Artificial Intelligence in Cancer, Biology and Oncology

Cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) are widely used in patients with hormone receptor-positive (HR+)/human epidermal growth factor receptor 2 negative (HER2−) advanced/metastatic breast cancer (ABC/MBC) in first line (1L), but little is known about their real-world use and clinical outcomes long-term, in Canada. This study used Pentavere’s previously validated artificial intelligence (AI) to extract real-world data on the treatment patterns and outcomes of patients receiving CDK4/6i+endocrine therapy (ET) for HR+/HER2− ABC/MBC at Sinai Health in Toronto, Canada. Between 1 January 2016 and 1 July 2021, 48 patients were diagnosed with HR+/HER2− ABC/MBC and received CDK4/6i + ET. A total of 38 out of 48 patients received CDK4/6i + ET in 1L, of which 34 of the 38 (89.5%) received palbociclib + ET. In 2L, 12 of the 21 (57.1%) patients received CDK4/6i + ET, of which 58.3% received abemaciclib. In 3L, most patients received chemotherapy (10/12, 83.3%). For the patients receiving CDK4/6i in 1L, the median (95% CI) time to the next treatment was 42.3 (41.2, NA) months. The median (95% CI) time to chemotherapy was 46.5 (41.4, NA) months. The two-year overall survival (95% CI) was 97.4% (92.4, 100.0), and the median (range) follow-up was 28.7 (3.4–67.6) months. Despite the limitations inherent in real-world studies and a limited number of patients, these AI-extracted data complement previous studies, demonstrating the effectiveness of CDK4/6i + ET in the Canadian real-world 1L, with most patients receiving palbociclib as CDK4/6i in 1L. Full article

Real-world evidence for patients with advanced EGFR-mutated non-small cell lung cancer (NSCLC) in Canada is limited. This study’s objective was to use previously validated DARWEN^TM artificial intelligence (AI) to extract data from electronic heath records of patients with non-squamous NSCLC at University Health Network (UHN) to describe EGFR mutation prevalence, treatment patterns, and outcomes. Of 2154 patients with NSCLC, 613 had advanced disease. Of these, 136 (22%) had common sensitizing EGFR mutations (cEGFRm; ex19del, L858R), 8 (1%) had exon 20 insertions (ex20ins), and 338 (55%) had EGFR wild type. One-year overall survival (OS) (95% CI) for patients with cEGFRm, ex20ins, and EGFR wild type tumours was 88% (83, 94), 100% (100, 100), and 59% (53, 65), respectively. In total, 38% patients with ex20ins received experimental ex20ins targeting treatment as their first-line therapy. A total of 57 patients (36%) with cEGFRm received osimertinib as their first-line treatment, and 61 (39%) received it as their second-line treatment. One-year OS (95% CI) following the discontinuation of osimertinib was 35% (17, 75) post-first-line and 20% (9, 44) post-second-line. In this real-world AI-generated dataset, survival post-osimertinib was poor in patients with cEGFR mutations. Patients with ex20ins in this cohort had improved outcomes, possibly due to ex20ins targeting treatment, highlighting the need for more effective treatments for patients with advanced EGFRm NSCLC. Full article

We developed machine and deep learning models to predict chemoradiotherapy in rectal cancer using ¹⁸F-FDG PET images and harmonized image features extracted from ¹⁸F-FDG PET/CT images. Patients diagnosed with pathologic T-stage III rectal cancer with a tumor size > 2 cm were treated with neoadjuvant chemoradiotherapy. Patients with rectal cancer were divided into an internal dataset (n = 116) and an external dataset obtained from a separate institution (n = 40), which were used in the model. AUC was calculated to select image features associated with radiochemotherapy response. In the external test, the machine-learning signature extracted from ¹⁸F-FDG PET image features achieved the highest accuracy and AUC value of 0.875 and 0.896. The harmonized first-order radiomics model had a higher efficiency with accuracy and an AUC of 0.771 than the second-order model in the external test. The deep learning model using the balanced dataset showed an accuracy of 0.867 in the internal test but an accuracy of 0.557 in the external test. Deep-learning models using ¹⁸F-FDG PET images must be harmonized to demonstrate reproducibility with external data. Harmonized ¹⁸F-FDG PET image features as an element of machine learning could help predict chemoradiotherapy responses in external tests with reproducibility. Full article

Photon absorption remote sensing (PARS) is a new laser-based microscope technique that permits cellular-level resolution of unstained fresh, frozen, and fixed tissues. Our objective was to determine whether PARS could provide an image quality sufficient for the diagnostic assessment of breast cancer needle core biopsies (NCB). We PARS imaged and virtually H&E stained seven independent unstained formalin-fixed paraffin-embedded breast NCB sections. These identical tissue sections were subsequently stained with standard H&E and digitally scanned. Both the 40× PARS and H&E whole-slide images were assessed by seven breast cancer pathologists, masked to the origin of the images. A concordance analysis was performed to quantify the diagnostic performances of standard H&E and PARS virtual H&E. The PARS images were deemed to be of diagnostic quality, and pathologists were unable to distinguish the image origin, above that expected by chance. The diagnostic concordance on cancer vs. benign was high between PARS and conventional H&E (98% agreement) and there was complete agreement for within-PARS images. Similarly, agreement was substantial (kappa > 0.6) for specific cancer subtypes. PARS virtual H&E inter-rater reliability was broadly consistent with the published literature on diagnostic performance of conventional histology NCBs across all tested histologic features. PARS was able to image unstained tissues slides that were diagnostically equivalent to conventional H&E. Due to its ability to non-destructively image fixed and fresh tissues, and the suitability of the PARS output for artificial intelligence assistance in diagnosis, this technology has the potential to improve the speed and accuracy of breast cancer diagnosis. Full article

Deep learning applications are emerging as promising new tools that can support the diagnosis and classification of different cancer types. While such solutions hold great potential for hematological malignancies, there have been limited studies describing the use of such applications in this field. The rapid diagnosis of double/triple-hit lymphomas (DHLs/THLs) involving MYC, BCL2 and/or BCL6 rearrangements is obligatory for optimal patient care. Here, we present a novel deep learning tool for diagnosing DHLs/THLs directly from scanned images of biopsy slides. A total of 57 biopsies, including 32 in a training set (including five DH lymphoma cases) and 25 in a validation set (including 10 DH/TH cases), were included. The DHL-classifier demonstrated a sensitivity of 100%, a specificity of 87% and an AUC of 0.95, with only two false positive cases, compared to FISH. The DHL-classifier showed a 92% predictive value as a screening tool for performing conventional FISH analysis, over-performing currently used criteria. The work presented here provides the proof of concept for the potential use of an AI tool for the identification of DH/TH events. However, more extensive follow-up studies are required to assess the robustness of this tool and achieve high performances in a diverse population. Full article

Muscle-invasive bladder cancer (MIBC) is a highly heterogeneous and costly disease with significant morbidity and mortality. Understanding tumor histopathology leads to tailored therapies and improved outcomes. In this study, we employed a weakly supervised learning and neural architecture search to develop a data-driven scoring system. This system aimed to capture prognostic histopathological patterns observed in H&E-stained whole-slide images. We constructed and externally validated our scoring system using multi-institutional datasets with 653 whole-slide images. Additionally, we explored the association between our scoring system, seven histopathological features, and 126 molecular signatures. Through our analysis, we identified two distinct risk groups with varying prognoses, reflecting inherent differences in histopathological and molecular subtypes. The adjusted hazard ratio for overall mortality was 1.46 (95% CI 1.05–2.02; z: 2.23; p = 0.03), thus identifying two prognostic subgroups in high-grade MIBC. Furthermore, we observed an association between our novel digital biomarker and the squamous phenotype, subtypes of miRNA, mRNA, long non-coding RNA, DNA hypomethylation, and several gene mutations, including FGFR3 in MIBC. Our findings underscore the risk of confounding bias when reducing the complex biological and clinical behavior of tumors to a single mutation. Histopathological changes can only be fully captured through comprehensive multi-omics profiles. The introduction of our scoring system has the potential to enhance daily clinical decision making for MIBC. It facilitates shared decision making by offering comprehensive and precise risk stratification, treatment planning, and cost-effective preselection for expensive molecular characterization. Full article

Intensity modulated radiation therapy (IMRT) is one of the most used techniques for cancer treatment. Using a linear accelerator, it delivers radiation directly at the cancerogenic cells in the tumour, reducing the impact of the radiation on the organs surrounding the tumour. The complexity of the IMRT problem forces researchers to subdivide it into three sub-problems that are addressed sequentially. Using this sequential approach, we first need to find a beam angle configuration that will be the set of irradiation points (beam angles) over which the tumour radiation is delivered. This first problem is called the Beam Angle Optimisation (BAO) problem. Then, we must optimise the radiation intensity delivered from each angle to the tumour. This second problem is called the Fluence Map Optimisation (FMO) problem. Finally, we need to generate a set of apertures for each beam angle, making the intensities computed in the previous step deliverable. This third problem is called the Sequencing problem. Solving these three sub-problems sequentially allows clinicians to obtain a treatment plan that can be delivered from a physical point of view. However, the obtained treatment plans generally have too many apertures, resulting in long delivery times. One strategy to avoid this problem is the Direct Aperture Optimisation (DAO) problem. In the DAO problem, the idea is to merge the FMO and the Sequencing problem. Hence, optimising the radiation’s intensities considers the physical constraints of the delivery process. The DAO problem is usually modelled as a Mixed-Integer optimisation problem and aims to determine the aperture shapes and their corresponding radiation intensities, considering the physical constraints imposed by the Multi-Leaf Collimator device. In solving the DAO problem, generating clinically acceptable treatments without additional sequencing steps to deliver to the patients is possible. In this work, we propose to solve the DAO problem using the well-known Particle Swarm Optimisation (PSO) algorithm. Our approach integrates the use of mathematical programming to optimise the intensities and utilizes PSO to optimise the aperture shapes. Additionally, we introduce a reparation heuristic to enhance aperture shapes with minimal impact on the treatment plan. We apply our proposed algorithm to prostate cancer cases and compare our results with those obtained in the sequential approach. Results show that the PSO obtains competitive results compared to the sequential approach, receiving less radiation time (beam on time) and using the available apertures with major efficiency. Full article

Glioma grading plays a pivotal role in guiding treatment decisions, predicting patient outcomes, facilitating clinical trial participation and research, and tailoring treatment strategies. Current glioma grading in the clinic is based on tissue acquired at the time of resection, with tumor aggressiveness assessed from tumor morphology and molecular features. The increased emphasis on molecular characteristics as a guide for management and prognosis estimation underscores is driven by the need for accurate and standardized grading systems that integrate molecular and clinical information in the grading process and carry the expectation of the exposure of molecular markers that go beyond prognosis to increase understanding of tumor biology as a means of identifying druggable targets. In this study, we introduce a novel application (GradWise) that combines rank-based weighted hybrid filter (i.e., mRMR) and embedded (i.e., LASSO) feature selection methods to enhance the performance of feature selection and machine learning models for glioma grading using both clinical and molecular predictors. We utilized publicly available TCGA from the UCI ML Repository and CGGA datasets to identify the most effective scheme that allows for the selection of the minimum number of features with their names. Two popular feature selection methods with a rank-based weighting procedure were employed to conduct comprehensive experiments with the five supervised models. The computational results demonstrate that our proposed method achieves an accuracy rate of 87.007% with 13 features and an accuracy rate of 80.412% with five features on the TCGA and CGGA datasets, respectively. We also obtained four shared biomarkers for the glioma grading that emerged in both datasets and can be employed with transferable value to other datasets and data-based outcome analyses. These findings are a significant step toward highlighting the effectiveness of our approach by offering pioneering results with novel markers with prospects for understanding and targeting the biologic mechanisms of glioma progression to improve patient outcomes. Full article

Introduction: Colorectal cancer (CRC) is a common cancer associated with poor outcomes, underscoring a need for the identification of novel prognostic and therapeutic targets to improve outcomes. This study aimed to identify genetic variants and differentially expressed genes (DEGs) using genome-wide DNA and RNA sequencing followed by validation in a large cohort of patients with CRC. Methods: Whole genome and gene expression profiling were used to identify DEGs and genetic alterations in 146 patients with CRC. Gene Ontology, Reactom, GSEA, and Human Disease Ontology were employed to study the biological process and pathways involved in CRC. Survival analysis on dysregulated genes in patients with CRC was conducted using Cox regression and Kaplan–Meier analysis. The STRING database was used to construct a protein–protein interaction (PPI) network. Moreover, candidate genes were subjected to ML-based analysis and the Receiver operating characteristic (ROC) curve. Subsequently, the expression of the identified genes was evaluated by Real-time PCR (RT-PCR) in another cohort of 64 patients with CRC. Gene variants affecting the regulation of candidate gene expressions were further validated followed by Whole Exome Sequencing (WES) in 15 patients with CRC. Results: A total of 3576 DEGs in the early stages of CRC and 2985 DEGs in the advanced stages of CRC were identified. ASPHD1 and ZBTB12 genes were identified as potential prognostic markers. Moreover, the combination of ASPHD and ZBTB12 genes was sensitive, and the two were considered specific markers, with an area under the curve (AUC) of 0.934, 1.00, and 0.986, respectively. The expression levels of these two genes were higher in patients with CRC. Moreover, our data identified two novel genetic variants—the rs925939730 variant in ASPHD1 and the rs1428982750 variant in ZBTB1—as being potentially involved in the regulation of gene expression. Conclusions: Our findings provide a proof of concept for the prognostic values of two novel genes—ASPHD1 and ZBTB12—and their associated variants (rs925939730 and rs1428982750) in CRC, supporting further functional analyses to evaluate the value of emerging biomarkers in colorectal cancer. Full article

Machine learning (ML) models have become capable of making critical decisions on our behalf. Nevertheless, due to complexity of these models, interpreting their decisions can be challenging, and humans cannot always control them. This paper provides explanations of decisions made by ML models in diagnosing four types of posterior fossa tumors: medulloblastoma, ependymoma, pilocytic astrocytoma, and brainstem glioma. The proposed methodology involves data analysis using kernel density estimations with Gaussian distributions to examine individual MRI features, conducting an analysis on the relationships between these features, and performing a comprehensive analysis of ML model behavior. This approach offers a simple yet informative and reliable means of identifying and validating distinguishable MRI features for the diagnosis of pediatric brain tumors. By presenting a comprehensive analysis of the responses of the four pediatric tumor types to each other and to ML models in a single source, this study aims to bridge the knowledge gap in the existing literature concerning the relationship between ML and medical outcomes. The results highlight that employing a simplistic approach in the absence of very large datasets leads to significantly more pronounced and explainable outcomes, as expected. Additionally, the study also demonstrates that the pre-analysis results consistently align with the outputs of the ML models and the clinical findings reported in the existing literature. Full article

Background: We investigated the feasibility of a deep learning algorithm (DLA) based on apparent diffusion coefficient (ADC) maps for the segmentation and discrimination of clinically significant cancer (CSC, Gleason score ≥ 7) from non-CSC in patients with prostate cancer (PCa). Methods: Data from a total of 149 consecutive patients who had undergone 3T-MRI and been pathologically diagnosed with PCa were initially collected. The labelled data (148 images for GS6, 580 images for GS7) were applied for tumor segmentation using a convolutional neural network (CNN). For classification, 93 images for GS6 and 372 images for GS7 were used. For external validation, 22 consecutive patients from five different institutions (25 images for GS6, 70 images for GS7) representing different MR machines were recruited. Results: Regarding segmentation and classification, U-Net and DenseNet were used, respectively. The tumor Dice scores for internal and external validation were 0.822 and 0.7776, respectively. As for classification, the accuracies of internal and external validation were 73 and 75%, respectively. For external validation, diagnostic predictive values for CSC (sensitivity, specificity, positive predictive value and negative predictive value) were 84, 48, 82 and 52%, respectively. Conclusions: Tumor segmentation and discrimination of CSC from non-CSC is feasible using a DLA developed based on ADC maps (b2000) alone. Full article

Accurate retroperitoneal lymph node metastasis (LNM) prediction in early-stage testicular germ cell tumours (TGCTs) harbours the potential to significantly reduce over- or undertreatment and treatment-related morbidity in this group of young patients as an important survivorship imperative. We investigated the role of computed tomography (CT) radiomics models integrating clinical predictors for the individualised prediction of LNM in early-stage TGCT. Ninety-one patients with surgically proven testicular germ cell tumours and contrast-enhanced CT were included in this retrospective study. Dedicated radiomics software was used to segment 273 retroperitoneal lymph nodes and extract features. After feature selection, radiomics-based machine learning models were developed to predict LN metastasis. The robustness of the procedure was controlled by 10-fold cross-validation. Using multivariable logistic regression modelling, we developed three prediction models: a radiomics-only model, a clinical-only model, and a combined radiomics–clinical model. The models’ performances were evaluated using the area under the receiver operating characteristic curve (AUC). Finally, decision curve analysis was performed to estimate the clinical usefulness of the predictive model. The radiomics-only model for predicting lymph node metastasis reached a greater discrimination power than the clinical-only model, with an AUC of 0.87 (±0.04; 95% CI) vs. 0.75 (±0.08; 95% CI) in our study cohort. The combined model integrating clinical risk factors and selected radiomics features outperformed the clinical-only and the radiomics-only prediction models, and showed good discrimination with an area under the curve of 0.89 (±0.03; 95% CI). The decision curve analysis demonstrated the clinical usefulness of our proposed combined model. The presented combined CT-based radiomics–clinical model represents an exciting non-invasive tool for individualised LN metastasis prediction in testicular germ cell tumours. Multi-centre validation is required to generate high-quality evidence for its clinical application. Full article

Background: The aim of this study was to develop and validate radiogenomic models to predict the MDM2 gene amplification status and differentiate between ALTs and lipomas on preoperative MR images. Methods: MR images were obtained in 257 patients diagnosed with ALTs (n = 65) or lipomas (n = 192) using histology and the MDM2 gene analysis as a reference standard. The protocols included T2-, T1-, and fat-suppressed contrast-enhanced T1-weighted sequences. Additionally, 50 patients were obtained from a different hospital for external testing. Radiomic features were selected using mRMR. Using repeated nested cross-validation, the machine-learning models were trained on radiomic features and demographic information. For comparison, the external test set was evaluated by three radiology residents and one attending radiologist. Results: A LASSO classifier trained on radiomic features from all sequences performed best, with an AUC of 0.88, 70% sensitivity, 81% specificity, and 76% accuracy. In comparison, the radiology residents achieved 60–70% accuracy, 55–80% sensitivity, and 63–77% specificity, while the attending radiologist achieved 90% accuracy, 96% sensitivity, and 87% specificity. Conclusion: A radiogenomic model combining features from multiple MR sequences showed the best performance in predicting the MDM2 gene amplification status. The model showed a higher accuracy compared to the radiology residents, though lower compared to the attending radiologist. Full article

Purpose: Develop and validate an accessible prediction model using machine learning (ML) to predict the risk of chemotherapy-induced myelosuppression (CIM) in children with Wilms’ tumor (WT) before chemotherapy is administered, enabling early preventive management. Methods: A total of 1433 chemotherapy cycles in 437 children with WT who received chemotherapy in our hospital from January 2009 to March 2022 were retrospectively analyzed. Demographic data, clinicopathological characteristics, hematology and blood biochemistry baseline results, and medication information were collected. Six ML algorithms were used to construct prediction models, and the predictive efficacy of these models was evaluated to select the best model to predict the risk of grade ? 2 CIM in children with WT. A series of methods, such as the area under the receiver operating characteristic curve (AUROC), the calibration curve, and the decision curve analysis (DCA) were used to test the model’s accuracy, discrimination, and clinical practicability. Results: Grade ? 2 CIM occurred in 58.5% (839/1433) of chemotherapy cycles. Based on the results of the training and validation cohorts, we finally identified that the extreme gradient boosting (XGB) model has the best predictive efficiency and stability, with an AUROC of up to 0.981 in the training set and up to 0.896 in the test set. In addition, the calibration curve and the DCA showed that the XGB model had the best discrimination and clinical practicability. The variables were ranked according to the feature importance, and the five variables contributing the most to the model were hemoglobin (Hgb), white blood cell count (WBC), alkaline phosphatase, coadministration of highly toxic chemotherapy drugs, and albumin. Conclusions: The incidence of grade ? 2 CIM was not low in children with WT, which needs attention. The XGB model was developed to predict the risk of grade ? 2 CIM in children with WT for the first time. The model has good predictive performance and stability and has the potential to be translated into clinical applications. Based on this modeling and application approach, the extension of CIM prediction models to other pediatric malignancies could be expected. Full article

The definitive diagnosis of histology specimen images is largely based on the radiologist’s comprehensive experience; however, due to the fine to the coarse visual appearance of such images, experts often disagree with their assessments. Sophisticated deep learning approaches can help to automate the diagnosis process of the images and reduce the analysis duration. More efficient and accurate automated systems can also increase the diagnostic impartiality by reducing the difference between the operators. We propose a FabNet model that can learn the fine-to-coarse structural and textural features of multi-scale histopathological images by using accretive network architecture that agglomerate hierarchical feature maps to acquire significant classification accuracy. We expand on a contemporary design by incorporating deep and close integration to finely combine features across layers. Our deep layer accretive model structure combines the feature hierarchy in an iterative and hierarchically manner that infers higher accuracy and fewer parameters. The FabNet can identify malignant tumors from images and patches from histopathology images. We assessed the efficiency of our suggested model standard cancer datasets, which included breast cancer as well as colon cancer histopathology images. Our proposed avant garde model significantly outperforms existing state-of-the-art models in respect of the accuracy, F1 score, precision, and sensitivity, with fewer parameters. Full article

This study aimed to investigate the potential of quantitative radiomic data extracted from conventional MR images in discriminating IDH-mutant grade 4 astrocytomas from IDH-wild-type glioblastomas (GBMs). A cohort of 57 treatment-naïve patients with IDH-mutant grade 4 astrocytomas (n = 23) and IDH-wild-type GBMs (n = 34) underwent anatomical imaging on a 3T MR system with standard parameters. Post-contrast T1-weighted and T2-FLAIR images were co-registered. A semi-automatic segmentation approach was used to generate regions of interest (ROIs) from different tissue components of neoplasms. A total of 1050 radiomic features were extracted from each image. The data were split randomly into training and testing sets. A deep learning-based data augmentation method (CTGAN) was implemented to synthesize 200 datasets from the training sets. A total of 18 classifiers were used to distinguish two genotypes of grade 4 astrocytomas. From generated data using 80% training set, the best discriminatory power was obtained from core tumor regions overlaid on post-contrast T1 using the K-best feature selection algorithm and a Gaussian naïve Bayes classifier (AUC = 0.93, accuracy = 0.92, sensitivity = 1, specificity = 0.86, PR_AUC = 0.92). Similarly, high diagnostic performances were obtained from original and generated data using 50% and 30% training sets. Our findings suggest that conventional MR imaging-based radiomic features combined with machine/deep learning methods may be valuable in discriminating IDH-mutant grade 4 astrocytomas from IDH-wild-type GBMs. Full article

After cardiovascular diseases, cancer is responsible for the most deaths worldwide. Detecting a cancer disease early improves the chances for healing significantly. One group of technologies that is increasingly applied for detecting cancer is artificial intelligence. Artificial intelligence has great potential to support clinicians and medical practitioners as it allows for the early detection of carcinomas. During recent years, research on artificial intelligence for cancer detection grew a lot. Within this article, we conducted a bibliometric study of the existing research dealing with the application of artificial intelligence in cancer detection. We analyzed 6450 articles on that topic that were published between 1986 and 2022. By doing so, we were able to give an overview of this research field, including its key topics, relevant outlets, institutions, and articles. Based on our findings, we developed a future research agenda that can help to advance research on artificial intelligence for cancer detection. In summary, our study is intended to serve as a platform and foundation for researchers that are interested in the potential of artificial intelligence for detecting cancer. Full article

Background: Preoperative prediction of short-term postoperative outcomes in spinal tumor patients can lead to more precise patient care plans that reduce the likelihood of negative outcomes. With this study, we aimed to develop machine learning algorithms for predicting short-term postoperative outcomes and implement these models in an open-source web application. Methods: Patients who underwent surgical resection of spinal tumors were identified using the American College of Surgeons, National Surgical Quality Improvement Program. Three outcomes were predicted: prolonged length of stay (LOS), nonhome discharges, and major complications. Four machine learning algorithms were developed and integrated into an open access web application to predict these outcomes. Results: A total of 3073 patients that underwent spinal tumor resection were included in the analysis. The most accurately predicted outcomes in terms of the area under the receiver operating characteristic curve (AUROC) was the prolonged LOS with a mean AUROC of 0.745 The most accurately predicting algorithm in terms of AUROC was random forest, with a mean AUROC of 0.743. An open access web application was developed for getting predictions for individual patients based on their characteristics and this web application can be accessed here: huggingface.co/spaces/MSHS-Neurosurgery-Research/NSQIP-ST. Conclusion: Machine learning approaches carry significant potential for the purpose of predicting postoperative outcomes following spinal tumor resections. Development of predictive models as clinically useful decision-making tools may considerably enhance risk assessment and prognosis as the amount of data in spinal tumor surgery continues to rise. Full article

Cytopathological examination is one of the main examinations for pleural effusion, and especially for many patients with advanced cancer, pleural effusion is the only accessible specimen for establishing a pathological diagnosis. The lack of cytopathologists and the high cost of gene detection present opportunities for the application of deep learning. In this retrospective analysis, data representing 1321 consecutive cases of pleural effusion were collected. We trained and evaluated our deep learning model based on several tasks, including the diagnosis of benign and malignant pleural effusion, the identification of the primary location of common metastatic cancer from pleural effusion, and the prediction of genetic alterations associated with targeted therapy. We achieved good results in identifying benign and malignant pleural effusions (0.932 AUC (area under the ROC curve)) and the primary location of common metastatic cancer (0.910 AUC). In addition, we analyzed ten genes related to targeted therapy in specimens and used them to train the model regarding four alteration statuses, which also yielded reasonable results (0.869 AUC for ALK fusion, 0.804 AUC for KRAS mutation, 0.644 AUC for EGFR mutation and 0.774 AUC for NONE alteration). Our research shows the feasibility and benefits of deep learning to assist in cytopathological diagnosis in clinical settings. Full article