Correlation between theoretical anatomical patterns ...

Eur J Nucl Med Mol Imaging DOI 10.1007/s00259-015-3256-6

ORIGINAL ARTICLE

Correlation between theoretical anatomical patterns of lymphatic drainage and lymphoscintigraphy findings during sentinel node detection in head and neck melanomas Mónica Vidal 1 & Sergi Vidal-Sicart 1,2 & Ferran Torres 3,4 & Diana Milena Ruiz 1 & Pilar Paredes 1,2 & Francesca Pons 1,2

Received: 21 July 2015 / Accepted: 5 November 2015 # Springer-Verlag Berlin Heidelberg 2015

Abstract Purpose In the diagnosis of head and neck melanoma, lymphatic drainage is complex and highly variable. As regional lymph node metastasis is one of the most important prognostic factors, lymphoscintigraphy can help map individual drainage patterns. The aim of this study was to compare the results of lymphoscintigraphy and sentinel lymph node (SLN) detection with theoretical anatomical patterns of lymphatic drainage based on the location of the primary tumour lesion in patients with head and neck melanoma. We also determined the percentage of discrepancies between our lymphoscintigraphy and the theoretical location of nodal drainage predicted by a large lymphoscintigraphic database, in order to explain recurrence and false-negative SLN biopsies. Methods In this retrospective study of 152 patients with head and neck melanoma, the locations of the SLNs on lymphoscintigraphy and detected intraoperatively were compared with the lymphatic drainage predicted by on-line software based on a large melanoma database. Results All patients showed lymphatic drainage and in all patients at least one SLN was identified by lymphoscintigraphy. Of the 152 patients, 4 had a primary lesion in areas that were not

* Mónica Vidal [email protected] 1

Nuclear Medicine Department, Hospital Clínic de Barcelona, Barcelona, Spain

2

Institut d’Investigacions Biomédiques Agustí Pi i Sunyer (IDIBAPS), Barcelona, Spain

3

Statistical of Biostatistics and Data Management Core Facility, IDIBAPS, Hospital Clínic Barcelona, Barcelona, Spain

4

Biostatistics Unit, Faculty of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain

described in the Sydney Melanoma Unit database, so agreement could only be evaluated in 148 patients. Agreement between lymphoscintigraphic findings and the theoretical lymphatic drainage predicted by the software was completely concordant in 119 of the 148 patients (80.4 %, 95 % CI 73.3 – 86 %). However, this concordance was partial (some concordant nodes and others not) in 18 patients (12.2 %, 95 % CI 7.8 – 18.4 %). Discordance was complete in 11 patients (7.4 %, 95 % CI 4.2 – 12.8 %). Conclusion In melanoma of the head and neck there is a high correlation between lymphatic drainage found by lymphoscintigraphy and the predicted drainage pattern and basins provided by a large reference database. Due to unpredictable drainage, preoperative lymphoscintigraphy is essential to accurately detect the SLNs in head and neck melanoma. Keywords Sentinel lymph node . Lymphoscintigraphy . Melanoma, head and neck

Introduction Melanoma is a potentially deadly disease that in recent decades has seen a steady increase in most Western countries. As there is currently no widely effective treatment for disseminated disease, early diagnosis gives the best chance of cure by surgical removal of the primary tumour [1, 2]. Usually, the earliest sign of metastatic spread beyond the primary melanoma site can be seen in regional lymph nodes [3]. The detection of metastatic melanoma in regional lymph nodes has important implications for treatment and prognosis. Sentinel lymph node (SLN) biopsy is used to detect whether melanoma cells have metastasized to a particular lymph node, the so-called SLN, which is defined as any lymph node receiving (or the first lymph node to receive) direct lymphatic drainage from

Eur J Nucl Med Mol Imaging

the primary tumour site [4]. The SLN can be localized by preoperative lymphoscintigraphy, which is also essential to provide greater accuracy in identifying transit and/or aberrant nodes (i.e. a nodal region outside the standard) [5]. Several studies have shown that when SLNs identified by lymphoscintigraphy are negative for melanoma metastasis, the remainder of the lymph node basin does not usually contain additional lymph nodes with metastases [6, 7]. Although this technique has been widely adopted by surgeons as an alternative to lymphadenectomy, there is still concern regarding head and neck melanomas, where lymphatic drainage is unpredictable and requires the use of anatomical patterns to improve the accuracy of lymphoscintigraphy. Lymphatic drainage from melanomas of the head and neck is highly complex, with SLNs commonly found in multiple node fields [8]. In addition, SLNs in the head and neck are usually small and often situated near the primary melanoma site, making their location more difficult [9]. Preoperative lymphoscintigraphy of the head and neck is technically difficult because of the local complexity of the lymphatic network, where more than 300 nodes join in an exceptionally compressed area [10]. However, the technique helps reveal lymphatic drainage patterns and the location of the SLN with respect to the primary tumour site (ipsilateral, contralateral and bilateral). This information is vital for planning surgery, as it helps the surgeon minimize the size of the incision and can improve the final cosmetic result [3]. In lateralized tumours, lymphatic mapping may show an unexpected drainage and can also detect aberrant or in-transit nodes, i.e. those that lie between the primary tumour and a regional drainage Basin [11]. Therefore, in order to establish the appropriate primary treatment, it is important that the anatomical sites be taken into account during evaluation of the drainage pattern. Since 1987 the American Head and Neck Society and the American Academy of Otolaryngology – Head and Neck Surgery have recommended the use of neck oncology levels I – VI. Additionally, another level (level VII) located in the upper mediastinum was included [12, 13]. The Sydney Melanoma Unit (now the Melanoma Institute Australia) database included 929 patients with melanoma of the head and neck, and the available data were analysed by the University of Auckland (New Zealand), creating a 3D computer model of lymphatic drainage of the skin using melanoma lymphoscintigraphy. This interactive tool includes a 3D depiction of drainage visualization coming from different points of the head and neck. The software shows based on database results the probability of regional lymphatic drainage according to location of the primary lesion [14, 15]. The aim of our study was to correlate the anatomical patterns of lymphatic drainage from the validated database and lymphoscintigraphy results during SLN procedures in our patients with head and neck melanoma.

Materials and methods Patients The study ran from June 1998 to June 2012 and retrospectively included 152 patients with head and neck melanoma comprising 63 women (41.4 %) and 89 men (58.6 %), age range 21 – 84 years. All patients underwent lymphoscintigraphy and intraoperative detection of SLN. The inclusion criteria for performing lymphatic mapping were a Breslow thickness in the range 0.75 – 1.0 mm, with at least one associated risk factor (ulceration, high mitotic rate, or Clark level IV or V) or melanomas greater than 1 mm. Melanoma in situ was only present in 1 of the 152 patients. In 127 of the 152 patients (83.6 %), the primary melanoma had been surgically removed and in 25 patients (16.4 %) was still present at the time of lymphoscintigraphy. The locations of the SLNs on lymphoscintigraphy and detected intraoperatively were compared with the lymphatic drainage clinically predicted using the software created by the University of Auckland, based on the location of the primary lesions [14, 15]. The regions used to establish the location of the primary melanoma are shown in Fig. 1. Lymph nodes in drainage areas outside the predicted patterns were considered aberrant lymph nodes.

Lymphoscintigraphy and sentinel node biopsy The radiotracer (Nanocoll; GE Healthcare) was always injected 5 – 10 mm away from the scar or the tumour margin. A maximum of four tuberculin syringes, each containing 37 MBq of 99mTc-nanocolloid in 0.1 mL, were used. The total activity administered ranged from 74 to 148 MBq. The needle was inserted as tangentially as possible to the skin surface, a few millimetres under the skin, just enough to produce a visible wheal in the skin. Care was taken to avoid any cutaneous contamination that could be confused with lymph node uptake. The injection site was covered with an adhesive plaster to prevent leakage of tracer through the puncture site. Immediately after radiotracer injection, a dynamic study (one frame/30 s for 10 min; matrix size 128 × 128) was acquired using a single-head digital gamma camera (E-Cam; Siemens; Erlangen, Germany). Several 5-min early static images (matrix size 256 × 256) were obtained after this dynamic study. Another set of static images (delayed) was obtained 2 h later. Based on lymphoscintigraphy, the main criteria used to identify lymph nodes as SLNs are the visualization of lymphatic ducts, the time of appearance, the lymph node basin and the intensity of lymph node uptake. Following these criteria,

Eur J Nucl Med Mol Imaging

Fig. 1 Distribution of the areas of the head and neck where the primary melanomas were located: a anterior view, b posterior view, c right view

visualized radioactive lymph nodes may be classified as follows [16]: 1. Definitely SLN: This includes all lymph nodes draining from the site of the primary tumour through their own lymphatic vessel or a single radioactive lymph node in a lymph node basin. 2. Highly probable SLN: This includes lymph nodes appearing between the injection site and a first draining node or nodes with increasing uptake appearing in other lymph node basins. 3. Non-SLN: Nodes with less uptake appearing distal from early draining node in a basin. Planar images were interpreted following the original report. When discrepancies appeared in relation to the theoretical drainage, the images were reviewed by two nuclear medicine physicians to reach a consensus. The anatomical location of SLNs was defined after obtaining all planar views to ensure accurate marking of the overlying skin. The localizations of the SLNs during surgery were indicated, assisted by the information related to the skin marking; however, not all SLNs could be surgically removed. The dissected tissue with the SLN was placed in 4 % formalin solution. No frozen sections were obtained. The SLN was separated from surrounding adipose tissue, bisected, sliced, and embedded. Serial sections were then made. Conventional staining with haematoxylin–eosin and immunohistological staining with S100 and human melanoma black-45 (HMB-45) were performed. Anatomical patterns of lymphatic drainage from the head and neck skin Lymphoscintigraphy studies from the Sydney Melanoma Unit database were analysed to obtain theoretical

lymphatic drainage patterns. The database includes lymphoscintigraphy studies obtained over a period of more than 15 years from more than 5,000 patients with cutaneous melanoma. Of these 5,000 studies, 929 were obtained from patients with head and neck melanoma. Based on these studies, the University of Auckland in New Zealand developed 3D interactive software showing the lymphatic drainage patterns according to the locations of the primary melanomas [14, 15]. This software is available at http:// sites.bioeng.auckland.ac.nz/hrey004/head/index.html. This software provides anatomically based geometric models of the skin and lymph nodes. A 3D finite element skin model was constructed using the Visible Human Project male dataset and a Sawbones head and neck model. A discrete lymph node model was also created using the Visible Human Project dataset. The Sydney Melanoma Unit database was mapped from 2D lymphoscintigraphic images onto the 3D anatomical model. Melanoma sites were mapped onto the skin model using free-form deformation and projection techniques. Spatial heat maps were created using field fitting to visualize the likelihood that any area of skin would drain to a particular node field, or a specified number of node fields. An interactive skin selection tool was also developed to provide dynamic predictions of the draining node fields from any region of skin. Skin on the head and neck was shown to drain more usually to two or more node fields, the most common of which were the cervical level II and preauricular node fields. For the current analysis, we grouped lymph nodes in the head and neck area into seven regions or levels based on anatomical landmarks (Fig. 2). Concordance Concordance was considered total when the pattern of lymphatic drainage observed on lymphoscintigraphy in our own patients was identical to that shown by the University of

Eur J Nucl Med Mol Imaging

Results Patients The histological types of head and neck melanoma in the 152 patients were: superficial spreading in 69 patients (45.4 %), nodular in 44 (28.9 %), lentigo maligna in 24 (15.8 %) and acral lentiginous in 15 (9.9 %). Primary melanoma locations are detailed in the Table 1. The median Breslow thickness was 2.19mm (25th – 75th percentile 1.3 – 3.85mm, mean±SD 2.97±2.23mm). In 40 % of the patients, the melanoma had a Breslow thickness in the range 1.1 – 2 mm. The pathology results in relation to Breslow thickness are shown in Table 2.

Lymphoscintigraphy and sentinel node biopsy Fig. 2 Six division levels of the head and neck. Level I Submental and submandibular nodes. Level II High jugular lymph node group: nodes adjacent to the internal jugular vein from the base of the skull to the lower border of the hyoid, stylohyoid behind and in front of the rear edge of the sternocleidomastoid. Level III Middle jugular lymph node group: nodes adjacent to the internal jugular vein located between the leading and trailing edges of the sternocleidomastoid, from the bottom edge of the hyoid to the lower border of the cricoid. Level IV Low jugular lymph node group: nodes adjacent to the internal jugular vein located between the leading and trailing edges of the sternocleidomastoid, under the lower edge of the cricoid and the clavicle. Level V Posterior triangle lymph node group: nodes located behind the rear edge of the sternocleidomastoid and in front of the anterior border of the trapezius, from the confluence of these muscles above the clavicle to the bottom. Level VI Anterior compartment group: nodes located between the two carotid arteries, in a space that extends from the hyoid above to below the sternum

Auckland software (based on the same primary melanoma locations). Concordance was considered partial when some areas of lymphatic drainage on lymphoscintigraphy were the same as those shown by the University of Auckland software and others were not. There was no concordance when the areas of lymphatic drainage on lymphoscintigraphy were different from those shown by the University of Auckland software.

Lymphoscintigraphy demonstrated lymphatic drainage from the intradermal injection site in all patients and visualized a total of 359 SLNs in the 152 patients (median 2.3 SLNs per Table 1

Right

Left

Data analysis and statistics Data are expressed as medians (25th – 75th percentiles) for continuous variables and frequencies and percentages for categorical variables, or as otherwise specified. Wilson’s method was used to estimate 95 % confidence intervals (CI) for percentages [17]. Cohens’ kappa was used to estimate intermethod concordance [18]. Concordance was rated as excellent when the kappa value was higher than 0.80. The analysis was performed using SAS version 9.2 (SAS Institute Inc., Cary, NC). Significance was established at the 5 % two-sided level.

Central

Locations of the head and neck primary melanomas Location

No. (%) of patients

Submandibular and submental region Anterior and lateral triangle of the neck Posterior triangle of the neck

2 (1.3) 4 (2.6) 10 (6.6)

Preauricular region Zygomatic and infraorbital region Supraciliary region Inferior eyelid Lower lip

5 (3.3) 10 (6.6) 3 (2.0) 2 (1.3) 1 (0.7)

Frontal region Temporal scalp

5 (3.3) 2 (1.3)

Parietal scalp Occipital scalp

6 (3.9) 3 (2.0)

Ear Submandibular and submental region Anterior and lateral triangle of the neck Posterior triangle of the neck Preauricular region Zygomatic and infraorbital region Supraciliary region Inferior eyelid Frontal region Temporal scalp Parietal scalp Occipital scalp Ear Nasal region Frontoparietal scalp

8 (5.3) 5 (3.3) 6 (3.9) 4 (2.6) 2 (1.3) 18 (11.8) 3 (2.0) 1 (0.7) 5 (3.3) 4 (2.6) 7 (4.6) 4 (2.6) 17 (11.2) 3 (2.0) 12 (7.9)

Eur J Nucl Med Mol Imaging Table 2

Pathology results in relation to Breslow thickness Breslow thickness (mm) 4

Positive pathology, n Negative pathology, n

1 8

4 56

6 42

9 21

SLNs not removed, n

2

1

1

1

All patients, n (%)

11 (7.2)

61 (40.1)

49 (32.2)

31 (20.3)

patient; Fig. 3). In 5 of these 152 patients (3.3 %), the SLNs could not be excised due to technical difficulties relating to surgical access, most often in the intraparotid region. According to the levels detailed in Fig. 2, 51 SLNs were found in level I, 119 in levels II, III and IV, 78 in level V, 103 in the preauricular region, and 3 in transit. In 39 patients (26 %) lymphoscintigraphy showed drainage to a single SLN, while in 113 patients (74 %) drainage was found to more than one SLN. Pathological analysis of the excised SLNs showed metastatic involvement in 20 of the 147 patients (14 %).

In the 119 patients who showed complete concordance, the main locations of the primary melanoma were the left zygomatic and infraorbital region in 19 patients (16 %), the left ear in 12 patients (10 %) and the right zygomatic and infraorbital region in 12 patients (10 %). In 3 of the 148 patients (2 %), in-transit or aberrant SLNs were visualized by lymphoscintigraphy, but among the 929 patients comprising the reference database only one such SLN was reported for the primary melanoma lesions (Fig. 4). Of the 11 discordant patients, 4 (36.4 %) showed metastatic involvement in the SLN. The clinical details of these 11 patients are shown in Table 3. In the 119 patients with complete concordance, the SLN were metastatic in 12 (10 %). Finally, in the 18 patients with partial concordance, the SLN were metastatic in 3 (17 %). The lymphatic drainage observed on lymphoscintigraphy performed in our department was ipsilateral in 129 patients (85 %), contralateral in 3 patients (2 %) and bilateral in 20 patients (13 %). The location of the primary melanoma in patients with contralateral drainage was the left ear, right parietal scalp and left cervical level I. In the patients with contralateral drainage, the theoretical pattern provided by the software tool showed the same drainage in one patient and was discordant in the other two patients.

Concordance Of the 152 patients, 4 (3 %) had primary melanoma in anatomical regions that were not registered in the Sydney Melanoma Unit database, so concordance could not be determined. Two of these four patients had primary melanoma in the nasal dorsum, one in the central frontoparietal scalp and another in the right lower lip. The remaining 148 patients were deemed suitable for concordance analysis. There was a total match between software-predicted drainage and drainage found in the patient in 119 of the 148 patients (80.4 %, 95 % CI 73.3 – 86 %). Concordance was partial (some concordant nodes and some not) in 18 of the 148 patients (12.2 %, 95 % CI 7.8 – 18.4 %). Considering total and partial concordance together, concordance was seen in 137 of 148 patients (92.6 %, 95 % CI 87.2 – 95.8 %). The concordance analysis comparing the theoretical tool and practice was excellent, yielding a kappa coefficient of 0.84 (0.79 – 0.91). Fig. 3 Difficulty in distinguishing the real SLN from second-tier nodes in a patient with a left malar melanoma: a early image, b delayed image

Discussion Sappey’s work on lymphatic drainage of skin was accepted as correct by the medical community for more than a century. In the 1970s, new information became available based on lymphoscintigraphic studies in melanoma patients. These studies showed that lymphatic drainage is unpredictable, especially in the trunk and in the head and neck areas. Some authors even showed that following Sappey’s guidelines would predict drainage to the wrong lymph node field in as many as 30 % of patients [3]. So lymphoscintigraphy now plays an important role in defining the pattern of potential metastatic spread through the lymphatic system from the primary lesion site and identifying the regional nodal basins at risk [4].

Eur J Nucl Med Mol Imaging

Fig. 4 Correlation with software-predicted drainage in a patient with a left frontal melanoma: a early image, b delayed image. c University of Auckland software prediction. Concordance with the predicted drainage was complete in this patient

The classical drainage patterns of cutaneous head and neck melanoma were described by O’Brien et al. [19] based on a

consecutive series of 183 neck dissection specimens. Many institutions adopted this scheme for planning therapeutic neck dissections. Currently, many lymphoscintigraphy studies have confirmed the clinically predicted drainage pattern according to the study of O’Brien et al. [19]. However, these studies also demonstrated unexpected drainage patterns in 8 – 43 % of patients, with the postauricular, suboccipital, contralateral, and distant nodal areas being the most common sites of discordance [20]. The staging of primary melanomas as intermediate thickness (1.2 – 3.5 mm) according to the results of SLN biopsy provides important prognostic information and identifies patients with nodal metastases whose survival can be prolonged by immediate lymphadenectomy [21]. Lymphoscintigraphy is an essential component of preoperative SLN identification in melanoma and breast cancer. Any lymph node with direct drainage from the primary tumour (i.e. the SLN) can be considered as a possible site of metastasis. Meticulously performed lymphoscintigraphy is essential for reliable SLN biopsy and serves several purposes: to point out the draining lymph node field at risk for metastatic disease, to indicate the number of SLNs, to help distinguish first-tier nodes from secondary nodes, to detect SLNs in unpredictable locations, and to mark the location of the SLN on the skin [7]. Since SLN biopsy using presurgical lymphoscintigraphy has been shown to enable accurate staging of the regional lymph nodes, it has been accepted that this imaging method is a reliable and reproducible approach for lymphatic assessment from a melanoma to the draining SLN. The SLN has been shown to be predictive of nodal involvement and its pathological status is a key prognostic factor in patients with primary melanoma. SLN biopsy identifies node-negative patients in whom further treatment may not be indicated, and thus reduces the number of unnecessary lymphadenectomies and avoids complications such as lymphoedema, delayed wound healing, infection and pain [9, 11]. Furthermore, SLN biopsy can identify clinically node-negative patients who actually have occult nodal metastases. SLN biopsy has substantially improved the accuracy of lymph node staging in patients with melanoma and has been shown in the large Multicentre Selective Lymphadenectomy Trial (MSLT-I) to increase disease-free survival. In this trial of 1,269 patients with primary melanoma of intermediate thickness, the mean 5-year disease-free survival rate was 78.3 % in the SLN group and 73.1 % in the observation group (hazard ratio, HR, for recurrence 0.74, 95 % CI 0.59 – 0.93; p=0.009). This trial also showed an improvement in overall survival in patients with metastatic nodal disease who had immediate complete surgical clearance of the entire lymph node field (i.e. all lymph nodes located within the region where the SLN was located, for example, the axilla or groin). In patients with nodal metastases, the 5-year survival rate was higher in those who had immediate lymphadenectomy compared with those in whom lymphadenectomy was delayed (72.3 % vs

Eur J Nucl Med Mol Imaging Table 3

Clinical details in 11 patients showing discordance between software-predicted lymphatic drainage and drainage found in practice

Patient Location of the primary melanoma Lymphatic drainage found

1

Left submandibular and submental region

2

Left submandibular and submental region

3 4

Right submandibular and submental region Right temporal scalp

5

Right temporal scalp

6

Left anterior triangle of the neck

7

Right anterior triangle of the neck

8

Left ear

9

Left ear

10

Left ear

11

Right parietal scalp

Left supraclavicular

Drainage options provided by the software

Right cervical levels III and IV Right supraclavicular Left cervical levels II, III, IV and V Submental Left supraclavicular Right cervical levels III and IV Right supraclavicular Left cervical levels II, III, IV and V Submental Right preauricular Right cervical levels II and IV Submental Right cervical level I Right cervical levels II, IV and V Right supraclavicular Right preauricular Right postauricular Right occipital Interval node Right cervical level I Right cervical levels II, IV and V Right supraclavicular Right preauricular Right postauricular Right occipital Interval node Left preauricular Left cervical levels I, II, III and IV Left supraclavicular Left axilla Right preauricular Right cervical levels II, III and IV Left cervical levels I and III Left supraclavicular Submental Right cervical levels I, II, III Left cervical levels I, II, III, IV and V Left supraclavicular Left occipital Left preauricular Left postauricular Right cervical levels I, II, III Left cervical levels I, II, III, IV and V Left supraclavicular Left occipital Left preauricular Left postauricular Right cervical levels I, II, III Left cervical levels I, II, III, IV and V Left supraclavicular Left occipital Left preauricular Left postauricular Right cervical level I Right cervical levels II, III, IV and V Right postauricular Right preauricular Interval node

52.4 %, HR for death 0.51, 95 % CI 0.32 – 0.81; p=0.004) [22]. Previous studies have shown a reproducibility rate in the range 80 – 96 % with the conclusion that lymphoscintigraphy is not always reproducible and thus may sometimes produce false-negative SLN results. In this regard, the greatest variability in lymphatic drainage pathways has been found in patients with head and neck or trunk cutaneous melanoma [23, 24]. SLN biopsy in the head and neck region is challenging due to the great number of lymphatic nodes (>300) that are found in this area. The huge network of lymph vessels and their routes also contribute to the complexity of lymph drainage in this area, with melanomas often draining to multiple,

Number of SLN Metastases

1

1 SLN negative

2

2 SLN negative

2

2 SLN negative

2

2 SLN negative

4

4 SLN negative

2

1 SLN positive 1 SLN negative

3

3 SLN negative

1

1 SLN positive

3

2 SLN positive 1 SLN negative

4

4 SLN negative

4

3 SLN positive 1 SLN negative

aberrant and bilateral lymph nodes [25, 26]. Furthermore, during lymphoscintigraphy, overprojection from radioactivity at the site of injection may obscure a SLN on the images. Because of the high number of lymph nodes in this region it is often hard to distinguish SLNs from second-echelon nodes. Thus, in a review of 32 different studies relating to more than 3,400 head and neck melanomas, the sensitivity of the SLN biopsy in this area was found to be in the range 80 – 100 % and the false-negative rate as high as 20 % [27]. These issues not only make the procedure technically difficult in this particular region, but also generate controversy about the extent of neck dissection in those patients with involved lymph

Eur J Nucl Med Mol Imaging

nodes. Of the 11 patients with discordant findings between lymphoscintigraphy and the drainage predicted by the Auckland University software, 8 patients were assessed in the first half of the study period (between 1998 and 2005) and the remaining 3 patients in the second half of the period (between 2006 and 2012). Lymphoscintigraphy not only provides a way of accurately identifying the SLN on a direct drainage pathway from the site of the primary melanoma, but can also highlight those nodes draining through unpredictable pathways [28]. Common to most nuclear medicine imaging methods, however, anatomical detail is lacking. The introduction of hybrid SPECT/CT offers a solution to this problem. In several recent studies looking into the application of SPECT/CT for lymphatic mapping in melanoma, additional SLNs not identified by planar imaging were detected [29, 30]. SPECT/CT provided additional anatomical information in 31 % of patients and the surgical approach was altered in 29 %. In a similar study, Vermeeren et al. identified additional SLNs in 8 % of patients and 30 % of those nodes contained melanoma metastases. SPECT/CT led to a change in the surgical approach in 35 % of patients [30]. Although in our centre we have been using SPECT/CT since 2007 in clinical practice, it was not the aim of this study to compare the SPECT/CT results with conventional planar images. Thus, SPECT/CT data were not taken into account. The inclusion of these data could have improved the current results and it might be of interest to compare in another study concordance rates based on 3D SPECT/CT information and the Sydney/Auckland model. In our study we used as reference the Sydney’s Melanoma Unit database, which has the largest collection of preoperative lymphoscintigraphy studies performed in over 5,000 patients with melanoma [3]. The Unit’s database includes the precise location of the primary melanoma and the location of every SLN in each patient, located by lymphoscintigraphy. These data have allowed drainage patterns to be tabulated and 2D displays to be generated relating melanoma sites to draining lymph node fields [6]. These models demonstrate quantitatively how the patterns of lymphatic drainage of the skin differ from traditional anatomical descriptions [14, 15, 31]. In our series, concordance was complete in 80.4 % of patients for all SLNs and partial in 12.2 % of patients. Discrepancy was observed in 7.4 % of patients. Thus, a strong point in favour of lymphoscintigraphy is that occasionally aberrant drainage can be reliably identified [32]. On the other hand, a hot spot does not always represent a lymph node. It can also be caused by accidental spillage from the injection site or a drop of radiotracer, as well as the well-known lymphatic lakes. So meticulous procedure during lymphoscintigraphy is mandatory for optimum results. When used in conjunction with SPECT/CT (for its detailed anatomical information) the procedure improves confidence and helps preoperative planning in the detection and sampling of metastatic nodes [32, 33].

Conclusion The use of a large validated database developed for melanoma of the head and neck enabled the highly accurate prediction of the potential lymphatic drainage coming from a definite location on the skin. Lymphoscintigraphy provides a particular added value to personalize lymphatic mapping in each patient. For this reason, it is essential to carry out preoperative lymphoscintigraphy to accurately detect the SLNs in head and neck melanoma, especially when faced with aberrant lymphatic drainage in unpredictable locations. Compliance with ethical standards Funding This work was funded in part by an Agència de Gestió d'Ajuts Universitaris i de Recerca (AGAUR) grant. AGAUR 2014 SGR 279. Conflicts of interest None. Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the principles of the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Informed consent Informed consent was obtained from all individual participants included in the study.

References 1.

2. 3.

4.

5.

6.

7.

8.

9.

Forschner A, Eigentler TK, Pflugfelder A, Leiter U, Weide B, Held L, et al. Melanoma staging: facts and controversies. Clin Dermatol. 2010;28:275–80. Thompson JF, Scolyer RA, Kefford RF. Cutaneous melanoma. Lancet. 2005;365:687–701. Thompson JF, Uren RF. Lymphatic mapping in management of patients with primary cutaneous melanoma. Lancet Oncol. 2005;6:877–85. Morton DL, Wen DR, Wong JH, Economou JS, Cagle LA, Storm FK, et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg. 1992;127:392–9. Balch CM, Morton DL, Gershenwald JE, McMasters KM, Nieweg OE, Powell B, et al. Sentinel node biopsy and standard of care for melanoma. J Am Acad Dermatol. 2009;60:872–5. Krag DN, Meijer SJ, Weaver DL, Loggie BW, Harlow SP, Tanabe KK, et al. Minimal-access surgery for staging of malignant melanoma. Arch Surg. 1995;130:654–8. Thompson JF, McCarthy WH, Bosch CM, O'Brien CJ, Quinn MJ, Paramaesvaran S, et al. Sentinel lymph node status as an indicator of the presence of metastatic melanoma in regional lymph nodes. Melanoma Res. 1995;5:255–60. Uren RF, Howman-Giles R, Thompson JF. Patterns of lymphatic drainage from the skin in patients with melanoma. J Nucl Med. 2003;44:570–82. Wilt JH, Thompson JF, Uren RF, Ka VS, Scolyer RA, McCarthy WH, et al. Correlation between preoperative lymphoscintigraphy and metastatic nodal disease sites in 362 patients with cutaneous melanomas of the head and neck. Ann Surg. 2004;239:544–52.

Eur J Nucl Med Mol Imaging 10.

11.

12.

13. 14.

15.

16.

17. 18. 19.

20.

21.

22.

23.

Chakera AH, Hesse B, Burak Z, Ballinger JR, Britten A, Caracò C, et al. EANM-EORTC general recommendations for sentinel node diagnostics in melanoma. Eur J Nucl Med Mol Imaging. 2009;36: 1713–42. Willis AI, Ridge JA. Discordant lymphatic drainage patterns revealed by serial lymphoscintigraphy in cutaneous head and neck malignancies. Head Neck. 2007;29:979–85. Ferlito A, Robbins KT, Shah JP, Medina JE, Silver CE, Al-Tamimi S, et al. Proposal for a rational classification of neck dissections. Head Neck. 2011;33:445–50. Rigual NR, Wiseman SM. Neck dissection: current concepts and future directions. Surg Oncol Clin N Am. 2004;13:151–66. Reynolds HM, Smith NP, Uren RF, Thompson JF, Dunbar PR. Three-dimensional visualization of skin lymphatic drainage patterns of the head and neck. Head Neck. 2009;31:1316–25. Reynolds HM, Dunbar PR, Uren RF, Thompson JF, Smith NP. Mapping melanoma lymphoscintigraphy data onto a 3D anatomically based model. Ann Biomed Eng. 2007;35:1444–57. Vidal Sicart S, Brouwer OR, Valdes O. Evaluation of the sentinel lymph node combining SPECT/CT with the planar image and its importance for the surgical act. Rev Esp Med Nucl. 2011;5:331–7. Wilson EB. Probable inference, the law of succession, and statistical inference. J Am Stat Assoc. 1927;22:209–12. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20:37–46. O’Brien CJ, Petersen-Schaefer K, Ruark D, Coates AS, Menzie SJ, Harrison RI. Radical, modified, and selective neck dissection for cutaneous malignant melanoma. Head Neck. 1995;17(3):232–41. Klop WM, Veenstra HJ, Vermeeren L, Nieweg OE, Balm AJ, Lohuis PJ. Assessment of lymphatic drainage patterns and implications for the extent of neck dissection in head and neck melanoma patients. J Surg Oncol. 2011;103:756–60. Morton DL, Thompson JF, Cochran AJ, Mozzillo N, Elashoff R, Essner R, et al. Sentinel-node biopsy or nodal observation in melanoma. N Engl J Med. 2006;355:1307–17. Morton DL, Cochran AJ, Thompson JF, Elashoff R, Essner R, Glass EC, et al. Sentinel node biopsy for early-stage melanoma: accuracy and morbidity in MSLT-I, an international multicenter trial. Ann Surg. 2005;242:302–13. Vidal M, Vidal-Sicart S, Torrents A, Perissinotti A, Navales I, P a r e d e s P, e t a l . A c c u r a c y a n d r e p r o d u c i b i l i t y o f

lymphoscintigraphy for sentinel node detection in patients with cutaneous melanoma. J Nucl Med. 2012;53:1193–9. 24. Uren RF, Howman-Giles R, Chung DK, Morton RL, Thompson JF. The reproducibility in routine clinical practice of sentinel lymph node identification by pre-operative lymphoscintigraphy in patients with cutaneous melanoma. Ann Surg Oncol. 2007;14:899–905. 25. Fincher TR, O’Brien JC, McCarty TM, Fisher TL, Preskitt JT, Lieberman ZH, et al. Patterns of drainage and recurrence following sentinel lymph node biopsy for cutaneous melanoma of the head and neck. Arch Otolaryngol Head Neck Surg. 2004;130:844–8. 26. Lin D, Franc BL, Kashani-Sabet M, Singer MI. Lymphatic drainage patterns of head and neck cutaneous melanoma observed on lymphoscintigraphy and sentinel lymph node biopsy. Head Neck. 2006;28:249–55. 27. De Rosa N, Lyman GH, Silbermins D, Valsecchi ME, Pruitt SK, Tyler DM, et al. Sentinel node biopsy for head and neck melanoma: a systematic review. Otolaryngol Head Neck Surg. 2011;145:375– 82. 28. Veenstra HJ, Klop WM, Lohuis PJ, Nieweg OE, van Velthuysen ML, Balm AJ. Cadaver study on the location of suboccipital lymph nodes: guidance for suboccipital node dissection. Head Neck. 2014;36:682–6. 29. Veenstra HJ, Klop MW, Lohuis PJ, Nieweg OE, Valdés Olmos RA. Lymphatic drainage of melanomas located on the manubrium sterni to cervical lymph nodes: a case series assessed by SPECT/CT. Clin Nucl Med. 2013;38:137–9. 30. Vermeeren L, Valdés Olmos RA, Klop WM, van der Ploeg IM, Nieweg OE, Balm AJ, et al. SPECT/CT for sentinel lymph node mapping in head and neck melanoma. Head Neck. 2011;33:1–6. 31. Reynolds HM, Dunbar PR, Uren RF, Blackett SA, Thompson JF, Smith NP. Three-dimensional visualization of lymphatic drainage patterns in patients with cutaneous melanoma. Lancet Oncol. 2007;8:806–12. 32. Leong SP, Morita ET, Südmeyer M, Chang J, Shen D, Achtem TA, et al. Heterogeneous patterns of lymphatic drainage to sentinel lymph nodes by primary melanoma from different anatomic sites. Clin Nucl Med. 2005;30:150–8. 33. Fairbairn N, Munson C, Khan ZA, Butterworth M. The role of hybrid SPECT/CT for lymphatic mapping in patients with melanoma. J Plast Reconstr Aesthet Surg. 2013;66:1248–55.