Inguinal CTV contouring in anal cancer: comparative validation of inguinal lymph nodes contouring guidelines and implications for consensus

Highlights

• •Garda ASCC guidelines achieved 90% inguinal nodal coverage with limited posterior irradiation.
• •Coverage gaps occurred in superolateral (7%) and inferior (3%) regions.
• •UK, AGITG and Chang improved coverage in uncovered areas; RTOG underperformed.
• •Results support disease-specific harmonization of inguinal CTV delineation in ASCC.

Abstract

Background and purpose

Contouring Inguinal lymph nodes (ILNs) in anal squamous cell cancer (ASCC) remains challenging. Accurate delineation is essential to optimize treatment efficacy with sparing of organs at risk (OaRs). Garda et al. proposed a contouring guideline based on the anatomical distribution of inguinal lymphadenopathy. This study evaluates their effectiveness in ensuring adequate ILNs coverage.

Materials and methods

We retrospectively analyzed ASCC patients with radiologically positive inguinal lymph nodes treated with IMRT in our institution between 2015 and 2024. ILNs were mapped relative to the femoral vessels and delineated based on the Garda’s guidelines. Descriptive analyses were conducted.

Results

Among 46 patients with positive ILNs, 27 (59%) had cT3-T4 stage and 27 (59%) had bilateral inguinal nodal involvement. Overall, 203 ILNs were analyzed: 88% (178 nodes) were anteromedial to the femoral vessels. Garda’s CTV encompassed 182 nodes (90%), while 22 nodes (11%) across 14 patients (30%) were not covered. Most of the uncovered nodes were in the superolateral region (n = 15, 7% of all nodes), whereas 6 nodes (3%) from six patients were identified in the inferior region. Only one node was found in the superomedial quadrant, while no nodes were located posteriorly.

Conclusions

These findings validate the proposed guidelines, confirming their effectiveness in reducing unnecessary irradiation of the posterolateral quadrant. However, coverage gaps highlight the need for further refinements to optimize target delineation. These findings offer disease-specific evidence to inform the harmonization of inguinal CTV delineation and support the development of standardized, consensus-based contouring protocols.

Introduction

The standard treatment for locoregionally confined squamous cell carcinoma of the anal canal (ASCC) consists of definitive external beam radiotherapy combined with concurrent chemotherapy (CRT).[1], [2], [3], [4], [5], [6].

Inguinal and femoral lymph node regions are common sites of spread, with involvement reported in approximately 6.4% of T1–T2 tumors and up to 16% in T3–T4 tumors[7], [8]. International guidelines consistently recommend elective irradiation of the ILNs during definitive treatment.[9], [10], [11], [12].

Omission of elective inguinal irradiation is associated with significantly higher rates of inguinal recurrence—up to 23%, compared to ≤ 2% in treated patients.[13], [14], [15], [16].

Intensity-modulated radiotherapy (IMRT) has improved conformity to target volumes and reduced radiation exposure to organs at risk (OARs), including the bowel, bladder, femoral heads, genitalia and anterior skin.[17], [18], [19], [20], [21].

However, these benefits depend on accurate clinical target volume (CTV) delineation, as inadequate definition can result in target undercoverage or increased toxicity. [17].

Several international guidelines have described the limits of the inguinal lymph node regions. They differ slightly in cranial and caudal extension, anterior and medial margins, and reference boundaries (skeletal, vascular, or fascial) [9], [10], [11], [22] (Table 1). The majority are based on expert consensus with the only ASCC-specific model, derived from positive lymph node mapping, being proposed by Garda et al., although external validation is still lacking.

Table 1.

Guidelines for delineation of ILNs clinical target volume (CTV) for definitive radiotherapy in anal cancer. HSIF: the horizontal superficial inguinal field, VSIF: the vertical superficial inguinal field, DIF: the deep inguinal field.

Consensus Group	Cranial	Caudal	Anterior	Posterior	Lateral	Medial
AGITG [9]	Level where external iliac artery leaves the bony pelvis to become the femoral artery	Lower edge of the ischial tuberosities	Minimum 20-mm margin on the inguinal vessels, inclusive of any visible lymph nodes or lymphoceles	Femoral bed formed by iliopsoas, pectineus, and adductor longus muscles	Medial edge of sartorius or iliopsoas	10 to 20 mm margin on the femoral vessels (approximated by medial third to half of the pectineus or adductor longus muscle)
RTOG [10]	Level of the caudad extent of the internal obturator vessels (approximated by upper edge of the superior pubic rami)	20 mm caudad to the saphenous-femoral junction	7 to 8 mm margin in soft tissue around the iliac vessels, consider larger (>10 mm) margin. Include any identified nodes	7 to 8 mm margin in soft tissue around the iliac vessels, excluding bone and muscle. Include any identified nodes	7 to 8 mm margin in soft tissue around the iliac vessels, consider larger (>10 mm) margin. Include any identified nodes	7 to 8 mm margin in soft tissue around the iliac vessels, excluding bone and muscle. Include any identified nodes
UK National Guidance for IMRT in Anal Cancer [23]	The external iliac nodal group	At the inferior slice demonstrating the lesser trochanter	Approximately 5 mm from the skin surface. All visible nodes and lymphoceles should be included	The pectineus, adductor longus, and iliopsoas. All visible nodes and lymphoceles should be included	The medial edge of sartorius or ilio-psoas. (include visible nodes)	To include all visible lymph nodes or lymphocoeles. The spermatic cord in men
Garda et al 2022 [22]	Level where external iliac vessels leave bony pelvis	12 mm superior to the inferiormost aspect of the primary tumor or 14 mm below inferior aspect of pubic symphysis	30 mm margin on femoral vessels, inclusive of all radiographically suspicious lymph nodes	Posterior border of femoral vessels	Lateral border of femoral vessels	26 mm margin on femoral vessels, inclusive of all radiographically suspicious lymph nodes
ECOG-ACRIN 2022 [24]	End of external iliac vessels	Lower edge of lesser trochanter or where the sartorius and adductor longus muscles cross	15–20 mm anterior to femoral vessels	Pectineus, adductor longus, iliopsoas	The medial edge of the sartorius or iliopsoas muscle.	Spermatic cord (men), medial third to half of pectineus or adductor longus muscle
NOAC 2023 [25]	The inguinal ligament (pelvic brim turns medially)	Lesser trochanter (extend below if LN present)	20 mm anterior to femoral vessels (≥5 mm from skin)	Upper inguinal: the posterior border of the femoral triangle. Lower inguinal: the pectineus muscle and the anterior border of the femoral vessels.	Sartorius or iliopsoas (include visible nodes)	Upper inguinal: the spermatic chord in men. 10–20 mm medial to the femoral vessels. Lower inguinal: at least 5–7 mm medial to the great saphenous vein.
Chang et al 2023 [26]	Acetabulum roof (HSIF/DIF); saphenofemoral junction (VSIF)	Ischial tuberosity (VSIF); saphenofemoral junction (DIF)	Camper’s fascia; ≤18 mm anterior to GSV (VSIF)	Scarpa’s fascia or anterior edge of muscle/vessels (HSIF); medial wall of femoral vein (DIF)	Medial edge of sartorius (HSIF); ≤20 mm lateral to GSV (VSIF)	5 mm from superior pubic ramus (HSIF); 10 mm medial to GSV (VSIF)

GSV: great saphenous vein

AGITG: Australasian Gastrointestinal Trials Group.

RTOG: Radiation Therapy Oncology Group.

NOAC: Nordin Anal Cancer Group.

EOCG ACRIN: Eastern Cooperative Oncology Group-American College of Radiology Imaging Network.

IMRT: Intensity Modulated Radiation Therapy.

This study aims to externally validate the inguinal CTV delineation guidelines proposed by Garda et al. [22] in ASCC patients treated with IMRT, compare their performance with other established models, and identify refinements that could better align with actual patterns of metastatic spread.

Materials and methods

This retrospective study included all consecutive patients with histological confirmed ASCC and radiological evident ILNs metastases at diagnosis, treated with IMRT in our institution between January 2015 and November 2024. Data were collected from institutional databases, including patients’ details, tumor characteristics, and treatment specifics.

Only patients with available pre-treatment diagnostic imaging were included.

All available imaging modalities, including computed tomography (CT), positron emission tomography (PET), and magnetic resonance imaging (MRI), were used by a diagnostic radiologist to determine positivity and location of involved ILNs.

Lymph nodes were considered malignant if they measured ≥ 15 mm in short axis on CT/MRI, or if smaller but with abnormal morphology (rounded shape, irregular contour, loss of fatty hilum, or necrosis),[5], [23] or if showing^18 Fluorodeoxyglucose (FDG) uptake greater than liver background on PET. Nodal positivity was assessed within a multidisciplinary tumor board.

All patients underwent a planning CT scan in the supine position with scanning margins from T12 to the proximal third of the femur and a slice thickness of 1 mm. No intravenous (IV) contrast was used in the design cohort as it was not mandated by institutional guidelines at that time. This may have reduced the anatomical accuracy of nodal localization, although multimodal fusion with PET and MRI partially mitigated this limitation.

We analyzed the spatial distribution of lymph nodes in the examined population.

The Lymph node location was recorded using a clock-face system relative to the femoral vessels (anteromedial: 12–3o’clock; posteromedial: 3–6; posterolateral: 6–9; anterolateral: 9–12), with all left-sided data mirrored to the right. Clock-face–based nodal distribution was independently assessed by three radiation oncologists and centrally reviewed by two senior experts in lower gastrointestinal nodal contouring, ensuring inter-observer consistency and minimizing subjective interpretation.

Subsequently ILNs were delineated based on CTV boundaries followed the Garda et al. anatomical definitions [22], [22] Cranial: exit of external iliac vessels; caudal: 12 mm above the most inferior tumor slice or 14 mm below the pubic symphysis; anterior: 30 mm from femoral vessels; medial: 26 mm from femoral vessels; posterior and lateral: directly abutting the femoral vessels. Volumes were adapted to patient anatomy and cropped 1 mm inside the body surface where necessary.

Descriptive statistics were used to summarize patient characteristics and nodal distribution. Coverage rates were expressed as proportions of total nodes, with subgroup analysis by location and patient factors. To assess whether nodes were included or missed, spatial relationships were evaluated directly within the treatment planning system on fused CT/MRI/PET images. Lymph nodes touching the contour boundary or extending ≤ 1 mm beyond it were considered included, consistent with clinical practice and Garda’s definitions for inclusion of suspicious nodes. Nodes located > 1 mm beyond the boundary were classified as uncovered.

For external validation, each anatomical boundary proposed by Garda et al. (superior, anterior, medial, lateral, inferior, and posterior) was tested on our dataset. Boundaries were considered applicable if they encompassed at least 95% of lymph nodes located in that specific direction. These coverage thresholds reflect prior nodal-mapping and CTV-validation studies, which commonly use ∼ 95% as the benchmark for an adequate elective margin, whereas optimized CTV definitions typically reach 97–98% to achieve near-complete anatomical inclusion without excessive expansion[22], [24], [25], [26], [27].

After validating coverage according to the Garda guidelines, we performed a secondary comparative analysis to evaluate how the uncovered nodes would have been encompassed by other pre-existing contouring guidelines (AGITG, RTOG, UK National Guidance, and Chang et al.), in order to contextualize our findings.

Results

Patient characteristics and inguinal distribution

Overall, 230 patients with ASCC were treated with radiotherapy at our institution; 46 (20%) presented with radiological involved ILNs at diagnosis and were included in the analysis.

The median age was 67 years (range, 35–83), 34 patients (74%) were female, and the median Body Mass Index (BMI) was 22.1 kg/m² (range, 18.4–33.6). Advanced primary tumors (T3–T4) were present in 27 patients (59%). Nodal involvement was bilateral in 27 patients (59%) and unilateral in 19 patients (41%). Detailed baseline characteristics are reported in Table 2.

Table 2.

Patient and tumor characteristics; *AJCC 9th Edition (2022).

Patient (N = 46) and tumor characteristics	Value (%)
Sex
Male n (%)	12 (26%)
Female n (%)	34 (74%)
Median Age (range)	67 (35–83)
Median BMI (range)(kg/m2)	22.1 (18.4–33.6)
cT stage *
TX	3 (6,5%)
T1	1 (2%)
T2	15 (33%)
T3	11 (24%)
T4	16 (35%)
cN stage
N1a	31 (67%)
N1c	15 (13%)
Total ILNs	203
Median ILNs/ patient, median (range)	3 (1––16)
ILNs laterality/patient
monolateral	19 (41%)
bilateral	27 (59%)
Skin infiltration by the primary tumor	12 (26%)

BMI: Body Mass Index.

ILNs: Inguinal Lymph Nodes.

A total of 203 positive ILNs were delineated and analyzed.

Using the clock-face system, the majority of nodes (88%, n = 178) were located anteromedial to the femoral vessels, with a complete absence of nodal disease in the posterior quadrant (5–9o’clock) (Fig. 1). Moreover, 178 lymph nodes were located between 12 and 5o’clock, 25 nodes between 9 and 12o’clock, and none between 5 and 9o’clock (supplementary table 1). This pattern is consistent with the assumptions of the Garda et al. guidelines and supports the exclusion of extensive posterior margins, as recommended by disease-specific protocols.

Fig. 1.

Clock-face graphical representation of lymph node distribution.

Inguinal contouring and external validation

When applying the anatomical boundaries proposed by Garda et al., 182 nodes (90%) were encompassed within the CTV. Applying a 95% threshold, all margins except the lateral were validated. When applying the stricter 98% coverage threshold the superior, anterior, medial, and posterior margins were validated. In contrast, both the lateral (92.6%) and inferior (97%) boundaries fell below the predefined cut-off. (Fig. 2).

Fig. 2.

Boundary coverage compared with 95% and 98% thresholds, according to Garda et al proposed limits. Bars show coverage (%) for each margin; dashed lines represent validation cut-offs at 95% and 98%.

Twenty-two nodes (11%) across 14 patients (30%) were located outside the delineated CTV (Fig. 3A, Fig. 3B, Fig. 3C; supplementary table 2). Most uncovered nodes (n = 15, 7% of total) were in the superolateral region, a site variably addressed by existing international guidelines while six nodes (3%) from 6 patients were located in the inferior region. Only 1 node (0.5%) was found outside the CTV in the superomedial quadrant.

Fig. 3A.

Example of supero-lateral coverage gap in the contouring of ILNs according to the guidelines of Garda et al.(purple line).

Fig. 3B.

Example of supero-medial coverage gap in the contouring of ILNs according to the guidelines of Garda et al.

Fig. 3C.

Example of inferior coverage gap in the contouring of ILNs according to the guidelines of Garda et al.

On reviewing the 14 patients with nodes not covered by the Garda guidelines, no consistent clinical or anatomical features were identified. The uncovered nodes were variably located (caudal, superolateral, and superomedial), and no association was found with T stage, N stage, skin involvement, or BMI. However, the limited number of uncovered nodes (n = 22) did not allow for a robust multivariate analysis.

Comparison with other contouring guidelines

We evaluated the uncovered area against existing guideline recommendations (supplementary table 3).

For inferior misses, coverage would have been achieved in 50% of cases using the UK guidelines and in 33% using AGITG, but in none with RTOG. For superolateral misses, both the UK and AGITG guidelines would have achieved full coverage (100%), whereas RTOG would have missed all. The fascia-based Chang et al. approach covered all superolateral nodes and 50% of inferior nodes, suggesting a potential advantage in anatomically complex areas.

Discussion

Contouring of ILNs in patients with ASCC remains a complex challenge in radiation oncology. Accurate target delineation is essential to optimize oncologic outcomes while minimizing radiation exposure to organs at risk (OARs), such as the femoral heads, genitalia and anterior skin. Several studies, including the dosimetric analysis by Dapper et al., have shown substantial variability among existing contouring guidelines, leading to inconsistent target coverage and potential geographic miss.[32].

In this context, our analysis builds upon the anatomical framework proposed by Garda et al. by evaluating its performance against pathologically involved ILNs in ASCC. Our findings largely confirm the original anatomical observations, with 88% of metastatic nodes located between 12:00 and 5:00o’clock, with no involvement of the posterior quadrant (5:00–9:00).

Application of the proposed boundaries resulted in a overall nodal coverage of 90%, with accurate posterior and anteromedial limits. However, 22 nodes (11%) across 14 patients (30%) were missed, predominantly in the superolateral (n = 15, 7%) and inferior (n = 6%) regions.

As illustrated in Fig. 3A, Fig. 3B, Fig. 3C, what appears as a superomedial coverage gap lies close to the proposed 26 mm medial expansion and may equally apply to superolateral and inferior regions. This finding highlights interpretative variability in anatomically complex regions and underscores the need to complement numerical margins with reproducible anatomical landmarks to ensure consistent and reliable CTV delineation.

Applied to achive a 95% nodal coverage threshold, the proposed margins appear generally adequate, with the exception of the lateral boundary, which is limited by superolateral nodes. Under a stricter 98% criterion, the inferior margin also fell short (97%), identifying two areas of suboptimal coverage. This stepwise analysis is consistent with validation strategies in other tumor sites, where high coverage cut-offs (≥ 95–98%) are considered clinically meaningful benchmarks to assess contouring adequacy.[25], [26], [27]. Specifically Mittal et al. defined inguinal CTV margins in penile cancer using a ≥ 95% cut-off, Keenan et al. validated para-aortic CTVs in cervical cancer at 97%, and Wang et al. optimized para-aortic delineation to reach 98%.

Despite differences in anatomical rationale, none of the existing countouring guidelines ensured complete inguinal nodal coverage in our cohort. Even the most inclusive approaches left some regions uncovered, highlighting the inherent limitations of fixed-margin models in anatomically complex areas. When compared across guidelines, superolateral nodes were fully covered by UK [23], AGITG [9], and Chang et al. [26], but were consistently missed by RTOG [10]. Inferior misses would have been covered in 50% of cases using the UK National Guidance [23] and in 33% using AGITG [9], but in none using RTOG [10]. These findings are consistent with those reported by Dapper et al. [28], [31], who also described inadequate coverage of inferior and radial regions with AGITG and RTOG definitions.

Taken together these discrepancies reflect fundamentally distinct anatomical rationales underlyng current guidelines. Vascular-based models such as those proposed by Garda [22] and Taylor et al. [34], emphasize lymphatic drainage patterns and aim to minimaze unnecessary posterior irradiation. In contrast, fascia-based approaches, such as the model by Chang et al.[26], incorporate Camper’s and Scarpa’s fascia as key anatomical landmarks and may offer advantages in anatomically complex regions. Rather than being mutually exclusive, these strategies appear complementary, suggesting that future consensus guidelines could benefit from integrating both approaches instead of relying on a single anatomical paradigm.

From a clinical perspective, areas of concern—particularly inferiorly—are often delineated using PET-CT guidance, which remains a critical tool for individualizing target volumes[33], [34]. Recent PET/CT-based mapping by Han et al. [35] supports margins of approximately 29 mm anteriorly, 28 mm medially, and caudal extension to the lesser trochanter or 26 mm below the pubic symphysis. These recommendations align with our observation that caudal extension may be warranted in selected high-risk patients, such as those with T3–T4 tumors or bilateral nodal involvement. Moreover, the extent of nodal disease and the potential for retrograde lymphatic flow further support the risk of spread beyond standard treatment fields, reinforcing the importance of adapting CTVs to patient-specific anatomy.

Our study has some limitations. Its retrospective design and the predominance of advanced-stage disease may limit generalizability. Moreover, the evaluated guidelines are primarily intended for elective irradiation in clinically node-negative patients, in whom the risk of inguinal recurrence is < 2% when the groins are treated [13], [14], [15], [16]. Nonetheless, mapping in a node-positive cohort provides important insights for refining elective CTV definitions and should be strengthened especially in locally advanced disease.

In conclusion, this study provides the first external validation of the Garda et al. guidelines using pathologically involved inguinal lymph nodes in ASCC, demonstrating that a vascular-based anatomical approach can achieve high nodal coverage while effectively limiting unnecessary posterior irradiation. Importantly, our findings also show that no existing contouring guideline ensures complete inguinal coverage, with residual gaps consistently observed in the superolateral and inferior regions.

These results carry a clear clinical and methodological implication: fixed-margin, single-paradigm models are insufficient to fully capture the complexity of inguinal nodal spread. Future consensus guidelines should therefore move toward integrated, anatomy-driven frameworks that combine vascular and fascial landmarks, incorporate disease-specific imaging, and allow for adaptive caudal extension in selected high-risk patients. Such an approach may represent a necessary step to harmonize target delineation, improve reproducibility across institutions, and optimize the balance between locoregional control and treatment-related toxicity.

Author contribution section

Manfrida S. Gambacorta MA conceived the presented idea.

Barogi N, De Luca V, Rinaldi RM, De Giacomo F developed the theory and collected the data.

Dinapoli N performed the analysis.

Manfrida S, Barogi N, De Luca V wrote the paper.

Macchia G, Hallemeier CL, Garda AE, Chiloiro G contributed to the final version of the manuscript.

Autorino R, Fionda B contributed to the implementation of the research.

Gambacorta MA supervised the findings of this work.

All authors discussed the results and contributed to the final manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: