Abstract
Objective:
To determine if intraoperative molecular imaging (IMI) can improve detection of malignant pulmonary nodules.
Background:
18Fluorodeoxyglucose positron emission tomography (PET) is commonly utilized in preoperative assessment of patients with solid malignancies; however, false negatives and false positives remain major limitations. Using patients with pulmonary nodules as a study model, we hypothesized that IMI with a folate receptor targeted near-infrared contrast agent (OTL38) can improve malignant pulmonary nodule identification when combined with PET.
Methods:
Fifty patients with pulmonary nodules with imaging features suspicious for malignancy underwent preoperative PET. Patients then received OTL38 prior to pulmonary resection. During resection, IMI was utilized to evaluate known pulmonary nodules and identify synchronous lesions. Tumor size, PET standardized uptake value (SUV), and IMI tumor-to-background ratio (TBR) were compared for known and synchronous nodules via paired and unpaired t-tests, when appropriate. Test characteristics of PET and IMI with OTL38 were compared.
Results:
IMI identified 56 of 59 (94.9%) malignant pulmonary nodules identified by preoperative imaging. IMI located an additional 9 malignant lesions not identified by PET. Nodules only detected by IMI were smaller than nodules detected by PET (0.5cm vs 2.4cm; p<0.01), but displayed similar fluorescence (TBR 3.3 and 3.1; p=0.50). Sensitivity of IMI and PET were 95.6% and 73.5% (p=0.001), respectively; and positive-predictive values were 94.2% and 89.3%, respectively (p>0.05). Additionally, utilization of IMI clinically upstaged 6 (12%) subjects and improved management of 15 (30% )of subjects.
Conclusions:
These data suggest that combining IMI with PET may provide superior oncologic outcomes for patients with resectable lung cancer.
Keywords: pulmonary adenocarcinoma, surgery, molecular imaging, folate receptor
Mini Abstract:
Positron emission tomography (PET) is commonly utilized preoperatively for patients with pulmonary nodules, however, inaccuracy exists. In this study, we hypothesize that intraoperative molecular imaging (IMI) can improve malignant pulmonary nodule identification when combined with PET. Fifty subjects with pulmonary nodules were enrolled in a clinical trial. We found that IMI impacted management of 30% of subjects. These data suggest that combining IMI with PET may provide superior oncologic outcomes for patients with resectable lung cancer.
Introduction:
Each year, pulmonary nodules are discovered in nearly 250,000 Americans [1–3]. Approximately 80,000 of these patients have nodules suspicious enough to warrant surgical intervention. Once deemed appropriate for surgical resection, patients frequently undergo 18Fluorodeoxyglucose positron emission tomography (PET). Guidelines supporting use of PET in the patient with pulmonary nodules are based on data demonstrating a sensitivity in malignancy detection greater than 90% and specificity approaching 80% [4]. Although frequently utilized, PET has known limitations including poor sensitivity in the subcentimeter range and in nodules with low metabolic activity (e.g. adenocarcinoma spectrum lesions) [5] and an inability to distinguish benign metabolically active processes such as inflammatory diseases (granuloma and infection) from malignancy [6, 7]. Furthermore, PET provides no intraoperative feedback to a surgeon in locating known small nodules or identifying synchronous disease at the time of resection.
Intraoperative molecular imaging (IMI) is an emerging technology that can characterize nodules and identify synchronous disease during oncologic resection. IMI incorporates systemic delivery of an optical contrast agent which accumulates in tissues, thus allowing for real-time optical imaging. Our group has investigated several intraoperative imaging probes over the last decade [8, 9] and has recently completed a phase I clinical trial involving a folate-receptor targeted, near infrared (NIR) contrast agent, OTL38 [10]. In this initial experience, we found that IMI can localize and characterize non-small cell lung cancer (NSCLC) nodules as small as 3 mm [10]. In addition to localizing preoperatively identified pulmonary malignancies, we noted that IMI identifies additional synchronous lesions in approximately 10% of patients.
Based on our initial experience with IMI, we hypothesized that IMI can be an important adjunct to PET by improving intraoperative identification of malignant pulmonary nodules. To test this hypothesis, we enrolled 50 subjects with pulmonary nodules in a study aimed to determine if the addition of IMI to PET can improve (1) identification of malignant pulmonary nodules found on preoperative imaging, and (2) identification of additional synchronous and metachronous lesions not identified preoperatively.
Materials and Methods:
Study Design:
A pilot intraoperative molecular imaging (IMI) trial was approved by the University of Pennsylvania Institutional Review Board. The primary objectives of this study were to assess the sensitivity of IMI with a folate receptor-targeted optical contrast agent (OTL38) in (1) identifying malignant pulmonary nodules found on preoperative imaging, and (2) identifying additional synchronous and metachronous lesions. Fifty subjects provided informed consent and were recruited between July 2015 and November 2016.
Included subjects had resectable pulmonary nodules identified by CT. Pulmonary nodules were either suspected or histologically confirmed to be malignant. All subjects also underwent preoperative dual modality PET-CT within 30 days of pulmonary resection. For imaging obtained outside the University of Pennsylvania Health System (UPHS), at least one UPHS thoracic radiologist and nuclear medicine specialist reviewed preoperative imaging to confirm the presence of pulmonary nodules, identify other suspicious nodules and confirm SUV by PET. A standardized uptake value (SUV) greater than 2.0 was considered metabolically active (positive). Based on limitations in depth of penetration recognized in our prior experience with NIR imaging, we selected subjects who had nodules within 3 cm of the visceral pleural surface on preoperative imaging [10].
Study participants received 0.025 mg/kg of intravenous OTL38 through a peripheral vein 3 to 6 hours prior to resection as previously described [10]. During resection, surgeons utilized white-light (also known as bright-field imaging) to determine fluorescence of known nodules. Additionally, surgeons interrogated the remainder of the ipsilateral chest using white-light thoracoscopy and molecular imaging to determine the presence of synchronous disease. All suspicious lesions were resected and imaged ex vivo on the back table prior to submitting for pathologic examination by a specialized pulmonary pathologist.
Study Drug:
OTL38 (chemical formula: C61H63N9Na4O17S4 (Tetrasodium salt); molecular weight: 1414.42 Da) is a folate analog conjugated to a NIR fluorescent dye. OTL38 maximally excites at a wavelength of 774–776 nm and has a peak emission of 794–796nm [10, 11]. OTL38 (>96% purity) was obtained from On Target Laboratories (West Lafayette, IN, USA). OTL38 was synthesized and manufactured at Nitto Denko Avecia Inc. (Cincinnati, OH) in compliance with Good Manufacturing Practices and formulated at Grand Rapids Aseptic Manufacturing (GRAM) at Grand Rapids, MI. OTL38 was stored at −20°C in vials containing 3.2 mg OTL38 in 1.6 mL PBS. Before administration, the frozen vials were thawed, vortexed, and then diluted with 5% dextrose in water (D5W) for intravenous infusion. OTL38 was diluted into 220 mL of D5W and was infused over 60 minutes. All subjects also received 25mg of intravenous diphenhydramine prior to OTL38 infusion as prophylaxis.
Imaging Device:
In situ, real-time fluorescent imaging was performed using an Iridium® system optimized for detection of OTL38 (Visionsense Corp, Philadelphia, PA). The Iridium® is a high definition, dual band (white light and NIR) camera system capable of emitting and detecting light in the NIR spectrum as previously described [10]. An excitation source with a wavelength of 785nm was utilized, and fluorescence was detected using a bandpass filter selective to light ranging from 800 to 835 nm. During VATS and robotic procedures, the Iridium® was equipped with a 5mm, 0- degree thoracoscope. For ex vivo back table evaluation of the specimen, a free-standing exoscope was utilized.
Pathologic Review of Specimen:
All excised specimens were examined by at least one thoracic oncology pathologist using standard hematoxylin/eosin (H/E) staining. Additionally, immunohistochemical (IHC) staining for FRα was performed a monoclonal antibody, 26B3.F2 (Morphotek Inc., PA). For IHC, human kidney and skeletal muscle were used as positive and negative controls, respectively.
Statistical Analysis:
Post hoc image analysis was performed to quantify the amount of fluorescence using region of interest (ROI) software within ImageJ (National Institute of Health; http://rsb.info.nih.gov/ij). A background fluorescence level was similarly obtained, and tumor-to-background (TBR) was calculated for all identified lesions. For disease identified only intraoperatively by IMI, preoperative imaging was reviewed post hoc by a pulmonary radiologist and an SUV was calculated from the ROI corresponding to identified nodules. Mean size, standardized uptake value (SUV), and tumor- to-background fluorescence ratio (TBR) were compared for preoperatively known and additional nodules found by IMI using paired t-tests. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for PET and IMI were calculated and compared. An SUV on PET of 2.0 was considered metabolically active (positive) and a TBR with IMI greater than 2.0 was considered fluorescent (positive). Histologic review by a pathologist was considered gold-standard. Results are expressed as mean (standard deviation) unless otherwise noted. All comparisons were made use Stata Statistical Software: Release 14 (College Station, TX: StataCorp LP). A p-value of 0.05 or less was considered statistically significant.
Results:
Study Population:
After meeting inclusion criteria, 50 subjects (n=35 male) with a mean age of 67.0 years (sd, 9.8 years) were referred for surgical evaluation of a primary pulmonary nodule between July 2015 and November 2016. Upon review of preoperative imaging, a total of 66 pulmonary nodules, with a mean of 1.3 nodules per subject, were identified. Thirty-nine (78%) subjects had a solitary pulmonary nodule, 11 (22%) subjects had multiple suspicious nodules identified in the ipsilateral hemithorax; 8 (16%) subjects had 2 pulmonary nodules, 2 (4%) subjects had 3 pulmonary nodules and 1 (2%) subject had 5 nodules. The mean nodule size was 2.5cm (sd, 2.1cm). In 7 subjects, a tissue confirmed diagnosis of NSCLC was obtained prior to resection. A full summary of subject characteristics is provided in Table 1.
Table 1:
Subject and Tumor Characteristics
| Gender (n,%) | Male | n=35 (70%) |
|---|---|---|
| Female | n=15 (30%) | |
| Age (years) mean (sd) | 67.0 (9.8) | |
| Number of Nodules per subject n,(%) | 1 | n=39 (78%) |
| 2 | n=8 (16%) | |
| 3 | n=2 (4%) | |
| 4 | n=0 (0%) | |
| 5 | n=1 (2%) | |
| Nodule Location (n,%) | RUL | n=20 (29.9%) |
| RML | n=5 (7.5%) | |
| RLL | n=12 (17.9%) | |
| LUL | n=18 (26.9%) | |
| LLL | n=11 (16.7%) | |
| Size (cm) (mean (,sd) | 2.5 (2.1) cm |
Positron emission tomography of pulmonary nodules:
As part of the preoperative staging evaluation, all 50 subjects underwent preoperative PET. Of the 66 pulmonary nodules identified by preoperative CT, 56 were found to be FDG-avid (mean SUV 6.5, sd 3.7) while 10 were non FDG-avid (mean SUV 1.2, sd 0.3). PET clinically upstaged 4 subjects by identifying hilar (N1) disease.
Intraoperative molecular imaging with OTL38:
During pulmonary resection, fluorescence was observed during IMI in 60 of the 66 (90.9%) pulmonary nodules identified by preoperative imaging (Figure 1). In addition to localizing preoperatively identified pulmonary nodules, IMI identified 9 additional pulmonary nodules that were undetectable by preoperative imaging or by traditional intraoperative palpation. These 9 synchronous nodules were found in 6 subjects: a solitary synchronous nodule was identified in 4 subjects, 2 synchronous nodules were identified in 1 subject, and 3 synchronous nodules were Of the 75 pulmonary nodules identified by preoperative imaging or IMI, 69 (92.0%) nodules displayed fluorescence with a mean TBR of 3.2 (sd 0.8) versus at TBR of 1.3 (sd, 0.2) in the 6 (8.0%) non-fluorescent nodules; p<0.001.
Figure 1:
Representative example of malignant pulmonary nodule with in vivo fluorescence. Subject 43 presented with a PET avid (SUV=3.0) 1.3cm right upper lobe nodule. At the time of resection, the tumor displayed strong in vivo fluorescence with a TBR of 3.4. Pathologic review demonstrated pulmonary adenocarcinoma with upregulation of the FRα by immunohistochemistry.
Final pathologic evaluation of all identified pulmonary nodules:
On final pathologic review of the 75 nodules identified preoperatively or by IMI, 68 (90.7%) were found to be primary lung cancers or pulmonary metastases, while the remaining 7 (9.3%) nodules were granulomas or inflammatory processes (Supplemental Table 1). Of note, no subjects were upstaged beyond their clinical stage on final pathologic evaluation.
The addition of IMI to preoperative PET improves diagnostic sensitivity:
Using pathologic evaluation as a gold standard, we compared test characteristics of IMI with OTL38 to PET using the 75 pulmonary nodules identified within our cohort. We noted that preoperative PET accurately characterized 50 of 68 malignant pulmonary nodules and 1 of 7 benign pulmonary nodules (Table 2).
Table 2:
Contingency Table with test characteristics of PET, IMI with OTL38, and PET combined with IMI for 75 pulmonary nodules identified by preoperative axial imaging or during molecular imaging.
| Malignant Nodule* | Benign Nodule* | ||
|---|---|---|---|
| PET | |||
| PET Avid (SUV ≥ 2) | 50 | 6 | |
| Non-PET Avid (SUV<2) | 18 | 1 | |
|
Sensitivity –73.5% (95% CI: 61.4% - 83.5%) Specificity – 14.3% (95% CI: 0.3% - 57.9%) PPV – 89.3% (95% CI: 78.1% - 96.0%) NPV – 5.26% (95% CI: 0.1% - 26.0%) |
|||
| IMI with OTL38 | |||
| Fluorescent (TBR ≥ 2) | 65 | 4 | |
| Non-fluorescent (TBR<2) | 3 | 3 | |
|
Sensitivity - 95.6% (95% CI: 87.6% - 99.1%) Specificity – 42.9% (95% CI: 9.9% - 81.6%) PPV – 94.2% (95% CI: 85.8% - 98.4%) NPV – 50% (95% CI: 11.8% - 88.2%) |
|||
Modality compared to gold-standard histologic evaluation by certified pathologist. PPV-positive predictive value, NPV-negative predictive value.
With regard to IMI, we observed fluorescence in 65 of 68 malignant pulmonary nodules and 4 of 7 benign pulmonary nodules (Table 2). Based on this data, sensitivity of IMI with OTL38 was found to be superior to PET with values of 95.6% versus 73.5%, respectively (p=0.001). Positive-predictive values of IMI and PET were 94.1% and 89.3%, respectively (p>0.05). Negative predictive value and specificity were also similar between PET and IMI (Table 2). Lastly, area under the receiver-operator curve (AUC) was significantly higher for IMI than with PET: 0.72 versus 0.47, respectively (p<0.01). Estimates of sensitivity, specificity, positive-predictive value and negative-predictive value are provided in Table 2.
Of note, in 9 (18%) subjects IMI was more accurate than PET in identifying malignant pulmonary nodules (Supplemental Table 2, Supplemental Figure 1). We noted that non-PET avid malignant nodules were smaller than those nodules which were PET avid (1.2cm vs 2.8cm; p=0.05); however, fluorescence was similar with mean TBRs of 3.1 and 3.0, respectively (p=0.61) (Supplemental Figure 1).
Synchronous disease identified by IMI upstages NSCLC patients and changes treatment:
Given an improved diagnostic yield with utilization of IMI, we sought to better understand how findings of IMI impacted patient management. As noted above, additional malignant nodules were identified in 6 (12%) of our subjects. All identified synchronous nodules proved to be malignant nodules after final pathologic evaluation and influenced intraoperative or postoperative patient care (Table 3). Identification of synchronous lesions upstaged 5 subjects, thus altering intraoperative decision making and postoperative adjuvant therapy as noted (representative examples provided in Figure 2 and 3). In the remaining subject (Subject 48), IMI identified an additional resectable pulmonary metastasis which improved pulmonary metastasectomy. A summary of all synchronous nodules detected by IMI and their clinical impact is noted in Table 3.
Table 3:
Summary of Synchronous Nodules Fluorescent with OTL38 IMI that were not identified by preoperative imaging
| ID | Location | Size (cm) | Histology | On preop CT | SUV by PET$ | TBR by IMI | Impact of Identification |
|---|---|---|---|---|---|---|---|
| 13 | LUL | 0.2 | AC | n | 0.6 | 4.2 | Known AC in 0.5cm LUL. IMI aided identification of otherwise undetectable AC also in LUL. Findings resulted in upstaging to Stage IIB (T3N0) from Stage IA (T1aN0). |
| 35 | Pleural studding | 0.3 | AC | n | 0.7 | 5.7 | Pleural implants noted which were not identifiable on preoperative imaging. Findings upstaged patient from Stage IIIA (T3N1M0) to Stage IV (M1a). |
| 37 | LLL | 0.8 | AC | n | 0.6 | 2.5 | Known 6.3cm LUL AC with ipsilateral hilar lymph nodes. IMI aided identification of otherwise undetectable 0.8cm AC in the LLL (Figure 2). Findings upstaged patient from Stage IIB (T2bN1) to Stage IIIA (T4N1). Case was aborted after identifying LLL cancer. |
| 41 | LLL LLL LLL |
0.8 0.6 0.2 |
AC AC AC |
n** n n |
0.8 0.8 0.7 |
3.7 3.5 2.7 |
In addition to known 2.0cm LUL nodule, IMI identified 3 subcentimeter nodules in LLL (Figure 3). Findings upstaged subject to IIIA (T4N0) from Stage IA (T1N0) and changed operative approach from LU lobectomy to wedge resection of all nodules. |
| 48 | RML RLL |
0.4 0.3 |
SCC SCC |
n n |
1.2 1.0 |
2.4 2.6 |
Preoperative imaging suspicious for head and neck SCC metastases (n=2). IMI aided identification of 2 otherwise undetectable nodules. IMI allowed for enhanced metastasectomy. |
| 50 | LUL* | 0.5 | AC | n | 0.7 | 2.6 | Known 2.5cm LUL AC and hilar lymphadenopathy. IMI aided identification of multiple otherwise undetectable subcentimeter nodules in LUL. Findings upstaged subject from Stage IIA (T1bN1) to Stage IIIA (T3N1). |
AC-Adenocarcinoma, SCC -Squamous Cell Carcinoma, RUL-right upper lobe, RML-right middle lobe, RLL-right lower lobe, LUL-Left Upper Lobe, LLL-left lower lobe
several nodules described, but exact number unclear based on path report.
Upon review of preoperative imaging, LLL nodules were identified by CT (non-avid on review of PET).
SUVs corresponding to the region of interest of identified nodules were calculated retrospectively by review of PET with nuclear medicine.
Figure 2:
IMI with OTL38 identifies synchronous pulmonary nodules. (a) Subject 37 is a representative subject in which IMI with OTL38 identified an otherwise undetectable malignant pulmonary nodule. Subject presented with a 6.3cm FDG-avid (SUV=6.2) pulmonary adenocarcinoma of the left upper lobe. At the time of resection, IMI identified a 0.8cm synchronous left lower lobe adenocarcinoma. Identification of left lower lobe disease upstaged the subject and resulted in a change of operative plan and postoperative adjuvant therapy course (see Table 3). Bar graphs comparing (b) size and (c) fluorescence of synchronous nodules to fluorescent pulmonary nodules discovered preoperatively. ns-not significant. * p<0.01
Figure 3:
Subject 41, a 71-year-old female presented with a 2.0cm FDG-avid (SUV8.6) located in her left upper lobe. During resection, three areas of fluorescence were noted in her left lower lobe. Wedge biopsy of all left lower lobe lesions revealed adenocarcinoma. These findings upstaged the patient and changed both the operative plan and the postoperative adjuvant treatment (see Table 3).
We observed that the average size of synchronous nodules identified only with IMI was 0.5cm (sd, 0.1cm) versus 2.5cm (sd, 2.1cm) in pulmonary nodules discovered by preoperative imaging; p<0.007 (Supplemental Figure 2a). Despite differences in size, mean TBR of synchronous nodules (3.4, sd 1.1) was similar to that observed in fluorescent pulmonary nodules which were identified preoperatively (3.1, sd 0.8); p=0.38 (Supplemental Figure 2b).
Fluorescence during IMI characterizes malignant subcentimeter pulmonary nodules:
Based the previous findings demonstrating strong fluorescence in smaller, non-PET avid nodules, as well known deficiencies of PET in the subcentimeter range, we sought to evaluate IMI after stratifying by nodule size (≤1cm versus >1cm) (Table 4). We identified 17 subcentimeter pulmonary nodules in our cohort; 15 being malignant and 2 being benign. We observed that 15 of 15 (100%) subcentimeter malignant pulmonary displayed fluorescence during IMI while only 4 (26.6%) were PET avid; p<0.001 (Table 5). The PPVs of IMI with OTL38 and PET were 93.8% and 80.0%, respectively (p>0.05). Both specificity and NPV were also similar between PET and IMI when evaluating subcentimeter pulmonary nodules (Table 5). Finally, area under the receiver-operator curve (AUC) was significantly higher for IMI than with PET; 0.75 versus 0.38 (p<0.001).
Table 4:
Pathology of 75 pulmonary nodules characterized by IMI and PET stratified by nodule size (≤1cm vs >1 cm).
| nodules ≤1cm (n=17) | nodules >1cm (n=58) | |||||||
|---|---|---|---|---|---|---|---|---|
| PET Avid | PET Avid | |||||||
| Yes | No | Yes | No | |||||
| Fluorescent by IMI | Yes | 4 malignant 1 benign |
11 malignant 0 benign |
Fluorescent by IMI | Yes | 43 malignant 3 benign |
7 malignant 0 benign |
|
| No | 0 malignant 0 benign |
0 malignant 1 benign |
No | 3 malignant 2 benign |
0 malignant 0 benign |
|||
Table 5:
Contingency Table with test characteristics of PET and IMI stratified by pulmonary nodule size (≤1cm vs >1cm) for 75 pulmonary nodules identified by preoperative axial imaging or during molecular imaging.
| Nodules ≤ 1cm (n=17) | Nodules > 1cm (n=58) | |||||
|---|---|---|---|---|---|---|
| Malignant Nodule* | Benign Nodule* | Malignant Nodule* | Benign Nodule* | |||
| PET | PET | |||||
| PET Avid | 4 | 1 | PET Avid | 43 | 3 | |
| Non-PET Avid | 11 | 1 | Non-PET Avid | 10 | 2 | |
|
Sensitivity – 26.7% (95% CI:7.8% - 55.1%) Specificity – 50.0% (95% CI: 1.26% - 98.7%) PPV – 80.0% (95% CI: 28.4% - 99.5%) NPV – 8.3% (95% CI: 0.2% - 38.5%) |
Sensitivity – 81.1% (95% CI: 68.0% - 90.6%) Specificity – 40.0% (95% CI: 5.27% - 85.3%) PPV – 93.5% (95% CI: 82.1% - 98.6%) NPV – 16.7% (95% CI: 2.09% - 48.4%) |
|||||
| IMI with OTL38 | IMI with OTL38 | |||||
| Fluorescent | 15 | 1 | Fluorescent | 50 | 3 | |
| Non-fluorescent | 0 | 1 | Non-fluorescent | 3 | 2 | |
|
Sensitivity - 100% (95% CI: 78.2 – 100%) Specificity – 50.0% (95% CI: 1.2% - 98.7%) PPV – 93.8% (95% CI: 69.8% - 99.8%) NPV – 100% (95% CI: 2.5% - 100%) |
Sensitivity – 94.3% (95% CI: 84.3% - 98.8%) Specificity – 40.0% (95% CI: 5.3% - 85.3%) PPV – 94.3% (95% CI: 84.3% - 98.8%) NPV – 40% (95% CI: 5.2% - 85.3%) |
|||||
Modalities compared to gold-standard histologic evaluation by certified pathologist. PPV-positive predictive value, NPV-negative predictive value.
Of note, in those 58 pulmonary nodules measuring greater than 1 cm, we observed no differences in sensitivity, specificity, NPV, PPV or AUC between IMI and PET (Table 5).
Discussion:
Preoperative PET has an important role in treatment planning for patients presenting with pulmonary nodules that are suspicious for malignancy. However, sensitivity and accuracy can be limited, particularly in the setting of subcentimeter lesions. Using patients with pulmonary nodules as a model, we present data suggesting that IMI with OTL38 can function in concert with PET by facilitating identification of preoperatively identified nodules and also drawing a surgeon’s attention to synchronous disease that is below the detection range of PET.
Although PET has been proposed as a diagnostic tool for a number of solid tumor histologies (e.g. colorectal cancer [11], breast cancer [12], esophageal cancer [13], pancreatic cancer [14], head and neck cancers [15], melanoma [16], and genitourinary cancers [17]), its role has become most well defined for the NSCLC staging[7, 18, 19]. For patients with NSCLC, PET has sensitivity and specificity estimates approaching 90% and 80%, respectively [4]. Incorporation of PET into the diagnostic work-up of pulmonary nodules provides information that influences management decisions in nearly 30% of patients [19]. The utility of PET is particularly limited for smaller nodules, and false negative rates approach 25% in nodules measuring less than 2.5 cm [20]. The false-negative rate is higher in patients presenting with subcentimeter pulmonary nodules [5], thus it is generally not useful for nodules measuring less than 1 cm due to the low spatial resolution of PET imaging [21].
Given the inability to reliably rule out malignancy, patients with pulmonary nodules are often taken to the operating room for a diagnostic wedge resection. During resection, accurate identification of pulmonary nodules may be challenging, especially with subcentimeter nodules which often display no parenchymal abnormalities and escape the reach of finger palpation. IMI has been developed to provide additional data which can help the surgeon overcome these diagnostic and therapeutic hurdles. Previous reports involving a variety of IMI agents suggest that this technology indeed not only enhances the surgeon’s ability to localize nodules, but also characterizes malignant potential in real-time [22–24]. In our Phase I experiences using folate-receptor targeted IMI in patients with NSCLC, we similarly observed accuracy rates approaching 95% and high levels of sensitivity which allowed for reproducible detection of lesions smaller than 5 mm [9, 25].
In this study, we further explore the potential of IMI with OTL38 to serve as an adjunct to standard-of-care for patients with suspicious pulmonary nodules. We observed that the sensitivity of IMI with OTL38 was greater than that of PET (95.6% versus 73.5%; p=0.001). The rate of malignancy in fluorescent nodules, either primary lung cancer or a pulmonary metastasis, was 94.2% which was similar to the positive predictive value of PET (89.3%). Specificity and negative predictive value were similar between the two compared modalities.
Upon subgroup analysis, it becomes clear that the greatest advantage to IMI with OTL38 is the excellent sensitivity for small pulmonary nodules. This was particularly impressive in the near- centimeter and subcentimeter range, a range that typically pushes the lower thresholds of detection of PET. In our study cohort, 9 (15.2%) of malignant nodules were not FDG-avid, yet displayed strong fluorescence during IMI (Supplemental Table 2). The average size of these nodules was 1.2cm. In all cases, malignancy was confirmed during final pathologic review. In addition, in 6 of 50 (12%) subjects IMI identified additional malignant pulmonary nodules which were not detected by PET (mean size 0.5 cm). As noted in Table 3, the identification and characterization of these nodules altered staging and operative planning.
Although these results need to be confirmed in larger studies, our results provide preliminary data suggesting that IMI and PET can function synergistically to help surgeons accurately characterize known pulmonary nodules and identify additional malignant lesions. More specifically, for patients undergoing resection for known subcentimeter disease, the presence of in situ fluorescence suggests that a nodule requires resection (Table 4, Table 5, and Supplemental Table 2) as the observed rate of malignancy is greater than 90%. In contrast, those subcentimeter nodules that do not fluoresce in situ are extremely unlikely to harbor malignancy, as demonstrated by our observed negative predictive value of 100% (Table 4, Table 5). Lastly, all synchronous malignancies that are identified by IMI should be resected as the positive predictive value of malignancy was 100% in our cohort (Tables 3). For preoperative nodules greater than 1cm in size, standard-of-care management should be followed.
It is important to place these findings in context of our study design. Our inclusion criteria required subjects be deemed surgical candidates prior to enrollment. This approach introduces a selection bias that accompanies all studies that evaluate a new technology in an “operative cohort”. To elaborate, all subjects were already triaged to resection given the surgeons’ high level of suspicion for NSCLC or pulmonary metastases, thus the high prevalence of malignancy observed may artificially elevate some of IMI’s test characteristics. Further the high prevalence of malignancy decreases opportunities for “true negatives”, thus specificities and negative predictive values are likely confounded. As we move forward in our planned multi-center Phase II evaluation, we have structured inclusion/exclusion criteria such that we will enroll a more diverse cohort which can improve our assessment of the diagnostic potential of IMI with OTL38. Despite these drawbacks, we feel this data demonstrates how IMI with OTL38 can be a powerful adjunct to standard preoperative PET imaging protocols.
We also feel obligated to note that the highly sensitive pattern of fluorescence was somewhat of a serendipitous finding. OTL38 was developed to capitalize on highly specific folate-receptor alpha (FRα) targeting which is pathologically expressed in approximately 85% of pulmonary adenocarcinomas [26]. Other malignancies express the FRα, albeit at much lower rates. The higher than anticipated sensitivity could be explained by the fact that only a small proportion of tumor cells need to express FRα for the tumor to appear fluorescent [27], or alternatively, that accumulation of FRβ-expressing tumor-associated macrophages also bind OTL38 and render the FRα-negative tumors fluorescent [25, 28]. The high sensitivity of folate-based probes may allow for a rapid “optical biopsy” as previously described by our group [9].
On a more general view of IMI, there remain several hurdles to overcome. The most obvious hurdle associated with IMI is the challenge of “depth of detection”. In our study, all enrolled subjects were required to have disease within 3 cm of the pleural surface as determined by axial imaging. This inclusion criterion was based on our previous IMI experiences which demonstrate that nodules deeper than this display fluorescence only upon ex vivo tumor bisection [8, 25]. Furthermore, all synchronous nodules were detected within 1.0 cm of the pleural surface. Based on previous reports and our own observations, it becomes more apparent that the current IMI prototypes will fall short in identifying deeper nodules. In efforts to overcome these hurdles, we are in collaborations to develop a NIR compatible flexible bronchoscope which can be used in addition to NIR capable thoracoscopic cameras. Alternative molecular imaging techniques, such as photoacoustic imaging, may provide additional approaches to detect deeper nodules and are currently being investigated in the setting of metastatic melanoma [29]. These advances have the potential to significantly increase disease identification and localization.
A second opportunity for advancement involves the ability to improve accuracy of IMI by simultaneously utilizing multiple targeted probes. To elaborate, in our study we found IMI with OTL38 to be highly sensitive for malignancy, but did not distinguish pulmonary adenocarcinomas from other NSCLC histologies or pulmonary metastases. Furthermore, 4 of 7 granulomas also accumulated OTL38 and fluoresced. As noted previously, this observed trend is likely the result of low levels of FRα expression in other malignancies or, more likely, the presence of FRβ expressing macrophages which are present in both tumors and inflammatory processes. In the future, specificity will likely be increased by including combinations of probes, each with a different receptor target. This information can aid the surgeon in not only localizing nodules, but also characterizing nodules via optical biopsy [9]. Preclinical and Phase I studies involving probes targeting the EGFR mutation and the prostate specific membrane antigen (upregulated in 70% of early lung cancers [30]) are underway [23, 30], and have the potential to enhance IMI with OTL38.
In conclusion, this data demonstrates that IMI with OTL38 can be a valuable tool useful for localizing malignant pulmonary nodules and identifying synchronous malignancies, particularly those within the subcentimeter size range. This data further suggests that IMI may be a reliable intraoperative tool that functions well in combination with standard-of-care preoperative imaging modalities such as CT and PET. These observations are the basis of the first phase II clinical trial of intraoperative tumor imaging in the United States. More broadly, these data set the stage for additional IMI studies that can better define the role of IMI in the context of standard-of-care diagnostics and therapeutics across a variety of malignancies.
Supplementary Material
Supplemental Figure 1: FDG-avid and Non-PET avid malignant pulmonary nodules display fluorescence during IMI with OTL38. Representative example of a non-PET avid malignant pulmonary nodule demonstrating fluorescence during resection. Subject 23, a 51-year-old female, presented with a 1.2cm left upper lobe nodule (a) which had no activity by PET (b). During resection (c) brightfield and (d) fluorescent overlays revealed tumor fluorescence with a TBR of 3.2. Pathology demonstrated adenocarcinoma. Bar graphs comparing (e) size and (f) fluorescence of PET avid and Non-PET avid malignant pulmonary nodules. ns-not significant. * p<0.05
Supplemental Figure 2: Bar graphs comparing (a) size and (b) fluorescence of synchronous nodules to fluorescent pulmonary nodules discovered preoperatively. ns-not significant. * p<0.01
Funding Sources:
JDP was supported by the American Philosophical Society, the NIH (1F32CA210409), and the Association for Academic Surgery Foundation. SS was supported by the NIH (R01 CA193556).
Footnotes
Conflicts of Interest: PL is a Board Member of On Target Laboratories, manufacturers of the study drug (OTL38). KAS is the Vice President of Research and Design at On Target Laboratories
References:
- 1.Ost D. and Fein A, Evaluation and management of the solitary pulmonary nodule. Am J Respir Crit Care Med, 2000. 162(3 Pt 1): p. 782–7. [DOI] [PubMed] [Google Scholar]
- 2.Henschke CI, et al. , Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet, 1999. 354(9173): p. 99–105. [DOI] [PubMed] [Google Scholar]
- 3.Wahidi MM, et al. , Evidence for the treatment of patients with pulmonary nodules: when is it lung cancer?: ACCP evidence-based clinical practice guidelines (2nd edition). Chest, 2007. 132(3 Suppl): p. 94S–107S. [DOI] [PubMed] [Google Scholar]
- 4.Gould MK, et al. , Accuracy of positron emission tomography for diagnosis of pulmonary nodules and mass lesions: a meta-analysis. JAMA, 2001. 285(7): p. 914–24. [DOI] [PubMed] [Google Scholar]
- 5.Kernstine KH, Grannis FW Jr., and Rotter AJ, Is there a role for PET in the evaluation of subcentimeter pulmonary nodules? Semin Thorac Cardiovasc Surg, 2005. 17(2): p. 110–4. [DOI] [PubMed] [Google Scholar]
- 6.Gallamini A, Zwarthoed C, and Borra A, Positron Emission Tomography (PET) in Oncology. Cancers (Basel), 2014. 6(4): p. 1821–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Deppen SA, et al. , Accuracy of FDG-PET to diagnose lung cancer in areas with infectious lung disease: a meta-analysis. JAMA, 2014. 312(12): p. 1227–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Keating J, et al. , Near-Infrared Intraoperative Molecular Imaging Can Locate Metastases to the Lung. Ann Thorac Surg, 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kennedy GT, et al. , The Optical Biopsy: A Novel Technique for Rapid Intraoperative Diagnosis of Primary Pulmonary Adenocarcinomas. Ann Surg, 2015. 262(4): p. 602–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Predina JDK,J; Gaughan C; Jarrar D; Pechet T; Kucharczuk J; Low P; Singhal S., A Phase I Clinical Trial of Targeted Intraoperative Molecular Imaging for Pulmonary Adenocarcinomas, in STS 53rd Annual Meeting. 2017: Houston, TX, USA. [Google Scholar]
- 11.Petersen RK, et al. , Clinical impact of FDG-PET/CT on colorectal cancer staging and treatment strategy. Am J Nucl Med Mol Imaging, 2014. 4(5): p. 471–82. [PMC free article] [PubMed] [Google Scholar]
- 12.Martoni AA, et al. , Early (18)F-2-fluoro-2-deoxy-d-glucose positron emission tomography may identify a subset of patients with estrogen receptor-positive breast cancer who will not respond optimally to preoperative chemotherapy. Cancer, 2010. 116(4): p. 805–13. [DOI] [PubMed] [Google Scholar]
- 13.Lutz MP, et al. , Highlights of the EORTC St. Gallen International Expert Consensus on the primary therapy of gastric, gastroesophageal and oesophageal cancer - differential treatment strategies for subtypes of early gastroesophageal cancer. Eur J Cancer, 2012. 48(16): p. 2941–53. [DOI] [PubMed] [Google Scholar]
- 14.Wang Z, et al. , FDG-PET in diagnosis, staging and prognosis of pancreatic carcinoma: a meta-analysis. World J Gastroenterol, 2013. 19(29): p. 4808–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Roh JL, et al. , Utility of 2-[18F] fluoro-2-deoxy-D-glucose positron emission tomography and positron emission tomography/computed tomography imaging in the preoperative staging of head and neck squamous cell carcinoma. Oral Oncol, 2007. 43(9): p. 887–93. [DOI] [PubMed] [Google Scholar]
- 16.Holder WD Jr., et al. , Effectiveness of positron emission tomography for the detection of melanoma metastases. Ann Surg, 1998. 227(5): p. 764–9; discussion 769–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Clarke T, et al. , Evaluation of follow-up strategies for patients with epithelial ovarian cancer following completion of primary treatment. Cochrane Database Syst Rev, 2014(9):p. CD006119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Schwarz-Dose J, et al. , Monitoring primary systemic therapy of large and locally advanced breast cancer by using sequential positron emission tomography imaging with [18F]fluorodeoxyglucose. J Clin Oncol, 2009. 27(4): p. 535–41. [DOI] [PubMed] [Google Scholar]
- 19.Fischer B, et al. , Preoperative staging of lung cancer with combined PET-CT. N Engl J Med, 2009. 361(1): p. 32–9. [DOI] [PubMed] [Google Scholar]
- 20.Bryant AS and Cerfolio RJ, The maximum standardized uptake values on integrated FDG-PET/CT is useful in differentiating benign from malignant pulmonary nodules. Ann Thorac Surg, 2006. 82(3): p. 1016–20. [DOI] [PubMed] [Google Scholar]
- 21.Tan BB, et al. , The solitary pulmonary nodule. Chest, 2003. 123(1 Suppl): p. 89S–96S. [DOI] [PubMed] [Google Scholar]
- 22.Koch M, et al. , Threshold analysis and biodistribution of fluorescently labeled bevacizumab in human breast cancer. Cancer Res, 2016. [DOI] [PubMed] [Google Scholar]
- 23.Rosenthal EL, et al. , Safety and Tumor Specificity of Cetuximab-IRDye800 for Surgical Navigation in Head and Neck Cancer. Clin Cancer Res, 2015. 21(16): p. 3658–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.van Dam GM, et al. , Intraoperative tumor-specific fluorescence imaging in ovarian cancer by folate receptor-alpha targeting: first in-human results. Nat Med, 2011. 17(10): p. 1315–9. [DOI] [PubMed] [Google Scholar]
- 25.Predina J,NA, Keating J, Dunbar A, Connolly C, Singhal S, Near-Infrared Intraoperative Molecular Imaging Can Enhance Malignant Pleural Mesothelioma Resection, in The 25th Annual Meeting of the Asian Society for Cardiovascular and Thoracic Surgery. 2017: Seoul, Korea. [Google Scholar]
- 26.Boogerd LS, et al. , Concordance of folate receptor-alpha expression between biopsy, primary tumor and metastasis in breast cancer and lung cancer patients. Oncotarget, 2016. 7(14): p. 17442–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Madajewski B, et al. , Intraoperative near-infrared imaging of surgical wounds after tumor resections can detect residual disease. Clin Cancer Res, 2012. 18(20): p. 5741–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Shen J, et al. , Assessment of folate receptor-beta expression in human neoplastic tissues. Oncotarget, 2015. 6(16): p. 14700–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Stoffels I, et al. , Metastatic status of sentinel lymph nodes in melanoma determined noninvasively with multispectral optoacoustic imaging. Sci Transl Med, 2015. 7(317): p. 317ra199. [DOI] [PubMed] [Google Scholar]
- 30.Kelderhouse LE, et al. , Development of tumor-targeted near infrared probes for fluorescence guided surgery. Bioconjug Chem, 2013. 24(6): p. 1075–80. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 1: FDG-avid and Non-PET avid malignant pulmonary nodules display fluorescence during IMI with OTL38. Representative example of a non-PET avid malignant pulmonary nodule demonstrating fluorescence during resection. Subject 23, a 51-year-old female, presented with a 1.2cm left upper lobe nodule (a) which had no activity by PET (b). During resection (c) brightfield and (d) fluorescent overlays revealed tumor fluorescence with a TBR of 3.2. Pathology demonstrated adenocarcinoma. Bar graphs comparing (e) size and (f) fluorescence of PET avid and Non-PET avid malignant pulmonary nodules. ns-not significant. * p<0.05
Supplemental Figure 2: Bar graphs comparing (a) size and (b) fluorescence of synchronous nodules to fluorescent pulmonary nodules discovered preoperatively. ns-not significant. * p<0.01