Vein-protection-first strategy revolutionizes submental artery perforator flap harvest success: the A-B-C technique and training assessment

Abstract

Background

Submental artery perforator flaps (SMAPF) represent valuable reconstructive options for head and neck defects, yet traditional artery-first dissection approaches inadequately protect complex submental vein (SMV) drainage variations, contributing to compromised flap outcomes. We systematically analyzed SMV drainage patterns to establish a standardized vein-protection-first technique.

Methods

We retrospectively analyzed venous drainage patterns in 408 successfully harvested SMAPF cases (December 2017-December 2022), identifying source and peripheral drainage systems. A clinical dissection-based A-B-C technique was developed: Point A (intersection of the facial nerve marginal mandibular branch and facial vein), Point B (confluence of SMV and facial vein), and Point C (confluence of facial vein with internal jugular vein/external jugular vein/anterior jugular vein). Following standardized training of twelve surgeons, we conducted a retrospective pre-post comparative study comparing the conventional surgery group (January 1, 2023, to January 1, 2024; n = 62) and the vein-optimized group (February 1, 2024, to February 1, 2025; n = 55). The primary endpoint was flap compromise.

Results

All 408 SMAPF cases carried source drainage systems; 29.2% carried peripheral drainage. A-B-C classification revealed A-B-single C (93.4%), A-B-double C (5.6%), and A-B-triple C (1.0%) configurations. Spatial analysis identified Type I inverted T-shaped (18.1%) and Type II Y-shaped (81.9%) patterns. The vein-optimized group demonstrated significantly reduced flap compromise rates (OR = 0.183, 95% CI:0.040–0.844, p = 0.029), representing 81.7% risk reduction. Harvest time decreased significantly (p = 0.008). Multivariable analysis confirmed A-B-C technique as an independent protective factor.

Conclusions

This novel A-B-C vein-optimized technique significantly reduces flap compromise and decreases operative duration. The standardized approach mitigates experience-dependent variability, providing a theoretical and practical framework for SMAPF clinical implementation and surgical training protocols.

Introduction

Head and neck cancers rank sixth among global malignancies [1–3]. Surgery remains the primary therapeutic modality for head and neck cancers [4, 5]. Moreover, reconstruction is essential for restoring form and function after tumor resection [6, 7]. Reconstruction typically employs two categories of flaps: free flaps [6, 8] and pedicled flaps [9–12]. Free flaps offer advantages of adequate tissue volume and large area, serving as the gold standard for reconstruction [13–16]. However, rising cancer incidence, population aging [17, 18], and emerging infectious diseases such as dengue fever and monkeypox [19–21] have strained healthcare resources significantly. There is a growing clinical need for safe, reliable, and cost-effective reconstruction options.

Pedicled flaps offer irreplaceable clinical advantages over free flaps: shorter operative time, lower costs and avoidance of a second surgical site. Additionally, pedicled flaps require no microsurgical anastomosis and are suitable for centers with limited microsurgical resources [22, 23]. Under precision medicine principles, surgeons should prioritize pedicled flaps for small-to-moderate defects, elderly patients with comorbidities, and patients with limited financial resources [16, 24]. The submental flap provides reliable vascularity, favorable tissue characteristics, and a donor site with acceptable cosmesis, with reported dimensions of up to 22 × 9 cm for head and neck reconstruction [9]. In 1993, Martin first systematically described the submental island flap harvest technique [25]. Due to the intimate relationship between the submental artery and the digastric muscle, the submental island flap typically incorporates the anterior belly of the digastric muscle and perivascular adipose tissue to ensure arterial supply to the flap [26, 27]. With the development and refinement of the perforator flap concept proposed by Koshima [28], Kim dissected the submental artery perforators without including the anterior belly of the digastric muscle and submental fat, leading to the development of the submental artery perforator flap (SMAPF) [29]. SMAPF enables precise vascular dissection to skeletonize the pedicle adequately, allowing thorough lymph node dissection and ensuring oncological safety [30, 31].

However, successful flap harvest depends heavily on surgeons’ dissection skills and experience. The submental artery has relatively constant anatomy. Traditional SMAPF harvest uses an “artery-first” dissection strategy [24, 27, 32, 33], dissecting the facial-submental artery trunk before identifying the venous drainage system. However, the complex embryological development of head and neck veins creates significant anatomical variations [34]. These variations are particularly pronounced in the facial vein (FV) system, particularly in submental vein (SMV) drainage patterns. While the SMV typically drains into the internal jugular vein (IJV), when the SMV drains to the anterior jugular vein (AJV) or external jugular vein (EJV), inadequate intraoperative identification and protection cause venous injury and compromise flap outcomes. Literature shows higher flap compromise rates in SMAPF than free flaps, with venous injury as a major cause of failure [22, 35–37]. These technical challenges and high failure risks limit widespread SMAPF adoption.

Previous anatomical studies investigated SMV drainage patterns. Rahpeyma and Lin identified that SMV also drains to the AJV in their cadaveric and clinical studies [27, 38–41], but the limited number of cases reviewed restricted systematic understanding of SMV drainage patterns. Therefore, identifying SMAPF venous drainage patterns and establishing a standardized vein-protection-based technique are crucial for successful flap harvest.

This study aims to: (1) summarize SMAPF venous drainage patterns through retrospective analysis (2) develop a vein-protection-first technique and establish a classification system (3) create a standardized SMAPF harvest protocol to guide clinical practice and education.

Methods

The study complied with the Declaration of Helsinki. It was reviewed and approved by the institutional review board of China Medical University (K2025015).

This study consisted of two phases. Phase 1 was a retrospective anatomical review (n = 408) to identify venous drainage patterns of the SMAPF. Phase 2 was a retrospective pre-post comparative study (n = 117) comparing conventional versus vein-optimized SMAPF harvesting after surgeon training.

Phase 1: retrospective analysis and protocol development

We retrospectively analyzed clinical data from 408 patients who underwent SMAPF reconstruction with flap survival between December 2017 and December 2022 at School and Hospital of Stomatology of China Medical University and First Affiliated Hospital of China Medical University. All patients had complete intraoperative documentation of the venous pedicle, and flap survival was confirmed at discharge. Flap venous drainage patterns were classified and served as the primary endpoint.

Based on venous drainage protection as the core strategy for safe flap harvest, we developed an approach utilizing centripetal venous dissection for vascular pedicle preparation. This approach first identified and protected venous drainage pathways prior to arterial dissection. Inspired by perforator localization techniques in deep artery perforator flaps [42], we termed this the vein-protection-first A-B-C technique. This technique involved dissecting three key anatomical points along venous drainage: Point A (intersection of the facial nerve marginal mandibular branch and FV), Point B (confluence of SMV and FV), and Point C (confluence of FV with IJV/EJV/AJV), while protecting arterial supply.

Phase 2: standardized flap harvesting training for surgeons

In January 2024, we conducted standardized training for 12 surgeons from the Department of Head and Neck Surgery (School and Hospital of Stomatology, China Medical University) and Department of Otolaryngology (First Affiliated Hospital, China Medical University). The training covered SMAPF venous drainage classification and the systematic A-B-C venous optimization technique. All participating surgeons met the following inclusion criteria: (1) qualified to perform SMAPF surgery independently; (2) no prior training in SMAPF venous drainage techniques; (3) completion of > 20 independent SMAPF procedures [43]; (4) voluntary participation with informed consent.

The training content included: (1) anatomical theory and A-B-C technique instruction; (2) dissection training for the identification of A-B-C anatomical points on perfused neck specimens. Following training, all surgeons passed standardized competency assessments including anatomical point identification and intraoperative technique application. Since February 2024, these surgeons have adopted the vein-protection-first technique for SMAPF harvesting.

Study subjects and clinical data collection

We retrospectively collected clinical data from patients who had planned SMAPF reconstruction performed by 12 surgeons from January 1, 2023 to February 1, 2025.

Patient inclusion criteria

1. Age ≥18 years;

2. Head and neck defect reconstruction using SMAPF;

3. Surgery performed by one of the 12 participating surgeons;

4. ASA physical status I-III.

Patient exclusion criteria

1. Previous neck surgery;

2. Psychiatric disorders affecting treatment compliance;

3. Incomplete medical data;

4. Alternative flap reconstruction required due to intraoperative extranodal extension.

Patients were divided into two groups based on surgical approach: (1) conventional surgery group (January 1, 2023, to January 1, 2024): surgeons used traditional harvest approaches, dissecting the submental artery along the facial artery while protecting adjacent drainage veins; (2) vein-optimized group (February 1, 2024, to February 1, 2025): surgeons used the A-B-C vein optimization technique for intraoperative venous dissection while protecting the arterial supply.

This study used flap compromise as the primary endpoint. Flap compromise assessment was performed by the operating surgeon and an independent senior surgeon not involved in the index procedure. In cases of disagreement, a third senior surgeon adjudicated the outcome. Flap compromise was defined as impaired flap perfusion identified by capillary refill testing and/or needle prick test that required medical intervention or surgical exploration [44, 45]. Secondary endpoints included flap harvesting time (minutes) and postoperative hospital stay (days). Factors potentially affecting flap compromise, harvest time, and postoperative hospital stay were analyzed, including: (1) Patient factors: age, sex, tumor location, BMI (kg/m²), tobacco use, alcohol use, hypertension (HTN), diabetes mellitus (DM), and coronary artery disease (CAD); (2) Surgical factors: surgical approach, type of neck dissection, and area (cm²); (3) Surgeon factors: clinical experience (years).

Statistical analysis

No missing data were identified for the primary endpoint. All 117 patients in the retrospective pre-post comparative study had complete records for primary and secondary endpoints. Statistical analysis was performed using SPSS 26.0 software. Hierarchical Linear Model (HLM) was used to evaluate potential clustering effects among surgeons. Intraclass Correlation Coefficients (ICC) were calculated to quantify clustering degree. ICC values were 0.05 for flap compromise and 0.015 for flap harvest time (both < 0.1), indicating weak clustering effects among surgeons (Tables 1 and 2). The following statistical methods were used for data analysis:

Table 1.

Random effects covariance for surgeon effects on flap compromise

Parameter	SD	Variance	Standard Error	z	p	95% CI
Intercept	0.070	0.005	0.021	0.233	0.815	-0.036 ~ 0.046
Residual	0.308	0.094

Intraclass Correlation Coefficient (ICC) = 0.050

Table 2.

Random effects covariance for surgeon effects on flap harvest time 

Parameter	SD	Variance	Standard Error	z	p	95% CI
Intercept	1.373	1.884	0.584	3.227	0.001	0.740 ~ 3.029
Residual	11.160	123.472

Intraclass Correlation Coefficient (ICC) = 0.015

Descriptive analysis

Normally distributed continuous variables are presented as mean ± standard deviation (SD), while non-normally distributed variables are presented as median (interquartile range). Since SMAPF approximates an elliptical shape, flap area was calculated using the elliptical formula (S=π×A×B/4, with π≈3.14).

Univariate analysis

To identify factors affecting flap outcomes, we used t-tests or Mann-Whitney U tests for continuous variables and Pearson chi-square tests or Fisher's exact tests for categorical variables.

Multivariable analysis

Multivariable logistic regression (stepwise backward elimination) was performed to analyze flap compromise, incorporating variables with p < 0.10 from univariate analysis. Multiple linear regression was used to analyze flap harvest time and postoperative hospital stay. Statistical significance was set at p < 0.05. Results are presented as odds ratios (OR) with 95% confidence intervals (CI).

To adjust for multiple comparisons, we performed a post-hoc Bonferroni correction for the two secondary endpoints (flap harvesting time and postoperative hospital stay), setting statistical significance at p < 0.025. The primary endpoint (flap compromise) was tested at p < 0.05. Original p-values were retained.

Results

SMAPF venous drainage anatomical review

Based on the clinical anatomical review of 408 SMAPF cases, we identified two distinct venous drainage systems in the submental region (Fig. 1): (1) source venous drainage system primarily consisting of the SMV; (2) peripheral venous drainage system primarily consisting of the AJV.

Fig. 1.

The diagram depicts two distinct venous drainage systems of the same SMAPF: A source venous drainage system; B peripheral venous drainage system includes both ipsilateral and contralateral AJVs

Source venous drainage system

All 408 SMAPF cases carried source venous drainage systems (100%), with 289 cases (70.8%) using only the source venous drainage system. SMV drainage pathways showed significant variations and were classified into 7 types based on final confluence sites (Fig. 2). Given this anatomical complexity, an overly detailed classification would offer limited practical surgical utility. Therefore, based on our vein-protection-first A-B-C technique, we established a simplified A-B-C classification system following the clinical anatomical sequence (Fig. 3 and Fig. 4). Points A, B, and C were present in all 408 SMAPF cases (100%). The source venous drainage system classification was simplified as follows: A-B-single C (381/408, 93.4%): SMV drains to one of the IJV, EJV, or AJV; A-B-double C (23/408, 5.6%): SMV drains to two of these veins; A-B-triple C (4/408, 1.0%): SMV drains to all three veins. Accurate identification of double-C and triple-C patterns is essential, as failure to identify all drainage sites may compromise venous outflow. Regarding anatomical localization, Points A and B maintained relatively constant anatomical positions with a mean distance of 2.3 cm. Point C exhibited notable anatomical variation, most commonly located within a 6.5 cm radius centered on Point B.

Fig. 2.

The diagrams illustrate seven types of drainage pathways in the source venous drainage system: A SMV drainage to IJV (206/408, 50.49%); B SMV drainage to EJV (172/408, 42.16%); C SMV drainage to AJV (3/408, 0.73%); D SMV drainage to EJV and IJV (8/408, 1.96%); E SMV drainage to IJV and AJV (7/408, 1.72%); F SMV drainage to EJV and AJV (8/408, 1.96%); G SMV drainage to IJV, EJV, and AJV (4/408, 0.98%)

Fig. 3.

The diagrams illustrate the A-B-C classification system in the source venous drainage system: A-B-single C: A, B, and C; A-B-double C: D, E, and F; A-B-triple C: G

Fig. 4.

The diagrams illustrate the anatomical points A, B, and C. Point A represents the intersection of marginal mandibular branch and FV; Point B represents the confluence of SMV with FV; Point C represents the confluence of FV with IJV, EJV, or AJV

To enhance spatial recognition of SMAPF for young surgeons, we classified SMV course patterns into two basic spatial configurations (Fig. 5). Type I inverted T-shaped configuration (74/408, 18.1%): The SMV runs along the submandibular gland surface, joins the anterior facial vein, then courses posteriorly along the mandibular border to drain into the external jugular vein, forming an inverted T-shaped pattern. Type II Y-shaped configuration (334/408, 81.9%): The SMV joins the anterior facial vein on the submandibular gland surface, then courses obliquely posteroinferiorly through or along the gland surface, joining the facial vein before draining into the IJV, EJV, or AJV.

Fig. 5.

The diagrams depict the source venous drainage system spatial configuration types. A and C Type I: T-shaped configuration; B and D Type II: Y-shaped configuration

Peripheral venous drainage system

Among 408 SMAPF cases, 119 cases (29.2%) carried peripheral venous drainage. AJVs were located within the superficial fascia of the anterior neck region, commonly existed in pairs, and descended along both sides of the midline. Notable anatomical variations were observed, including unilateral absence of the AJV in 7 cases (1.7%) and ipsilateral AJV origin at or below the hyoid level in 9 cases (2.2%), which did not directly drain the submental region.

Inspired by dual venous anastomosis systems in free flaps [46] and circuit redundancy principles, backup venous pathways maintain drainage when primary routes become obstructed. We routinely preserved the ipsilateral anterior jugular vein as a “safeguard vein” during flap harvesting. If AJV traction limited transfer, vascular clamping tests were used to assess flap circulation and determine AJV preservation (Fig. 6). When inadvertent SMV injury caused drainage obstruction, the AJV provided a crucial “salvage” function through hybrid SMV-AJV anastomosis, thereby rescuing flap drainage. For inadequate ipsilateral drainage in large flaps or when the ipsilateral AJV is unavailable, surgeons could utilize the contralateral AJV (Fig. 7).

Fig. 6.

The diagram illustrates the use of vascular clamping tests to assess flap color changes to determine whether to preserve the AJV. After clamping the anterior jugular vein, flap color changes indicate that the AJV should be preserved

Fig. 7.

The diagrams illustrate the peripheral venous drainage system. A The AJV ipsilateral to the tumor. B The AJV contralateral to the tumor

Head and neck surgeon training: pre- and post-training case comparison

Twelve surgeons participated in the training program: 9 men and 3 women, including 3 attending surgeons, 4 associate chief surgeons, and 5 chief surgeons (mean age 40.9 ± 7.1 years). Six surgeons (50%) had ≥ 20 years of clinical experience, and six (50%) had < 20 years (mean experience 18.5 ± 7.5 years). We included 117 patients with a mean age of 59.1 ± 9.6 years. Among surgeons with ≥ 20 years of clinical experience, the flap compromise rate was 10.0% (3/30) with the conventional approach and 3.6% (1/28) with the vein-optimized approach after training. Among surgeons with < 20 years of clinical experience, the flap compromise rate was 25.0% (8/32) with the conventional approach and 7.4% (2/27) with the vein-optimized approach after training.

Univariate analysis results

Pearson chi-square tests or Fisher’s exact tests demonstrated significant associations between flap compromise and surgical approach (p = 0.041), hypertension (p = 0.002), and coronary artery disease (p = 0.054) (Table 3). Mann-Whitney U tests demonstrated significant associations between flap compromise and patient age (p = 0.015) and surgeon clinical experience (p = 0.022) (Table 4). An independent samples t-test showed no significant difference in flap harvest time between flap compromise and non-compromise patients (p > 0.05) (Table 5).

Table 3.

Pearson chi-square test / Fisher’s exact test for flap compromise

Variable	Category	Flap Compromise		Total	χ2	p
		No(%)	Yes(%)
Sex	Female	35(33.98)	4(28.57)	39(33.33)	—	0.772†
	Male	68(66.02)	10(71.43)	78(66.67)
Total		103	14	117
Tumor location	Base of tongue	4(3.88)	0(0.00)	4(3.42)	—	0.461†
	Buccal mucosa	17(16.50)	1(7.14)	18(15.38)
	Buccal mucosa and Mandibular gingiva	1(0.97)	0(0.00)	1(0.85)
	Facial	1(0.97)	0(0.00)	1(0.85)
	Floor of mouth	15(14.56)	1(7.14)	16(13.68)
	Floor of mouth and Tongue	2(1.94)	0(0.00)	2(1.71)
	Mandible	8(7.77)	0(0.00)	8(6.84)
	Mandibular gingiva	12(11.65)	2(14.29)	14(11.97)
	Maxilla	2(1.94)	0(0.00)	2(1.71)
	Maxillary gingiva	3(2.91)	0(0.00)	3(2.56)
	Oropharynx	13(12.62)	4(28.57)	17(14.53)
	Parotid gland	1(0.97)	0(0.00)	1(0.85)
	Soft palate	1(0.97)	1(7.14)	2(1.71)
	Sublingual region	0(0.00)	1(7.14)	1(0.85)
	Tongue	22(21.36)	4(28.57)	26(22.22)
	Tongue and Floor of mouth	1(0.97)	0(0.00)	1(0.85)
Total		103	14	117
Surgical approach	Conventional	51(49.51)	11(78.57)	62(52.99)	4.177	0.041*
	Vein-optimized	52(50.49)	3(21.43)	55(47.01)
Total		103	14	117
Tobacco use	Never	57(55.34)	5(35.71)	62(52.99)	1.906	0.167
	Current or former	46(44.66)	9(64.29)	55(47.01)
Total		103	14	117
Alcohol use	Never	66(64.08)	7(50.00)	73(62.39)	1.041	0.308
	Current or former	37(35.92)	7(50.00)	44(37.61)
Total		103	14	117
HTN	No	74(71.84)	4(28.57)	78(66.67)	—	0.002*†
	Yes	29(28.16)	10(71.43)	39(33.33)
Total		103	14	117
DM	No	92(89.32)	14(100.00)	106(90.60)	—	0.356†
	Yes	11(10.68)	0(0.00)	11(9.40)
Total		103	14	117
CAD	No	98(95.15)	11(78.57)	109(93.16)	—	0.054*†
	Yes	5(4.85)	3(21.43)	8(6.84)
Total		103	14	117
Type of neck dissection	Bilateral MRND	12(11.65)	1(7.14)	13(11.11)	—	0.860†
	MRND	50(48.54)	10(71.43)	60(51.28)
	MRND SND	1(0.97)	0(0.00)	1(0.85)
	None	1(0.97)	0(0.00)	1(0.85)
	RND	11(10.68)	1(7.14)	12(10.26)
	RND MRND	4(3.88)	0(0.00)	4(3.42)
	SND	24(23.30)	2(14.29)	26(22.22)
Total		103	14	117

* p < 0.10, † Fisher’s exact test

Table 4.

Mann-Whitney U test for flap compromise

	Flap compromise Median M(P25, P75)		Mann-Whitney Test U Value	Mann-Whitney Test z Value	p
	No compromise (n = 103)	Compromise (n = 14)
Age	60.000(52.0,66.0)	68.000(59.3,71.3)	432.000	-2.429	0.015*
BMI (kg/m²)	25.500(23.7,27.2)	26.550(24.1,28.1)	554.000	-1.403	0.161
Area (cm²)	33.380(25.1,42.4)	31.420(21.2,49.1)	664.500	-0.475	0.634
Clinical experience (years)	22.000(14.0,24.0)	11.000(8.0,22.5)	451.500	-2.291	0.022*

* p < 0.10

Table 5.

t-test for flap compromise

	Flap compromise(Mean ± SD)		t	p
	No Compromise (n = 103)	Compromise (n = 14)
Flap harvesting time (min)	136.94 ± 11.81	139.43 ± 10.49	-0.748	0.456

* p < 0.10

Multivariable analysis results

Logistic regression analysis of flap compromise

Multivariable logistic regression analysis using stepwise selection was performed to identify independent predictors of flap compromise. Variables with p < 0.10 in univariate analysis were entered into the model. The final model retained four significant predictors: surgeon clinical experience ≥ 20 years (p = 0.043), age ≥ 70 years (p = 0.033), surgical approach (p = 0.029), and hypertension (p = 0.006) (Table 6). Model adequacy was confirmed by the Hosmer-Lemeshow goodness-of-fit test (χ² = 1.869, p = 0.931) and Nagelkerke R² = 0.361. The A-B-C vein optimization technique was significantly associated with reduced flap compromise risk. After controlling for confounding factors including surgeon experience, patient age, and hypertension, patients who underwent surgery with the A-B-C vein optimization technique had significantly lower odds of developing flap compromise compared with conventional techniques (OR = 0.183, 95% CI: 0.040–0.844, p = 0.029), indicating an 81.7% reduction in the odds of flap compromise. This protective effect remained significant in both univariate and multivariable analyses, confirming that the A-B-C technique serves as an independent protective factor against flap compromise.

Table 6.

Multivariable logistic regression analysis for flap compromise

Variable	Regression Coefficient (β)	Standard Error	z	Wald χ2	p	OR	OR 95% CI
Age group_≥70	1.586	0.744	2.132	4.544	0.033	4.884	1.136 ~ 20.991
Clinical experience group_≥20	-1.530	0.757	-2.020	4.082	0.043	0.216	0.049 ~ 0.955
Surgical approach	-1.700	0.781	-2.177	4.741	0.029	0.183	0.040 ~ 0.844
HTN	2.031	0.740	2.746	7.540	0.006	7.622	1.788 ~ 32.480
Intercept	-2.194	0.586	-3.747	14.037	0.000	0.111	0.035 ~ 0.351

Dependent variable = Flap compromise

Nagelkerke R ² = 0.361

McFadden R ² = 0.284

Cox & Snell R ² = 0.188

Multiple linear regression analysis of flap harvesting time

We conducted multiple linear regression analysis with surgical approach, flap area, surgeon clinical experience, neck dissection type, BMI, tumor location, and age as independent variables, and flap harvesting time as the dependent variable. The regression analysis showed statistical significance (F = 3.304, p = 0.003, R² = 0.175). Model diagnostics confirmed no multicollinearity (all VIF < 5) and no autocorrelation (D-W statistic ≈ 2). Among all included factors, flap area (β = 0.280, p = 0.001) and surgical approach (β = -5.748, p = 0.008) emerged as significant predictors (Table 7). The results demonstrated that the A-B-C vein optimization technique significantly reduces flap harvesting time and serves as an independent protective factor compared to conventional techniques.

Table 7.

Linear regression analysis for flap harvesting time (n=117)

	Unstandardized Coefficients		Standardized Coefficients	t	p	Collinearity Diagnostics
	B	Standard Error	Beta			VIF	Tolerance
Constant	150.549	13.605	-	11.066	0.000**	-	-
Surgical approach	-5.748	2.122	-0.247	-2.708	0.008**	1.103	0.907
Age	-0.049	0.108	-0.040	-0.451	0.653	1.060	0.943
BMI (kg/m²)	-0.511	0.492	-0.093	-1.038	0.302	1.070	0.934
Tumor location	-0.014	0.220	-0.006	-0.065	0.948	1.060	0.943
Type of neck dissection	-0.524	0.472	-0.101	-1.111	0.269	1.099	0.910
Area (cm²)	0.280	0.084	0.301	3.347	0.001**	1.068	0.937
Clinical experience (years)	-0.135	0.146	-0.082	-0.928	0.356	1.042	0.960
R²	0.175
Adjusted R²	0.122
F	F (7,109) = 3.304, p = 0.003
D-W	2.140

Dependent variable = Flap harvesting time (min)

* p < 0.05 ** p < 0.01

Multiple linear regression analysis of postoperative hospital stay

We conducted multiple linear regression modeling to investigate factors affecting postoperative hospital stay, incorporating 13 potential factors as independent variables: surgical approach, flap area, neck dissection type, BMI, tumor location, age, sex, tobacco use, alcohol use, hypertension, diabetes mellitus, coronary artery disease, and flap compromise. The model failed the F-test (F = 0.754, p = 0.706 > 0.05), indicating the overall model lacked statistical significance (Table 8). None of the 13 included variables significantly influenced postoperative hospital stay.

Table 8.

Linear regression analysis for postoperative hospital stay (n=117)

	Unstandardized Coefficients		Standardized Coefficients	t	p	Collinearity Diagnostics
	B	Standard Error	Beta			VIF	Tolerance
Constant	11.424	2.035	-	5.614	0.000**	-	-
Age	-0.030	0.016	-0.190	-1.836	0.069	1.209	0.827
Surgical approach	-0.323	0.326	-0.107	-0.991	0.324	1.317	0.759
Sex	0.412	0.399	0.129	1.034	0.303	1.761	0.568
Tumor location	0.008	0.032	0.025	0.255	0.799	1.110	0.901
Tobacco use	-0.284	0.373	-0.094	-0.763	0.447	1.724	0.580
Alcohol use	0.157	0.391	0.051	0.402	0.689	1.786	0.560
HTN	0.371	0.338	0.116	1.098	0.275	1.268	0.788
DM	-0.031	0.554	-0.006	-0.056	0.956	1.301	0.769
CAD	0.449	0.671	0.075	0.669	0.505	1.431	0.699
Type of neck dissection	0.024	0.068	0.036	0.355	0.723	1.148	0.871
BMI (kg/m²)	0.009	0.074	0.013	0.122	0.903	1.232	0.812
Flap compromise	0.511	0.519	0.110	0.985	0.327	1.412	0.708
Area (cm²)	0.007	0.012	0.062	0.627	0.532	1.090	0.917
R²	0.087
Adjusted R²	-0.028
F	F (13,103) = 0.754, p = 0.706
D-W	2.206

Dependent variable = Postoperative hospital stay(days)

* p < 0.05 ** p < 0.01

Discussion

SMAPF offers shorter operative time and reduced overall cost compared with free flaps, and is widely used for reconstructing oral, maxillofacial, oropharyngeal, and nasopharyngeal defects, with defect areas ranging from 12.57 to 155.51 cm² [47–49], though it is contraindicated for patients with a primary tumor infiltrating the mylohyoid muscle, or a level I cervical lymph node demonstrating extranodal extension or measuring > 1.5 cm in diameter [30, 50].

Traditional SMAPF harvest focuses on arterial anatomy while venous protection is often overlooked [33]. The complex embryological development of head and neck veins results in significant anatomical variation in the SMV drainage system, which is not adequately addressed by conventional artery-first approaches. Through systematic analysis of 408 SMAPF cases, we identified two major drainage systems: source venous and peripheral venous systems, and developed the “vein-protection-first” A-B-C technique accordingly. Inadequate venous identification during conventional harvest may account for the wide variation in flap failure rates reported across studies [35–37, 47, 51]. When the SMV drains to the EJV or AJV, surgeons may overlook anterior drainage pathways to the AJV and the direct posterior drainage to the EJV in inverted T-shaped configurations. When double C or triple C drainage patterns exist, protecting only one drainage vein while ligating others may compromise venous outflow.

Among 408 SMAPF cases, SMV drainage to the EJV occurred in 47.06% (192 cases). This rate significantly exceeds the 27.1% reported by Lin et al. [38]. SMV drainage to the AJV occurred in 5.39% (22 cases). The higher prevalence of EJV drainage in our cohort underscores the risk of venous injury when surgeons assume a standard IJV drainage pattern. These findings support the need for systematic venous identification and highlight the clinical rationale for Point C localization in the A-B-C technique.

The A-B-C technique follows the “vein-protection-first” concept. It simplifies complex venous anatomy into sequential identification of three key points: Point A along the marginal mandibular branch, Point B at the SMV-FV confluence approximately 2.3 cm distal to Point A, and Point C at the terminal drainage site. Point C is the most anatomically variable, commonly located within a 6.5 cm radius of Point B, and requires particular attention to identify double C or triple C configurations. This stepwise approach provides clear operative guidance and reduces the risk of inadvertent venous injury. Venous preservation is equally critical to arterial supply in flap design. Inadvertent injury to the primary venous drainage during harvest may result in irreversible flap compromise. The AJV serves as a “safeguard vein” providing dual protection: preventive protection when large flap designs exceed source venous drainage capacity, and rescue protection when inadvertent SMV injury occurs through hybrid anastomosis. When the ipsilateral AJV is unavailable or its preservation would limit flap transfer, the contralateral AJV may serve as an alternative drainage option.

Traditional SMAPF harvest techniques depend heavily on surgeons’ clinical experience [27, 37].This limits its widespread adoption. The A-B-C venous optimization technique standardizes dissection by simplifying complex venous anatomy into sequential identification of three key points, allowing surgeons at different experience levels to follow a unified protocol. The venous and spatial classification systems provide a structured cognitive framework, enabling surgeons to anticipate drainage variation before and during dissection. Flap compromise rates decreased in both experience groups following training, with a more pronounced reduction among less experienced surgeons. Multivariable analysis identified the A-B-C technique as an independent protective factor against flap compromise. The A-B-C technique also significantly shortened harvest time, and flap area emerged as an additional independent predictor of harvest duration. These results support the clinical value of a systematic venous-first dissection approach in SMAPF harvest.

This study has several limitations. First, venous drainage patterns were analyzed exclusively in Chinese patients (n = 408), which may limit generalizability across different ethnic populations. Second, the retrospective pre-post comparative design cannot fully exclude temporal bias arising from unmeasured changes in surgical practice or instrumentation between study periods. The modest sample size in Phase 2 (n = 117) yielded wide confidence intervals, warranting further validation in larger prospective cohorts. Third, outcome assessment was not blinded to surgical approach, introducing potential ascertainment bias. A prospective multicenter randomized controlled trial is currently being planned to further validate these findings.

Conclusion

Through systematic analysis of 408 SMAPF cases, we developed a “vein-protection-first” strategy incorporating a standardized A-B-C venous optimization technique and a three-type drainage classification (A-B-single/double/triple C). The A-B-C technique was associated with significantly reduced flap compromise rates and shorter harvest time across surgeons of varying experience levels. These findings suggest that this structured approach can help address experience-dependent variability in SMAPF harvest, offering a practical framework to support both clinical implementation and surgical training.

Abbreviations

SMAPF

Submental artery perforator flaps

SMV

Submental vein

FV

Facial vein

IJV

Internal jugular vein

AJV

Anterior jugular vein

EJV

External jugular vein

Authors' contributions

Conceptualization：Z.Z.；Data curation：Z.Z.；Formal analysis：Z.Z.；Funding acquisition：X.T.；N.Z.；Investigation：Z.Z.；Methodology：Z.Z.；X.X.；X.T.；N.Z.；Project administration：Z.Z.；X.X.；X.T.；N.Z.；Resources：Z.Z.；X.T.；N.Z.；Software：Z.Z.；Supervision：Z.Z.；X.X.；X.T.；N.Z.；Validation：Z.Z.；X.T.；N.Z.；Visualization：Z.Z.；Writing – original draft：Z.Z.；X.X.；Writing – review & editing：Z.Z.；X.X.；X.T.；N.Z.

Funding

The authors received no financial support for the research, authorship, and publication of this article.

Data availability

The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study complied with the Declaration of Helsinki. The study was reviewed and approved by the institutional review board of School and Hospital of Stomatology, China Medical University (K2025015). Informed consent was obtained from all of the patients who participated in this research.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.