Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
1 SUPPLEMENTARY MATERIALS AND METHODS / FIGURE LEGENDS 2 Trial Design 3 Tracer accumulation was defined as fluorescence activity assessed either by detection by the 4 optical imaging system in vivo, ex vivo of the excised specimen, or ex vivo fluorescence flatbed 5 scanning. The study was to be suspended if any undesired SAE occurred throughout the study or 6 if no accumulation of bevacizumab-IRDye800CW could be detected in any of the first seven 7 patients. The full study aimed at evaluating a total of 20 patients for bevacizumab-IRDye800CW 8 accumulation, with replacement of 10 patients if needed, amounting to a maximum of 30 patients 9 who would receive a bevacizumab-IRDye800CW dose. Patients were eligible if they were ≥18 10 years of age and had a World Health Organisation (WHO) performance score of 0 - 2. Exclusion 11 criteria were the presence of another invasive malignancy, another serious medical condition (as 12 determined by the research team), pregnancy or lactation, prior neo-adjuvant chemotherapy, prior 13 radiotherapy of the involved area, major surgery within 28 days before initiation of the study 14 procedures, prior allergic reaction to immunoglobulins, or the presence of a breast prosthesis in 15 the breast with cancer. In the first five patients, a tumor size of at least 15 mm, according to 16 standard anatomical imaging data (i.e. mammography, ultrasound and/or dynamic contrast 17 enhanced magnetic resonance imaging (DCE-MRI)), was required. From the sixth patient 18 onwards, 5 mm diameter was considered sufficient, as based on the data provided by the 89Zr- 19 bevacizumab patient study executed at the UMCG at the same time and as such approved by the 20 Institutional Review Board (IRB) (19). Routine tumor staging included mammography and 21 ultrasound. MRI was performed if better tumor extent definition was needed as standard of care 22 in the Netherlands. The Institutional Review Board (IRB) of the University Medical Center 23 Groningen (UMCG) approved the study, with local agreement of the University Medical Center 24 Utrecht (UMCU). All patients gave written informed consent. 1 1 Bevacizumab-IRDye800CW Preparation and Injection 2 The development (i.e. implementation of the manufacturing process and analytical methods 3 according to current Good Manufacturing Practice (cGMP), formulation studies, extended 4 characterization and stability testing) and safety testing (extended single dose toxicity study in 5 mice) was reported for clinical grade bevacizumab-IRDye800CW (27). Both GMP produced 6 bevacizumab and IRDye800CW were commercially obtained from the manufacturers. The costs 7 of one vial of 4,5 mg bevacizumab-IRDye800CW in a non-commercial academic GMP 8 production setting were estimated to be $550. Briefly, bevacizumab and IRDye800CW were 9 reacted in a molar ratio of 1:4 in phosphate-buffered saline (PBS, pH 8.5). Thereafter, the product 10 was purified by gel filtration using PD-10 Desalting Columns (GE Healthcare). Labeling 11 efficiency and purity were determined by validated size-exclusion high-performance liquid 12 chromatography. The labeling efficiency was 80-85%, resulting in an average labeling ratio of 13 3.3 dyes per antibody with a purity of > 95% after purification. The stability of the tracer was 14 verified by storing the product in 0.9% NaCl for 7 days at various conditions (-80, -20 and 4C), 15 for which storage at -20C proved to be sufficient. According to FDA/EMEA guidelines (20) we 16 adhered to the microdosing concept of imaging studies, as defined by a dose of ≤30 nmol for 17 proteins and imaging studies, also described by Kummar et al and the Task Force on 18 Methodology for the Development of Innovative Cancer Therapies (MDICT) (21). The molecular 19 weight of the protein bevacizumab is 149 kDa. The fluorescent dye has a molecular weight of 20 1,166 kDa. With an average conjugation ratio of 1:4 protein:IRDye800CW, the total molecular 21 weight is 153,7 kDa. This means for our 4.5 mg tracer dose bevacizumab-IRDye800CW, a 22 systemic injection of 26 nmol of bevacizumab-IRDye800CW. 2 1 After release of the final product by the certified qualified person at the UMCG GMP facility, 2 patients received 4.5 mg bevacizumab-IRDye800CW at a concentration of 1 mg/ml as 3 intravenous bolus injection 3 days before scheduled surgery, based on the preclinical 4 pharmacokinetic data obtained in breast cancer tumor bearing mice injected with bevacizumab- 5 IRDye800CW (26). After injection, the infusion line was flushed with 0.9% NaCl and patients 6 were observed and vital signs were monitored for the occurrence of adverse events for 4 hours 7 (patient 1-9) or 1 hour (patient 10-20) after tracer administration. Immediately after injection, as 8 well as on day 3 (directly prior to surgery) and on day 10-14 after surgery, blood was drawn for 9 bevacizumab-IRDye800CW whole blood measurement (immediately after injection, day 3, and 10 day 10-14). 11 Optical Imaging System 12 At both participating centers, a similar near-infrared (NIR) optical imaging system was used for 13 intra-operative and fluorescence imaging of the excised specimen. The optical imaging system 14 was developed by the Technical University Munich (TUM)/Helmholtz Center) and has been 15 described previously (23,26,30). In brief, the excitation wavelength was created by a 750 nm- 16 300mW continuous wave diode laser (BWF2-750-0, B&W Tek) guided through a widening lens 17 (F260SMA-B, Thorlabs Inc) and a diffuser (DG10-220, Thorlabs Inc). The field of view (FOV) 18 was illuminated by a 250 W halogen lamp (KL-2500LCD, Schott AG), guided through a short 19 pass filter (E700SP-2P, Chroma) to prevent illumination crosstalk to the fluorescence channel. 20 The FOV was imaged by a motorized lens (CVO GAZ11569M, Goyo Optical), separated in 21 visible and NIR fluorescence parts by a dichroic beam splitter (T760lpxr, Chroma) and relayed to 22 appropriate CCD sensors accordingly (MAP10100100, Thorlabs Inc). The NIR fluorescence 23 image was guided through a band pass filter (ET810/90, Chroma) for eliminating non- 24 fluorescence parts of the spectrum and acquired by a scientific deep-cooled (-70°C) back3 1 illuminated CCD Camera (iXon DU-888, Andor Technology). The color-reflection image was 2 guided through a shortpass filter (E700SP-2P, Chroma) and acquired by a 12-bit bayer-patterned 3 color camera (pixelfly QE 275xs PCO AG). For mitigation of specula reflections on wet tissue 4 surfaces in the color image the white-light illumination and color camera reflection channel were 5 equipped with linear-polarizers (21003b, Chroma) in a cross-polarization geometry. The heat 6 disposal for the deep-cooled CCD camera was achieved by additional liquid cooling. Both, color 7 and fluorescence images where acquired, processed, stored and displayed simultaneously at video 8 rate. The acquisition was performed by a standard PC where image correlation, de-bayering, 9 mapping to visualization dynamic range and image fusion are performed at video rates using the 10 Graphics Processing Unit (GPU). All imaging parameters including the zoom, iris and focus of 11 the motorized lens are controllable by a lightweight, user-friendly, GPU-accelerated acquisition 12 and visualization software. The color reflectance image, the raw florescence image or a fused 13 false-color overlay image are displayed instantly in video rate and latency-free at the build-in 14 screens of the operational theater. A synchronized multi-CPU system (Dell Computers) controls 15 the camera system and can process raw data and image registration and rendering simultaneously. 16 The field-of-view (FOV) can be varied from 3 x 3 cm to 15 x 15 cm, with corresponding 17 resolution varying from 30 x 30 µm to 150 x 150 µm per pixel for a total of 512 x 512 pixels per 18 fluorescence image. Color and fluorescence images were simultaneously collected by the two 19 cameras and processed. Custom made software was used for imaging processing and post- 20 processing. 21 Sterile transparent drapes (Carl Zeiss Vision BV) covered the camera-head and the arm 22 during surgery. As the system was a prototype, no CE marking was obtained. Instead, the device 23 was verified for laser safety and leak currents as tested at the TUM and by the Medical Devices 24 Unit at the UMCG and UMCU. Both imaging devices were approved for intraoperative 4 1 application in humans. For standardization purposes of imaging data between both hospitals, a 2 standard concentration series of indocyanin green (ICG, ICG-PULSION, PULSION Medical 3 Systems) for each imaging session was evaluated, which made comparison between both systems 4 and centers possible. 5 6 Ex Vivo Analyses of Tumor Samples 7 Imaging fresh surgical specimen. After bread-loaf slicing of the fresh surgical breast 8 specimen and before formalin fixation, all slices were imaged with the optical imaging system for 9 determination of fluorescence signal in tumor and surrounding healthy breast tissue. 10 Subsequently, small (2-6 mm) biopsies from tumor tissue and normal breast parenchyma were 11 taken from the bread-loaf slices, snap frozen and stored in dark conditions at −80C, provided the 12 tumor was large enough and no interference with standard processing procedures was anticipated. 13 Subsequently, the specimen was fixed in formalin and embedded in paraffin (FFPE) as a routine 14 procedure. 15 Imaging FFPE blocks and slides. FFPE blocks were available for all patients, including 16 tumor tissue and surrounding tissue. We used the Odyssey® CLX fluorescence scanning system 17 (LI-COR Biosciences Inc.) for assessment of fluorescence in the NIR region and location in the 18 FFPE blocks and/or tissue slides. Additionally, whole blocks were imaged by the intra-operative 19 real-time optical imaging camera to compare both techniques. 20 Bevacizumab-IRDye800CW and VEGF-A quantification 21 For bevacizumab-IRDye800CW quantification in tumor tissue, adjacent margin tissue and 22 normal breast tissue, each biopsy was weighed and subsequently disrupted with a TissueLyser II 23 system (Qiagen) using pre-cooled Eppendorf holders, 5 mm stainless steel beads, and 5 1 RIPA buffer (50 mM Tris–HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% 2 SDS) supplemented with a complete EDTA-free mini tablet protease inhibitor cocktail (Roche 3 Applied Science). The homogenates were then diluted in series (1:2 dilution steps) in 96-well 4 plates, with the same buffer used for homogenization of the biopsies. In parallel, the 5 dilution series of the probe were also made using RIPA buffer. The intensity of the 6 NIR fluorescence of the samples and standard probe was detected at 800 nm with the Odyssey 7 scanner using the following settings: intensity 7, focus offset 3.5, quality: medium, resolution 169 8 μm. Subsequently, the concentration of the probe present in the homogenates was extrapolated 9 from the calibration curves made with the standard probe, using the software Prism 6 (GraphPad 10 Software Inc.). These concentration values were used to calculate the concentration of the tracer 11 per mg of tissue biopsy (ng/ml/mg) (31). 12 For quantification of VEGF-A expression (pg/mL) in the surgical specimen, a VEGF-A 13 Enzyme-Linked Immuno Sorbent (ELISA) assay was performed according to the manufacturer 14 instructions (Quantikine). One to three random samples were collected from the surgical 15 specimen of the primary tumor, margin tumor-surrounding tissue, and surrounding (normal) 16 tissue. Tissue was lyzed manually using RIPA buffer (50 mM Tris–HCl pH 7.4, 150 mM NaCl, 1 17 mM EDTA, 1% Triton X-100, 0.1% SDS) supplemented with a complete EDTA-free mini tablet 18 protease inhibitor cocktail (Roche Applied Science). Thereafter, whole lysate mixtures were first 19 measured on the Odyssey scanner at the 800 nm channel to determine the intensity of the 20 NIR fluorescence of each sample using the following settings: intensity 7, focus offset 3.5, 21 quality: medium, resolution 169 μm. Subsequently, lysates were centrifuged to spin down cellular 22 debris, and supernatants were used to determine VEGF-A levels by ELISA. Results of VEGF-A 23 measurement were normalized for the total weight of the tissue sample before sample analysis. 24 Tumor lysates of three patients were also analyzed by sodium dodecyl sulfate polyacrylamidegel 6 1 electrophoresis (SDS-PAGE), to ensure the complete compound bevacizumab-IRDye800CW 2 was present in the primary breast tumor. 3 Immunohistochemistry 4 FFPE blocks of tumor tissue and normal healthy breast tissue were available for all patients and 5 negative controls. Secondary use of excised breast cancer specimen from four patients 6 undergoing breast-conserving surgery as negative controls was approved by the IRB and the 7 Department of Pathology. Paraffin sections were cut at 4 μm, mounted on silane-coated slides 8 and dried overnight at 37 °C. All specimens were examined by one pathologist using 9 standard H/E staining. Additionally, (I)HC staining was performed for VEGF, CD34 and 10 collagen. Before (I)HC staining, slides were deparaffinized using xylene, and the 800 nm NIR 11 fluorescence signal was measured on the Odyssey scanner. The tissue blocks and slides were 12 scanned with the following settings. For blocks: wavelength 800 nm, intensity 6, focus offset 1, 13 quality: highest, resolution 21 μm. For slides (4 μm and 10 μm where appropriate): wavelength 14 800 nm, intensity 9, focus offset 0, quality: highest, resolution 21 μm. Subsequently, CD34 was 15 stained using monoclonal antibody anti-CD34 clone Qbend10 (Immunotech) and the OptiView 16 detection kit on the Ventana Benchmark Ultra. Collagen was stained with Sirius red. For VEGF- 17 A staining we used the following protocol. First, slides were rehydrated via graded 18 alcohols. Endogenous peroxidase blocking was performed during 15 minutes with a phosphate 19 citrate buffer containing 5% hydrogen peroxide, followed by heat induced antigen retrieval in 0.1 20 M 95-99 °C sodium citrate (pH 6.0) for 20 minutes on a hot plate. After washing steps, sections 21 were incubated overnight at 4°C with polyclonal rabbit anti-human VEGF-A (RB9031, 22 ThermoScientific) in a concentration of 0.2 µg/mL (1:1000). Tissue was stained using poly-HRP- 23 anti Ms/Rb/Rt IgG (BrightVision) for 30 minutes at room temperature, followed by visualization 24 with diaminobenzidine for 10 minutes. PBS was used throughout for washing. Finally, sections 7 1 were counterstained with Mayer’s hematoxylin, dehydrated in alcohol and xylene and 2 coverslipped. Resection margins were defined according to the Dutch guidelines: a free margin is 3 that tumor does not reach into any of the surgical margins in an adequately processed sample. 4 There is focal spread in a surgical margin if the tumor (invasive carcinoma and/or DCIS) reaches 5 into a single limited area (≤ 4 mm) in an inked margin (33). 6 Fluorescence microscopy 7 Additional microscopic evaluation of NIR fluorescence signal was performed on 8 additional 4 μm sections, with counterstaining of the nuclei (Hoechst staining, 33258; Invitrogen) 9 and mounting with Kaiser’s glycerine (modified), using an inverted laser-scanning microscope 10 (DMI6000B, Leica Biosystems GmbH). To optimize NIR visualization, the setup was equipped 11 with additional accessories, including a highly sensitive monochrome DFC365 FX fluorescence 12 camera (1.4M Pixel CCD, Leica Biosystems GmbH), an optimal NIR filter set (two band-pass 13 filters 850-90m-2p and a long-pass emission filter HQ800795LP; Chroma Technology Corp®), a 14 highly sensitive NIR LED light source ranging up to 900 nm (X-Cite® 200DC, Excelitas™ 15 Technologies) and LAS-X software (Leica Biosystems GmbH). 16 Multiplex Advanced Pathology Imaging (MAPI) methodology 17 For ex vivo correlation of the fluorescence scans obtained by the optical imaging system 18 and the Odyssey imaging system, a workflow methodology was developed for cross-correlation 19 of fluorescence scans with IHC for VEGF-A, CD34 (microvessel density) and presence of 20 collagen (Sirius Red staining) in a multiplex overlay (Supplemental Fig. S6 and Supplemental 21 Movie S1). After whole slide NIR fluorescence scanning on the Odyssey imaging system, IHC 22 stained slides were scanned on a digital slide scanner (ScanScope XT) at 20x resolution (0.50 23 µm/pixel). Color deconvolution with vectors H DAB (ImageJ version 1.48) was used to separate 24 DAB and hematoxylin color layers. DAB color layers were converted to red-black images using 8 1 auto threshold. NIRF Odyssey images of the same slide were converted green-black images also 2 with auto threshold. NIRF images and IHC images of the same slide where registered to show 3 expression of bevacuzimab-IRDye800CW in the H/E stained histology slides and identify co- 4 localization of the bevacuzimab-IRDye800CW NIRF signal and VEGF-A or collagen. 5 Statistical Analysis 6 Given the early phase nature of this exploratory study, the emphasis was on a high power (1-β) to 7 detect potentially clinically relevant bevacizumab-IRDye800CW breast tumor accumulation in 8 patients with primary breast cancer. Tracer accumulation was defined as fluorescence activity 9 assessed either by detection by the optical imaging system in vivo, or ex vivo of the excised breast 10 cancer specimen, bread-loaf slices, paraffin embedded tissue blocks, or fluorescence flatbed 11 scanning of tissue slides. This was done to avoid erroneous dismissal of a successful compound 12 in the early stage of its evaluation. Accordingly, if no accumulation was detected in any of the 13 first seven consecutive patients of stage 1, we could dismiss a true accumulation rate of 35% or 14 larger with a β of 5% or smaller, and a true accumulation rate of 48% or larger at the ≤ 1% β 15 level. Similarly, the study was powered at 99.99% to continue after the first seven patients given 16 a true accumulation rate of 75%. If the study continued to its full size in the second stage with a 17 total of 20 evaluable patients for accumulation of the bevacizumab-IRDye800CW tracer, the 18 margin of error of the accumulation rate estimate (as calculated by the Score method using α = 19 5%) would be at most 20% for an observed accumulation rate of 50%, yielding an accumulation 20 rate estimate of 0.50 with a 95% confidence interval (CI) of 0.30 to 0.70. As one patient in our 21 study did not show accumulation, this related to an accumulation rate of 0.95 (95%CI: 76-99%). 22 Similarly, for an observed accumulation rate of 75%, the margin of error would be 18%, yielding 23 an accumulation rate estimate of 0.75 (95% CI: 0.53-0.89). The full-study sample size thus 24 allowed reasonably precise estimates of the bevacizumab-IRDye800CW accumulation rate. 9 1 Sample size considerations pertained only to bevacizumab-IRDye800CW tumor accumulation, as 2 the expected incidence of SAEs was so low that the study size needed to formally address and 3 estimate SAE incidence associated with the compound was prohibitively large. Nevertheless, 4 given a true SAE incidence rate of 2.5%, the chance of observing an SAE in the study 5 approached 50%, which is close to 75% for a true SAE incidence rate of 5.0%. Data-analysis of 6 secondary endpoints was exploratory in nature. The secondary outcome measures of the clinical 7 study included the detection ability of the optical imaging system camera of the fluorescent signal 8 from bevacizumab-IRDye800CW during surgery and to assess and compare the presence of a 9 fluorescent signal in breast cancer tissue and normal tissue by assessing the images during and 10 after surgery, the accumulation, tissue distribution, localization and quantification of the tracer in 11 resected tumor tissue, and surrounding healthy tissue, as measured by fluorescence imaging in 12 vivo and ex vivo. Correlations of intraoperative imaging to ex vivo quantitative measures of 13 fluorescence, tracer distribution, and IHC analyses of surgical specimens were investigated for 14 the purposes of co-localization and correlation (Pearson correlation) and for determination of 15 MFI obtained by fluorescence flatbed scans in regions of tumor, stroma and healthy surrounding 16 tissue (paired t-test or Wilcoxon signed rank test for paired data and the unpaired t-test for 17 unpaired data). Data are presented as a mean ± standard deviation (SD). A two-sided P value of 18 less than 0.05 was considered significant (SPSS, version 19.0; IBM). 19 20 21 22 23 24 10 1 SUPPLEMENTAL FIGURE LEGENDS / MOVIE 2 3 Supplemental Figure S1. Near-Infrared Optical Imaging of Positive Tumor Margin. In one 4 of the two patients with a positive margin (see also Fig. 1), the bevacizumab-IRDye800CW tracer 5 could be detected ex vivo by optical imaging of the excised lump (panel A white-light, panel B 6 fluorescence, panel C overlay). Bread-loaf slicing (panel D-F) and the paraffin block (panel G-I) 7 is shown for the corresponding fluorescent signal. Hematoxylin/eosin (H/E) stain (panel J), 8 fluorescence flatbed scanning (panel K) and image overlay (panel L) is shown for the 4 μm slide 9 (red/white/black dashed line = tumor outline, white/black arrow = positive tumor margin). 10 11 12 Supplemental Figure S2. Representative Fluorescence Flatbed Scans of Paraffin Blocks of all 13 Patients (N=20). Of all included patients, representative paraffin blocks are depicted, including 5 14 negative controls of patients with breast cancer (i.e. tumor present but without injection of 15 bevacizumab-IRDye800CW, last column). Tumor areas are demarcated by dotted lines. * 16 (UMCU4) = skin, ** (UMCG9) = fibroadenoma. Fluorescence is depicted by mean fluorescence 17 intensity (MFI). Macroscopic Segmentation of Fluorescence Tissue Localization and Target- 18 to-Background Ratios (TBR). The mean TBR is depicted (** = P < 0.001) in a separate panel. 19 11 1 Supplemental Figure S3. Microscopic Biodistribution of Bevacizumab-IRDye800CW. Tissue 2 slides of 4 μm were scanned on a flatbed scanner for mean fluorescence intensity, followed by 3 hematoxylin/eosin (H/E) staining and microscopy (2x panel A and 20x panel B) for co- 4 localization at the microscopic level of fluorescence intensities (panel C and D) in a tumor sprout 5 surrounded by blood vessels and collagen-rich stroma (rectangular). Color and fluorescence 6 images were superimposed (panel E and F). Localization of Fluorescence Bevacizumab- 7 IRDye800CW in Human Primary Breast Cancer. Near-infrared fluorescence microscopy of 4 8 µm tissue slides (panel G: [BLUE] = Hoechst stain, [RED] = NIR channel / bevacizumab- 9 IRDye800CW, white arrows = tumor margin) and subsequently stained for hematoxylin/eosin 10 (panel H, black arrows = tumor margin). 11 12 Supplemental Figure S4. Co-localization of Bevacizumab-IRDye800CW, hematoxylin/eosin 13 (H/E), VEGF-A, Collagen and CD34 in Breast Cancer and a Satellite Lesion. A fresh excised 14 breast tumor specimen with a satellite lesion was processed with multiplex advanced pathology 15 imaging (MAPI) for co-localization of white light images (white-light panel A, H/E panel E, 16 VEGF-A staining panel I, collagen panel M, CD34 panel Q), black-and-white bevacizumab- 17 IRDye800CW fluorescence (panel B and green pseudocolor in panel C, G, K, O and S), 18 segmentation (panel F), deconvolution (VEGF-A panel J, collagen panel N, CD34 panel R) and 19 overlay images (panel D: white-light + fluorescence, panel H: H/E and fluorescence, panel L: 20 VEGF-A and fluorescence, panel P: collagen and fluorescence and panel T: CD34). In panel T an 21 overlay of CD34 (blue color) bevacizumab-IRDye800CW fluorescence (green) and VEGF-A 22 (red) is depicted. 23 12 1 Supplemental Figure S5. Co-localization of Bevacizumab-IRDye800CW, hematoxylin / eosin 2 (H/E), VEGF-A, Collagen and CD34 in Ductal Carcinoma In Situ (DCIS). A fresh excised 3 breast tumor specimen with DCIS was processed with multiplex advanced pathology imaging 4 (MAPI) for co-localization of white light images (white-light panel A, H/E panel E, VEGF-A 5 staining panel I, collagen panel M, CD34 panel Q), black-and-white bevacizumab-IRDYe800CW 6 fluorescence (panel B and green pseudocolor in panels C, G, K, O and S), segmentation (panel 7 F), deconvolution (VEGF-A panel J, collagen panel N, CD34 panel R) and overlay images (panel 8 D: white-light + fluorescence, panel H: H/E and fluorescence, panel L: VEGF-A and 9 fluorescence, panel P: collagen and fluorescence and panel T: CD34). In panel T and overlay of 10 CD34 (blue color) bevacizumab-IRDye800CW fluorescence (green) and VEGF-A (red) is 11 depicted. 12 13 Supplemental Figure S6. Intraoperative and Ex Vivo Optical Imaging of Axillary Lymph 14 Nodes. Intraoperative optical imaging of the axillary dissection (white-light panel A, black-and- 15 white fluorescence panel B (lymph nodes encircled white), pseudocolor overlay panel C). 16 Paraffin-embedded tissue block (white light panel D, pseudocolor fluorescence panel E, 17 hematoxylin/eosin staining panel F). VEGF-A IHC staining (panel G), red deconvolution (panel 18 H), bevacizumab-IRDye800CW flatbed scanning (panel I), green deconvolution (panel J) and 19 VEGF-A / bevacizumab-IRDye800CW overlay (panel K). 20 13 1 Supplemental Figure S7. Co-localization of Hematoxylin / Eosin (H/E), VEGF-A, and 2 Bevacizumab-IRDye800CW in Tumor-negative Lymph Nodes. Two excised lymph nodes 3 were processed for co-localization of white light images (panel A), pseudocolor fluorescence 4 (panel B) and Hematoxylin/eosin (H/E, panel C). VEGF-A staining (panel D) red deconvolution 5 (panel E), bevacizumab-IRDye800CW flatbed scanning (panel F), green deconvolution (panel 6 G), and overlay VEGF-A / bevacizumab-IRDye800CW (panel H). 7 8 Supplemental Movie S1. Explanatory Movie of Multiplex Advanced Pathology Imaging 9 (MAPI). Multiplex pathology imaging of white-light microscopy, bevacizumab-IRDye800CW 10 fluorescence imaging (green pseudocolor), VEGF-A (red pseudocolor), collagen (red 11 pseudocolor) and CD34 (blue pseudocolor) in breast cancer, including overlay images is shown 12 in a stepwise manner for reasons of clarity. 13 14 14