806 http://oncology.thelancet.com Vol 8 September 2007 Articles Three-dimensional visualisation of lymphatic drainage patterns in patients with cutaneous melanoma Hayley M Reynolds, P Rod Dunbar, Roger F Uren, Shane A Blackett, John F Thompson, Nicolas P Smith Summary Background Lymphoscintigraphy accurately maps lymphatic drainage from sites of cutaneous melanoma to the draining sentinel lymph nodes. The Sydney Melanoma Unit has accumulated lymphoscintigraphy data from over 5000 patients with cutaneous melanoma over more than 15 years, collectively revealing patterns of skin lymphatic drainage. We aimed to map these data onto a three-dimensional computer model to provide improved visualisation and analysis of lymphatic drainage from sites of cutaneous melanoma. Methods Lymphoscintigraphy data from 5239 patients with cutaneous melanoma were collected between July 27, 1987 and Dec 16, 2005. 4302 of these patients had primary melanoma sites below the neck, and were included in this analysis. From these patients, two-dimensional lymphoscintigraphy data were mapped onto an anatomically based three-dimensional computer model of the skin and lymph nodes. Spatial analysis was done to visualise the relation between primary melanoma sites and the locations of sentinel lymph nodes. Findings We created three-dimensional, colour-coded heat maps that showed the drainage patterns from melanoma sites below the neck to individual lymph-node fields and to many lymph-node fields. These maps highlight the inter- patient variability in skin lymphatic drainage, and show the skin regions in which highly variable drainage can occur. To enable interactive and dynamic analysis of these data, we also developed software to predict lymphatic drainage patterns from melanoma skin sites to sentinel lymph-node fields. Interpretation The heat maps confirmed that the commonly used Sappey’s lines are not effective in predicting lymphatic drainage. The heat maps and the interactive software could be a new resource for clinicians to use in preoperative discussions with patients with melanoma and other skin cancers that can metastasise to the lymph nodes, and could be used in the identification of sentinel lymph-node fields during follow-up of such patients. Introduction Cutaneous melanoma is a potentially fatal disease which has increased in incidence over recent decades in most people of European background. 1 Detection of metastatic melanoma in the regional lymph nodes has major implications for treatment and prognosis. Sentinel lymph-node biopsy (SLNB) 2 is used to detect whether melanoma cells have metastasised to the sentinel lymph nodes (SLNs), which are defined as any lymph node receiving direct lymphatic drainage from a primary tumour site. SLNs are located by preoperative lympho- scintigraphy, which involves imaging the lymphatic drainage from a primary melanoma site to the SLNs through a radioactive tracer injected into the skin. 2 SLNB 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 improve disease-free survival. 3 In this trial of 1269 patients with primary melanoma of intermediate thickness, the mean 5-year disease-free survival was 78·3% (SE 1·6) in the SLNB group and 73·1% (SE 2·1) 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 for patients with metastatic nodal disease who had an immediate complete surgical clearance of the entire lymph-node field 3 (ie, all lymph nodes located within the region where the SLN is located, for example, the axilla or groin). In patients with nodal metastases, the 5-year survival was higher in those who had immediate lymphadenectomy compared with those in whom lymphadenectomy was delayed (72·3% [SE 4·6] vs 52·4% [SE 5·9]; HR for death 0·51; 95% CI 0·32–0·81; p=0·004). However, not all centres have access to lympho- scintigraphy, and clinical follow-up in patients with melanoma then relies on predictions of lymphatic drainage based on historical assumptions that are probably incorrect in 30% of individuals. 4 Lymphoscintigraphy studies have confirmed that lymphatic drainage of the skin is highly variable between patients, with very few areas of the skin from which lymphatic drainage is clinically predictable. 5 For over 100 years, patterns of lymphatic drainage from the skin were predicted from the work of Sappey, 6 whose 1874 atlas stated that lymphatic drainage never crossed the midline of the body nor a theoretical horizontal line drawn around the waist through the umbilicus. These concepts were challenged in the 1970s, 1980s, and 1990s, largely due to work in patients with melanoma by the use of lymphoscintigraphy, which showed that lymphatic drainage frequently occurs across Sappey’s lines, and that, although skin sites usually drain to ipsilateral lymph-node fields, contralateral drainage is not uncommon. 7–12 Lancet Oncol 2007: 8: 806–12 Published Online July 20, 2007 DOI:10.1016/S1470- 2045(07)70176-6 See Reflection and Reaction page 755 Bioengineering Institute (H M Reynolds BE, S A Blackett ME, N P Smith PhD), and School of Biological Sciences, University of Auckland, Auckland, New Zealand (P R Dunbar PhD); Sydney Melanoma Unit, Royal Prince Alfred Hospital, Camperdown, Sydney, NSW, Australia (Prof R F Uren MD, Prof J F Thompson MD); Discipline of Medicine, University of Sydney, Sydney, NSW, Australia (Prof R F Uren); Nuclear Medicine and Diagnostic Ultrasound, RPAH Medical Centre, Sydney, NSW, Australia (Prof R F Uren); Discipline of Surgery, University of Sydney, Sydney, NSW, Australia (Prof J F Thompson); and University Computing Laboratory, University of Oxford, Oxford, UK (N P Smith PhD) Correspondence to: Ms Hayley M Reynolds, Bioengineering Institute, University of Auckland, Auckland, New Zealand h.reynolds@auckland.ac.nz