Refining the boundaries of the nodal CTV in the middle neck for nasopharyngeal carcinoma Authors: Yue Deng, Yi Liu, Dan Wang, Yue Zhao, Wenjun Liao, Li Yin, Jianxia Lyu, Guotai Wang, Shichuan Zhang Published: 2026 ABSTRACT AbstractPurposeIn our previous research, we found that the subcompartments formed by the cervical fascia in the supraclavicular region (levels IV, Vb, and lower VI) effectively define the boundaries of the nodal clinical target volume (CTV) for nasopharyngeal carcinoma (NPC). This study aimed to determine whether the same anatomical approach could be applied to the middle neck region (levels III, Va, and upper VI).Materials and methodsWe enrolled 100 patients with NPC who had positive lymph nodes throughout the neck, including the supraclavicular region. Their pretreatment images were systematically reviewed. We analyzed lymph node distribution patterns in relation to both the 2013 International Consensus node‐level definitions and the cervical fascia‐based subcompartments proposed in our previous study. Based on our findings, we recommend several modifications to the 2018 international CTV consensus guidelines.ResultsOur analysis confirmed that the cervical fascial subcompartments effectively define nodal CTV borders. Based on the observed lymph node distribution patterns, we propose several specific optimizations for the middle neck. The lateral border of level Va can be retracted from the skin and platysma to the line connecting the posterior margin of the sternocleidomastoid muscle (SCM) and the anterior margin of the trapezius muscle (TM). Medially, the border can be safely retracted to an anterior‐posterior line bisecting the center of the common carotid artery (CCA), as the space medial to this line is consistently devoid of nodes. In addition, the anterior boundary of the CTV may be defined by the lymph nodes located anterior to the carotid sheath; otherwise, the carotid sheath itself serves as the boundary.ConclusionsCervical fascial anatomy provides valuable insight into the distribution patterns of lymph nodes in the middle neck region in patients with NPC. Based on these anatomical considerations, appropriate modifications to the nodal CTV defined by the 2018 consensus can be implemented for the middle neck region. INTRODUCTION Prophylactic nodal irradiation represents a pivotal element in definitive radiotherapy for nasopharyngeal carcinoma (NPC). Although international guidelines for clinical target volume (CTV) delineation in NPC, including nodal CTV, were established in 2018 (the 2018 consensus) 1 and subsequently revised in 2025 (the 2025 consensus), 2 the exact delineation boundaries continue to warrant further scrutiny and optimization. In our previous study, we demonstrated that the deep cervical fascia delineates multiple longitudinal compartments extending from the skull base to the supraclavicular fossa, 3 thereby facilitating precise demarcation of the nodal CTV in NPC. Specifically, within the supraclavicular region, these compartments were identified as the para‐esophageal space (PES), carotid sheath space (CSS), sub‐thyroid pre‐tracheal space (STPTS), pre‐vascular space (PVS), and vascular lateral spaces (VLS) I and II. Our findings showed that lymph node metastases in NPC predominantly involve the CSS, VLS I, and VLS II, while sparing the PES, STPTS, and PVS. 4 Building upon this foundation, the current study investigates lymph node distribution in the middle neck regions (levels III, Va, and upper VI) and their spatial relationships with the cervical fascia in patients with NPC. Based on these observations, we propose refined approaches for CTV delineation in these anatomical regions. MATERIALS AND METHODS Patients and images Imaging data from consecutive patients with histologically confirmed NPC treated at Sichuan Cancer Hospital between January 1, 2020, and December 31, 2021, were retrospectively evaluated. From this cohort, 100 patients with positive lymph nodes in the upper, middle, and lower cervical regions were selected. The middle cervical region was defined as the anatomical zone extending from the caudal border of the hyoid bone to the caudal border of the cricoid cartilage. Patients with prior therapeutic interventions or concurrent secondary malignancies were excluded. Lymph node positivity was determined based on the following criteria: (1) a short‐axis diameter exceeding 10 mm, (2) the presence of central necrosis irrespective of nodal size, and (3) clusters of three or more lymph nodes with a maximum short‐axis diameter of at least 8 mm. 4 Patients presenting with enlarged supraclavicular nodes without concomitant metastases in levels II or III were excluded, given the rarity of skipped cervical metastases in NPC. 5 , 6 Pretreatment contrast‐enhanced computed tomography (CT) and magnetic resonance imaging (MRI) were performed to assess lymph node distribution. Analysis of the distribution patterns of lymph nodes in the middle neck To comprehensively delineate potential metastatic pathways, both positive lymph nodes and all additional nodes measuring ≥ 3 mm in diameter—consistent with the axial slice thickness of CT and MRI—were identified and annotated. This approach enabled a detailed characterization of nodal topography within the middle cervical region. The nodes were subsequently stratified into three size categories: 3–8 mm, 8–10 mm, and ≥ 10 mm, reflecting increasing risks of malignant involvement. In relation to surrounding anatomical landmarks, including vascular structures, the thyroid gland, and scalene muscles, each lymph node was represented as a 3‐mm square and projected onto a standardized CT template for head and neck malignancies. Nodal positioning was determined using the geometric center of each node. Cervical nodal levels (stations) in the middle cervical region were defined according to the 2013 International Consensus Guidelines for Cervical Node Levels in Head and Neck Cancer. 7 Definition of middle cervical sub‐compartments based on fascial anatomy and lymph node distribution patterns Building on the six subcompartments of the supraclavicular space delineated in our earlier research, we evaluated their extension into the middle cervical region. Five of these compartments extend cranially into this region; notably, the PES transitions into the parapharyngeal space (PPS), whereas the STPTS is anatomically absent. The prevalence of lymph nodes within each subcompartment was also quantified. In addition, superficial lymph nodes located between the superficial and deep cervical fasciae were documented and overlaid onto representative CT images. RESULTS Clinical characteristics of the patients Pretreatment imaging data from patients with NPC admitted between 2020 and 2021 were systematically reviewed until 100 patients with positive lymph nodes in the upper, middle, and lower cervical regions were selected. The patient demographics and clinical features are detailed in Supplementary Table 1. Node distribution pattern in the middle cervical region and its spatial relationship to node levels defined in the 2013 consensus In total, 570 lymph nodes were identified and stratified by size as follows: 314 nodes (3–8 mm), 85 nodes (8–10 mm), and 171 nodes (≥10 mm). The distribution patterns and their associations with cervical nodal levels, as per the 2013 consensus, are outlined in Supplementary Table 2 and depicted in Figure 1 A1–8 (encompassing all axial planes of the middle cervical region). Our findings underscore the fact that levels III and Va constitute high‐risk zones for metastasis in patients with NPC. Among the nodes ≥3 mm, 99.47% were situated within these levels; similarly, 99.61% of nodes ≥8 mm and 99.42% of nodes ≥10 mm were confined to the same areas. Two diminutive nodes (3–8 mm) were detected at level VI, neither of which were altered in size post‐treatment or during surveillance, implying no linkage to NPC. One metastatic node was noted posterior to level Va in the zone posterior to the anterior margin of the TM, previously termed the posterior lateral volume (PLV) by Jiang et al. 8 (Figure 1 A5–8 and Supplementary Table 2). Nodes in this compartment may exhibit heightened metastatic potential in the presence of Va involvement. Distribution of lymph nodes at the middle neck in NPC, displayed in eight transverse sections. A1‐8, the distribution pattern of nodes based on the 2013 International Consensus on Cervical Node Level; B1‐8, the distribution pattern of nodes based on the proposed five fascia anatomy‐based sub‐compartments. CSS, carotid sheath space; PPS, para‐pharyngeal space; PVS, pre‐vascular space; VLS, vascular lateral space. Certain anatomical zones demonstrated sparse or absent node presence. Within level Va, no clinically relevant lymph nodes were found in the subcutaneous compartment between the skin and external jugular vein. The region medial to the carotid sheath uniformly lacked nodes; nodal occurrence was extremely infrequent in the precarotid sheath area, and the limited subset of nodes identified was closely adjacent to the carotid sheath, with a complete absence of lymph nodes in the proximity of the strap muscles. Proposal of sub‐compartments in the middle neck based on cervical fascial anatomy As outlined earlier, five subcompartments, as defined by the deep cervical fascia, were extrapolated from the supraclavicular region (levels IV, Vb, and lower VI) to the middle neck, with the PES supplanted by the PPS. The delineations of these subcompartments are specified in Supplementary Table 3 and are visualized in Figure 2. Anatomy of the cervical fascia in the middle cervical region and depiction of the proposed five sub‐compartments based on fascia anatomy in a typical transverse section. Even though the cervical fascia cannot be seen on CT or MRI scans, its presence can be inferred by following the outlines of the structures it surrounds. Right side shows cervical fascia anatomy; left side displays the proposed sub‐compartments. It's important to note that on the left side, the lateral section of the middle layer of the deep cervical fascia is depicted with a dotted line due to insufficient documentation in the literature. Distribution of lymph nodes in proposed fascial spaces Overall, 99.47% of lymph nodes resided within the five proposed fascial compartments, irrespective of their dimensions. Among the NPC cases, nodes of 8–10 mm and ≥10 mm were primarily concentrated in the CSS, VLS I, and VLS II, comprising 97.65% of the 8–10 mm cohort and 98.83% of the ≥10 mm group (Figure 1 B1–8 and Supplementary Table 4). Fifteen nodes were observed in the PVS, whereas none were observed in the PPS. Three nodes were designated as superficial, positioned between the superficial and deep cervical fasciae, and categorized as levels III and VI according to the 2013 nodal classification framework. Suggested optimization of CTV for middle neck Figure 3 shows the proposed delineation of the nodal CTV for the middle neck. Although the 2025 consensus updated the guidelines for CTV delineation, refinements regarding the anterior border were specifically limited to retracting the lower neck (level IVa) to the posterior edge of the infrahyoid muscles, without addressing the middle neck. For a more direct comparison, we combined the 2013, 2018 and 2025 consensus guidelines to delineate an exemplary CTV for the middle neck, which served as a reference template for comparison with the proposed CTV. Comparison of nodal CTV delineation between international guidelines (left) and our proposals (right) in two representative planes. Note the first plane is right below the hyoid bone and there is a small node located just before the common carotid, which can serve as the anterior border of CTV. The second plane is the lower part of middle neck.The anterior‐posterior line bisecting the center of the carotid artery can serve as the medial border of CTV. The anterior edge of carotid sheath, generally the common carotid, constitutes the anterior edge of CTV. External jugular vein can be seen in both planes. The lateral border of CTV should not extend beyond it at middle neck. Selective combination of fascial sub‐compartments forms CTV Based on these observations, the prophylactic irradiation CTV for the middle cervical region in NPC should incorporate the PVS, CSS, VLS I, and VLS II. The exclusion of PPS from the prophylactic CTV appears justifiable for patients with NPC. Moreover, only one positive node (>10 mm) was detected in the PLV (Figure 1, A7). Thus, in scenarios in which VLS II exhibits minimal involvement risk (i.e., the absence of positive nodes in the proposed VLS I and CSS), the PLV segment of VLS II can be prudently omitted from the prophylactic CTV. (1)The anterior boundary of the CTV may be defined as the lymph nodes located anterior to the carotid sheath; otherwise, the carotid sheath itself serves as the boundary.Fifteen lymph nodes (2.63%) were documented in the PVS in 14 patients (14% of the cohort). Of these, 73.33% (11/15) were diminutive (3–8 mm) and 86.67% (13/15) resided in the superior portion of the PVS, aligned with the extension of level IIA. The majority (14/15) were positioned adjacent to the carotid sheath and lateral to the common carotid artery (CCA), indicating an affiliation with the deep jugular chain proximal to the internal jugular vein (IJV). Under physiological conditions, the lymphatic flow from this chain proceeds caudally along the IJV via divergent eschewing routes (Figure 4). Consequently, the nodes in the PVS denote the anterior extremities of the pathway. According to the 2013 consensus guidelines, the anterior boundary of Level III is anatomically defined as the anterior edge of the sternocleidomastoid muscle or the posterior third of the thyrohyoid muscle. This definition encompasses a broad region anterior to the carotid sheath. Although the 2025 consensus has already optimized and retracted the anterior border for the supraclavicular region, the middle neck has not been similarly addressed. Given that the lymph nodes in the middle neck are generally sparse, and only a minimal number exist in the PVS anterior to the carotid sheath, we suggest that the anterior boundary of the middle neck CTV should be retracted to the anterior margin of the carotid sheath. In specific instances in which nodal presence is identified in the PVS, the boundary may be defined to encompass these lymph nodes.(2)The medial border can be set as an anterior‐posterior line crossing the center of the CCA instead of the medial edge of the CCA.The 2013 consensus guidelines explicitly define the medial borders of levels III and IV as the medial edges of the CCA and scalene muscles. The current study confirmed that PPS was completely devoid of lymph nodes. Importantly, the lymph nodes in both PVS and CSS were invariably lateral to the CCA. We observed that, even when the boundary was shifted laterally to bisect the CCA, the region medial to this midpoint remained entirely free of lymph nodes (Figure 4). Consequently, the medial border can be prudently retracted to an anterior‐posterior line crossing the center of the CCA. This approach effectively positioned all the implicated nodes laterally, thereby potentially mitigating radiation doses to the midline structures by safely sparing the medial CCA.(3)The lateral border may be retracted from the skin/platysma to a line connecting the posterior margin of the SCM and anterior margin of the TM.Regarding cervical fascia anatomy, the SCM and TM are surrounded by the superficial layer of the deep cervical fascia (SLDCF). The deep cervical lymph nodes related to NPC are typically situated within the adipose tissue beneath this layer. The line connecting the SCM and TM approximates the SLDCF trajectory on the axial scans. The external jugular vein (EJV) also functions as a landmark for SLDCF localization. In the middle neck, the EJV lies between the platysma and the SLDCF, perforating the latter to access deeper planes in the inferior neck. The proposed line effectively includes all the detected nodes medially. The EJV is uniformly superficial to this line, reinforcing its utility as a proxy for SLDCF in this region. (1)The anterior boundary of the CTV may be defined as the lymph nodes located anterior to the carotid sheath; otherwise, the carotid sheath itself serves as the boundary.Fifteen lymph nodes (2.63%) were documented in the PVS in 14 patients (14% of the cohort). Of these, 73.33% (11/15) were diminutive (3–8 mm) and 86.67% (13/15) resided in the superior portion of the PVS, aligned with the extension of level IIA. The majority (14/15) were positioned adjacent to the carotid sheath and lateral to the common carotid artery (CCA), indicating an affiliation with the deep jugular chain proximal to the internal jugular vein (IJV). Under physiological conditions, the lymphatic flow from this chain proceeds caudally along the IJV via divergent eschewing routes (Figure 4). Consequently, the nodes in the PVS denote the anterior extremities of the pathway. According to the 2013 consensus guidelines, the anterior boundary of Level III is anatomically defined as the anterior edge of the sternocleidomastoid muscle or the posterior third of the thyrohyoid muscle. This definition encompasses a broad region anterior to the carotid sheath. Although the 2025 consensus has already optimized and retracted the anterior border for the supraclavicular region, the middle neck has not been similarly addressed. Given that the lymph nodes in the middle neck are generally sparse, and only a minimal number exist in the PVS anterior to the carotid sheath, we suggest that the anterior boundary of the middle neck CTV should be retracted to the anterior margin of the carotid sheath. In specific instances in which nodal presence is identified in the PVS, the boundary may be defined to encompass these lymph nodes. The anterior boundary of the CTV may be defined as the lymph nodes located anterior to the carotid sheath; otherwise, the carotid sheath itself serves as the boundary. Fifteen lymph nodes (2.63%) were documented in the PVS in 14 patients (14% of the cohort). Of these, 73.33% (11/15) were diminutive (3–8 mm) and 86.67% (13/15) resided in the superior portion of the PVS, aligned with the extension of level IIA. The majority (14/15) were positioned adjacent to the carotid sheath and lateral to the common carotid artery (CCA), indicating an affiliation with the deep jugular chain proximal to the internal jugular vein (IJV). Under physiological conditions, the lymphatic flow from this chain proceeds caudally along the IJV via divergent eschewing routes (Figure 4). Consequently, the nodes in the PVS denote the anterior extremities of the pathway. According to the 2013 consensus guidelines, the anterior boundary of Level III is anatomically defined as the anterior edge of the sternocleidomastoid muscle or the posterior third of the thyrohyoid muscle. This definition encompasses a broad region anterior to the carotid sheath. Although the 2025 consensus has already optimized and retracted the anterior border for the supraclavicular region, the middle neck has not been similarly addressed. Given that the lymph nodes in the middle neck are generally sparse, and only a minimal number exist in the PVS anterior to the carotid sheath, we suggest that the anterior boundary of the middle neck CTV should be retracted to the anterior margin of the carotid sheath. In specific instances in which nodal presence is identified in the PVS, the boundary may be defined to encompass these lymph nodes. (2)The medial border can be set as an anterior‐posterior line crossing the center of the CCA instead of the medial edge of the CCA.The 2013 consensus guidelines explicitly define the medial borders of levels III and IV as the medial edges of the CCA and scalene muscles. The current study confirmed that PPS was completely devoid of lymph nodes. Importantly, the lymph nodes in both PVS and CSS were invariably lateral to the CCA. We observed that, even when the boundary was shifted laterally to bisect the CCA, the region medial to this midpoint remained entirely free of lymph nodes (Figure 4). Consequently, the medial border can be prudently retracted to an anterior‐posterior line crossing the center of the CCA. This approach effectively positioned all the implicated nodes laterally, thereby potentially mitigating radiation doses to the midline structures by safely sparing the medial CCA. The medial border can be set as an anterior‐posterior line crossing the center of the CCA instead of the medial edge of the CCA. The 2013 consensus guidelines explicitly define the medial borders of levels III and IV as the medial edges of the CCA and scalene muscles. The current study confirmed that PPS was completely devoid of lymph nodes. Importantly, the lymph nodes in both PVS and CSS were invariably lateral to the CCA. We observed that, even when the boundary was shifted laterally to bisect the CCA, the region medial to this midpoint remained entirely free of lymph nodes (Figure 4). Consequently, the medial border can be prudently retracted to an anterior‐posterior line crossing the center of the CCA. This approach effectively positioned all the implicated nodes laterally, thereby potentially mitigating radiation doses to the midline structures by safely sparing the medial CCA. (3)The lateral border may be retracted from the skin/platysma to a line connecting the posterior margin of the SCM and anterior margin of the TM.Regarding cervical fascia anatomy, the SCM and TM are surrounded by the superficial layer of the deep cervical fascia (SLDCF). The deep cervical lymph nodes related to NPC are typically situated within the adipose tissue beneath this layer. The line connecting the SCM and TM approximates the SLDCF trajectory on the axial scans. The external jugular vein (EJV) also functions as a landmark for SLDCF localization. In the middle neck, the EJV lies between the platysma and the SLDCF, perforating the latter to access deeper planes in the inferior neck. The proposed line effectively includes all the detected nodes medially. The EJV is uniformly superficial to this line, reinforcing its utility as a proxy for SLDCF in this region. The lateral border may be retracted from the skin/platysma to a line connecting the posterior margin of the SCM and anterior margin of the TM. Regarding cervical fascia anatomy, the SCM and TM are surrounded by the superficial layer of the deep cervical fascia (SLDCF). The deep cervical lymph nodes related to NPC are typically situated within the adipose tissue beneath this layer. The line connecting the SCM and TM approximates the SLDCF trajectory on the axial scans. The external jugular vein (EJV) also functions as a landmark for SLDCF localization. In the middle neck, the EJV lies between the platysma and the SLDCF, perforating the latter to access deeper planes in the inferior neck. The proposed line effectively includes all the detected nodes medially. The EJV is uniformly superficial to this line, reinforcing its utility as a proxy for SLDCF in this region. Lymph drainage pathways at middle neck and its significance in determining CTV coverage. Lymph flow from the nasopharynx normally goes through the internal jugular chain anteriorly and the accessory chain posteriorly (red arrows). Lymph flow from the larynx and thyroid goes laterally to the internal jugular chain (yellow arrows). Lymphatic spread of NPC theoretically has very low odds to involve lymph system medial to the carotid sheath. Covering lateral half of the common carotid artery is enough to include all nodes at risk of NPC involvement. DISCUSSION The middle neck (encompassing levels III, Va, and upper VI) exhibited a transitional lymphatic distribution pattern, bridging the configurations observed in the upper neck (level II) and lower neck (levels IV, Vb, and lower VI). In the upper neck, lymph nodes are commonly positioned medial (retropharyngeal), anterior, and lateral to the carotid sheath, including those in the jugulodigastric group. These structures facilitate lymphatic drainage from the nasopharynx, oropharynx, oral cavity, larynx, and hypopharynx (Figure 4). Efferent vessels from these nodes predominantly converge into the internal jugular chain within the middle neck, which parallels the IJV and ultimately drains into the venous angle. 8 , 9 This pathway explains the predominance of lymph nodes located posterior and lateral to the carotid sheath at level III. A limited subset of nodes was identified anterior to the carotid sheath in the middle neck, primarily in its cranial segment. These nodes likely accommodate drainage from levels Ib and IIa, as supported by the identification of two NPC‐related positive nodes in this region. Nevertheless, they may also contribute to lymphatic drainage from the superior thyroid and supraglottic regions, given their proximity to the superior thyroid vessels. 10 In contrast, the absence of lymph nodes in the PPS of the middle neck, compared with their presence in the para‐esophageal space of the lower neck, warrants further investigation. The 2013 guidelines for cervical nodal stations do not include this compartment. Furthermore, studies on lymph node dissection for hypopharyngeal carcinoma have not reported nodes in this region. 11 The 2013 consensus defined the medial margin of the CCA as the medial boundary of level III, which was adopted by the 2018 consensus as the medial border for nodal CTV. In the present study, we observed that this boundary can be shifted laterally to a sagittal plane bisecting the CCA, as no nodes were identified medial to this line. A contemporary study by Lu et al. 12 demonstrated that excluding the CCA from the CTV, while allowing partial inclusion through planning target volume (PTV) margins, does not compromise regional control in NPC, which is consistent with our findings. However, it is important to acknowledge the complexity of lymphatic drainage in patients with regionally advanced disease. In clinical practice, patients with extensive nodal involvement or obstruction of normal pathways may exhibit retrograde lymphatic drainage. The recommended CTV boundaries in this study are based on typical drainage patterns, and their application should be limited to cases without substantial extranodal extension or prior radiotherapy. Moreover, adequate PTV margins must always be incorporated to account for uncertainties in setup and positioning. CONCLUSIONS By examining lymph node spatial patterns in the middle cervical region, this investigation substantiates the utility of a fascial compartment‐oriented framework in refining nodal CTV delineations for patients with NPC. These modifications curtail the irradiated volume, potentially alleviating radiation‐induced adverse effects without compromising the coverage of susceptible lymphatic areas. Future prospective studies are warranted to validate the clinical safety of these refined boundaries. AUTHOR CONTRIBUTIONS Yue Deng: Data curation; formal analysis; investigation; validation; visualization; methodology; and writing—original draft. Yi Liu: Data curation; investigation; visualization; validation; writing the original draft; writing the review, and editing. Dan Wang: Investigation. Yue Zhao: Conceptualization; methodology; writing—review and editing. Wenjun Liao: Conceptualization; methodology; writing—review and editing. Li Yin: Investigation. Jianxia Lyu: Investigation; project administration. Guotai Wang: Methodology. Shichuan Zhang: Conceptualization; resources; formal analysis; supervision; funding acquisition; writing—review and editing. CONFLICT OF INTEREST STATEMENT The authors declare no competing interests. ETHICS APPROVAL The Ethics Committee of Sichuan Cancer Hospital approved this retrospective study. (approval number SCCSMC‐01–2024‐114, and the requirement for individual informed consent was waived) Supporting information Supporting Information