Quantitative image analysis pipeline for detecting circulating hybrid cells in immunofluorescence images with human-level accuracy.

Circulating hybrid cells (CHCs) are a newly discovered, tumor-derived cell population found in the peripheral blood of cancer patients and are thought to contribute to tumor metastasis. However, identifying CHCs by immunofluorescence (IF) imaging of patient peripheral blood mononuclear cells (PBMCs) is a time-consuming and subjective process that currently relies on manual annotation by laboratory technicians. Additionally, while IF is relatively easy to apply to tissue sections, its application to PBMC smears presents challenges due to the presence of biological and technical artifacts. To address these challenges, we present a robust image analysis pipeline to automate the detection and analysis of CHCs in IF images. The pipeline incorporates quality control to optimize specimen preparation protocols and remove unwanted artifacts, leverages a ß-variational autoencoder (VAE) to learn meaningful latent representations of single-cell images, and employs a support vector machine (SVM) classifier to achieve human-level CHC detection. We created a rigorously labeled IF CHC data set including nine patients and two disease sites with the assistance of 10 annotators to evaluate the pipeline. We examined annotator variation and bias in CHC detection and provided guidelines to optimize the accuracy of CHC annotation. We found that all annotators agreed on CHC identification for only 65% of the cells in the data set and had a tendency to underestimate CHC counts for regions of interest (ROIs) containing relatively large amounts of cells (>50,000) when using the conventional enumeration method. On the other hand, our proposed approach is unbiased to ROI size. The SVM classifier trained on the ß-VAE embeddings achieved an F1 score of 0.80, matching the average performance of human annotators. Our pipeline enables researchers to explore the role of CHCs in cancer progression and assess their potential as a clinical biomarker for metastasis. Further, we demonstrate that the pipeline can identify discrete cellular phenotypes among PBMCs, highlighting its utility beyond CHCs.

Nerve-Sparing Gynecologic Surgery Enabled by A Near-Infrared Nerve-Specific Fluorophore Using Existing Clinical Fluorescence Imaging Systems.

Patients undergoing gynecological procedures suffer from lasting side effects due to intraoperative nerve damage. Small, delicate nerves with complex and nonuniform branching patterns in the female pelvic neuroanatomy make nerve-sparing efforts during standard gynecological procedures such as hysterectomy, cystectomy, and colorectal cancer resection difficult, and thus many patients are left with incontinence and sexual dysfunction. Herein, a near-infrared (NIR) fluorescent nerve-specific contrast agent, LGW08-35, that is spectrally compatible with clinical fluorescence guided surgery (FGS) systems is formulated and characterized for rapid implementation for nerve-sparing gynecologic surgeries. The toxicology, pharmacokinetics (PK), and pharmacodynamics (PD) of micelle formulated LGW08-35 are examined, enabling the determination of the optimal imaging doses and time points, blood and tissue uptake parameters, and maximum tolerated dose (MTD). Application of the formulated fluorophore to imaging of female rat and swine pelvic neuroanatomy validates the continued clinical translation and use for real-time identification of important nerves such as the femoral, sciatic, lumbar, iliac, and hypogastric nerves. Further development of LGW08-35 for clinical use will unlock a valuable tool for surgeons in direct visualization of important nerves and contribute to the ongoing characterization of the female pelvic neuroanatomy to eliminate the debilitating side effects of nerve damage during gynecological procedures.

Clinically relevant dual probe difference specimen imaging (DDSI) protocol for freshly resected breast cancer specimen staining.

BACKGROUND: Re-excision rates following breast conserving surgery (BCS) remain as high as ~ 35%, with positive margins detected during follow-up histopathology. Additional breast cancer resection surgery is not only taxing on the patient and health care system, but also delays adjuvant therapies, increasing morbidity and reducing the likelihood of a positive outcome. The ability to precisely resect and visualize tumor margins in real time within the surgical theater would greatly benefit patients, surgeons and the health care system. Current tumor margin assessment technologies utilized during BCS involve relatively lengthy and labor-intensive protocols, which impede the surgical work flow. METHODS: In previous work, we have developed and validated a fluorescence imaging method termed dual probe difference specimen imaging (DDSI) to accurately detect benign and malignant tissue with direct correlation to the targeted biomarker expression levels intraoperatively. The DDSI method is currently on par with touch prep cytology in execution time (~ 15-min). In this study, the main goal was to shorten the DDSI protocol by decreasing tissue blocking and washing times to optimize the DDSI protocol to < 10-min whilst maintaining robust benign and malignant tissue differentiation. RESULTS: We evaluated the utility of the shortened DDSI staining methodology using xenografts grown from cell lines with varied epidermal growth factor receptor (EGFR) expression levels, comparing accuracy through receiver operator characteristic (ROC) curve analyses across varied tissue blocking and washing times. An optimized 8-min DDSI methodology was developed for future clinical translation. CONCLUSIONS: Successful completion of this work resulted in substantial shortening of the DDSI methodology for use in the operating room, that provided robust, highly receptor specific, sensitive diagnostic capabilities between benign and malignant tissues.

Assessment of human pancreas cancer tissue and precursor lesions via a fluorophore with inherent PDAC selectivity.

The current five-year survival rate of <5% for pancreatic ductal adenocarcinoma (PDAC) is compounded by late diagnosis, a lack of PDAC-specific intraoperative guidance to ensure complete resection, and the ineffectiveness of current therapies. Previously, utilizing compound 1, a fluorophore with inherent PDAC selectivity, PDAC was visualized both in vivo and ex vivo in a murine model. In the current study, human PDAC tissue is targeted. Compound 1 selectively stains ducts of the adenocarcinoma versus the surrounding stroma, enabling the imaging of PDAC in frozen tissue sections with high contrast. To enhance the potential of 1 for intraoperative applications, the ex vivo staining protocol was optimized for rapid margin assessment, with a final staining time of ~15 min. To measure diagnostic performance, the area under a receiver operating characteristic (ROC) curve was measured for the identification of ductal adenocarcinoma vs. stroma. The bright fluorescence contrast enabled quantitative determination of PDAC (or precancerous PanIN lesions) versus healthy pancreas tissue in human tissue array samples.

Simultaneous Multicolor Single-Molecule Tracking with Single-Laser Excitation via Spectral Imaging.

Single-molecule tracking (SMT) offers rich information on the dynamics of underlying biological processes, but multicolor SMT has been challenging due to spectral cross talk and a need for multiple laser excitations. Here, we describe a single-molecule spectral imaging approach for live-cell tracking of multiple fluorescent species at once using a single-laser excitation. Fluorescence signals from all the molecules in the field of view are collected using a single objective and split between positional and spectral channels. Images of the same molecule in the two channels are then combined to determine both the location and the identity of the molecule. The single-objective configuration of our approach allows for flexible sample geometry and the use of a live-cell incubation chamber required for live-cell SMT. Despite a lower photon yield, we achieve excellent spatial (20-40 nm) and spectral (10-15 nm) resolutions comparable to those obtained with dual-objective, spectrally resolved Stochastic Optical Reconstruction Microscopy. Furthermore, motions of the fluorescent molecules did not cause loss of spectral resolution owing to the dual-channel spectral calibration. We demonstrate SMT in three (and potentially more) colors using spectrally proximal fluorophores and single-laser excitation, and show that trajectories of each species can be reliably extracted with minimal cross talk.

Near-infrared fluorescent digital pathology for the automation of disease diagnosis and biomarker assessment.

Hematoxylin-eosin (H&E) staining of tissue has been the mainstay of pathology for more than a century. However, the learning curve for H&E tissue interpretation is long, whereas intra- and interobserver variability remain high. Computer-assisted image analysis of H&E sections holds promise for increased throughput and decreased variability but has yet to demonstrate significant improvement in diagnostic accuracy. Addition of biomarkers to H&E staining can improve diagnostic accuracy; however, coregistration of immunohistochemical staining with H&E is problematic as immunostaining is completed on slides that are at best 4 µm apart. Simultaneous H&E and immunostaining would alleviate coregistration problems; however, current opaque pigments used for immunostaining obscure H&E. In this study, we demonstrate that diagnostic information provided by two or more independent wavelengths of near-infrared (NIR) fluorescence leave the H&E stain unchanged while enabling computer-assisted diagnosis and assessment of human disease. Using prostate cancer as a model system, we introduce NIR digital pathology and demonstrate its utility along the spectrum from prostate biopsy to whole mount analysis of H&E-stained tissue.

Polymeric Micelles as Carriers for Nerve-Highlighting Fluorescent Probe Delivery.

Nerve damage during surgery is a common morbidity experienced by patients that leaves them with chronic pain and/or loss of function. Currently, no clinically approved imaging technique exists to enhance nerve visualization in the operating room. Fluorescence image-guided surgery has gained in popularity and clinical acceptance over the past decade with a handful of imaging systems approved for clinical use. However, contrast agent development to complement these fluorescence-imaging systems has lagged behind with all currently approved fluorescent agents providing untargeted blood pool information. Nerve-specific fluorophores are known, however translations of these agents to the clinic has been complicated by their lipophilic nature, which necessitates specialized formulation strategies for successful systemic administration. To date the known nerve-specific fluorophores have only been demonstrated preclinically due to the necessity of a dimethyl sulfoxide containing formulation for solubilization. In the current study, a polymeric micellar (PM) formulation strategy was developed for a representative nerve-specific fluorophore from the distyrylbenzene family, BMB. The PM formulation strategy was able to solubilize BMB and demonstrated improved nerve-specific accumulation and fluorescence intensity when the same fluorophore dose was administered to mice utilizing the previous formulation strategy. The success of the PM formulation strategy will be important for moving toward clinical translation of these novel nerve-specific probes as it is nontoxic and biodegradable and has the potential to decrease the necessary dose for imaging while also improving the safety profile.

Ultra high content analyses of circulating and tumor associated hybrid cells reveal phenotypic heterogeneity.

Persistently high, worldwide mortality from cancer highlights the unresolved challenges of disease surveillance and detection that impact survival. Development of a non-invasive, blood-based biomarker would transform survival from cancer. We demonstrate the functionality of ultra-high content analyses of a newly identified population of tumor cells that are hybrids between neoplastic and immune cells in patient matched tumor and peripheral blood specimens. Using oligonucleotide conjugated antibodies (Ab-oligo) permitting cyclic immunofluorescence (cyCIF), we present analyses of phenotypes among tumor and peripheral blood hybrid cells. Interestingly, the majority of circulating hybrid cell (CHC) subpopulations were not identified in tumor-associated hybrids. These results highlight the efficacy of ultra-high content phenotypic analyses using Ab-oligo based cyCIF applied to both tumor and peripheral blood specimens. The combination of a multiplex phenotypic profiling platform that is gentle enough to analyze blood to detect and evaluate disseminated tumor cells represents a novel approach to exploring novel tumor biology and potential utility for developing the population as a blood-based biomarker in cancer.

In situ single-cell therapeutic response imaging facilitated by the TRIPODD fluorescence imaging platform.

Purpose: Small molecule drugs such as tyrosine kinase inhibitors (TKIs) targeting tumoral molecular dependencies have become standard of care for numerous cancer types. Notably, epidermal growth factor receptor (EGFR) TKIs (e.g., erlotinib, afatinib, osimertinib) are the current first-line treatment for non-small cell lung cancer (NSCLC) due to their improved therapeutic outcomes for EGFR mutated and overexpressing disease over traditional platinum-based chemotherapy. However, many NSCLC tumors develop resistance to EGFR TKI therapy causing disease progression. Currently, the relationship between in situ drug target availability (DTA), local protein expression and therapeutic response cannot be accurately assessed using existing analytical tools despite being crucial to understanding the mechanism of therapeutic efficacy. Procedure: We have previously reported development of our fluorescence imaging platform termed TRIPODD (Therapeutic Response Imaging through Proteomic and Optical Drug Distribution) that is capable of simultaneous quantification of single-cell DTA and protein expression with preserved spatial context within a tumor. TRIPODD combines two complementary fluorescence imaging techniques: intracellular paired agent imaging (iPAI) to measure DTA and cyclic immunofluorescence (cyCIF), which utilizes oligonucleotide conjugated antibodies (Ab-oligos) for spatial proteomic expression profiling on tissue samples. Herein, TRIPODD was modified and optimized to provide a downstream analysis of therapeutic response through single-cell DTA and proteomic response imaging. Results: We successfully performed sequential imaging of iPAI and cyCIF resulting in high dimensional imaging and biomarker assessment to quantify single-cell DTA and local protein expression on erlotinib treated NSCLC models. Pharmacodynamic and pharmacokinetic studies of the erlotinib iPAI probes revealed that administration of 2.5 mg/kg each of the targeted and untargeted probe 4 h prior to tumor collection enabled calculation of DTA values with high Pearson correlation to EGFR, the erlotinib molecular target, expression in the tumors. Analysis of single-cell biomarker expression revealed that a single erlotinib dose was insufficient to enact a measurable decrease in the EGFR signaling cascade protein expression, where only the DTA metric detected the presence of bound erlotinib. Conclusion: We demonstrated the capability of TRIPODD to evaluate therapeutic response imaging to erlotinib treatment as it relates to signaling inhibition, DTA, proliferation, and apoptosis with preserved spatial context.

Detection of neoplastic-immune hybrid cells with metastatic properties in uveal melanoma.

BACKGROUND: Uveal melanoma is the most common non-cutaneous melanoma and is an intraocular malignancy affecting nearly 7,000 individuals per year worldwide. Of these, approximately 50% will progress to metastatic disease for which there are currently no effective curative therapies. Despite advances in molecular profiling and metastatic stratification of uveal melanoma tumors, little is known regarding their underlying biology of metastasis. Our group has identified a disseminated neoplastic cell population characterized by co-expression of immune and melanoma proteins, circulating hybrid cells (hybrids), in patients with uveal melanoma. Compared to circulating tumor cells, which lack expression of immune proteins, hybrids are detected at an increased prevalence in peripheral blood and can be used as a non-invasive biomarker to predict metastatic progression. METHODS: To ascertain mechanisms underlying enhanced hybrid cell dissemination we identified hybrid cells within primary uveal melanoma tumors using single cell RNA sequencing (n = 8) and evaluated their gene expression and predicted ligand-receptor interactions in relation to other melanoma and immune cells within the primary tumor. We then verified expression of upregulated hybrid pathways within patient-matched tumor and peripheral blood hybrids (n = 4) using cyclic immunofluorescence and quantified their protein expression relative to other non-hybrid tumor and disseminated tumor cells. RESULTS: Among the top upregulated genes and pathways in hybrid cells were those involved in enhanced cell motility and cytoskeletal rearrangement, immune evasion, and altered cellular metabolism. In patient-matched tumor and peripheral blood, we verified gene expression by examining concordant protein expression for each pathway category: TMSB10 (cell motility), CD74 (immune evasion) and GPX1 (metabolism). Both TMSB10 and GPX1 were expressed on significantly higher numbers of disseminated hybrid cells compared to circulating tumor cells, and CD74 and GPX1 were expressed on more disseminated hybrids than tumor-resident hybrids. Lastly, we identified that hybrid cells express ligand-receptor signaling pathways implicated in promoting metastasis including GAS6-AXL, CXCL12-CXCR4, LGALS9-P4HB and IGF1-IGFR1. CONCLUSION: These findings highlight the importance of TMSB10, GPX1 and CD74 for successful hybrid cell dissemination and survival in circulation. Our results contribute to the understanding of uveal melanoma tumor progression and interactions between tumor cells and immune cells in the tumor microenvironment that may promote metastasis.

Onward to Better Surgery - the Critical Need for Improved Ex Vivo Testing and Training Methods.

Guided surgery has demonstrated significant improvements in patient outcomes in some disease processes. Interest in this field has led to substantial growth in the technologies under investigation. Most likely no single technology will prove to be "best," and combinations of macro- and microscale guidance-using radiological imaging navigation, probes (activatable, perfusion, and molecular-targeted; large- and small-molecule), autofluorescence, tissue intrinsic optical properties, bioimpedance, and other characteristics-will offer patients and surgeons the greatest opportunity for high-success/low-morbidity medical interventions. Problems are arising, however, from the lack of valid testing formats; surgical training simulators suffer the same problems. Small animal models do not accurately recreate human anatomy, especially in terms of tissue volume. Large animal models are expensive and have difficulty replicating many pathological states, particularly when molecular specificity for individual cancers is required. Furthermore, the sheer number of technologies and the potential for synergistic combination leads to exponential growth of testing requirements that is unrealistic for in vivo testing. Therefore, critical need exists to expand the ex vivo/in vitro testing platforms available to investigators and, once validated, a need to increase the acceptance of these methods for funding and regulatory endpoints. Herein is a review of the available ex vivo/in vitro testing formats for guided surgery, a review of their advantages/disadvantages, and consideration for how our field may safely and more swiftly move forward through stronger adoption of these testing and validation methods.

Nerve Visualization using Phenoxazine-Based Near-Infrared Fluorophores to Guide Prostatectomy.

Fluorescence-guided surgery (FGS) is poised to revolutionize surgical medicine through near-infrared (NIR) fluorophores for tissue- and disease-specific contrast. Clinical open and laparoscopic FGS vision systems operate nearly exclusively at NIR wavelengths. However, tissue-specific NIR contrast agents compatible with clinically available imaging systems are lacking, leaving nerve tissue identification during prostatectomy a persistent challenge. Here, it is shown that combining drug-like molecular design concepts and fluorophore chemistry enabled the production of a library of NIR phenoxazine-based fluorophores for intraoperative nerve-specific imaging. The lead candidate readily delineated prostatic nerves in the canine and iliac plexus in the swine using the clinical da Vinci Surgical System that has been popularized for minimally invasive prostatectomy procedures. These results demonstrate the feasibility of molecular engineering of NIR nerve-binding fluorophores for ready integration into the existing surgical workflow, paving the path for clinical translation to reduce morbidity from nerve injury for prostate cancer patients.

Topical dual-stain difference imaging for rapid intra-operative tumor identification in fresh specimens.

Assessing tumor margin status during surgery is critical to ensure complete resection of cancer tissue; however, current approaches are ineffective and often result in repeat surgery. We present an optical imaging approach for margin assessment using topical application of two fluorescent stains, one targeted to a tumor biomarker and the other a nontargeted reference, to freshly excised specimens. Computing a normalized difference image from fluorescence images of the targeted and untargeted stains suppresses the confounding effects of nonspecific uptake. Applying this approach in excised breast tumor models produced promising tumor-to-normal tissue contrasts that were significantly higher than single-targeted-stain imaging.

Improving precision surgery: A review of current intraoperative nerve tissue fluorescence imaging.

Iatrogenic nerve injury represents one of the most feared surgical complications and remains a major morbidity across many surgical specialties. Currently, no clinically approved technique can directly enhance intraoperative nerve visualization, where intraoperative nerve identification continues to challenge even experienced surgeons. Fluorescence-guided surgery (FGS) has been successfully integrated into clinical medicine to improve safety and efficacy in the surgical arena. A number of tissue- and disease-specific contrast agents are in the clinical translation pipeline for future FGS integration. Within this context, a diverse repertoire of fluorescent tracers have been developed to improve surgeons' intraoperative vision. This review aims to convey the recent developments for nerve-specific FGS and its potential for clinical translation.

Use of Freshly Amputated Human Limbs for Pre-Clinical Evaluation of Molecular-Targeted Fluorescent Probes.

We have co-developed a first-in-kind model of fluorophore testing in freshly amputated human limbs. Ex vivo human tissue provides a unique opportunity for the testing of pre-clinical fluorescent agents, collection of imaging data, and histopathologic examination in human tissue prior to performing in vivo experiments. Existing pre-clinical fluorescent agent studies rely primarily on animal models, which do not directly predict fluorophore performance in humans and can result in wasted resources and time if an agent proves ineffective in early human trials. Because fluorophores have no desired therapeutic effect, their clinical utility is based solely on their safety and ability to highlight tissues of interest. Advancing to human trials even via the FDA's phase 0/microdose pathway still requires substantial resources, single-species pharmacokinetic testing, and toxicity testing. In a recently concluded study using amputated human lower limbs, we were able to test successfully a nerve-specific fluorophore in pre-clinical development. This study used systemic administration via vascular cannulization and a cardiac perfusion pump. We envision that this model may assist with early lead agent testing selection for fluorophores with various targets and mechanisms.

Robust biomarker discovery through multiplatform multiplex image analysis of breast cancer clinical cohorts.

Spatial profiling of tissues promises to elucidate tumor-microenvironment interactions and enable development of spatial biomarkers to predict patient response to immunotherapy and other therapeutics. However, spatial biomarker discovery is often carried out on a single patient cohort or imaging technology, limiting statistical power and increasing the likelihood of technical artifacts. In order to analyze multiple patient cohorts profiled on different platforms, we developed methods for comparative data analysis from three disparate multiplex imaging technologies: 1) cyclic immunofluorescence data we generated from 102 breast cancer patients with clinical follow-up, in addition to publicly available 2) imaging mass cytometry and 3) multiplex ion-beam imaging data. We demonstrate similar single-cell phenotyping results across breast cancer patient cohorts imaged with these three technologies and identify cellular abundance and proximity-based biomarkers with prognostic value across platforms. In multiple platforms, we identified lymphocyte infiltration as independently associated with longer survival in triple negative and high-proliferation breast tumors. Then, a comparison of nine spatial analysis methods revealed robust spatial biomarkers. In estrogen receptor-positive disease, quiescent stromal cells close to tumor were more abundant in good prognosis tumors while tumor neighborhoods of mixed fibroblast phenotypes were enriched in poor prognosis tumors. In triple-negative breast cancer (TNBC), macrophage proximity to tumor and B cell proximity to T cells were greater in good prognosis tumors, while tumor neighborhoods of vimentin-positive fibroblasts were enriched in poor prognosis tumors. We also tested previously published spatial biomarkers in our ensemble cohort, reproducing the positive prognostic value of isolated lymphocytes and lymphocyte occupancy and failing to reproduce the prognostic value of tumor-immune mixing score in TNBC. In conclusion, we demonstrate assembly of larger clinical cohorts from diverse platforms to aid in prognostic spatial biomarker identification and validation.

Flexible Cyclic Immunofluorescence (cyCIF) Using Oligonucleotide Barcoded Antibodies.

Advances in our understanding of the complex, multifaceted interactions between tumor epithelia, immune infiltrate, and tumor microenvironmental cells have been driven by highly multiplexed imaging technologies. These techniques are capable of labeling many more biomarkers than conventional immunostaining methods. However, multiplexed imaging techniques suffer from low detection sensitivity, cell loss-particularly in fragile samples-, and challenges with antibody labeling. Herein, we developed and optimized an oligonucleotide antibody barcoding strategy for cyclic immunofluorescence (cyCIF) that can be amplified to increase the detection efficiency of low-abundance antigens. Stained fluorescence signals can be readily removed using ultraviolet light treatment, preserving tissue and fragile cell sample integrity. We also extended the oligonucleotide barcoding strategy to secondary antibodies to enable the inclusion of difficult-to-label primary antibodies in a cyCIF panel. Using both the amplification oligonucleotides to label DNA barcoded antibodies and in situ hybridization of multiple fluorescently labeled oligonucleotides resulted in signal amplification and increased signal-to-background ratios. This procedure was optimized through the examination of staining parameters including staining oligonucleotide concentration, staining temperature, and oligonucleotide sequence design, resulting in a robust amplification technique. As a proof-of-concept, we demonstrate the flexibility of our cyCIF strategy by simultaneously imaging with the original oligonucleotide conjugated antibody (Ab-oligo) cyCIF strategy, the novel Ab-oligo cyCIF amplification strategy, as well as direct and indirect immunofluorescence to generate highly multiplexed images.

Dual probe difference specimen imaging for prostate cancer margin assessment.

Significance: Positive margin status due to incomplete removal of tumor tissue during radical prostatectomy for high-risk localized prostate cancer requires reoperation or adjuvant therapy, which increases morbidity and mortality. Adverse effects of prostate cancer treatments commonly include erectile dysfunction, urinary incontinence, and bowel dysfunction, making successful initial curative prostatectomy imperative. Aim: Current intraoperative tumor margin assessment is largely limited to frozen section analysis, which is a lengthy, labor-intensive process that is obtrusive to the clinical workflow within the operating room (OR). Therefore, a rapid method for prostate cancer margin assessment in the OR could improve outcomes for patients. Approach: Dual probe difference specimen imaging (DDSI), which uses paired antibody-based probes that are labeled with spectrally distinct fluorophores, was shown herein for prostate cancer margin assessment. The paired antibody-based probes consisted of a targeted probe to prostate-specific membrane antigen (PSMA) and an untargeted probe, which were used as a cocktail to stain resected murine tissue specimens including prostate tumor, adipose, muscle, and normal prostate. Ratiometric images (i.e., DDSI) of the difference between targeted and untargeted probe uptake were calculated and evaluated for accuracy using receiver operator characteristic curve analysis with area under the curve values used to evaluate the utility of the DDSI method to detect PSMA positive prostate cancer. Results: Targeted and untargeted probe uptake was similar between the high and low PSMA expressing tumor due to nonspecific probe uptake after topical administration. The ratiometric DDSI approach showed substantial contrast difference between the PSMA positive tumors and their respective normal tissues (prostate, adipose, muscle). Furthermore, DDSI showed substantial contrast difference between the high PSMA expressing tumors and the minimally PSMA expressing tumors due to the ratiometric correction for the nonspecific uptake patterns in resected tissues. Conclusions: Previous work has shown that ratiometic imaging has strong predictive value for breast cancer margin status using topical administration. Translation of the ratiometric DDSI methodology herein from breast to prostate cancers demonstrates it as a robust, ratiometric technique that provides a molecularly specific imaging modality for intraoperative margin detection. Using the validated DDSI protocol on resected prostate cancers permitted rapid and accurate assessment of PSMA status as a surrogate for prostate cancer margin status. Future studies will further evaluate the utility of this technology to quantitatively characterize prostate margin status using PSMA as a biomarker.

Quantitative image analysis pipeline for detecting circulating hybrid cells in immunofluorescence images with human-level accuracy.

Circulating hybrid cells (CHCs) are a newly discovered, tumor-derived cell population identified in the peripheral blood of cancer patients and are thought to contribute to tumor metastasis. However, identifying CHCs by immunofluorescence (IF) imaging of patient peripheral blood mononuclear cells (PBMCs) is a time-consuming and subjective process that currently relies on manual annotation by laboratory technicians. Additionally, while IF is relatively easy to apply to tissue sections, its application on PBMC smears presents challenges due to the presence of biological and technical artifacts. To address these challenges, we present a robust image analysis pipeline to automate the detection and analyses of CHCs in IF images. The pipeline incorporates quality control to optimize specimen preparation protocols and remove unwanted artifacts, leverages a ß-variational autoencoder (VAE) to learn meaningful latent representations of single-cell images and employs a support vector machine (SVM) classifier to achieve human-level CHC detection. We created a rigorously labeled IF CHC dataset including 9 patients and 2 disease sites with the assistance of 10 annotators to evaluate the pipeline. We examined annotator variation and bias in CHC detection and then provided guidelines to optimize the accuracy of CHC annotation. We found that all annotators agreed on CHC identification for only 65% of the cells in the dataset and had a tendency to underestimate CHC counts for regions of interest (ROI) containing relatively large amounts of cells (>50,000) when using conventional enumeration methods. On the other hand, our proposed approach is unbiased to ROI size. The SVM classifier trained on the ß-VAE encodings achieved an F1 score of 0.80, matching the average performance of annotators. Our pipeline enables researchers to explore the role of CHCs in cancer progression and assess their potential as a clinical biomarker for metastasis. Further, we demonstrate that the pipeline can identify discrete cellular phenotypes among PBMCs, highlighting its utility beyond CHCs.

Evaluating Receptor-Specific Fresh Specimen Staining for Tumor Margin Detection in Clinical Breast Specimens.

PURPOSE: Reliable and rapid identification of tumor in the margins of breast specimens during breast-conserving surgery to reduce repeat surgery rates is an active area of investigation. Dual-stain difference imaging (DDSI) is one of many approaches under evaluation for this application. This technique aims to topically apply fluorescent stain pairs (one targeted to a receptor-of-interest and the other a spectrally distinct isotype), image both stains, and compute a normalized difference image between the two channels. Prior evaluation and optimization in a variety of preclinical models produced encouraging diagnostic performance. Herein, we report on a pilot clinical study which evaluated HER2-targeted DDSI on 11 human breast specimens. PROCEDURES: Gross sections from 11 freshly excised mastectomy specimens were processed using a HER2-receptor-targeted DDSI protocol shortly after resection. After staining with the dual-probe protocol, specimens were imaged on a fluorescence scanner, followed by tissue fixation for hematoxylin and eosin and anti-HER2 immunohistochemical staining. Receiver operator characteristic curves and area under the curve (AUC) analysis were used to assess diagnostic performance of the resulting images. Performance values were also compared to expression level determined from IHC staining. RESULTS: Eight of the 11 specimens presented with distinguishable invasive ductal carcinoma and/or were not affected by an imaging artifact. In these specimens, the DDSI technique provided an AUC = 0.90 ± 0.07 for tumor-to-adipose tissue and 0.81 ± 0.15 for tumor-to-glandular tissue, which was significantly higher than AUC values recovered from images of the targeted probe alone. DDSI values and diagnostic performance did not correlate with HER2 expression level, and tumors with low HER2 expression often produced high AUC, suggesting that even the low expression levels were enough to help distinguish tumor. CONCLUSIONS: The results from this preliminary study of rapid receptor-specific staining in human specimens were consistent with prior preclinical results and demonstrated promising diagnostic potential.