A whirl of radiomics-based biomarkers in cancer immunotherapy, why is large scale validation still lacking?

The search for understanding immunotherapy response has sparked interest in diverse areas of oncology, with artificial intelligence (AI) and radiomics emerging as promising tools, capable of gathering large amounts of information to identify suitable patients for treatment. The application of AI in radiology has grown, driven by the hypothesis that radiology images capture tumor phenotypes and thus could provide valuable insights into immunotherapy response likelihood. However, despite the rapid growth of studies, no algorithms in the field have reached clinical implementation, mainly due to the lack of standardized methods, hampering study comparisons and reproducibility across different datasets. In this review, we performed a comprehensive assessment of published data to identify sources of variability in radiomics study design that hinder the comparison of the different model performance and, therefore, clinical implementation. Subsequently, we conducted a use-case meta-analysis using homogenous studies to assess the overall performance of radiomics in estimating programmed death-ligand 1 (PD-L1) expression. Our findings indicate that, despite numerous attempts to predict immunotherapy response, only a limited number of studies share comparable methodologies and report sufficient data about cohorts and methods to be suitable for meta-analysis. Nevertheless, although only a few studies meet these criteria, their promising results underscore the importance of ongoing standardization and benchmarking efforts. This review highlights the importance of uniformity in study design and reporting. Such standardization is crucial to enable meaningful comparisons and demonstrate the validity of biomarkers across diverse populations, facilitating their implementation into the immunotherapy patient selection process.

Comprehensive evaluation of surrogate endpoints to predict overall survival in trials with PD1/PD-L1 immune checkpoint inhibitors plus chemotherapy.

BACKGROUND: PD1/PD-L1 immune checkpoint inhibitors (ICI) have revolutionized cancer treatment. Although there is controversy about the accuracy of surrogate endpoints in the ICI setting to predict overall survival (OS), these endpoints are commonly used in confirmatory trials. Here we aimed to explore the validity of classical and novel surrogate endpoints in randomised controlled trials (RCT) that combine ICI plus chemotherapy (CT) in the first-line setting. MATERIAL AND METHODS: A systematic review was conducted to identify RCT investigating anti-PD1/PD-L1 drugs plus CT versus CT alone. We performed (i) arm-level analysis to evaluate predictors of median OS (mOS) and (ii) comparison-level analysis for OS hazard ratio (HR) estimations. Linear regression models weighted by trial size were fitted and adjusted R2 values were reported. RESULTS: Thirty-nine RCTs involving 22,341 patients met the inclusion criteria (17 non-small cell lung, 9 gastroesophageal and 13 in other cancers) with ten different ICI under study. Overall, ICI plus CT improved OS (HR = 0.76; 95%CI: 0.73-0.80). In the arm-level analysis, the best mOS prediction was obtained with a new endpoint that combines median duration of response and ORR (mDoR-ORR) and with median PFS (R2 = 0.5 both). In the comparison-level analysis, PFS HR showed a moderate association with OS HR (R2 = 0.52). Early OS read-outs were highly associated with final OS outcomes (R2 = 0.80). CONCLUSIONS: The association between surrogate endpoints and OS in first-line RCT combining anti-PD1/PD-L1 and CT is moderate-low. Early OS read-outs showed a good association with final OS HR while the mDOR-ORR endpoint could help to better design confirmatory trials after single-arm phase II trials.

Physical exercise and fatigue in advanced gastrointestinal cancer during chemotherapy.

OBJECTIVES: This study was aimed to analyse the effect of a patient-oriented modality of physical exercise (programmed and directed physical exercise (PDPE)) on cancer-related fatigue (CRF) and quality of life (QoL). The secondary aim was to evaluate changes in body composition and skeletal muscle function during the study in patients with and without PDPE. METHODS: A prospective randomised study was conducted to analyse the impact of PDPE on CRF and QoL. Patients were selected before the development of CRF to set the intervention before its appearance. A high probability CRF population was chosen: patients with advanced gastrointestinal cancer undergoing chemotherapy with weight loss (≥5%) over the last 6 months. PDPE consisted of a programme of exercise delivered weekly and adjusted to patients' medical conditions. Four visits were planned (weeks 0, 4, 8 and 12). QoL, CRF, body composition and skeletal muscle function were evaluated in each visit. RESULTS: From 101 patients recruited, 64 were considered evaluable, with three or four visits completed (n=30 control, n=34 PDPE group). Satisfactory compliance of ≥50% to the PDPE programme was seen in 47%. A reduction in the severity of fatigue was detected in the PDPE group (p=0.019), being higher in the subgroup of satisfactory compliance (p<0.001). This latter group showed better results of QoL in comparison with the control group (p=0.0279). A significant increase in endurance was found in the PDPE group (p<0.001). CONCLUSION: PDPE reduced the severity of fatigue and improved QoL. The difference in endurance would explain the results seen in the severity of fatigue.

Clinical implications of homologous recombination repair mutations in prostate cancer.

Prostate cancer is a disease with significant interpatient genomics, with a proportion of patients presenting mutations in key homologous recombination repair (HRR) gene aberrations, particularly in late-stage disease. A better understanding of the genomic landscape of prostate cancer and the prognostic and predictive value of HRR mutations could lead to more precise care for prostate cancer patients. BRCA1/2 mutations are associated with a more aggressive disease course and higher risk of developing lethal prostate cancer, but also identify patients who could benefit from directed therapeutic strategies with PARP inhibitors. Other HRR mutations are also frequent but their prognostic and predictive value for prostate cancer patients is less clear. Moreover, a proportion of these mutations are associated with inherited germline defects, being relevant for the patients' risk of second malignancies but also to inform their relatives' risk of cancer through cascade testing. In this manuscript, we review current knowledge of the prognostic and predictive value for different HHR alterations across the different prostate cancer disease states. Additionally, we assess the challenges to implement genomic testing in clinical practice for prostate cancer patients.

Practical foundations of machine learning for addiction research. Part I. Methods and techniques.

Machine learning assembles a broad set of methods and techniques to solve a wide range of problems, such as identifying individuals with substance use disorders (SUD), finding patterns in neuroimages, understanding SUD prognostic factors and their association, or determining addiction genetic underpinnings. However, the addiction research field underuses machine learning. This two-part narrative review focuses on machine learning tools and concepts, providing an introductory insight into their capabilities to facilitate their understanding and acquisition by addiction researchers. This first part presents supervised and unsupervised methods such as linear models, naive Bayes, support vector machines, artificial neural networks, and k-means. We illustrate each technique with examples of its use in current addiction research. We also present some open-source programming tools and methodological good practices that facilitate using these techniques. Throughout this work, we emphasize a continuum between applied statistics and machine learning, we show their commonalities, and provide sources for further reading to deepen the understanding of these methods. This two-part review is a primer for the next generation of addiction researchers incorporating machine learning in their projects. Researchers will find a bridge between applied statistics and machine learning, ways to expand their analytical toolkit, recommendations to incorporate well-established good practices in addiction data analysis (e.g., stating the rationale for using newer analytical tools, calculating sample size, improving reproducibility), and the vocabulary to enhance collaboration between researchers who do not conduct data analyses and those who do.

Practical foundations of machine learning for addiction research. Part II. Workflow and use cases.

In a continuum with applied statistics, machine learning offers a wide variety of tools to explore, analyze, and understand addiction data. These tools include algorithms that can leverage useful information from data to build models; these models can solve particular tasks to answer addiction scientific questions. In this second part of a two-part review on machine learning, we explain how to apply machine learning methods to addiction research. Like other analytical tools, machine learning methods require a careful implementation to carry out a reproducible and transparent research process with reliable results. This review describes a workflow to guide the application of machine learning in addiction research, detailing study design, data collection, data pre-processing, modeling, and results communication. How to train, validate, and test a model, detect and characterize overfitting, and determine an adequate sample size are some of the key issues when applying machine learning. We also illustrate the process and particular nuances with examples of how researchers in addiction have applied machine learning techniques with different goals, study designs, or data sources as well as explain the main limitations of machine learning approaches and how to best address them. A good use of machine learning enriches the addiction research toolkit.

Effect of nitric oxide inhibition in Bacillus Calmette-Guerin bladder cancer treatment.

BACKGROUND: Bacillus Calmette-Guerin (BCG) is the standard treatment for patients with high-risk non-muscle invasive bladder cancer (BC). Despite its success, about 30-50% of patients are refractory. It was reported that inducible nitric oxide synthase (iNOS) tumor expression is presented in 50% of human BC, associated with bad prognosis and BCG failure. OBJECTIVE: to evaluate in human bladder tumors the association between iNOS expression and the tumor microenvironment focusing on the immunosuppressive protein S100A9. Also, investigate in a preclinical murine MB49-BC model the tumor immunoresponse induced by BCG in combination with the nitric oxide production inhibitor l-NAME. RESULTS: In human bladder tumors, we detected a positive association between iNOS and S100A9 tumor expression, suggesting a relationship between both immunomodulatory proteins. We also found a positive correlation between iNOS tumor expression and the presence of S100A9+ tumor-infiltrating cells, suggesting an immunosuppressive tumor microenvironment induced by the nitric oxide production. Using the subcutaneous murine BC model, we show that similarly to the human pathology, MB49 tumors constitutively expressed iNOS and S100A9 protein. MB49 tumor-bearing mice presented an immunosuppressive systemic profile characterized by fewer cytotoxic cells (CD8+ and NK) and higher suppressor cells (Treg and myeloid-derived suppressor cells -MDSC-) compared to normal mice. BCG treatment reduced tumor growth, increasing local CD8+-infiltrating cells and induced a systemic increase in CD8+ and a reduction in Treg. BCG combined with l-NAME, significantly reduced tumor growth compared to BCG alone, diminishing iNOS and S100A9 tumor expression and increasing CD8+-infiltrating cells in tumor microenvironment. This local response was accompanied by the systemic increase in CD8+ and NK cells, and the reduction in Treg and MDSC, even more than BCG alone. Similar results were obtained using the orthotopic BC model, where an increase in specific cytotoxicity against MB49 tumor cells was detected. CONCLUSION: The present study provides preclinical information where NO inhibition in iNOS-expressing bladder tumors could contribute to improve BCG antitumor immune response. The association between iNOS and S100A9 in human BC supports the hypothesis that iNOS expression is a negative prognostic factor and a promising therapeutic target.