Molecular profiling of tumors can help determine patient eligibility for immunotherapy treatments. Immunohistochemistry is widely used to detect biomarker proteins, but it can face challenges for multiplexed screening. As an alternative immunostaining strategy, our team has been exploring the use of antibody-bioconjugated nanoparticles. With side-illumination darkfield microscopy, the colored light scattered by individual nanoparticles attached to cancer cells can be observed with high resolution and quantified as a measure of protein expression. Here, we explored using Protein A/G to control antibody orientation on gold nanoparticles that are employed as immunoplasmonic nanoprobes. We used these nanoprobes to screen for biomarker proteins that are relevant to lung and breast cancers, namely, CD44, PD-L1, and HER2. No further optimization was required when we switched to other types of antibodies that display high binding affinity toward Protein A/G. We observed a concordance of results for immunoplasmonic and immunofluorescence measurements, suggesting that our proposed strategy has high potential for the screening of biomarkers expressed on cellular membranes.

Colorectal cancer is one of the most prevalent malignancies worldwide, with liver metastases (CLM) critically impacting patient outcomes. Current prediction models for CLM typically rely on whole slide images (WSIs) and clinical data, rather than fully integrating heterogeneous data sources such as pathology reports, which limits their ability to predict patient outcomes accurately. We introduce a novel multimodal framework for the survival and recurrence prediction of CLM that leverages latent diffusion models (LDMs) and large language models (LLMs), combining visual and textual information. Our approach integrates dual-resolution WSIs with detailed pathology reports to predict critical clinical endpoints, namely Time-to-Recurrence (TTR) and overall survival (OS). In the visual branch, a Mean Teacher model processes WSIs to generate precise classification maps that identify key Regions of Interest (ROIs), which are further refined using an advanced LDM to extract robust, highlevel features. Concurrently, pathology reports are analyzed with multiple LLMs to obtain dense semantic embeddings that capture essential clinical parameters, including tumor grade, histopathological patterns, and biomarker profiles. These complementary features are fused using a learned weighting strategy and incorporated into a Cox proportional hazards model for survival prediction. Evaluated on a dataset of 2,118 histological slides from 326 patients, our model achieves a concordance index (C -index) of $0.72 \pm 0.02$ for TTR and $0.78 \pm 0.02$ for OS. Ablation experiments demonstrate the importance of LLM-derived textual features to improve the prognostic performance, improving by 5 % and 6 % in TTR and OS prediction accuracy, respectively, to state-of-the-art methods. These results underscore the potential of our approach to enhance risk stratification and guide personalized treatment planning. Our findings motivate further validation in a large prospective study to confirm the translational impact of this multimodal framework.

Supplementary Table from The Movember Global Action Plan 1 (GAP1): Unique Prostate Cancer Tissue Microarray Resource

BACKGROUND AND OBJECTIVE:Prostate cancer (PCa) is hormone dependent, with UDP-glucuronosyltransferase 2B17 (UGT2B17) playing a central role in androgen inactivation. This study aimed to evaluate whether UGT2B17 expression in prostatectomy specimens can serve as a prognostic marker for lethal PCa. METHODS:A prespecified hypothesis posited that UGT2B17 expression (>25%) in primary tumors is associated with an aggressive disease phenotype, leading to metastasis, castration resistance (castration-resistant PCa [CRPC]), and mortality in men initially diagnosed with localized disease. Two high-density prostate tumor tissue microarray datasets were analyzed: the first from the Canadian Prostate Cancer Biomarker Network biobank (n = 1454) and the second from the PROCURE cohort (n = 1562). Kaplan-Meier and Cox proportional hazard ratio analyses were used to evaluate metastasis-free survival, CRPC, and PCa-specific mortality. Steroid levels were measured in plasma samples by mass spectrometry, and a linear regression model was used to evaluate variations in hormone levels based on tumoral UGT2B17 expression. KEY FINDINGS AND LIMITATIONS:UGT2B17 was associated with prognostic factors and linked to elevated levels of androsterone glucuronide (60%), the major circulating androgen-inactive metabolite, which is inactivated by UGT2B17. Kaplan-Meier and multivariable Cox analyses revealed that higher tumoral UGT2B17 is associated with an increased risk of progression to metastatic/CRPC stages and with PCa-specific mortality. CONCLUSIONS AND CLINICAL IMPLICATIONS:UGT2B17 expression influences hormone levels and identifies a subset of patients at an increased risk of progression to an incurable disease stage. Findings support the notion that enhanced UGT2B17, through increased androgen inactivation, creates a low-androgen tumor environment that drives tumor progression to a more aggressive phenotype.

We present a new method for lung pathology detection in blood plasma, including lung cancer staging. Raman spectroscopy uses inelastically scattered laser light to obtain molecular information in a reagent-free manner. Obtaining Raman spectral data from liquid samples has long proven challenging, but we have developed a novel tool for obtaining spectra from 60 μl liquid samples within two minutes: Raman of Well-based Samples (ROWS). With a low-cost ROWS device, we analyzed 372 blood plasma samples from a national biobank, including controls (n=92), patients with stage I-II lung cancer (n=99), stage III-IV cancer (n=46), benign tumours (n=36) and other lung conditions (n=99). Machine learning models were built to assess lung cancer stage and lung pathology presence. ROWS achieves up to 94% sensitivity, 90% specificity and 93% accuracy depending on classification. ROWS proves a robust method for rapid, low-cost, user-friendly, point-of-care lung pathology analysis in small quantities of blood plasma. ### Competing Interest Statement Frederic Leblond, Francois Daoust and Nassim Ksantini are shareholders of Reveal Life Science. Francois Daoust, Juliette Selb and Nassim Ksantini are employees of Reveal Life Science. ### Funding Statement This research was funded by a National Science and Engineering Research Council (NSERC) Alliance Grant in collaboration with Reveal Life Science (previously Exclaro-Tridan) (FL). ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The Research Ethics Board of the Centre Hospitalier de l'universite de Montreal (CHUM) gave ethical approval for this work. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data cannot be shared due to patient privacy and agreements with the AIRS Network Tissue Biobank. Data processing code is available through Github. Machine learning models are described in detail in Ember et al. Scientific Reports 2024. Code repository for model training, analysis and validation is publicly available in the paper "Open-sourced Raman spectroscopy data processing package implementing a novel baseline removal algorithm validated from multiple datasets acquired in human tissue and biofluids" Sheehy et al., Journal of Biomedical Optics, (2023) and also on Github (https://github.com/mr-sheg/orpl).

Prostate cancer (PC) rarely expresses aberrant immunohistochemical markers such as CK7, CK20, CDX2, GATA3 and TTF1. This study evaluates whether expression of CK7, CK20, CDX2, GATA3 and TTF1 is increased in hormone-resistant (HR) PC compared to hormone-sensitive (HS) disease. 64 patients undergoing transurethral resection of the prostate (TURP) for PC were included: 34 with HS disease, 22 with HR disease, and 8 whose status changed from HS to HR on a subsequent TURP (HS-HR). Overall, CK20 was the most frequently expressed aberrant marker (33.3 % of HS, 60.0 % of HR), followed by CK7 (16.7 % of HS, 13.3 % of HR) and CDX2 (11.9 % of HS, 16.7 % of HR). Compared to HS cases, HR tumors significantly overexpressed CK20 (p = 0.02). Positivity for aberrant markers was usually sparse and heterogeneous within ≤20 % of tumor cells. HR PC was significantly more likely to express aberrant markers among >20 % neoplastic cells than HS tumors (2.4 % and 20.0 % respectively, p = 0.008). The expression of an aberrant marker at >20 % positivity was also associated with loss of expression of ≥1 marker of prostatic origin (PSA, PSMA, P501S or NKX3), p = 0.01. For 21 patients with ≥2 TURPs separated in time, ≥1 aberrant marker was gained over time in 1/7 HS, 2/6 HR and 3/8 HS-HR patients. These results suggest that HR PC has increased likelihood of CK20 positivity and aberrant marker positivity in >20 % of tumor cells compared to HS cases, and that the aberrant immunohistochemical expression in a given PC patient may increase over time.

For BRCA mutation ( BRCA+ ) carriers, the risk of ovarian cancer can be as high as 59% compared with 1.4% for the general population. While the impact of BRCA mutations on epithelial cell transformation has been extensively studied, we hypothesize that these mutations cause structural changes that prematurely transform the ovary into a rich metastatic niche that supports the early onset of ovarian cancer. Analysis of collagen content and organization in human ovaries revealed increased coherence associated with fibrosis in premenopausal BRCA + ovaries relative to those without a BRCA mutation. Brca1 deficiency in murine ovarian fibroblasts triggered the expression of hallmarks of senescence, including Cdkn2a (p16) and acidic β-galactosidase activity. Brca1 -deficient fibroblasts also acquired an antigen-presenting myofibroblastic phenotype, characterized by expression of MHC-II molecules, α-SMA and extracellular matrix components, suggesting the capacity to modulate immune activity and drive structural changes resembling fibrosis. These results provide insight into the mechanisms contributing to accelerated ovarian aging in BRCA+ carriers. Teaser: BRCA mutation promotes fibroblast hyperactivity and senescence that changes the stromal architecture of the ovarian niche. ### Competing Interest Statement The authors have declared no competing interest. Canadian Institutes of Health Research, E-419853

Intraductal carcinoma of the prostate (IDC-P) is a very aggressive histopathological subtype of prostate cancer (PCa) that is strongly associated with poor clinical outcomes but for which no accurate biomarkers exist. Here, we demonstrate a novel application of texture analysis-based machine learning alongside multimodal nonlinear optical imaging using second-harmonic generation (SHG) and stimulated Raman scattering (SRS) at 1450 cm-1 and 1668 cm-1 Raman shifts to distinguish IDC-P from regular PCa and benign prostate. Images from each tissue type were analyzed to extract the first-order statistics and texture-based second-order statistics derived from the gray-level co-occurrence matrix of the images. A machine learning model was constructed using support vector machine (SVM) to classify the prostate tissue based on these statistics. Our results demonstrate that SVM models trained on either SHG or SRS images accurately classify IDC-P as well as high-grade PCa, low-grade PCa, and benign tissue with a mean classification accuracy exceeding 89%. Moreover, a mean classification accuracy of 98% was achieved using an SVM model trained on combined SHG and SRS images. Our study demonstrates that multimodal nonlinear optical imaging using SHG and SRS can be combined with texture analysis-based SVM classification to provide pathologists with a reliable biomarker of IDC-P.

Significance:Prostate cancer (PCa) confirmation during needle-based procedures is limited by the lack of intraoperative diagnostic tools. Raman spectroscopy (RS), combined with classification models, offers a promising solution for real-time tissue characterization, potentially improving sampling accuracy and therapy guidance. However, such models require tissue- and organ-specific data, making deployment in studies challenging due to limited data availability. Aim:The aim is to develop a one-dimensional convolutional neural network (1D-CNN) for real-time PCa detection using RS on prospectively collected ex vivo data, leveraging multi-organ pre-training and evaluating two domain adaptation strategies. Approach:A ResNet-based 1D-CNN was trained for binary cancer/normal tissue classification. We implemented a pre-training strategy using retrospective RS data from brain, breast, and prostate (202 patients), along with pre-trained bacterial models, followed by efficient fine-tuning and test-time adaptation (TTA) to adapt to unseen domains. Results:Prospective RS data were acquired using a robotic system from 10 PCa patients (two to five biopsies each). The fine-tuned model achieved 0.76 area under the receiver operating characteristic curve, 0.79 accuracy, 0.83 sensitivity, and 0.72 specificity, outperforming support vector machines. TTA improved predictions when labels were unavailable. Conclusions:Pre-trained 1D-CNNs combined with efficient fine-tuning or TTA enable accurate PCa detection in small-cohort settings using real-time RS.

Malignancy grading of prostate cancer (PCa) is fundamental for risk stratification, patient counseling, and treatment decision-making. Deep learning has shown potential to improve the expert consensus for tumor grading, which relies on the Gleason score/grade grouping. However, the core problem of interobserver variability for the Gleason grading system remains unresolved. We developed a novel grading system for PCa and utilized artificial intelligence (AI) and multi-institutional international datasets from 2647 PCa patients treated with radical prostatectomy with a long follow-up of ≥10 years for biochemical recurrence and cancer-specific death. Through survival analyses, we evaluated the novel grading system and showed that AI could develop a tumor grading system with four risk groups independent from and superior to the current five grade groups. Moreover, AI could develop a scoring system that reflects the risk of castration resistant PCa in men who have experienced biochemical recurrence. Thus, AI has the potential to develop an effective grading system for PCa interpretable by human experts.

Introduction High prostate eicosapentaenoic fatty acid (EPA) levels have been associated with a significant reduction of prostate cancer upgrading to grade group (GG) ≥2 in men with GG1 prostate cancer on active surveillance. The current phase IIb randomized pre-prostatectomy placebo-controlled trial assessed the effect of a monoacylglyceride-EPA (MAG-EPA) supplement on prostate cancer aggressiveness in 130 men diagnosed with prostate cancer. Methods Men diagnosed with GG ≥2 prostate cancer and undergoing radical prostatectomy between 2015-2017 were randomized to either 3g/day of MAG-EPA (n=65) or placebo (n=65) for seven weeks prior to radical prostatectomy and for up to one year after surgery (NCT02333435). The primary outcome was the cancer proliferation index quantified by automated image analysis of tumor nuclear Ki-67 expression using standardized prostatectomy tissue microarrays. One exploratory clinical outcome was grade reclassification from baseline biopsy at prostatectomy. Stool samples were collected in a sub-group of consent patients (n=42) for gut microbiome and fecal short-chain fatty acid analyses, using 16srRNA sequencing and targeted metabolomics, respectively. Results Men randomized to MAG-EPA had four-fold higher EPA levels in prostate tissues compared to those on placebo. The primary outcome was the cancer proliferation index measured by Ki-67 expression which was not statistically different between intervention (3.10%) and placebo (2.85%) groups. In the per protocol analyses, the adjusted estimated effect of MAG-EPA was greater but remained non-significant. However, there was a significant increase in size and proliferative index of tumor lymphoid aggregates in MAG-EPA treated prostate cancer, suggesting an immune mediated effect. In exploratory analyses, the MAG-EPA group had more cancer pathological downgrade and less cancer upgrade at prostatectomy, compared to the placebo group (p=0.024). Gut microbiota analysis revealed that the cancer up-grading reduction in pre-prostatectomy prostate cancer patients taking MAG-EPA was associated with a reduction of gut Ruminococaceae and fecal butyrate levels. Conclusions Our results suggest that lowering gut butyrate, a known immune modulator, may partly explain the beneficial effect of MAG-EPA on prostate cancer aggressiveness. More studies are needed to better understand the biological and clinical outcomes following this concentrated EPA supplementation and determine if and how it can benefit prostate cancer patients.

ObjectivesTo evaluate the International Society of Urological Pathology (ISUP) 5-tier grade grouping (GG) system of prostate cancers as well as previously proposed optimizations.Patients and methodsThe PROCURE biobank is a prospective cohort study of patients with localized prostate cancer who underwent radical prostatectomy in Quebec province between 2005 and 2013. Surgical specimens were graded by experienced genitourinary pathologists using 2019 ISUP criteria. Follow-up was conducted until November 2021. The current 5-tier and a proposed 6-tier GG system were evaluated, the latter having two changes: 1) Gleason 3+4 and 4+3 tumors with minor/tertiary Gleason 5 patterns were upgraded to GG 3 and 4, respectively; and 2) patients in GG5 were separated based on primary Gleason pattern (4 or 5). Cox proportional hazards models and Harrell’s concordance (C) indices were used for statistical analyses.Results2003 patients were included (median follow-up: 8.7 years). The current 5-tier GG system predicted time to recurrence (hazard ratio [HR] 2.12, 95% confidence interval [95%CI] 1.99-2.25, C 0.717), androgen-deprivation therapy (HR 2.58, 95%CI 2.38-2.80, C 0.790), metastasis (HR 2.48, 95%CI 2.17-2.83, C 0.806), castration-resistant prostate cancer (HR 2.67, 95%CI 2.28-3.13, C 0.829), and cancer-specific mortality (HR 2.80, 95%CI 2.27-3.44, C 0.835). Goodness-of-fit further improved with the proposed 6-tier GG system, with Harrell’s C of 0.733, 0.807, 0.827, 0.853, and 0.853, respectively.ConclusionsThe 5-tier GG system predicted short- and long-term outcomes for patients with localized prostate cancer, and the proposed 6-tier GG system further improved its accuracy.

We present a rapid, portable optical system for label-free detection of COVID-19. Raman spectra from an entire liquid drop of saliva supernatant can be obtained within 6 minutes, and the sample is classified as COVID-19 positive or negative using artificial intelligence (AI). 293 COVID negative and 49 COVID positive saliva supernatant samples were analyzed. Positive samples were from hospitalized patients (non-critical and critical) and non-hospitalized testing clinic volunteers (symptomatic and asymptomatic). Our Raman/AI system has an 82% accuracy detecting people with COVID-19 of any severity with any symptom presentation, and 89% accuracy when detecting COVID-19 in hospitalized patients alone. Rapid label-free analysis of biofluids for viruses could provide a low-cost screening solution that could be adapted to respond to viral mutations. This could be invaluable for future pandemics and for reducing infections in hospitals, care homes and workplaces.

With greater population density, the likelihood of viral outbreaks achieving pandemic status is increasing. However, current viral screening techniques use specific reagents, and as viruses mutate, test accuracy decreases. Here, we present the first real-time, reagent-free, portable analysis platform for viral detection in liquid saliva, using COVID-19 as a proof-of-concept. We show that vibrational molecular spectroscopy and machine learning (ML) detect biomolecular changes consistent with the presence of viral infection. Saliva samples were collected from 470 individuals, including 65 that were infected with COVID-19 (28 from hospitalized patients and 37 from a walk-in testing clinic) and 251 that had a negative polymerase chain reaction (PCR) test. A further 154 were collected from healthy volunteers. Saliva measurements were achieved in 6 minutes or less and led to machine learning models predicting COVID-19 infection with sensitivity and specificity reaching 90%, depending on volunteer symptoms and disease severity. Machine learning models were based on linear support vector machines (SVM). This platform could be deployed to manage future pandemics using the same hardware but using a tunable machine learning model that could be rapidly updated as new viral strains emerge. Raman spectroscopy and machine learning is used in combination to detect COVID-19 positive saliva in liquid form.

Colorectal liver metastases (CLM) develop in almost half of patients with colon cancer. Response to systemic chemotherapy is the main determinant of patient survival. Due to the importance of assessing treatment response of CLM to chemotherapy for patient prognosis in the early stages, there is a need to classify tumor response grade (TRG) based on whole slide images (WSI). However, annotating WSI for training neural networks is a time-consuming task and inter-observer variability among pathologists presents a significant challenge in the accurate classification of TRG. In this work, we introduce an end-to-end pipeline for the prediction of TRG in CLM. The pipeline begins with color normalization of WSI via a Generative Adversarial Network (GAN), followed by precise nuclei segmentation combining a GAN with a Vision Transformer (ViT). The segmented nuclei are then clustered using a Graph Neural Network (GNN) to map the intricate cellular structures and interactions. Finally, TRG prediction is achieved through a Support Vector Machine (SVM) that integrates GNN outputs and clinical data, demonstrating enhanced accuracy compared to existing models. The proposed method achieves encouraging results at all stages. The GAN-based normalization and ViT-enhanced segmentation of WSIs yielded a high average Dice score of 82.5%. The GNN clustering of nuclei demonstrated precision, as evidenced by low Davies-Bouldin Index (DBI) scores, particularly for immune cells (0.40), cancer cells (0.45), and fibroblasts (0.42). Most notably, in TRG prediction, our model is able to stratify patients into two TRG classes (1-2 vs 3-5) with an accuracy of 91.3%.

Colorectal liver metastases (CLM) affect almost half of all colon cancer patients and the response to systemic chemotherapy plays a crucial role in patient survival. While oncologists typically use tumor grading scores, such as tumor regression grade (TRG), to establish an accurate prognosis on patient outcomes, including overall survival (OS) and time-to-recurrence (TTR), these traditional methods have several limitations. They are subjective, time-consuming, and require extensive expertise, which limits their scalability and reliability. Additionally, existing approaches for prognosis prediction using machine learning mostly rely on radiological imaging data, but recently histological images have been shown to be relevant for survival predictions by allowing to fully capture the complex microenvironmental and cellular characteristics of the tumor. To address these limitations, we propose an end-to-end approach for automated prognosis prediction using histology slides stained with Hematoxylin and Eosin (H&E) and Hematoxylin Phloxine Saffron (HPS). We first employ a Generative Adversarial Network (GAN) for slide normalization to reduce staining variations and improve the overall quality of the images that are used as input to our prediction pipeline. We propose a semi-supervised model to perform tissue classification from sparse annotations, producing segmentation and feature maps. Specifically, we use an attention-based approach that weighs the importance of different slide regions in producing the final classification results. Finally, we exploit the extracted features for the metastatic nodules and surrounding tissue to train a prognosis model. In parallel, we train a vision Transformer model in a knowledge distillation framework to replicate and enhance the performance of the prognosis prediction. We evaluate our approach on an in-house clinical dataset of 258 CLM patients, achieving superior performance compared to other comparative models with a c-index of 0.804 (0.014) for OS and 0.735 (0.016) for TTR, as well as on two public datasets. The proposed approach achieves an accuracy of 86.9% to 90.3% in predicting TRG dichotomization. For the 3-class TRG classification task, the proposed approach yields an accuracy of 78.5% to 82.1%, outperforming the comparative methods. Our proposed pipeline can provide automated prognosis for pathologists and oncologists, and can greatly promote precision medicine progress in managing CLM patients.