Multimodal hybrid imaging agents for sentinel node mapping as a means to (re)connect nuclear medicine to advances made in robot-assisted surgery

Abstract

Purpose: Radical prostatectomy and complementary extended pelvic lymph node dissection (ePLND) of sentinel lymph nodes (SNs) and non-sentinel lymph nodes (LNs) at risk of containing metastases are increasingly being performed using high-tech robot-assisted approaches. Although this technological evolution has clear advantages, the physical nature of robotic systems limits the integrated use of routine radioguided surgery technologies. Hence, engineering effort in robotics are focused on the integration of fluorescence guidance technologies. Using the hybrid SN tracer indocyanine green-(99m)Tc-nanocolloid (radioactive and fluorescent), for the first time in combination with a robot-integrated laparoscope, we investigated whether the robot-assisted approach affects the accuracy of fluorescence detection of SNs identified preoperatively using nuclear medicine.

Methods: The study included 55 patients (Briganti nomogram-based risk >5 % on LN metastases) scheduled for robot-assisted radical prostatectomy, SN biopsy and ePLND. Following indocyanine green-(99m)Tc-nanocolloid injection, preoperative nuclear imaging (lymphoscintigraphy and SPECT/CT) was used to locate the SN(s). The fluorescence laparoscope was used intraoperatively to identify the SN(s) with standard fluorescence settings (in 50 patients) and with customized settings (in 5 patients). The number and location of the SNs, the radioactive, fluorescence (both in vivo and ex vivo) and tumour status of the resected SNs/LNs, and postoperative complications were recorded and analysed.

Results: Combined, preoperative lymphoscintigraphy and SPECT/CT imaging identified 212 SNs (median 4 per patient). Intraoperative fluorescence imaging using standard fluorescence settings visualized 80.4 % (148/184 SNs; 50 patients; ex vivo 97.8 %). This increased to 85.7 % (12/14 SNs; 5 patients; ex vivo 100 %) with customized fluorescence settings. SPECT/CT images provided guidance towards the residual SNs. Ex vivo all removed SNs were radioactive. SNs were tumour-positive in 25.4 % of patients (14/55; false-negative rate 7 %, 1/14 patients). In ten patients, the SN was the only tumour-positive LN. Surgical complications were minimal.

Conclusion: Directly linking 3D preoperative nuclear imaging information on SNs to a robot-integrated fluorescence laparoscope improved the surgeon's use of the technology and did not influence the sensitivity or morbidity of the procedure. To our surprise, however, the detection rates with the current fluorescence camera did not improve.

Keywords: Fluorescence-guided surgery; Prostate cancer; Robot-assisted surgery; SPECT/CT; Sentinel node.

Fig. 1

Fluorescence-based SN identification: a white light image; b fluorescence-based image

Fig. 2

Locations of SNs and LNs detected intraoperatively. a, b SNs located inside and outside the ePLND area (green in vivo fluorescent SNs, yellow ​ex vivo identified SNs n = 50). c SNs that could not be removed (red, n = 55). d, e SNs located in the additionally included five patients inside and outside the ePLND area (green in vivo fluorescent SNs, yellow ​ex vivo identified SNs). f Location of tumour-positive SNs (black right-sided SNs, grey left-sided SNs). The images were generated using Visible Body software (Argosy Publishing, Newton Upper Falls, MA)

Fig. 3

Evaluation of adjustable fluorescence settings. SNs identified with various fluorescence settings