TO THE EDITOR:
I read with interest the continuing education article by Uren et al. (1) stating that clinical prediction of lymphatic drainage from the skin is not possible and that the old clinical guidelines based on Sappey’s lines therefore should be abandoned. To the best of my knowledge, my former group at the Hospital of the Frankfurt Goethe University was the first ever to standardize scintigraphic mapping of lymphatic drainage in cutaneous tumors, particularly malignant melanoma, in the late 1970s and early 1980s (2–4). We were able to clearly document that not only tumors located inside but also outside lymphatic watersheds of the skin show an ambiguous lymphatic drainage, which is practically unpredictable by conventional anatomic guidelines in individual patients. We concluded that the anatomic thesis of lymphatic watersheds should be revised. In more than 90% of our patients with skin lesions on the trunk, one or both axillary lymph node groups were found to be involved in lymphatic drainage, either solely or combined with inguinal, supraclavicular, posterior cervical, parasternal, or other node-bearing areas or in-transit lymph nodes; hence, the axillary lymph node groups as the “center in lymphatic drainage from the truncal skin in man” should attract our greatest attention in melanomas or other cutaneous tumors of the trunk independent of their topographic position (3). Our data on the lymphatic drainage patterns in skin tumors of trunk, head and neck, and upper and lower limbs published some 20 y ago were proven to be true (1,5).
Detection and localization of “true” sentinel lymph nodes, permitting correct staging of regional lymph nodes, is essential for management and prognostic assessment in malignant melanoma. In 43 of the 100 melanoma patients examined prospectively, additional information was obtained by simple temporary lead shielding of hot spots in lymphatic drainage areas, applied in combination with dynamic acquisition in various views: In 7 patients, the exact course of lymph vessels could be mapped only after shielding; in 3 patients, hot spots in the drainage area proved to be lymph vessels, lymph vessel intersections, or lymph vessel ectasias; in 33 patients, 1 or 2 additional sentinel lymph nodes that showed less tracer accumulation or were smaller (<1.5 cm) were detected after shielding by visualization of their own lymph vessels (7% sentinel lymph node metastases) (6). Preliminary data from another prospective study on 276 melanoma patients indicated that the time of scintigraphic appearance of sentinel lymph nodes is a clinically relevant factor for prediction of metastatic spread to sentinel lymph nodes, provided the time of appearance is assessed under standardized conditions (7). However, larger numbers of patients need to be examined to truly evaluate the benefit of the time of scintigraphic appearance compared with other predictors of sentinel lymph node tumor positivity.
Finally, we have created a classification of the lymphatic drainage status of primary tumors that preferably metastasize via their draining lymph vessels (8). The classification is based on the number of sentinel lymph nodes and their locations (node group or in-transit node) and comprises 4 classes (D-class I–IV) and distinct subclasses (A–E): For example, D-IA means 1 draining node location (NL) and 1 sentinel lymph node (SN); D-IIA means 2 NL, 2 × 1 SN; D-IIIB means 3 NL, 1 × >1 SN; and D-IVE means ≥4 NL, ≥4 × 1 SN. The classification is easy to learn and reliably reproducible using various approaches (e.g., γ-camera imaging, γ-probe detection, or dye mapping). We are currently testing its diagnostic, prognostic, and therapeutic value in prospective studies on melanoma and breast cancer patients and encourage others to join us.
REFERENCES
REPLY:
The group from the Goethe University in Frankfurt have reported what many who have studied the lymphatic drainage of the skin have found, and that is the variability of drainage from one person to another. In the late 1700s Mascagni (1) observed lymph drainage across the midline of the body, and in 1903 Delamere et al. (2) described “accessory channels” draining the trunk to supraclavicular nodes from the anterior trunk and drainage from the upper back over the shoulders to neck nodes.
After the development of lymphoscintigraphy by Sherman and Ter-Pogossian (3) in 1953, this physiologic approach to lymphatic mapping was applied to individual patients with melanoma. Fee et al. (4) in 1978 and Meyer et al. (5) in 1979 described lymphatic mapping using lymphoscintigraphy in melanoma patients to determine the pattern of drainage in individual patients and thus to determine which lymph node field to dissect. Many others over the years, including Munz and Hör (6) from Frankfurt, have continued this work. More recently, completely new lymphatic drainage pathways from the skin have been discovered (7).
The challenge today is to apply the techniques carefully in individual patients so that all true sentinel nodes are located for surgical removal and careful histologic examination. An understanding of the possible drainage pathways from each area of skin will make this more likely.
It is interesting that the Frankfurt group has published data (8) that suggest that the speed of lymph flow through lymphatic collecting vessels influences the likelihood that metastases will be found in the draining sentinel nodes (SNs). We have measured the speed of lymph flow, in centimeters per minute, on dynamic imaging in 198 patients with melanoma (9), and though we found that lymph flow rates vary systematically throughout the body, we did not find this variance to have any influence on the incidence of metastasis in the draining SNs. The fastest flow occurred from the foot and leg, with an average flow rate of 10.2 cm/min in our study, yet the incidence of metastasis in groin SNs is the same as in other node fields.
We agree that the best method of identifying a true SN on lymphoscintigraphy is to visualize the lymphatic vessel passing directly to the SN on dynamic imaging. Some find star artifacts a problem when the injection site is in the field of view, and this problem can be overcome by shielding; however, we find this cumbersome because we perform many studies every day. We have found that using a super-high-resolution collimator with a septal penetration of less than 1% at 140 keV solves the problem. Lymphatic vessels can clearly be visualized without the need to shield the injection site.
Finally, we have found that all SNs, regardless of their location, can contain metastatic disease and that the incidence varies with the thickness of the primary melanoma and presence or absence of ulceration. We have not found that the number of sentinel nodes at each site or their exact location has any effect on this incidence. We therefore suspect that use of a rather complex classification system based on SN numbers and location will not provide useful prognostic or therapeutic information.