Feasibility of Indocyanine Green–Guided Localization of Pulmonary Nodules in Children with Solid Tumors

Abstract

Background:

Clearing all pulmonary metastases is essential for curing pediatric solid tumors. However, intraoperative localization of such pulmonary nodules can be challenging. Therefore, an intraoperative tool that localizes pulmonary metastases is needed to improve diagnostic and therapeutic resections. Indocyanine green (ICG) real-time fluorescence imaging is used for this purpose in adult solid tumors, but its utility pediatric solid tumors has not been determined.

Methods:

A single-center, open-label, nonrandomized, prospective clinical trial (NCT04084067) was conducted to assess the ability of ICG to localize pulmonary metastases of pediatric solid tumors. Patients with pulmonary lesions who required resection, either for therapeutic or diagnostic intent, were included. Patients received a 15-min intravenous infusion of ICG (1.5 mg/kg), and pulmonary metastasectomy was performed the following day. A near-infrared spectroscopy Iridium system was optimized to detect ICG, and all procedures were photo-documented and recorded.

Results:

ICG-guided pulmonary metastasectomies were performed in 12 patients (median age, 10.5 years). A total of 79 nodules were visualized, 13 of which were not detected by preoperative imaging. Histologic examination confirmed the following histologies: hepatoblastoma (n=3), osteosarcoma (n=2), and one each of rhabdomyosarcoma, Ewing sarcoma, inflammatory myofibroblastic tumor, atypical cartilaginous tumor, neuroblastoma, adrenocortical carcinoma, and papillary thyroid carcinoma. ICG guidance failed to localize pulmonary metastases in five (42%) patients who had inflammatory myofibroblastic tumor, atypical cartilaginous tumor, neuroblastoma, adrenocortical carcinoma, or papillary thyroid carcinoma.

Conclusions:

ICG-guided identification of pulmonary nodules is not feasible for all pediatric solid tumors. However, it may localize most metastatic hepatic tumors and high-grade sarcomas in children.

Introduction

The lung is the most common site of metastases for pediatric solid tumors.¹ Reliable localization of pulmonary nodules is crucial for accurate staging, diagnosis, relapse, and therapeutic intervention. Localization techniques for pulmonary nodules include computed tomography (CT)-guided placement of a localizer (e.g., a hook, wire, or coil). However, these approaches require transferring anesthetized patients from the interventional radiology (IR) suite to the operating room.² Moreover, IR localization is associated with the risks of positioning inaccuracy, migration of the localizer, pneumothorax, and pain.^2–8

Indocyanine green (ICG) is a water-soluble tricarbocyanine fluorophore that has an excellent safety profile and is approved by the U.S. Food and Drug Administration.⁹ Near-infrared (NIR) visualization of ICG can be performed either immediately after injection to assess the perfusion of tissues or in a delayed manner to exploit the differential retention of the ICG–plasma protein complex between tumor and normal tissue at 24 hours. Tumor retention of ICG at 24 hours is thought to be caused by the enhanced permeability and retention (EPR) properties of newly developed tumor vasculature, which has wide fenestrations that hold the ICG–plasma protein complex in the tumor mass.^10,11

Preclinical trials and adult clinical trials have shown that second-window imaging improves the precision of ICG-guided localization of pulmonary nodules.^12–21 More recently, pediatric studies have demonstrated promising sensitivity and specificity of this technique for localizing hepatoblastoma pulmonary metastases.^22–30 However, little is known about the feasibility of ICG-guided localization of other metastatic solid tumors. Therefore, we conducted a nonrandomized pilot study of the feasibility of ICG-guided localization of pulmonary metastases in a pediatric oncology population.

Materials and Methods

Study design and participants

The ICGLOW protocol was a prospective, nonrandomized, open-label, single-arm, single-center clinical trial (NCT04084067) that included subjects in eight histologic groups. The study was approved by the Institutional Review Board of St. Jude Children’s Research Hospital, and written informed consent was obtained from a parent/legal guardian and/or patient, as appropriate.

Patients with pulmonary metastases who required resection for therapeutic or diagnostic purposes were included. Patients with non–oncology-related lung pathology(i.e., benign pathologies) were excluded. The clinical trial used the second-window approach. Specifically, patients received a 15-min intravenous infusion of ICG (1.5 mg/kg), and thoracoscopic or open pulmonary metastasectomy was performed the following day. Patients with Wilms tumor metastases were excluded from the study because of the “inverse” fluorescence pattern observed in our previous study, in which there was a hypo-fluorescence of the primary tumor and high ICG avidity of surrounding healthy kidney tissue.³¹

After the surgeon identified the pulmonary nodules by visual and/or tactile methods, the infrared camera and excitation probe with NIR wavelength were used. ICG was visualized using an Iridium system (Visionsense Corp, Philadelphia, PA) optimized for detecting the fluorophore. Images of the tumor field were displayed on a video screen, and the iridium system software measured the spectral reading from the nodule by capturing the NIR signal it emitted.

The entire resection procedure was photo-documented and recorded. ICG NIR fluorescence was evaluated and validated by two surgeons. All intraoperative and postoperative monitoring of patients was prospectively recorded to analyze any ICG-related adverse events.

Statistics

The primary endpoint of the trial was the ability of ICG NIR imaging to detect the presence of metastatic nodules in the lungs. A detection rate of 75% was considered standard (null), and 90% was desirable. Simon’s two-stage minimax design was adopted with 80% power and a 5% significance level. In the first stage, the accrual of 22 patients was planned. If true-positive nodules were detected (i.e., by ICG avidity and histology confirmation) in 17 or more cases, then the trial would progress to the second stage. However, if fewer than 17 cases had true-positive nodules, then the study would be stopped for futility.

Results

Twelve patients (7 female and 5 male, median age 10.5 years [range <1–23 years]) underwent ICG-guided localization of pulmonary metastases. We detected 79 nodules intraoperatively. Histologic examination confirmed the following diagnoses: hepatoblastomas (n=3), osteosarcomas (n=2), and one each of rhabdomyosarcoma, Ewing sarcoma, inflammatory myofibroblastic tumor, atypical cartilaginous tumor, neuroblastoma, adrenocortical carcinoma, and papillary thyroid carcinoma.

ICG guidance failed to localize pulmonary metastases in five (42%) patients, triggering a futility-stopping rule and indicating that the method is unlikely to achieve the desired detection rate of 90%. ICG guidance successfully localized all cases of metastatic hepatoblastoma (3 patients) and high-grade sarcomas (4 patients), including osteosarcoma, Ewing sarcoma, and rhabdomyosarcoma. However, the following tumors were not ICG avid: inflammatory myofibroblastic tumor, atypical cartilaginous tumor, neuroblastoma, adrenocortical carcinoma, and papillary thyroid carcinoma (Table 1). ICG failed to detect 21 nodules, which were detected intraoperatively via direct visualization and/or palpation, without any fluorescent signal. The lesions were removed and then analyzed by a pathologist, who confirmed them as true-positive nodules.

Table 1.

ICG-guided localization of pulmonary metastases in 12 pediatric patients with solid tumors

Histology of pulmonary metastases	No. of patients	No. of nodules	Purpose of procedure	No. of true-positive nodulesa (No. patients)	No. of nodules not detected by preoperative CT scans but detected by ICG NIR imaging (No. of patients)	Mean nodule size in cms (range)	Tumor necrosis (%)
Hepatoblastoma	3	50	Tx	48 (3)	13 (1)	0.5 (0.1–1.8)	0–90
Osteosarcoma	2	4	Tx	4 (2)	0	2 (0.2–4)	10–30
Ewing sarcoma	1	1	Dx	1 (1)	0	1.4	0
Rhabdomyosarcoma	1	3	Dx	2 (1)	0	0.3b	0
Atypical cartilaginous tumor	1	1	Tx	0	0	0.9	0
Inflammatory myofibroblastic tumor	1	1	Dx	0	0	1.4	0
Neuroblastoma	1	17	Tx	0	0	0.5 (0.4–1.1)	0–5
Adrenocortical carcinoma	1	1	Tx	0	0	1.2	0
Papillary thyroid carcinoma	1	1	Dx	0	0	0.4	0

^aDetection rate at the patient level = 53.84% (95% CI: 25.63%–82.04%); that at the nodule level = 69.62% (95% CI: 59.84%–79.76%).

^bBoth true-positive rhabdomyosarcoma nodules were 0.3 cm.

Abbreviations: CI, confidence interval; CT, computed tomography; cms, centimeters; Dx, diagnosis; ICG, indocyanine green; NIR, near-infrared imaging; No., number; Tx, treatment

During the first stage, we observed a high false-negative rate, and it was not possible to achieve 17 or more true positives to reach statistical power. Therefore, after five false-negative results among the first 12 patients, we closed the metastatic pulmonary deposit histology group. For this current investigation, the accuracy of ICG positivity to locate an extrapulmonary metastatic deposit was determined by detection rates with 95% confidence intervals calculated at the metastatic nodule level and at the overall patient level.

We found no significant difference between the size of ICG-avid nodules (i.e., hepatoblastoma and sarcoma) and that of the other nodules that were not ICG avid. Preoperative CT imaging did not detect 13 of the 50 (26%) nodules related to primary liver tumors; however, those nodules were ICG avid, resulting in true-positive nodules at histologic analysis. No adverse events related to the use of ICG occurred in any of the patients.

Discussion

This prospective clinical trial evaluated whether ICG NIR imaging consistently localizes metastatic pulmonary disease in pediatric patients with solid tumors. Accrual was stopped early because our interim data analysis revealed that ICG accurately detected 75% of positive nodules; our target accuracy was 90%.

Patients enrolled in this trial could have any pediatric solid tumor histology, and the study failed to meet the threshold for detection in a small cohort with different histologies due to the absence of ICG avidity in several nodules with different pathologies. Therefore, ICG administered at a dose of 1.5 mg/kg and in a second-window strategy appeared to be unreliable for localizing metastatic pulmonary nodules in this relatively small cohort with various pediatric solid tumors. To localize pulmonary deposits that are challenging because of their small nodule size or depth, an alternative approach (e.g., IR-guided localization) may be required. However, some solid tumors, including hepatoblastoma and high-grade sarcomas, may be amenable to ICG-guided localization, regardless of their size or depth.

Detailed analysis of the study data showed that all enrolled patients with metastatic hepatoblastoma demonstrated ICG-avid pulmonary nodules, and many nodules that were detected by ICG were not detected by preoperative CT scans. Furthermore, all four patients with high-grade sarcomas had ICG-avid metastatic pulmonary nodules. Therefore, we suggest wisely applying ICG NIR visualization during resections of pulmonary metastatic disease in pediatric patients with solid tumors, because of the potential avidity discrepancy across histologies. However, we recommend that this modality be considered in patients with hepatoblastoma or high-grade sarcoma. Future studies evaluating the utility of ICG-guided pulmonary nodule localization should focus on hepatoblastoma and high-grade sarcoma histologies.

The ICG–plasma protein complex is retained in the tumor microenvironment due to its size and in the context of leaky tumor vasculature and defective tumor lymphatic clearance (i.e., the EPR effect). Different tumor histologies may have variable vascular permeability and interstitial fluid pressure, which in turn, would result in variable EPRs.^32,33 In addition, the variable rate of tumor neoangiogenesis after chemotherapy of different nonhepatic tumor–related metastases suggests the failure of ICG guidance is a histology-related event in our study or the effectiveness for some nonhepatic tumor metastases, including high-grade sarcomas.^34,35

Multiple retrospective studies have shown a high success rate of localizing pulmonary metastases of hepatic primary tumors in pediatric populations. Those studies reported increased ICG uptake and impaired excretion of the fluorophore due to the tumor’s biliary characteristics. ^22–30 However, the literature lacks information about nonhepatic tumor–related metastases and the feasibility and effectiveness of ICG application in those cases.

Adult studies have shown the feasibility of ICG-guided localization of high-grade sarcomas.^17,35–39 Similarly, our trial suggests that IGC-guided localization of pediatric high-grade sarcomas should be further investigated.

Earlier pediatric studies included variable doses and timing of ICG administration, with 0.2–0.75 mg/kg ICG intravenously infused 24–96 hours prior to surgery.^29,36 This lack of standardization could jeopardize the use of ICG NIR imaging in children; thus, the aim of this study was to investigate the feasibility of ICG NIR imaging with a standardized approach for pediatric pulmonary metastasectomy. This study confirmed the safety of ICG in a pediatric oncology population. Although ICG side effects, including hypersensitivity and toxic reactions, have been reported in a few cases,⁴⁰ no adverse events have been reported, even with a higher dose than that used in the available data on lung metastasectomy in children. ³⁶

The results of this trial merit further studies of ICG NIR imaging in pediatric oncology patients to improve our understanding of the impact of tumor histology, prior treatment, tumor microenvironment, and tumor necrosis on the EPR phenomena and thereby ICG fluorescence. Targeting pulmonary cancer by utilizing tumor-specific receptors conjugated to a fluorophore has shown promising results in adult oncology studies,^41,42 but the utility of this technique in the pediatric population is yet to be examined. Our study was designed to determine the feasibility of using ICG NIR imaging to detect pulmonary nodules of various histology groups.

This study has some limitations. The variability in the histologies of nonhepatic tumor metastases was the main limitation. Our results suggested the cautious application of ICG NIR imaging in these patients, depending on the histology of their tumor. Therefore, our conclusions are limited to histology-specific feasibility due to the study design and the relatively small number, overall, of enrolled patients. The inclusion of all pediatric solid tumors was another limitation of the study, with a subsequent limitation associated with the high false-negative rate. In addition, the study design proved that standardization of an ICG protocol is feasible and effective in a limited category of tumors and should be adapted and modified for different tumors, according to their potential EPR and histology modifications caused by chemotherapy. The utility of ICG-guided, histology-specific pulmonary nodule localization during metastectomy should be tested in prospective trials adopting universally applied gold standards.²⁴

Conclusion

When a technical advance in pediatric surgical oncology is introduced, a prospective protocol should be conducted to determine the method’s feasibility, safety, and effectiveness and to highlight potential limitations. In this pilot trial, we found that ICG-guided localization of metastatic pulmonary nodules was not successful for all pediatric solid tumors; however, this approach may be useful for metastatic hepatic tumors and high-grade sarcomas. Therefore, intraoperative ICG NIR imaging to detect pulmonary nodules warrants further investigation in prospective trials of larger cohorts with specific solid tumors.

Figure 1.

A. Thoracoscopic view of hepatoblastoma lung nodule with white light.

B.Thoracoscopic view of hepatoblastoma lung nodule highlighted by indocyanine green.

Abbreviations

ICG

Indocyanine green

IR

Interventional radiology

NIR

Near-infrared

EPR

Enhanced permeability and retention

CT

Computed Tomography