence (LRR). LRR was defined as the time from the date of surgery to the date of first documented appearance of new local or regional disease after the establishment of an imaging-confirmed disease-free state of three months. Patients with positive surgical margins who had gross residual disease confirmed on immediate post-operative imaging were classified as having persistent disease and were not considered at risk for LRR events. Patients who died without evidence of LRR were censored at the time of death in standard survival analysis, but were treated as competing events in the competing risk analysis.
Secondary outcomes included:
Overall survival (OS): Defined as the time from surgery to death from any cause.
Pattern of recurrence: The anatomical site of first recurrence was recorded, with specific attention paid to recurrences occurring within level I or in the subcutaneous tract of the flap, which could suggest iatrogenic tumor seeding.
Flap-related complications graded as minor (I–II) and major (≥ III) according to the Clavien-Dindo classification [13]: The incidence of total or partial flap loss, vascular compromise requiring re-exploration, hematoma, seroma, wound infection, and donor site morbidities such as marginal mandibular nerve weakness or hematoma.
Postoperative follow-up was conducted according to institutional protocols. Patients typically underwent clinical examination and imaging (ultrasound, contrast-enhanced CT, or MRI) every 3–6 months for the first two years, and every 6–12 months thereafter. Data were collected until December 31, 2025, or until the date of death or last known follow-up.
Statistical analysis
Categorical variables were presented as frequencies and percentages, while continuous variables were presented as mean with standard deviation or median with interquartile range (IQR), depending on their distribution. Baseline characteristics between the NICT and US groups were compared using the Chi-square test or Fisher’s exact test for categorical variables, and the independent samples t-test or Mann-Whitney U test for continuous variables.
To reduce the impact of treatment selection bias and potential confounding inherent in this observational study, a propensity score matching (PSM) analysis was performed. Propensity scores were estimated using a multivariable logistic regression model with the receipt of NICT as the dependent variable. NICT and US patients were then matched 1:1 using a nearest-neighbor matching algorithm with a caliper width of 0.2 of the standard deviation of the logit of the propensity score. Standardized mean differences (SMD) were used to assess covariate balance between the two groups in the matched cohort, with an SMD < 0.1 indicating acceptable balance; variance ratios were also examined for continuous variables.
The primary outcome of LRR was analyzed using a competing risks framework. In standard Kaplan-Meier survival analysis, patients who die without experiencing a LRR are censored at the time of death. This censoring assumes that these patients would have eventually experienced the event of interest had they lived, an assumption that is often violated in head and neck oncology where mortality from competing causes is non-negligible. Treating death as a non-informative censoring event in this context can lead to an overestimation of the cumulative incidence of LRR. Therefore, to minimize this bias, we employed a competing risks model where LRR was the event of interest and death without prior recurrence was defined as the competing event. We calculated the cumulative incidence function (CIF) for each cohort, which estimates the marginal probability of recurrence accounting for the fact that a subject cannot experience recurrence after death. Differences in CIFs between the NICT and US groups were compared using Gray’s test. For multivariable analysis of LRR, we applied the Fine-Gray subdistribution hazard model, which directly models the effect of covariates on the CIF. The proportional subdistribution hazards assumption was verified by examining Schoenfeld-type residuals and testing for correlation with time. For OS, survival curves were estimated using the Kaplan-Meier method and compared between the NICT and US groups using the log-rank test. Variables with a p-value < 0.10 in the univariate analysis, along with clinically important factors (pathological N-stage, ENE, margin status, and receipt of adjuvant radiotherapy), were entered into a multivariable model. The adjusted hazard ratio (HR) for the effect of NICT, along with its 95% confidence interval (CI), was calculated. The proportional hazards assumption was tested using Schoenfeld residuals.
Subgroup analyses were planned a priori to evaluate the effect of NICT on LRR with False Discovery Rate (FDR) correction in:
Patients with pathologically confirmed lymph node metastasis (pN+).
Patients according to their response to NICT (responders [CR/PR] vs. non-responders [SD/PD]).
Patients treated exclusively in the immunotherapy era (2020–2024).
NICT group l by specific anti-PD-1 agent and by number of NICT cycles.
All statistical tests were two-sided, and a p-value of less than 0.05 was considered statistically significant. Statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY) and R version 4.5.2 (R Foundation for Statistical Computing, Vienna, Austria).
Results
Patient characteristics and cohort derivation
A total of 378 patients were initially screened for eligibility. Following the application of the exclusion criteria, 99 patients were excluded. The reasons for exclusion were: history of prior radiotherapy to the head and neck (n = 24), history of prior neck dissection (n = 18), presence of a second primary malignancy (n = 15), distant metastasis at presentation (n = 8), incomplete follow-up data (n = 22), and incomplete pathological data regarding level I lymph nodes (n = 12). Consequently, a total of 279 patients were included in the final analysis. Based on their preoperative treatment, 135 patients were assigned to the NICT group, and 144 patients were assigned to the US group.
A 1:1 PSM was performed. The model included age, sex, smoking status, drinking status, Charlson Comorbidity Index (CCI), body mass index (BMI), primary tumor site, clinical T-stage, clinical N-stage, and tumor grade. After matching, 85 patients were included in the NICT group and 85 in the US group. The mean age was 58.7 ± 10.1 years, and the cohort included 118 males and 52 females. Covariate balance in the matched cohort was significantly improved, with all SMDs falling below 0.1 (range, 0.012–0.081). Variance ratios for continuous variables ranged from 0.85 to 1.18, confirming excellent covariate balance (Table 1 and Supplementary Fig. 1).
Table 1.
Baseline characteristics of patients before and after propensity score matching
| Characteristic | Before PSM | After PSM | ||||
|---|---|---|---|---|---|---|
| NICT (n = 135) |
US (n = 144) |
p | NICT (n = 85) |
US (n = 85) |
SMD | |
| Age (years), mean (SD) | 58.4 (10.2) | 59.1 (11.5) | 0.591 | 58.9 (9.8) | 58.5 (10.5) | 0.039 |
| Sex, n (%) | 0.742 | 0.047 | ||||
| Male | 92 (68.1) | 101 (70.1) | 58 (68.2) | 60 (70.6) | ||
| Female | 43 (31.9) | 43 (29.9) | 27 (31.8) | 25 (29.4) | ||
| Smoking Status, n (%) | 0.512 | 0.055 | ||||
| Never Smoker | 51 (37.8) | 60 (41.7) | 35 (41.2) | 33 (38.8) | ||
| Former Smoker | 28 (20.7) | 33 (22.9) | 18 (21.2) | 20 (23.5) | ||
| Current Smoker | 56 (41.5) | 51 (35.4) | 32 (37.6) | 32 (37.6) | ||
| Drinking Status, n (%) | 0.428 | 0.048 | ||||
| Never Drinker | 47 (34.8) | 58 (40.3) | 33 (38.8) | 31 (36.5) | ||
| Former Drinker | 25 (18.5) | 29 (20.1) | 16 (18.8) | 18 (21.2) | ||
| Current Drinker | 63 (46.7) | 57 (39.6) | 36 (42.4) | 36 (42.4) | ||
| CCI, n (%) | 0.912 | 0.015 | ||||
| < 2 | 100 (74.1) | 106 (73.6) | 64 (75.3) | 63 (74.1) | ||
| ≥ 2 | 35 (25.9) | 38 (26.4) | 21 (24.7) | 22 (25.9) | ||
| BMI, n (%) | 0.972 | 0.081 | ||||
| < 18.5 | 35 (25.9) | 39 (27.1) | 18 (21.2) | 20 (23.5) | ||
| 18.5–24 | 76 (56.3) | 80 (55.6) | 55 (64.7) | 58 (68.2) | ||
| > 24 | 24 (17.8) | 25 (17.4) | 12 (14.1) | 7 (8.2) | ||
| Primary Tumor Site, n (%) | 0.381 | 0.072 | ||||
| Oral Tongue | 68 (50.4) | 80 (55.6) | 46 (54.1) | 48 (56.5) | ||
| Floor of Mouth | 22 (16.3) | 19 (13.2) | 13 (15.3) | 12 (14.1) | ||
| Buccal Mucosa | 18 (13.3) | 24 (16.7) | 12 (14.1) | 13 (15.3) | ||
| Gingiva | 15 (11.1) | 12 (8.3) | 8 (9.4) | 7 (8.2) | ||
| Other | 12 (8.9) | 9 (6.3) | 6 (7.1) | 5 (5.9) | ||
| Clinical T-stage, n (%) | < 0.001 | 0.024 | ||||
| cT1-2 | 54 (40.0) | 118 (81.9) | 41 (48.2) | 42 (49.4) | ||
| cT3-4 | 81 (60.0) | 26 (18.1) | 44 (51.8) | 43 (50.6) | ||
| Clinical N-stage, n (%) | < 0.001 | 0.012 | ||||
| cN0 | 42 (31.1) | 115 (79.9) | 33 (38.8) | 34 (40.0) | ||
| cN+ | 93 (68.9) | 29 (20.1) | 52 (61.2) | 51 (60.0) | ||
| Tumor Grade, n (%) | 0.225 | 0.061 | ||||
| Well differentiated | 31 (23.0) | 42 (29.2) | 20 (23.5) | 22 (25.9) | ||
| Moderately differentiated | 77 (57.0) | 81 (56.3) | 49 (57.6) | 46 (54.1) | ||
| Poorly differentiated | 27 (20.0) | 21 (14.6) | 16 (18.8) | 17 (20.0) | ||
| Adjuvant Radiotherapy | 92 (68.1) | 48 (33.3) | < 0.001 | 48 (56.5) | 46 (54.1) | 0.048 |
Data are presented as mean (standard deviation) for continuous variables and number (percentage) for categorical variables
Abbreviations: PSM Propensity score matching, CCI Charlson Comorbidity Index, BMI Body mass index, NICT Neoadjuvant immunochemotherapy, US Upfront surgery, SMD Standardized mean difference
P-values were calculated using t-tests or chi-square tests as appropriate. An SMD < 0.1 indicates acceptable balance after matching
Treatment details and response
In the NICT group, all 135 patients received a neoadjuvant regimen consisting of an anti-PD-1 agent, including sintilimab (n = 78, 57.8%), toripalimab (n = 41, 30.4%), and camrelizumab (n = 16, 11.9%). The median number of NICT cycles administered was 3 (IQR, 2–4), and the median time interval between the last dose of immunotherapy and definitive surgery was 28 days (IQR, 21–35 days).
The objective response rate was 90.4%, with 41 patients (30.4%) achieving a CR and 81 patients (60.0%) achieving a PR. SD was observed in 13 patients (9.6%), and no patient experienced PD during the neoadjuvant period. Patients were dichotomized into responders (CR/PR, n = 122) and non-responders (SD/PD, n = 13). Regarding adjuvant treatment, patients in the NICT group received adjuvant radiotherapy more frequently than those in the US group before propensity score matching (68.1% vs. 33.3%), consistent with their more advanced clinical stage at presentation. After matching, the rates of adjuvant radiotherapy were well-balanced between the NICT and US groups (56.5% vs. 54.1%, respectively) (Table 1 and Table 2).
Table 2.
Pathological outcomes in the propensity score-matched cohort
| Characteristic | NICT Group (n = 85) | US Group (n = 85) |
|---|---|---|
| Pathological T-stage, n (%) | ||
| ypT0 | 28 (32.9) | — |
| ypT1-2 | 40 (47.1) | — |
| ypT3-4 | 17 (20.0) | — |
| pT1-2 | — | 41 (48.2) |
| pT3-4 | — | 44 (51.8) |
| Surgical Margin Status, n (%) | ||
| Clear (≥ 5 mm) | 77 (90.6) | 66 (77.6) |
| Close (< 5 mm) | 8 (9.4) | 14 (16.5) |
| Positive (tumor at ink) | 0 (0.0) | 5 (5.9) |
| Pathological N-stage, n (%) | ||
| ypN0 | 60 (70.6) | — |
| ypN+ | 25 (29.4) | — |
| pN0 | — | 31 (36.5) |
| pN+ | — | 54 (63.5) |
| Total Lymph Node Yield | ||
| Median (IQR) | 33 (28–41) | 34 (28–42) |
| Number of Positive Lymph Nodesa | ||
| Median (IQR) | 1 (1–2) | 2 (1–4) |
| Extranodal Extension (ENE), n (%) | ||
| Present | 13 (15.3) | 27 (31.8) |
| Absent | 72 (84.7) | 58 (68.2) |
| Level I Analysis | ||
| Level I Lymph Node Yield | ||
| Median (IQR) | 7 (5–9) | 7 (5–10) |
| Patients with Positive Level I Nodes, n (%) | 2 (8.0) | 6 (11.1) |
Data are presented as median (interquartile range) for continuous variables and number (percentage) for categorical variables
Abbreviations: NICT Neoadjuvant immunochemotherapy, US Upfront surgery, ENE Extranodal extension
aAmong patients with positive lymph nodes
Loco-regional recurrence
In the PSM cohort of 170 patients, the median follow-up time was 38 months (IQR, 24–52 months). LRR was observed in 10 patients (11.8%) in the NICT group and 26 patients (30.6%) in the US group. The NICT group demonstrated a significantly lower cumulative incidence of LRR compared to the US group (Gray’s test, p = 0.004, Fig. 1). Further multivariable analysis revealed that NICT remained an independent protective factor against LRR (adjusted SHR = 0.38; 95% CI: 0.20–0.72; p = 0.003), corresponding to a 62% relative risk reduction. In addition, the presence of extranodal extension (adjusted SHR = 1.89; 95% CI: 1.05–3.40; p = 0.034) and positive surgical margins (adjusted SHR = 3.71; 95% CI: 1.52–9.05; p = 0.004) were independently associated with an increased risk of LRR (Table 3).
Fig. 1.
Cumulative incidence of loco-regional recurrence in the propensity score-matched cohort. Data are presented as cumulative incidence functions (CIF) with 95% confidence bands. The NICT group demonstrated a significantly lower cumulative incidence of loco-regional recurrence compared to the US group (Gray’s test, p = 0.004). Abbreviations: NICT, neoadjuvant immunochemotherapy; US, upfront surgery
Table 3.
Univariate and multivariable competing risks analysis for loco-regional recurrence
| Characteristic | Patients (n) |
Events (n) |
Univariate Analysis | Multivariable Analysis | ||
|---|---|---|---|---|---|---|
| SHR (95% CI) | p-value | Adjusted SHR (95% CI) |
p-value | |||
| Treatment Group | ||||||
| Upfront Surgery | 85 | 26 | 1.00 (Reference) | 1.00 (Reference) | ||
| NICT | 85 | 10 | 0.35 (0.19–0.64) | < 0.001 | 0.38 (0.20–0.72) | 0.003 |
| Age (years) | ||||||
| Mean (SD) | 58.7 (10.1) | – | 0.99 (0.97–1.02) | 0.591 | 1.01 (0.98–1.04) | 0.512 |
| Sex | ||||||
| Female | 52 | 10 | 1.00 (Reference) | 1.00 (Reference) | ||
| Male | 118 | 26 | 1.24 (0.71–2.17) | 0.448 | 1.15 (0.63–2.10) | 0.651 |
| Smoking Status | ||||||
| Never Smoker | 68 | 12 | 1.00 (Reference) | 1.00 (Reference) | ||
| Former Smoker | 38 | 8 | 1.31 (0.64–2.68) | 0.462 | 1.18 (0.55–2.53) | 0.672 |
| Current Smoker | 64 | 16 | 1.98 (1.08–3.63) | 0.027 | 1.52 (0.79–2.92) | 0.208 |
| Drinking Status | ||||||
| Never Drinker | 64 | 11 | 1.00 (Reference) | 1.00 (Reference) | ||
| Former Drinker | 34 | 7 | 1.28 (0.58–2.82) | 0.540 | 1.09 (0.47–2.53) | 0.842 |
| Current Drinker | 72 | 18 | 1.89 (1.04–3.43) | 0.036 | 1.34 (0.70–2.57) | 0.381 |
| CCI | ||||||
| < 2 | 127 | 24 | 1.00 (Reference) | 1.00 (Reference) | ||
| ≥ 2 | 43 | 12 | 1.52 (0.83–2.78) | 0.174 | 1.38 (0.73–2.61) | 0.321 |
| BMI | ||||||
| < 18.5 | 38 | 11 | 1.48 (0.78–2.81) | 0.228 | 1.35 (0.69–2.64) | 0.382 |
| 18.5–24 | 113 | 20 | 1.00 (Reference) | 1.00 (Reference) | ||
| > 24 | 19 | 5 | 1.21 (0.52–2.82) | 0.651 | 1.15 (0.48–2.75) | 0.751 |
| Primary Tumor Site | ||||||
| Oral Tongue | 94 | 18 | 1.00 (Reference) | 1.00 (Reference) | ||
| Floor of Mouth | 25 | 6 | 1.42 (0.65–3.10) | 0.380 | 1.21 (0.53–2.76) | 0.652 |
| Buccal Mucosa | 25 | 5 | 1.08 (0.46–2.54) | 0.861 | 0.95 (0.39–2.31) | 0.912 |
| Gingiva | 15 | 4 | 1.53 (0.61–3.84) | 0.364 | 1.32 (0.50–3.48) | 0.575 |
| Other | 11 | 3 | 1.61 (0.56–4.63) | 0.378 | 1.41 (0.47–4.23) | 0.540 |
| Clinical T-stage | ||||||
| cT1-2 | 83 | 12 | 1.00 (Reference) | 1.00 (Reference) | ||
| cT3-4 | 87 | 24 | 2.15 (1.24–3.73) | 0.006 | 1.48 (0.79–2.77) | 0.222 |
| Clinical N-stage | ||||||
| cN0 | 67 | 8 | 1.00 (Reference) | 1.00 (Reference) | ||
| cN+ | 103 | 28 | 2.89 (1.54–5.42) | 0.001 | 1.52 (0.73–3.16) | 0.262 |
| Tumor Grade | ||||||
| Well differentiated | 42 | 6 | 1.00 (Reference) | 1.00 (Reference) | ||
| Moderately differentiated | 95 | 18 | 1.46 (0.68–3.13) | 0.334 | 1.22 (0.55–2.71) | 0.622 |
| Poorly differentiated | 33 | 12 | 2.89 (1.28–6.52) | 0.010 | 1.71 (0.71–4.12) | 0.228 |
| Extranodal Extension (ENE) | ||||||
| Absent | 130 | 22 | 1.00 (Reference) | 1.00 (Reference) | ||
| Present | 40 | 14 | 2.41 (1.41–4.12) | 0.001 | 1.89 (1.05–3.40) | 0.034 |
| Surgical Margin Status | ||||||
| Clear (≥ 5 mm) | 143 | 26 | 1.00 (Reference) | 1.00 (Reference) | ||
| Close (< 5 mm) | 22 | 6 | 1.60 (0.79–3.24) | 0.191 | 1.33 (0.63–2.81) | 0.455 |
| Positive | 5 | 4 | 5.32 (2.52–11.23) | < 0.001 | 3.71 (1.52–9.05) | 0.004 |
| Adjuvant Radiotherapy | ||||||
| No | 76 | 16 | 1.00 (Reference) | 1.00 (Reference) | ||
| Yes | 94 | 20 | 1.12 (0.65–1.93) | 0.678 | 0.89 (0.50–1.58) | 0.691 |
Data are presented as subdistribution hazard ratio (SHR) with 95% confidence interval. Bold values indicate statistical significance at p < 0.05
Abbreviations: CCI Charlson Comorbidity Index, BMI Body mass index, NICT Neoadjuvant immunochemotherapy
Recurrence pattern
The distribution of first recurrence sites also differed. In the NICT group, recurrences were local in 4 patients (4.7%), regional in 5 patients (5.9%), and distant in 1 patient (1.2%). In the US group, recurrences were local in 9 patients (10.6%), regional in 13 patients (15.3%), and distant in 4 patients (4.7%) (Table 4).
Table 4.
Patterns of first recurrence and flap-related complications in the propensity score-matched cohort
| Outcome | NICT Group (n = 85) | US Group (n = 85) | p Value |
|---|---|---|---|
| Patterns of Recurrence | |||
| Total Patients with Recurrence, n (%) | 10 (11.8) | 26 (30.6) | 0.003 |
| Site of First Recurrence, n (%) | |||
| Local (Primary Site) | 4 (4.7) | 9 (10.6) | 0.248 |
| Regional (Neck Nodes) | 5 (5.9) | 13 (15.3) | 0.080 |
| Distant Metastasis | 1 (1.2) | 4 (4.7) | 0.369 |
| Critical Safety Analysis | |||
| Recurrence in Level I, n (%) | 0 (0.0) | 0 (0.0) | > 0.999 |
| Recurrence in Flap Subcutaneous Tract, n (%) | 0 (0.0) | 0 (0.0) | > 0.999 |
| Flap-related Complications | |||
| Any Complication, n (%) | 16 (18.8) | 18 (21.2) | 0.701 |
| Minor Complications (Clavien-Dindo I–II) | |||
| Hematoma (Grade I) | 2 (2.4) | 1 (1.2) | 0.560 |
| Seroma (Grade I) | 3 (3.5) | 3 (3.5) | > 0.999 |
| Wound Infection (Grade II) | 4 (4.7) | 5 (5.9) | 0.732 |
| Marginal Mandibular Nerve Weakness (Grade I) | 6 (7.1) | 7 (8.2) | 0.774 |
| Major Complications (Clavien-Dindo III–V) | |||
| Partial Flap Loss | 1 (1.2) | 2 (2.4) | 0.560 |
| Total Flap Loss | 0 (0.0) | 0 (0.0) | > 0.999 |
| Grade IV or V Complications | 0 (0.0) | 0 (0.0) | > 0.999 |
Data are presented as number (percentage). Complications were graded according to the Clavien-Dindo classification. P-values were calculated using Chi-square or Fisher’s exact test as appropriate. Bold values indicate statistical significance at p < 0.05
Abbreviations: NICT Neoadjuvant immunochemotherapy, US Upfront surgery
Flap-related complication
In the NICT group, there were 15 minor complications including two hematomas, three seromas, four wound infections, and six cases of marginal mandibular nerve weakness. Additionally, One patient (1.2%) experienced a major complication. In the US group, minor complications included one hematoma, three seromas, five wound infections, and seven cases of marginal mandibular nerve weakness. Two patients (2.4%) experienced major complications. The incidence of major complications was comparable between groups (1.2% vs. 2.4%; p = 0.560) (Table 4).
OS
At the time of analysis, there were 13 deaths (15.3%) in the NICT group and 22 deaths (25.9%) in the US group. The 3-year OS rates for the NICT group were 91.8%, compared to 85.7% for the US group (p = 0.041, Fig. 2). After multivariable adjustment, NICT remained an independent protective factor for OS (adjusted HR = 0.51; 95% CI: 0.27–0.96; p = 0.038), corresponding to a 49% relative risk reduction in mortality compared to US. Additionally, extranodal extension (adjusted HR = 2.44; 95% CI: 1.28–4.65; p = 0.007), positive surgical margins (adjusted HR = 5.03; 95% CI: 1.98–12.78; p = 0.001) and poorly differentiated tumor grade (adjusted HR = 2.38; 95% CI: 1.00–5.66; p = 0.0496) were independently associated with significantly worse overall survival (Table 5).
Fig. 2.
Overall survival in the propensity score-matched cohort. Data are presented as Kaplan-Meier survival estimates. Tick marks indicate censored observations. The 3-year overall survival rate was 91.8% in the NICT group compared to 85.7% in the US group (log-rank p = 0.041). Abbreviations: NICT, neoadjuvant immunochemotherapy; US, upfront surgery
Table 5.
Univariate and multivariable cox regression analysis for overall survival
| Characteristic | Patients (n) |
Events (n) |
Univariate Analysis | Multivariable Analysis | ||
|---|---|---|---|---|---|---|
| HR (95% CI) | p-value | Adjusted HR (95% CI) |
p-value | |||
| Upfront Surgery | 85 | 22 | 1.00 (Reference) | 1.00 (Reference) | ||
| NICT | 85 | 13 | 0.54 (0.30–0.98) | 0.042 | 0.51 (0.27–0.96) | 0.038 |
| Age (years) | ||||||
| < 60 | 89 | 16 | 1.00 (Reference) | 1.00 (Reference) | ||
| ≥ 60 | 81 | 19 | 1.35 (0.69–2.63) | 0.381 | 1.18 (0.59–2.36) | 0.638 |
| Sex | ||||||
| Female | 52 | 10 | 1.00 (Reference) | 1.00 (Reference) | ||
| Male | 118 | 25 | 1.15 (0.61–2.17) | 0.665 | 1.09 (0.56–2.12) | 0.802 |
| Smoking Status | ||||||
| Never Smoker | 68 | 10 | 1.00 (Reference) | 1.00 (Reference) | ||
| Former Smoker | 38 | 8 | 1.52 (0.67–3.45) | 0.315 | 1.34 (0.57–3.15) | 0.503 |
| Current Smoker | 64 | 17 | 2.24 (1.12–4.48) | 0.022 | 1.58 (0.75–3.33) | 0.228 |
| Drinking Status | ||||||
| Never Drinker | 64 | 10 | 1.00 (Reference) | 1.00 (Reference) | ||
| Former Drinker | 34 | 7 | 1.41 (0.59–3.37) | 0.441 | 1.19 (0.48–2.95) | 0.708 |
| Current Drinker | 72 | 18 | 2.12 (1.08–4.16) | 0.029 | 1.45 (0.70–3.00) | 0.318 |
| CCI | ||||||
| < 2 | 127 | 22 | 1.00 (Reference) | 1.00 (Reference) | ||
| ≥ 2 | 43 | 13 | 1.84 (1.01–3.35) | 0.046 | 1.62 (0.86–3.05) | 0.135 |
| BMI | ||||||
| < 18.5 | 38 | 12 | 1.72 (0.91–3.25) | 0.095 | 1.48 (0.76–2.88) | 0.248 |
| 18.5–24 | 113 | 17 | 1.00 (Reference) | 1.00 (Reference) | ||
| > 24 | 19 | 6 | 1.38 (0.60–3.17) | 0.448 | 1.25 (0.53–2.95) | 0.612 |
| Primary Tumor Site | ||||||
| Oral Tongue | 94 | 16 | 1.00 (Reference) | 1.00 (Reference) | ||
| Floor of Mouth | 25 | 6 | 1.51 (0.65–3.51) | 0.338 | 1.28 (0.53–3.09) | 0.584 |
| Buccal Mucosa | 25 | 5 | 1.19 (0.47–3.01) | 0.712 | 1.05 (0.40–2.75) | 0.921 |
| Gingiva | 15 | 4 | 1.67 (0.62–4.50) | 0.309 | 1.41 (0.51–3.90) | 0.507 |
| Other | 11 | 4 | 2.58 (1.02–6.53) | 0.045 | 1.89 (0.71–5.03) | 0.202 |
| Clinical T-stage | ||||||
| cT1-2 | 83 | 12 | 1.00 (Reference) | 1.00 (Reference) | ||
| cT3-4 | 87 | 23 | 2.31 (1.22–4.37) | 0.010 | 1.52 (0.76–3.04) | 0.238 |
| Clinical N-stage | ||||||
| cN0 | 67 | 8 | 1.00 (Reference) | 1.00 (Reference) | ||
| cN+ | 103 | 27 | 3.12 (1.56–6.24) | 0.001 | 1.68 (0.78–3.62) | 0.184 |
| Tumor Grade | ||||||
| Well differentiated | 42 | 6 | 1.00 (Reference) | 1.00 (Reference) | ||
| Moderately differentiated | 95 | 16 | 1.28 (0.56–2.92) | 0.558 | 1.18 (0.51–2.73) | 0.701 |
| Poorly differentiated | 33 | 13 | 3.24 (1.38–7.61) | 0.007 | 2.38 (1.00–5.66) | 0.0496 |
| Extranodal Extension (ENE) | ||||||
| Absent | 130 | 18 | 1.00 (Reference) | 1.00 (Reference) | ||
| Present | 40 | 17 | 3.62 (2.07–6.33) | < 0.001 | 2.44 (1.28–4.65) | 0.007 |
| Surgical Margin Status | ||||||
| Clear (≥ 5 mm) | 143 | 24 | 1.00 (Reference) | 1.00 (Reference) | ||
| Close (< 5 mm) | 22 | 6 | 1.71 (0.78–3.75) | 0.180 | 1.28 (0.56–2.92) | 0.559 |
| Positive | 5 | 5 | 9.16 (4.15–20.22) | < 0.001 | 5.03 (1.98–12.78) | 0.001 |
| Adjuvant Radiotherapy | ||||||
| No | 76 | 16 | 1.00 (Reference) | 1.00 (Reference) | ||
| Yes | 94 | 19 | 0.96 (0.54–1.71) | 0.894 | 0.76 (0.41–1.41) | 0.386 |
Data are presented as hazard ratio (HR) with 95% confidence interval. Age is presented as both a categorical variable (< 60 vs. ≥60 years) and a continuous measure. Bold values indicate statistical significance at p < 0.05
Abbreviations: CCI Charlson Comorbidity Index, BMI Body mass index, NICT Neoadjuvant immunochemotherapy, ENE Extranodal extension
Subgroup analyses
In total, 25 patients (29.4%) in the NICT group were pN+, compared to 54 patients (63.5%) in the US group. Among these pN+ patients, LRR was observed in 4 patients (16.0%) in the NICT group and 20 patients (37.0%) in the US group. The NICT group demonstrated a significantly lower CIF of LRR compared to the US group among this high-risk subset (Gray’s test, unadjusted p = 0.041; FDR-adjusted q = 0.082).
Patients (n = 122) who achieved a radiologic response (CR/PR) to NICT demonstrated a 3-year cumulative incidence of LRR of 9.1%. In contrast, patients (n = 13) with SD had a significantly higher 3-year cumulative incidence of LRR of 33.3% (Gray’s test, unadjusted p = 0.022; FDR-adjusted q = 0.066). Patients achieving a pathologic complete response at the primary site suffered from no LRR.
Outcomes by specific anti-PD-1 agent and by number of NICT cycles showed consistent LRR rates across all subgroups (sintilimab: 11.5%, toripalimab: 12.2%, camrelizumab: 12.5%, Gray’s test, p = 0.735; 2 cycles: 12.5%, 3 cycles: 11.4%, 4 cycles: 12.0%; Gray’s test, p = 0.666; all FDR-adjusted q > 0.20) (Fig. 3).
Fig. 3.
Forest plot of treatment effect on loco-regional recurrence across subgroups. Data are presented as subdistribution hazard ratios (SHR) with 95% confidence intervals. The vertical dashed line represents an SHR of 1.0 (no effect). Values to the left of the line favor NICT, and values to the right favor US. Abbreviations: NICT, neoadjuvant immunochemotherapy; US, upfront surgery; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease
Sensitivity analysis
In analysis restricted to the immunotherapy era (2020–2024), the beneficial effect of NICT on LRR remained significant (NICT group vs. US group: cumulative incidence 10.9% vs. 28.4%; Gray’s test, p = 0.011).
In the 3-month landmark analysis restricted to patients alive and disease-free at that time point, the NICT benefit persisted for late recurrences (3-year CIF: 10.2% vs. 28.8%; Gray’s test, p = 0.008; adjusted SHR = 0.38; 95% CI: 0.19–0.76; p = 0.006).
First, we re-analyzed LRR using a standard cause-specific Cox proportional hazards model. In this model, NICT remained significantly associated with less LRR (cause-specific HR = 0.41; 95% CI: 0.21–0.80; p = 0.009). Second, we conducted an analysis treating the composite endpoint of LRR or death as disease-free survival. The NICT group demonstrated a superior 3-year disease-free survival rate compared to the US group (81.2% vs. 65.9%; p = 0.022).
Discussion
The contemporary management of OSCC requires the integration of advanced reconstructive techniques for functional preservation with NICT to optimize oncologic outcomes. The efficacy of this systemic approach is corroborated by the phase 3 KEYNOTE-689 trial [10], which established that incorporating neoadjuvant and adjuvant pembrolizumab into standard care regimens significantly prolongs event-free survival in patients with locally advanced head and neck squamous cell carcinoma. Furthermore, the trial demonstrated that the neoadjuvant administration of pembrolizumab neither impeded surgical completion rates nor introduced novel safety signals. Within the context of these evolving systemic therapies, the utilization of the SIF has remained controversial due to persistent concerns regarding its oncologic safety within the level I lymph node basin [14]. The present PSM analysis of 170 patients indicates that SIF reconstruction is both oncologically sound and clinically viable in the NICT era. Our findings suggest that the administration of NICT prior to SIF reconstruction is associated with significantly enhanced survival, without a statistically significant increase in flap-related complications or an observed increase in level I recurrences.
The most persistent concern regarding the SIF has been the potential for oncologic compromise. Critics have argued that harvesting a flap from level I could interfere with the completeness of the neck dissection or, worse, inadvertently transplant occult metastatic disease into the oral defect [15, 16]. In our PSM cohort, we observed no recurrences within level I or along the subcutaneous tract of the flap in either the NICT or US group. While this finding is reassuring and suggests the risk is low when using meticulous surgical technique, it must be tempered by the study’s sample size, which limits the power to detect very rare events. The absence of observed tract seeding does not equate to the absence of risk, particularly in a larger population. But our data align with the largest and most definitive studies on this topic. Thomas et al. [15], in their analysis of 229 cases, concluded that the SIF is oncologically safe, with survival outcomes comparable to standard care, provided it is avoided in patients with obvious nodal disease. Similarly, a large retrospective study by Wang et al. [17] found no significant difference in locoregional recurrence rates between pN0 and pN+ patients undergoing SIF reconstruction, concluding that the flap itself does not increase recurrence risk. Our study builds upon this foundation by extending these conclusions into the NICT setting, where tumor downstaging could theoretically alter surgical planes and nodal characteristics. While this finding is reassuring, it should be interpreted with appropriate caution given the sample size and follow-up duration; the absence of observed events does not definitively exclude a low but non-zero risk of level I recurrence or tract seeding.
The introduction of NICT adds another layer of complexity, as it could theoretically obscure the assessment of level I nodes or create a hostile inflammatory environment that might complicate healing [18]. However, our data strongly suggest that NICT is a complementary, not complicating, factor. In fact, the NICT group in our study demonstrated significantly better survival after matching. Multivariable analysis identified NICT as a factor independently associated with less LRR and improved OS. This benefit was particularly pronounced in high-risk subgroups, such as patients with pathologically confirmed nodal disease (pN+). This finding is consistent with the growing body of evidence supporting the efficacy of NICT in head and neck cancer. A recent meta-analysis by Pedroso et al. [19] highlighted the potential of combination immunotherapy to achieve high partial response rates and reduce disease progression, although they noted variability in surgical complication rates. Our study provides observational data suggesting that this improved oncologic response, as measured by radiographic and pathologic parameters, is also associated with superior post-surgical outcomes when combined with a reliable reconstructive option like the SIF.
Our findings stand in contrast to the more cautious recommendations from some authors. Miao et al. [20], in a study focusing specifically on oral tongue cancer, recommended against using the SIF in patients with pN+ disease at level I, citing an increased risk of “SIF-related locoregional recurrence.” This apparent discrepancy likely highlights the critical importance of surgical technique. As demonstrated by the anatomical study by Philip et al. [16], a skeletonized, perforator-based SIF can be raised that is largely devoid of lymph nodes (73% in their series). This suggests that the “traditional” SIF, which may inadvertently include more nodal tissue, carries a higher oncologic risk than the meticulously dissected, skeletonized flap used in our series and advocated by others [21]. Our technique, which prioritizes complete level I nodal clearance as a separate step before flap harvest, may directly address the concerns raised by Miao et al. [20] and is consistent with the findings of Sittitrai et al. [22], who concluded that the SIF can be used safely even in selected patients with level I metastases when harvested meticulously. The very low rate of positive level I nodes in our NICT group after treatment is consistent with the potent downstaging effect of immunotherapy, further associated with a mitigation of the theoretical risk of nodal transfer.
A second major concern regarding the combination of NICT and major surgery is the potential for increased perioperative morbidity. The theoretical risk is that immunotherapy-induced inflammation and fibrosis could impair wound healing or increase flap complications. Early case series, such as that by Hwang et al. [23], reported severe wound complications like osteoradionecrosis in patients on anti-PD-1 therapy, although these were largely in the context of prior radiation. A more recent comprehensive review by Kuo et al. [24] concluded that while most patients on immunotherapy heal without issue, neoadjuvant use before complex surgeries may warrant careful patient selection. Our study provides reassuring, high-quality data on this front. In our matched cohort, flap complication rates were low and comparable between the NICT and US groups, with no total flap losses in either arm. There were no significant differences in rates of partial necrosis, hematoma, infection, or marginal mandibular nerve weakness. This finding aligns with the large retrospective study by Philips et al. [25], which found no increase in surgical complications after adjusting for confounders in patients receiving neoadjuvant immunochemotherapy. It also echoes the findings from studies in other surgical oncology disciplines, such as cytoreductive nephrectomy after immunotherapy, where surgery was found to be safe and feasible [24]. The key to these good outcomes, we believe, lies in patient selection and surgical timing. In our cohort, surgery was performed a median of 28 days after the last NICT dose, allowing any acute, high-grade inflammatory response to subside before the surgical insult. This pragmatic approach, as suggested by Kuo et al. [24], appears to mitigate the risk of wound healing problems while preserving the oncologic benefit of the NICT. Our data, therefore, suggest that NICT administration was not associated with a statistically significant increase in surgical risk, and its known downstaging effects were observed without compromising flap outcomes when surgery was performed by experienced teams using meticulous technique.
Perhaps the most exciting implication of our study is the validation of a “response-adapted” surgical strategy. The significant reduction in the need for mandibulectomy and free flap reconstruction in the NICT group suggests that effective neoadjuvant treatment can successfully downstage tumors, allowing for less morbid, function-preserving surgery. This concept is at the heart of modern de-escalation strategies. Our subgroup analysis, which showed that patients achieving a radiologic response to NICT had a 3-year LRR rate of only 9.1% compared to 33.3% in non-responders, strongly supports tailoring the surgical approach to the treatment response. This finding is corroborated by a recent and highly relevant study by Fang et al. [26], who specifically evaluated response-adapted surgery (RAS) after NICT in OSCC. They demonstrated that RAS achieved equivalent oncologic outcomes to traditional surgery while significantly reducing the need for mandibulectomy and free flap reconstruction, leading to better quality of life. Our results are perfectly aligned with this emerging paradigm. By achieving high response rates (90.4% objective response in our NICT group), immunotherapy enables surgeons to perform a more precise, less extensive resection, thereby maximizing the benefits of a versatile, pedicled flap like the SIF and minimizing functional deficits. This represents a true shift from “maximal” to “optimal” surgery, guided by the tumor’s biology and response to therapy. However, this optimistic finding likely stems from a combination of factors unique to our study design. First, strict patient selection inherently influenced our cohort; inclusion required patients to be candidates for pedicled SIF reconstruction, which may have inadvertently selected for tumors of a specific size, location, or biological behavior that are more amenable to both downstaging and regional flap coverage compared to massive, deeply infiltrative tumors necessitating large free flaps. Second, as a high-volume tertiary referral center, institutional factors such as highly standardized multidisciplinary care protocols may have optimized patient compliance and treatment delivery.
We acknowledge several limitations in this study. First, its retrospective, non-randomized design carries a high risk of selection bias and residual confounding, despite our use of PSM. The decision to administer NICT was made by a multidisciplinary tumor board, raising the possibility that medically fitter patients or those with tumors anticipated to respond better to immunotherapy were preferentially selected. PSM cannot account for these unmeasured clinical judgments. Additionally, several important oncologic and demographic variables were unavailable for inclusion in our matching model; therefore, the cohorts may not be truly comparable despite balanced observed covariates. Second, temporal bias is a concern, as NICT was introduced in 2020 while the US cohort includes historical controls from 2018. Although our sensitivity analysis restricted to the 2020–2024 era showed a persistent benefit, era-related confounding cannot be fully excluded. Third, our subgroup analyses are limited by very small sample sizes and are underpowered. Despite FDR correction, the wide confidence intervals and post-hoc nature of these analyses render them exploratory and hypothesis-generating only. Fourth, the median follow-up of 38 months is insufficient to establish definitive oncologic safety for rare events such as flap tract or delayed level I recurrence. The absence of such events may reflect limited cohort size and follow-up rather than genuinely negligible risk, and should not be interpreted as conclusive proof of oncologic safety. Fifth, standardized functional outcome measures and quality-of-life assessments were not available, and pathological evaluations were performed without blinding, introducing potential interpretation bias. Pathological response was not systematically quantified using standard criteria such as major pathological response or residual viable tumor scoring. Sixth, our study lacks comparator groups reconstructed with free or other regional flaps, precluding conclusions about the relative superiority or equivalence of SIF reconstruction in the NICT era. Finally, as a single-center study, generalizability may be limited. Future multicenter, prospective trials with blinded pathological assessment, standardized response scoring, longer follow-up, and comparator arms are warranted to validate these findings.
In conclusion, this PSM study demonstrates that in patients with OSCC treated with NICT, SIF reconstruction was not associated with any observed level I or flap tract recurrences. Crucially, NICT significantly improved loco-regional recurrence and overall survival without increasing flap-related complications. Therefore, SIF could be integrated into modern NICT-based treatment algorithms. Given the observational, single-center design of this study, future prospective and multi-institutional studies are warranted to fully confirm these results.
Supplementary Information
Supplementary Material 2. Supplementary Fig. 1: Love plot demonstrating covariate balance before and after propensity score matching. After PSM, NICT and US groups showed acceptable covariate balance.
Acknowledgements
None declared.
Authors’ contributions
YD: Conceptualization, Methodology, Data curation, Formal analysis, Writing – original draft. QY: Conceptualization, Methodology, Supervision, Writing – review & editing, Project administration. Both authors read and approved the final manuscript.
Funding
None declared.
Data availability
The datasets analyzed during the current study are not publicly available due to institutional patient privacy regulations but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by The First Affiliated Hospital of Zhengzhou University Institutional Research Committee, and written informed consent for medical research was obtained from every patient prior to initial treatment. All methods were performed in accordance with the Declaration of Helsinki.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Celidonio J, Chinta S, de Armas JS, Roden D. Genetic Landscape of Oral Cavity Squamous Cell Carcinoma. OTO Open. 2026;10:e70194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dunn LA, Ho AL, Pfister DG. Head and Neck Cancer: A Review. JAMA. 2026;335:531–41. [DOI] [PubMed] [Google Scholar]
- 3.Becker P, Raguse JD, Rau A, Wermker K, Beck-Broichsitter B, Weber M, Bouffleur F, Zeller AN, Pabst A. De-escalation strategies in the treatment of oral squamous cell carcinoma: A cross-sectional study in oral and maxillofacial surgery in Germany, Austria, and Switzerland. J Craniomaxillofac Surg. 2026;54:104478. [DOI] [PubMed] [Google Scholar]
- 4.Mosquera C, Ramirez C. Narrative Review: Submental Artery Island Pedicled Flap, Indications, Tips, and Pitfalls. Craniomaxillofac Trauma Reconstr. 2024;17:244–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Aldrich AR, Dorneden AM, Boyd NH, Olson GT, Syme NP. Preoperative Imaging Review Predicts the Complex Venous Anatomy of the Submental Island Flap. Ann Otol Rhinol Laryngol. 2025;134:900–6. [DOI] [PubMed] [Google Scholar]
- 6.Ruenmarkkaew D, Sangiamporn O, Klibngern H, Ariyanon T, Watcharatsiriyuth W, Sittitrai P. Submental island flap reconstruction in oral cavity cancer patients with pathological level I lymph node metastasis: Flap site recurrence and survival analysis. J Stomatol Oral Maxillofac Surg. 2025;126:102309. [DOI] [PubMed] [Google Scholar]
- 7.Qin Q, Chu T, Zhao Y, Fu Y, Yang H, Liu Q. Facial artery-based flap for oral and maxillofacial defects: a narrative review. Clin Oral Investig. 2025;29:382. [DOI] [PubMed] [Google Scholar]
- 8.Zhou S, Qin Y, Lei A, Li Y, Yang P, Liu H, Sun Y, Zhang J, Deng C, Chen Y. Neoadjuvant and Adjuvant Immunotherapy in the Treatment of Oral Squamous Cell Carcinoma. J Biochem Mol Toxicol. 2025;39:e70199. [DOI] [PubMed] [Google Scholar]
- 9.Liu C, Li M, Liu X, Shi T, Wang Y, Sui C, Zhang W, Wang B. Evaluating the efficacy and safety of different neoadjuvant immunotherapy combinations in locally advanced HNSCC: a systematic review and meta-analysis. Front Immunol. 2024;15:1467306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Uppaluri R, Haddad RI, Tao Y, Le Tourneau C, Lee NY, Westra W, Chernock R, Tahara M, Harrington KJ, Klochikhin AL, Braña I, Vasconcelos Alves G, Hughes BGM, Oliva M, Pinto Figueiredo Lima I, Ueda T, Rutkowski T, Schroeder U, Mauz PS, Fuereder T, Laban S, Oridate N, Popovtzer A, Mach N, Korobko Y, Costa DA, Hooda-Nehra A, Rodriguez CP, Bell RB, Manschot C, Benjamin K, Gumuscu B, Adkins D. KEYNOTE-689 Investigators. Neoadjuvant and Adjuvant Pembrolizumab in Locally Advanced Head and Neck Cancer. N Engl J Med. 2025;393:37–50. [DOI] [PubMed] [Google Scholar]
- 11.Mays AC, Yarlagadda B, Achim V, Jackson R, Pipkorn P, Huang AT, Rajasekaran K, Sridharan S, Rosko AJ, Orosco RK, Coughlin AM, Wax MK, Shnayder Y, Spanos WC, Farwell DG, McDaniel LS, Hanasono MM. American Head and Neck Society Reconstructive Microsurgery Committee. Examining the relationship of immunotherapy and wound complications following flap reconstruction in patients with head and neck cancer. Head Neck. 2021;43:1509–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Seymour L, Bogaerts J, Perrone A, Ford R, Schwartz LH, Mandrekar S, Lin NU, Litière S, Dancey J, Chen A, Hodi FS, Therasse P, Hoekstra OS, Shankar LK, Wolchok JD, Ballinger M, Caramella C, de Vries EGE. RECIST working group. iRECIST: guidelines for response criteria for use in trials testing immunotherapeutics. Lancet Oncol. 2017;18:e143–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ridout R, Ahmad R, Rai P, McMahon JD. Predictors of wound complication in patients undergoing major head and neck cancer surgery who require free tissue transfer. Br J Oral Maxillofac Surg. 2025;63:208–13. [DOI] [PubMed] [Google Scholar]
- 14.Taseen S, Qureshi MFH, Abbas M, Ghaloo SK, Wasif M, Iftikhar H, Pasha HA, Dhanani R. Oncological Safety of the Submental Artery Island Flap for Oral Reconstruction: A Systematic Review and Meta-Analysis. Head Neck. 2026. 10.1002/hed.70252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Thomas S, Varghese BT, Ganesh SA, Desai KP, Iype EM, Balagopal PG, Sebastian P. Oncological Safety of Submental Artery Island Flap in Oral Reconstruction - Analysis of 229 Cases. Indian J Surg Oncol. 2016;7:420–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Philip J, Mizen K, Singh M, Karzai L, Sanders K, Bazira PJ. Submental island flap - should we be worried about lymph node transfer? An anatomical and histological study. Br J Oral Maxillofac Surg. 2023;61:221–6. [DOI] [PubMed] [Google Scholar]
- 17.Wang J, Tan Y, Shen Y, Lv M, Li J, Sun J. Oncological safety of submental island flap for reconstruction of pathologically node-negative and node-positive T1-2 oral squamous cell carcinoma-related defects: A retrospective study and comparison of outcomes. Oral Oncol. 2020;102:104507. [DOI] [PubMed] [Google Scholar]
- 18.Zhang W, Lu P, Li X, Fang Q. Hypothyroidism following immunotherapy predicts more postoperative complication in oral squamous cell carcinoma. BMC Cancer. 2025;25:643. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pedroso CM, de Pauli Paglioni M, Normando AGC, Chaves ALF, Kowalski LP, de Castro Júnior G, Matos LL, Willian Junior WN, de Oliveira TB, de Marchi P, Harada G, Mak MP, Lima CSP, Viani GA, Moraes FY, Gouveia AG, Santos-Silva AR, Marta GN, Latin American Cooperative Oncology Group (LACOG). - Head and Neck and Brazilian Group of Head and Neck Cancer (GBCP). Preoperative neoadjuvant chemotherapy or immunotherapy in head and neck cancer: A systematic review and meta-analysis of surgical risk and pathologic response. Crit Rev Oncol Hematol. 2025;212:104742. [DOI] [PubMed] [Google Scholar]
- 20.Miao HJ, Sun SK, Tian YY, Yang YQ, Wang SH, Bai S, Chen W, Mao C, Liang SX, Yan YB. Oncologic safety of the pedicled submental island flap for reconstruction in oral tongue squamous cell carcinoma: An analysis of 101 cases. Oral Oncol. 2023;140:106395. [DOI] [PubMed] [Google Scholar]
- 21.Qayyum MU, Keerio AA, Zaheer R, Mushtaq U, Baig V, Hussain R, Faisal M. The Oncological Safety of Submental Flap Reconstruction in Head-and-Neck Cancers: An Extended Follow-up Study. J Cancer Allied Spec. 2024;10:573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sittitrai P, Srivanitchapoom C, Reunmakkaew D, Yata K. Submental island flap reconstruction in oral cavity cancer patients with level I lymph node metastasis. Br J Oral Maxillofac Surg. 2017;55:251–5. [DOI] [PubMed] [Google Scholar]
- 23.Hwang V, Mendez E, Chow LQM, Futran ND, Andersen P, Li R, Divi V, Rodriguez CP. Wound Complications in Head and Neck Squamous Cell Carcinomas After Anti-PD-1 Therapy. Laryngoscope. 2019;129:E428–33. [DOI] [PubMed] [Google Scholar]
- 24.Kuo PJ, Lin PC, Hsieh CH. Wound healing in cancer patients under immunotherapy. Int J Surg. 2025;111:7087–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Philips R, Alnemri A, Amin D, Patel J, Topf MC, Johnson JM, BarAd V, Axelrod R, Argiris A, Fundakowski C, Luginbuhl AJ, Cognetti DM, Curry JM. Effect of preoperative programmed death-1 or programmed death ligand-1 immune check point inhibition on complications after surgery for primary head and neck cancer. Cancer. 2024;130:863–75. [DOI] [PubMed] [Google Scholar]
- 26.Fang Q, Yuan J, Zhang X, Huang T, Guo L. Response-adapted surgery after neoadjuvant immunochemotherapy in oral squamous cell carcinoma. Oral Oncol. 2025;171:107769. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 2. Supplementary Fig. 1: Love plot demonstrating covariate balance before and after propensity score matching. After PSM, NICT and US groups showed acceptable covariate balance.
Data Availability Statement
The datasets analyzed during the current study are not publicly available due to institutional patient privacy regulations but are available from the corresponding author on reasonable request.