Use of indocyanine green for detecting the sentinel lymph node in breast cancer patients: from preclinical evaluation to clinical validation

Abstract

Assessment of the sentinel lymph node (SLN) in patients with early stage breast cancer is vital in selecting the appropriate surgical approach. However, the existing methods, including methylene blue and nuclides, possess low efficiency and effectiveness in mapping SLNs, and to a certain extent exert side effects during application. Indocyanine green (ICG), as a fluorescent dye, has been proved reliable usage in SLN detection by several other groups. In this paper, we introduce a novel surgical navigation system to detect SLN with ICG. This system contains two charge-coupled devices (CCD) to simultaneously capture real-time color and fluorescent video images through two different bands. During surgery, surgeons only need to follow the fluorescence display. In addition, the system saves data automatically during surgery enabling surgeons to find the registration point easily according to image recognition algorithms. To test our system, 5 mice and 10 rabbits were used for the preclinical setting and 22 breast cancer patients were utilized for the clinical evaluation in our experiments. The detection rate was 100% and an average of 2.7 SLNs was found in 22 patients. Our results show that the usage of our surgical navigation system with ICG to detect SLNs in breast cancer patients is technically feasible.

Figure 1. Schematic diagram of the surgical navigation system and its application in surgery.

a. The principle figure of the surgical navigation system clarifying the operation course of the system. When the LED light illuminated the surgical area, the ICG dye emitted NIR light. The emission and reflection of the halogen light went through the lens to the prism. Then, the light was equally divided into two beams by the prism. One beam went through the filter to the color CCD and the other to the EMCCD. All of the data collected from the CCD were transferred to the computer, and the computer controlled the CCD. b. The hardware of the surgical navigation system. c. User interface of the software offering exposure time and auto capture interval time parameter settings. d. Image acquisition interface as an example of the capture mode results. e. Preoperative preparation in the operating room. f. Intraoperative diagnosis with a surgical navigation system carried out during the surgery.

Figure 2. Pharmacokinetic experiments on rabbits.

Five pharmacokinetic experiments using ICG with five concentrations were done on rabbits with the injection doses of 0.1 ml in the areolar area. Our software to evaluate the changes of the light intensity value at each time point tested the light intensity of SLN in rabbits. The results are shown in Figures a-e. The error bars mean the variance of light intensity from 3 rabbit experiments per group at a certain concentration and time point. Then, we entered data into the computer program Prism 5.0 (GraphPad-Prism). The light intensity statistics were performed using the software. From the results we could obtain the time point of the surgery and the effective time of the operation.

Figure 3. The SLN resection experiments in nude mice.

There were two groups of figures directly acquired by the CCD cameras. The last column images were obtained after the processing of the two images in the front. When the ICG solution with a concentration of 1mg/ml was injected into the third areola of a nude mouse, ten minutes later we got a. fluorescent image for mapping the SLN. At the same time, we got b. the color image. According to the software computation, the pseudo-green fluorescent image was overlaid on top of the color image. The result of the image fusion was c. the overlay image. After dissection of the SLN, d. fluorescent image and e. color image were acquired simultaneously. From the f. overlay image, we could clearly see the fluorescence information in the color image. All dissections were sent in for pathological examination. All of the tissue sections were judged to be the SLN.

Figure 4. The SLN resection experiments in rabbits.

When the SLN was dissected, we took a photo as shown in a. fluorescent image. The b. color image was acquired at the same time. c. The overlay image also showed the fluorescent position in the color image. After resection, the SLN was put on medical gauze. It was shining brightly as d. the fluorescent image. Although in the e. color image there was no difference in the light, the f. overlay image showed that it was illuminated. All of the dissections were sent in for pathological examination. All of the tissue sections were judged to be the SLN.

Figure 5. ICG-guided intraoperative detection and resection of the SLN in humans.

According to the preclinical trials, 22 cases of patients were taken from the SLNB surgery. In the beginning, the ICG solution was injected into the areolar region. About 3 minutes later, the lymphatic drainage and SLN would be clearly displayed on the monitor as shown in a. the fluorescent image in vivo. Because near-infrared light is not visible, there was no light information in the b. color image in vivo. Through the software of the surgical navigation system, the location of the SLN is shown in c. where the overlay image in vivo could be distinguished accurately. According to the guidelines of the fluorescent image, the surgery could quickly find the location of the SLN. d. The fluorescent image was captured before dissection. From the e. color image and the f. overlay image, SLN could be located with tweezers. The SLN was carefully removed and put on gauze. With the near-infrared light irradiation, the SLN was bright as shown in g. the fluorescent image during dissection. Such a visible image was displayed in h. the color image during dissection. Finally, the merged image of the pseudo-green fluorescence image and the color image is shown in i. the overlay image during dissection. All dissections were sent in for pathological examination. All of the tissue sections were judged to be the SLN.

Figure 6. The normal and tumor-metastasis pathology slices of SLN dissected by ICG-guided surgery.

All of the dissected SLNs were sent in for pathological examination. After the conventional Hematoxylin-Eosin (HE) staining, the results proved that all of the dissected tissue specimens were lymph nodes. Figure a. shows normal sentinel lymph node cells with no cancer metastasis. Figure b. shows infiltrating ductal breast cancer.