Combined analysis of phase III trials evaluating [⁹⁹mTc]tilmanocept and vital blue dye for identification of sentinel lymph nodes in clinically node-negative cutaneous melanoma

Abstract

Background: [(99m)Tc]Tilmanocept is a CD206 receptor-targeted radiopharmaceutical designed for sentinel lymph node (SLN) identification. Two nearly identical nonrandomized phase III trials compared [(99m)Tc]tilmanocept to vital blue dye.

Methods: Patients received [(99m)Tc]tilmanocept and blue dye. SLNs identified intraoperatively as radioactive and/or blue were excised and histologically examined. The primary end point, concordance, was the proportion of blue nodes detected by [(99m)Tc]tilmanocept; 90 % concordance was the prespecified minimum concordance level. Reverse concordance, the proportion of radioactive nodes detected by blue dye, was also calculated. The prospective statistical plan combined the data from both trials.

Results: Fifteen centers contributed 154 melanoma patients who were injected with both agents and were intraoperatively evaluated. Intraoperatively, 232 of 235 blue nodes were detected by [(99m)Tc]tilmanocept, for 98.7 % concordance (p < 0.001). [(99m)Tc]Tilmanocept detected 364 nodes, for 63.7 % reverse concordance (232 of 364 nodes). [(99m)Tc]Tilmanocept detected at least one node in more patients (n = 150) than blue dye (n = 138, p = 0.002). In 135 of 138 patients with at least one blue node, all blue nodes were radioactive. Melanoma was identified in the SLNs of 22.1 % of patients; all 45 melanoma-positive SLNs were detected by [(99m)Tc]tilmanocept, whereas blue dye detected only 36 (80 %) of 45 (p = 0.004). No positive SLNs were detected exclusively by blue dye. Four of 34 node-positive patients were identified only by [(99m)Tc]tilmanocept, so 4 (2.6 %) of 154 patients were correctly staged only by [(99m)Tc]tilmanocept. No serious adverse events were attributed to [(99m)Tc]tilmanocept.

Conclusions: [(99m)Tc]Tilmanocept met the prespecified concordance primary end point, identifying 98.7 % of blue nodes. It identified more SLNs in more patients, and identified more melanoma-containing nodes than blue dye.

Figures

Fig. 1

Chemical structure of [99mTc]tilmanocept. [99mTc]Tilmanocept is composed of a dextran backbone (black) to which are attached multiple units of mannose (green) and DTPA (blue). The mannose units provide a molecular mechanism by which [99mTc]tilmanocept avidly binds to a receptor specific to reticuloendothelial cells (CD206), and the DTPA units provide a highly stable means to radiolabel tilmanocept with 99mtechnetium (red). The molecular weight of [99mTc]tilmanocept is approximately 19,000 g/mol; the molecular diameter is 7.1 nm (adapted from Wallace et al.16)

Fig. 2

Preoperative lymphoscintigraphy with [99mTc]tilmanocept. Images were obtained over 24 h in a 72-year-old melanoma patient undergoing SLN biopsy for a 0.95-mm nonulcerated melanoma. The patient received four intradermal injections, each consisting of 0.49 mCi of [99mTc]tilmanocept, at the primary site on the left forearm (first image, top row) and one axillary SLN was visualized at 20 min (second through fourth images, top row), along with activity in lymphatic channels of the upper arm seen superior and distal to the SLN. The same SLN visualized at 20 min was observed at 24 h (bottom row). Intraoperatively, approximately 1 h after the final image was obtained, the tissue background 10-second count was 50 (3σ = 71, see text) and a single blue axillary node with 10-s counts of 1,521 was found corresponding to the SLN on the images. This node had no histologic evidence of melanoma

Fig. 3

The 3σ rule. a The equation for determining whether the in vivo radioactive counts in a lymph node are estimated to be at least three standard deviations above the normal tissue background count. b Example of 3σ calculation for a protocol patient with four SLNs identified intraoperatively. All in vivo node counts were greater than the 3σ value, and all in vivo counts were >250. All were therefore considered hot per the protocol definition. c Schematic of injections of vital blue dye (blue syringe) and [99mTc]tilmanocept (gold syringe) intradermally at the primary tumor site (brown circle) with lymphatic mapping with a handheld gamma detection probe. In this example, nodes 1 through 3 are shown as hot and blue (i.e., per lymph node concordance 3/3 = 100 %; per patient concordance = yes), while node 4 is only hot (i.e., reverse concordance 3/4 = 75 %)

Fig. 4

Distribution of excised lymph nodes and concordance of [99mTc]tilmanocept with vital blue dye. a Lymph node distribution, based on a total of 379 excised lymph nodes. Twelve lymph nodes removed from nine patients were not blue and did not meet the protocol definition for radioactivity above background; none contained melanoma. b Per lymph node concordance, based on a total of 235 blue lymph nodes. The statistical threshold for concordance was prospectively set as 90 % of blue nodes would be hot. (In b and c, the red line marks the 90 % concordance threshold.) c Per patient concordance, based on 138 patients with at least one blue lymph node. d Per lymph node reverse concordance, based on a total of 364 hot nodes. Only 232 (63.7 %) were also blue. e Tumor-positive lymph node distribution, based on 45 lymph nodes found to contain melanoma by routine histology and/or immunohistochemistry. No tumor-positive nodes were blue but not hot