500 Consecutive Pulmonary Resections Guided by Intraoperative Molecular Imaging: Lessons from a Single Institutional Experience
Gregory T. Kennedy1, Feredun Azari1, Bilal Nadeem1, Ashley Chang1, Alix Segil1, Elizabeth Bernstein1, Azra Din1, Isvita Marfatia1, Olugbenga Okusanya5, Jane Keating4, Jarrod Predina3, Andrew Newton2, John Kucharczuk1, Sunil Singhal1
1University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, United States, 2MD Anderson Cancer Center, Houston, Texas, United States, 3Massachusetts General Hospital, Boston, Massachusetts, United States, 4University of Connecticut School of Medicine, Hartford, Connecticut, United States, 5Thomas Jefferson University, Philadelphia, Pennsylvania, United States
Objectives: Intraoperative molecular imaging (IMI) using tumor-targeted optical contrast agents has been developed in recent years to aid tumor resection in real-time. Existing trials of IMI consist of small cohorts with short-term follow up. Here, we report our institutional experience with IMI for lung cancer resection, including 500 consecutive patients over a decade.
Methods: Between December 2011 and November 2021, patients with pulmonary nodules suspicious for cancer were preoperatively infused with one of four different IMI tracers: EC17 (emission wavelength[?em]=520 nm), ICG (?em=814 nm), OTL38 (?em=794 nm), or SGM-101 (?em=694 nm). During resection, IMI was used to evaluate known nodules, identify tumor margins, and detect synchronous lesions. The primary endpoint was any change in clinical management due to IMI findings, termed a clinically significant event (CSE).
Results: Five hundred patients underwent resection of 677 pulmonary lesions. Most patients underwent IMI with OTL38 (n=291, 58.2%), followed by ICG (n=154, 30.8%), EC17 (n=51, 10.2%), and SGM-101 (n=4, 0.8%). All tracers demonstrated a high safety profile. Mean lesion size was 2.3 cm and mean depth from the pleural surface was 0.6 cm. The most frequent diagnosis was an adenocarcinoma-spectrum lesion (n=332, 49.0%). IMI identified 503 (74.3%) lesions. In reviewing our experience, we developed several key conclusions. First, near-infrared (NIR) tracers permit greater depth of detection compared to non-NIR tracers (1.7 vs 0.9 cm, p<0.001). Second, the choice of tracer should vary by clinical scenario and can be tailored based on suspected tumor type. We found that receptor-targeted tracers were most effective for primary pulmonary cancers and subpleural ground glass opacities (mean signal-to-background ratio [SBR] for targeted tracers: 3.2 vs 2.1 for non-targeted tracers, p=0.01) and non-targeted tracers were more effective for metastatic disease (SBR 4.1 vs. 3.4, p=0.04). Third, the CSE for different operations varies: the most common CSE for sublobar pulmonary resection is identifying a visually occult, non-palpable lesion, whereas the most common CSE for a pulmonary metastasectomy is locating synchronous lesions not appreciated on preoperative imaging. Fourth, a major source of CSE is back table IMI of specimens, which can accurately determine margin distance as compared to pathology (R2=0.87). Finally, failure of IMI due to false positive fluorescence is due to benign inflammatory lesions (mean SBR=6.2) and false negative fluorescence is seen in mucinous adenocarcinomas (mean SBR=1.8), heavy smokers (>30 pack years, SBR=1.9), and tumors greater than 2.0 cm from the pleural surface (SBR=1.3).
Conclusions: IMI is safe and effective in guiding resection of pulmonary cancers. Choice of IMI tracer should vary by surgical indication and certain lesions may be unlikely to benefit from IMI. Future work in IMI should focus on stewarding this new technology for the correct indications.
Back to 2022 Abstracts