Hasil untuk "Neoplasms. Tumors. Oncology. Including cancer and carcinogens"

Qi Zhang, Wencui Yang, Lina Tan et al.

Proton therapy, by leveraging its unique physical characteristic of the Bragg peak, enables high-precision dose delivery to the tumor target while effectively protecting surrounding normal tissues, and has become an important representative of advanced radiotherapy. This review aims to systematically summarize key technological breakthroughs in recent years that have driven the progress of proton therapy, including compact superconducting accelerators, pencil beam scanning (PBS), image-guided proton therapy (IGPT), and the transformative ultra-high dose rate FLASH radiotherapy, while highlighting the role of artificial intelligence (AI) in advancing proton therapy toward real-time adaptive precision radiotherapy. The article also explores the global distribution and development status of proton centers, with a specific analysis of China’s notable advancements as an emerging market in center construction, equipment localization, and the treatment of characteristic local tumor types. Moving forward, it is essential to continue promoting technological integration and innovation, strengthen high-quality clinical research, and develop a more accessible, intelligent, and personalized proton therapy system to achieve broader clinical application and patient benefit.

Brian Isett, Rebekah Dadey, Aofei Li et al.

Accurate localization of tumor regions from hematoxylin and eosin-stained whole-slide images is fundamental for translational research including spatial analysis, molecular profiling, and tissue architecture investigation. However, deep learning-based tumor detection trained within specific cancers may exhibit reduced robustness when applied across different tumor types. We investigated whether balanced training across cancers at modest scale can achieve high performance and generalize to unseen tumor types. A multi-cancer tumor localization model (MuCTaL) was trained on 79,984 non-overlapping tiles from four cancers (melanoma, hepatocellular carcinoma, colorectal cancer, and non-small cell lung cancer) using transfer learning with DenseNet169. The model achieved a tile-level ROC-AUC of 0.97 in validation data from the four training cancers, and 0.71 on an independent pancreatic ductal adenocarcinoma cohort. A scalable inference workflow was built to generate spatial tumor probability heatmaps compatible with existing digital pathology tools. Code and models are publicly available at https://github.com/AivaraX-AI/MuCTaL.

Karol Gawalski, Weronika Przybyszewska, Jaromir Hunia et al.

Messenger ribonucleic acid (mRNA) technology is a promising platform for cancer immunotherapy. Unlike traditional vaccines that prevent infectious diseases, mRNA’s role in oncology is to stimulate or enhance the immune response against tumor antigens. This review provides an overview of mRNA’s historical development, from its discovery in 1961 to recent clinical trials and Nobel Prize-winning breakthroughs. Therapeutic mRNA flexibility allows the alteration of diverse tumor antigens. Key targets include tumor-associated antigens, which are present on both tumor cells and some healthy cells, as well as tumor-specific antigens unique to cancer cells, such as antiviral antigens and neoantigens arising from tumor mutations. Various approaches to protect mRNA from degradation, including protamine-complexed mRNA, lipoplexes, and lipid nanoparticles, as well as several administration routes, are currently being tested in clinical trials. They are focused on malignancies like melanoma, non-small cell lung cancer, prostate cancer, or pancreatic ductal adenocarcinoma, one of the most challenging cancers. While many trials are in early phases, some have advanced to phase 3 and have shown promising results in both safety and efficacy. However, due to the complexity and heterogeneity of tumors, even among patients presenting the same subgroup of neoplasm, fully universal mRNA-based cancer vaccine seems to be elusive. Personalized mRNA cancer vaccines targeting neoantigens unique to an individual’s tumor have gained traction as a feasible and promising solution. Technological advances in bioinformatics, AI, and machine learning now allow for more accurate identification of immunogenic neoepitopes. The combination this type of therapy with other treatment such as immune checkpoint inhibitors may become one of new solutions in oncology.

A. Laffi, A. Prinzi, C. di Dato et al.

Background Ectopic Adrenocorticotropic Hormone (ACTH) Syndrome (EAS) is a complex disorder caused by ACTH-producing tumors located outside the pituitary gland. EAS is most commonly associated with neuroendocrine neoplasms (NENs), rare malignancies category. Due to the nonspecific symptoms, EAS is often misdiagnosed, contributing to increased morbidity and complicating clinical management. In Italy, access to diagnostic and therapeutic resources for EAS and NENs varies significantly by region. As part of the 2024–2025 NIKE (Neuroendocrine Tumors, Innovation in Knowledge and Education) initiative, a multidisciplinary group, including endocrinologists, oncologists, pathologists, and nuclear medicine experts, designed a national survey to assess awareness, diagnostic approaches, and management of EAS in Italian centers. Methods A 50-items structured questionnaire was developed, covering 3 sections: respondents’ profile, diagnostic approaches, and treatment strategies. The survey was distributed as an anonymized form via email, with data collected from April to June 2025. Results Sixteen Italian centers with NEN and EAS expertise participated. Most experts worked in European referral centers for rare tumors where the majority have an in-house, NEN-dedicated multidisciplinary team. Initial points of contact occurred most frequently in oncology (37.5%) and endocrinology (31.5%) clinics. A diagnostic delay was reported by 56% of respondent centers; hypokalemia was the most common presenting sign (93.8%). In 56.3% of centers, respondents reported that EAS was more commonly diagnosed before the detection of the underlying NEN, most frequently lung carcinoids or small/large cell cancers (87.5%). Regarding diagnostic practices, 56.3% of centers indicated the use of the 1 mg dexamethasone suppression test (DST), followed by the high-dose DST. The desmopressin test was considered outdated or replaceable by 43.8% of respondents. Regarding therapeutic approaches, respondents reported that upfront surgery was performed in up to 50% of centers, with preoperative bridging pharmacological therapy used to achieve eucortisolism. Osilodrostat was the most frequently preferred first-line treatment. Conclusion This survey provides a valuable snapshot of EAS care in Italy, highlighting both strengths and areas for improvement. The findings underscore the need for a national, more structured referral network to ensure timely diagnosis and access to specialized care. These insights may guide national protocol harmonization in EAS management and better alignment with international standards.

Claire Delehouzé, Melodie Mallais, Arnaud Comte et al.

Abstract In the past two decades, various non-apoptotic pathways of regulated cell death have been identified; a small subset of these, including necroptosis and ferroptosis, manifests the phenotypic features of necrotic death. These two regulated necroses are being extensively studied because of their putative roles in severe acute and chronic pathologies. Moreover, as these regulated necrotic pathways are coactivated in a number of common pathologies, the development of multi-target directed ligands (that is, the use of a polypharmacological strategy) is a path-breaking avenue of research. In this study, we determined that the 7-azaindole derivative, sibiriline, inhibited both RIPK1-driven necroptosis (induced by Tumor Necrosis Factor-α) and ferroptosis (triggered by various classes of ferroptosis inducers), with EC50s against each in the µM range. We next performed a combined large-scale transcriptomic study in order to determine the molecular mechanisms of action of sibiriline. We identified the stress response protein heme oxygenase-1 (HMOX1) as the main biomarker of ferroptosis inhibition by sibiriline. We hypothesized that this compound reacts as an antioxidant to block ferroptosis; indeed, we found that sibiriline inhibits lipid peroxidation by trapping phospholipid-derived peroxyl radicals as a radical-trapping antioxidant (RTA). Taken together, these results show that sibiriline is a new dual inhibitor of necroptosis and ferroptosis cell death pathways; it works by inhibition of both RIPK1 kinase and (phospho)lipid peroxidation. We also demonstrate the in vitro efficacy of sibiriline to inhibit cell death in cell-based models of Parkinson’s disease and cystic fibrosis. These findings shed light on the high therapeutic potency of RIPK1 inhibitors with RTA activity.

Rocio Roldan-Testillano, Rafael Sanchez-Salas

Hritika D. Pai MPT, K Vijaya Kumar PhD, Prasanna Mithra MD, DNB et al.

Yuhui Tao, Zhongwei Zhao, Zilong Wang et al.

The non-invasive assessment of increasingly incidentally discovered renal masses is a critical challenge in urologic oncology, where diagnostic uncertainty frequently leads to the overtreatment of benign or indolent tumors. In this study, we developed and validated RenalCLIP using a dataset of 27,866 CT scans from 8,809 patients across nine Chinese medical centers and the public TCIA cohort, a visual-language foundation model for characterization, diagnosis and prognosis of renal mass. The model was developed via a two-stage pre-training strategy that first enhances the image and text encoders with domain-specific knowledge before aligning them through a contrastive learning objective, to create robust representations for superior generalization and diagnostic precision. RenalCLIP achieved better performance and superior generalizability across 10 core tasks spanning the full clinical workflow of kidney cancer, including anatomical assessment, diagnostic classification, and survival prediction, compared with other state-of-the-art general-purpose CT foundation models. Especially, for complicated task like recurrence-free survival prediction in the TCIA cohort, RenalCLIP achieved a C-index of 0.726, representing a substantial improvement of approximately 20% over the leading baselines. Furthermore, RenalCLIP's pre-training imparted remarkable data efficiency; in the diagnostic classification task, it only needs 20% training data to achieve the peak performance of all baseline models even after they were fully fine-tuned on 100% of the data. Additionally, it achieved superior performance in report generation, image-text retrieval and zero-shot diagnosis tasks. Our findings establish that RenalCLIP provides a robust tool with the potential to enhance diagnostic accuracy, refine prognostic stratification, and personalize the management of patients with kidney cancer.

Saurabh Bhandari, Michael J. Daniels, Chenguang Wang

In early-phase cancer clinical trials, the limited availability of data presents significant challenges in developing a framework to efficiently quantify treatment effectiveness. To address this, we propose a novel utility-based Bayesian approach for assessing treatment effects in these trials, where data scarcity is a major concern. Our approach synthesizes tumor burden, a key biomarker for evaluating patient response to oncology treatments, and survival outcome, a widely used endpoint for assessing clinical benefits, by jointly modeling longitudinal and survival data. The proposed method, along with its novel estimand, aims to efficiently capture signals of treatment efficacy in early-phase studies and holds potential for development as an endpoint in Phase 3 confirmatory studies. We conduct simulations to investigate the frequentist characteristics of the proposed estimand in a simple setting, which demonstrate relatively controlled Type I error rates when testing the treatment effect on outcomes.

T. I. Mughal, J. Mascarenhas, R. Rampal et al.

Despite the study of BCR::ABL1‐positive and ‐negative myeloproliferative neoplasms (MPNs) providing seminal insights into cancer biology, tumor evolution and precision oncology over the past half century, significant challenges remain. MPNs are clonal hematopoietic stem cell‐derived neoplasms with heterogenous clinical phenotypes and a clonal architecture which impacts the often‐complex underlying genetics and microenvironment. The major driving molecular abnormalities have been well characterized, but debate on their role as disease‐initiating molecular lesions continues. The introduction of the ABL1 tyrosine kinase inhibitors have been extremely successful in the treatment of chronic myeloid leukemia with most patients having a near‐normal life expectancy. Similar success has, however, not been achieved for BCR::ABL1‐negative MPNs in terms of disease course modification and most patients remain incurable. In both disease categories, genomic instability seems to increase the risk of disease progression to accelerated/blast phase, which is resistant/refractory to conventional treatment and associated with a poor prognosis. To address some of these issues, the late John Goldman and Tariq Mughal founded a scientific and clinical platform in 2006, the Post‐American Society of Hematology (ASH) MPN workshop, to appraise novel cancer biology, candidate therapeutic targets, treatments and other clinical challenges and pay tribute to all the many scientists and clinicians around the world instrumental to the progress made and continuing advances being made. This paper summarizes some of the recent data discussed at the 18th edition of the workshop and includes reference to some data presented or published after the workshop, including the 26th John Goldman CML conference.

Jawad S. Alnajjar, Faisal A Al-Harbi, A. K. Alsaif et al.

Background: Appendiceal neoplasms account for less than 1% of gastrointestinal cancers but are increasing in incidence worldwide. Their marked histological variations and differences create multiple challenges for prognosis and management planning, as current staging systems are limited in certain aspects for capturing the entire disease complexity. Methods: We synthesized data from 18 large observational studies, including 67,001 patients diagnosed between 1973 and 2024. Using advanced computational modeling, we combined multiple statistical methods and machine learning techniques to improve risk stratification, survival prediction, treatment optimization, and forecasting. A novel overlap-aware weighting methodology was applied to prevent double-counting across overlapping registries. Results: Our multi-dimensional risk model outperformed TNM staging (C-index 0.758 vs. 0.689), identifying five prognostic groups with five-year overall survival ranging from 88.7% (low-risk neuroendocrine tumors (NETs)) to 27.3% (high-risk signet-ring cell carcinomas (SRCC)). Hierarchical survival analysis demonstrated marked variation across histological variants, with goblet cell adenocarcinoma showing the most favorable outcomes. Causal inference confirmed the survival benefit of hyperthermic intraperitoneal chemotherapy (HIPEC) in stage IV disease (five-year overall survival (OS) 87.4%) and highlighted disparities in outcomes by race and institutional volume. Time-series forecasting projected a 25% to 50% increase in incidence by 2030, highlighting the growing risk of global burden. Conclusions: By integrating multi-database evidence with advanced modeling and statistical methodologies, our findings demonstrate valuable insights and implications for individualized prognosis, better management decision-making, and health system planning. Our proposed approach and demonstrated methodologies are warranting better progression and advancements in precision oncology and utilization of computational modeling techniques in big data as well as digital health progression landscape.

Dennise A. De Jesús-Díaz, Aaron D Goldberg, Esteban Abella et al.

Background: Nicotinamide phosphoribosyltransferase (NAMPT) controls how cells use energy and is central to human biology. NAMPT is up-regulated in many cancers, including hematologic malignancies. Cancer cells have a heightened dependency on NAMPT-driven NAD flux that makes them particularly sensitive to NAMPT inhibitors. However, since healthy cells also require NAD for survival, previous clinical attempts at NAMPT inhibition have resulted in dose-limiting toxicities preventing their development. To circumvent this, we developed RPT1G, a first-in-class hyperbolic NAMPT inhibitor with an enhanced therapeutic window that enables therapeutically active NAD reduction in cancer cells while allowing sufficient NAD production in healthy tissues (Crimmins, ASH 2023). Data from a first-in-human, Phase 1, randomized, double-blind, placebo-controlled, SAD and MAD study in healthy volunteers shows that oral administration of RPT1G is safe and well-tolerated with a favorable PK profile. Human target engagement results are consistent with RPT1G inhibiting NAMPT at doses predicted to be therapeutically relevant in oncology. Study Design and Methods: RPT1G is being investigated in an international, Phase 1, multi-center, open-label clinical trial (NCT07107126) for treatment of Relapsed/Refractory Acute Myeloid Leukemia (R/R AML) and High-Risk Myelodysplastic Syndromes/Neoplasms (HR-MDS). The study will use a standard 3+3 design with escalating oral monotherapy doses of RPT1G, twice daily for 14-day cycles, with potential to explore other dosing schedules based on observed pharmacokinetics (PK) parameters, PK/pharmacodynamic (PD) relationships, safety, and tolerability data to identify the maximum tolerated dose (MTD) or presumptive biologically effective dose (BED) and select the recommended phase 2 dose (RP2D). The starting dose, derived from healthy volunteer study, achieves NAMPT target engagement at levels predicted to reduce tumor burden. Primary objectives include: 1) defining safety and tolerability; 2) determining the RP2D, optimal schedule and/or BED. Key secondary objectives include: 1) evaluating PK; 2) assessing preliminary efficacy by European LeukemiaNet (ELN) 2022, including overall response rate (ORR), duration of response (DoR), and hematologic improvement (HI), and clinical benefit by transfusion independence and the International Working Group 2023 HR-MDS response criteria. The study will enroll adults with a histological confirmation of R/R-AML as defined by the ELN 2022 criteria or HR-MDS as defined by the International Consortium for MDS 2023 criteria that have received appropriate standard of care therapy(s) in the opinion of the investigator or declined receipt of these. Adequate organ function is required. Patients will be excluded if they have ongoing AEs from prior therapies; have received radiation within 14 days of the first dose of study drug; have known active infections; or have uncontrolled cardiac issues, or other medical comorbidities that will preclude safety evaluation. Approximately 18 patients will be enrolled across 5-10 sites across the US and Australia. RPT1G is the first NAMPT hyperbolic inhibitor to be studied for the treatment of R/R-AML and HR-MDS patients.

F. Tucci, M. G. Filippone, R. Bonfanti et al.

Advances in the personalized management of bladder cancer (BCa) have been hampered by the lack of predictive biomarkers and targeted therapies. In this study, we highlight a previously uncharacterized role for the loss of the tumor suppressor, NUMB, as a hallmark of biologically and clinically aggressive BCa. Retrospective analyses of longitudinal patient cohorts show that low NUMB expression in the primary tumor associates with increased risk of muscle invasion progression in non-muscle invasive BCa (NMIBC), and correlates with advanced stage at diagnosis, vascular invasion, and worse overall survival in post-cystectomy muscle-invasive BCa (MIBC). The causal role of NUMB loss to bladder tumorigenesis is highlighted in a transgenic mouse model where targeted Numb gene deletion in the basal layer of the bladder mucosa is per se sufficient to drive malignant transformation of the normal urothelium and sensitizes it to carcinogenic insults, accelerating tumor onset and progression. Mechanistically, through integrative transcriptomic and functional analyses in mouse and human BCa models, we reveal that the molecular hallmark associated with NUMB loss is the inhibition of the Hippo cascade with ensuing activation of the YAP oncogene via a RHOA/ROCK-dependent pathway. This ultimately results in a YAP-dependent transcriptional activation of an EMT/plasticity program and acquisition of uniquely aggressive proliferative and invasive phenotypes in BCa cells. Through pharmacological or genetic inhibition studies, we provide formal proof-of-evidence of the actionability of the RHOA/ROCK/YAP-TEAD axis to selectively inhibit proliferation and invasion of NUMB-deficient primary mouse and established human BCa cells in 3D-Matrigel organoid models, and to curb their in vivo tumorigenic potential in xenograft studies. We thereby highlight RHOA, ROCK and YAP-TEAD as selective vulnerable targets for therapeutic intervention in highly aggressive NUMB-deficient human BCa. We also describe a minimal signature composed of 27 up/down-regulated genes, characteristic of the NUMB-defective condition (the “NUMBLESS signature) that behaves as an independent prognostic predictor of risk of MIBC transition in a validation cohort study including 535 NMIBC patients enrolled in the UROMOL trial, where the NUMBLESS signature also associates with favorable response to BCG treatment. In summary, our study identifies NUMB loss as a key determinant of an underlying aggressive biology in BCa, and highlights its potential as a clinically actionable biomarker for prediction of NMIBC progression to life-threatening MIBC and response to standard therapies (i.e. BCG) or, even more importantly, to innovative anti-RHOA/ROCK/YAP-TEAD targeted treatments, thereby advancing a framework for precision oncology for high-risk NMIBC patients. Francesco A. Tucci, Maria Grazia Filippone, Roberta Bonfanti, Elisa Guerrera, Luca Acerbi, Giuseppe Renne, Gianluca Vago, Daniela Tosoni, Salvatore Pece. Loss of the tumor suppressor NUMB is a hallmark of biologically and clinically aggressive bladder cancer and leads to hyperactivation of an actionable RHOA-ROCK-YAP circuitry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4309.

Fruzsina Fazekas, Zsuzsanna Ujfaludi, Krisztina Bíró et al.

BACKGROUND Testicular cancer is the most common solid malignancy among men aged 15-35. Radical orchiectomy and platinum-based chemotherapy (BEP) are curative in the majority of patients, including advanced, metastatic cases. According to current urooncology guidelines all non-seminoma patients harbouring post-chemotherapy residual masses of ≥ 1cm should undergo salvage retroperitoneal lymph node dissection (RPLND). However, only 10% of residual tumors contain viable disease. OBJECTIVE To assess patient outcomes and complications considering different treatment regimens and clinical characteristics. MATERIALS AND METHODS In a retrospective cross-sectional study patients (n=127) who underwent postchemotherapy RPLND between 2007 and 2023 at our referral center were evaluated. The patients received systemic treatment at various oncology centers. The number of BEP cycles received were occasionally different from standard. Only patients with normal postchemotherapy serum tumor markers and primary testicular or extragonadal germ cell neoplasms were included. Treatment groups were established according to the number of BEP cycles received, and the extent of RPLND (bilateral or modified template). Treatment outcomes and complications were assessed. RESULTS Standard 3-4 courses of BEP were received by 100 (78,7%) patients, while 11 (8,7%) patients underwent less, and 16 (12,6%) more courses than standard. On histopathologic evaluation viable germ cell tumor, teratoma, and necrosis/fibrosis was present in 26 (20,5%), 67 (52,7%) and 34 (26,8%) of specimen, respectively. In the 5-6 BEP series subgroup high rate of viable disease (37,5%) was found and significantly more nephrectomies were performed, than other chemotherapy subgroups. Extratesticular GCT, viable disease in residual mass or progression after RPLND indicated lower survival. Mild (Clavien-Dindo I-II) or no postoperative complications were reported in 93,7% of cases. CONCLUSIONS The study suggests no significant benefit from exceeding 3-4 courses of BEP. Timely salvage RPLND should be performed in high volume centers for optimal treatment outcomes with acceptable complication rates. Adherence to the Heidenreich criteria is advisable where practical.

M. Hashem, Elaheh Mohandesi Khosroshahi, Melika Aliahmady et al.

Bladder cancer (BC) is a highly frequent neoplasm in correlation with significant rate of morbidity, mortality, and cost. The onset of BC is predominantly triggered by environmental and/or occupational exposures to carcinogens, such as tobacco. There are two distinct pathways by which BC can be developed, including non-muscle-invasive papillary tumors (NMIBC) and non-papillary (or solid) muscle-invasive tumors (MIBC). The Cancer Genome Atlas project has further recognized key genetic drivers of MIBC along with its subtypes with particular properties and therapeutic responses; nonetheless, NMIBC is the predominant BC presentation among the suffering individuals. Radical cystoprostatectomy, radiotherapy, and chemotherapy have been verified to be the common therapeutic interventions in metastatic tumors, among which chemotherapeutics are more conventionally utilized. Although multiple chemo drugs have been broadly administered for BC treatment, cisplatin is reportedly the most effective chemo drug against the corresponding malignancy. Notwithstanding, tumor recurrence is usually occurred following the consumption of cisplatin regimens, particularly due to the progression of chemo-resistant trait. In this framework, non-coding RNAs (ncRNAs), as abundant RNA transcripts arise from the human genome, are introduced to serve as crucial contributors to tumor expansion and cisplatin chemo-resistance in bladder neoplasm. In the current review, we first investigated the best-known ncRNAs, i.e. microRNAs (miRNAs), long ncRNAs (lncRNAs), and circular RNAs (circRNAs), correlated with cisplatin chemo-resistance in BC cells and tissues. We noticed that these ncRNAs could mediate the BC-related cisplatin-resistant phenotype through diverse cellular processes and signaling mechanisms, reviewed here. Eventually, diagnostic and prognostic potential of ncRNAs, as well as their therapeutic capabilities were highlighted in regard to BC management.

T. Nazarov, P. Shcheplev, N. Naumov et al.

Introduction. Penile cancer (PC) is one of the rare tumors, accounting for 1-2% of all neoplasms of the genitourinary system. The clinical course of the cancer process raises many questions among both urologistsoncologists in hospitals and outpatient doctors. A retrospective analysis of articles indicates that this process has not been sufficiently studied and requires more clinical observations to determine the diagnosis and correct treatment tactics for such patients. Materials and methods. Clinical observations of the development of human cancer in patients are presented. An analysis of original articles was also carried out in the following databases: Pubmed, Scopus, Scopus, Web of Science from 2007 to 2022, dedicated to this disease. Results. When a tumor of the penis was detected, one patient underwent surgical intervention to the extent of amputation of the glans penis. In the second patient, total damage to the penis was revealed, which required more radical surgical treatment including: Emasculation with urethroplasty. The third patient underwent radiation therapy and subsequently circumcision of the foreskin. The postoperative period in patients was without complications. No relapse was observed. Conclusion. Penile cancer, despite its rarity, is a pressing problem in modern oncology. Even with proper examination and treatment, it is difficult to predict the further development of the disease. These clinical cases force specialists to examine patients more carefully so as not to miss pathology that can cripple the patient and change his future lifestyle, and also shows different options for treatment outcomes for penile cancer

Tiago Gonçalves, Dagoberto Pulido-Arias, Julian Willett et al.

The interactions between tumor cells and the tumor microenvironment (TME) dictate therapeutic efficacy of radiation and many systemic therapies in breast cancer. However, to date, there is not a widely available method to reproducibly measure tumor and immune phenotypes for each patient's tumor. Given this unmet clinical need, we applied multiple instance learning (MIL) algorithms to assess activity of ten biologically relevant pathways from the hematoxylin and eosin (H&E) slide of primary breast tumors. We employed different feature extraction approaches and state-of-the-art model architectures. Using binary classification, our models attained area under the receiver operating characteristic (AUROC) scores above 0.70 for nearly all gene expression pathways and on some cases, exceeded 0.80. Attention maps suggest that our trained models recognize biologically relevant spatial patterns of cell sub-populations from H&E. These efforts represent a first step towards developing computational H&E biomarkers that reflect facets of the TME and hold promise for augmenting precision oncology.

Amirehsan Ghasemi, Soheil Hashtarkhani, David L Schwartz et al.

With the advances in artificial intelligence (AI), data-driven algorithms are becoming increasingly popular in the medical domain. However, due to the nonlinear and complex behavior of many of these algorithms, decision-making by such algorithms is not trustworthy for clinicians and is considered a black-box process. Hence, the scientific community has introduced explainable artificial intelligence (XAI) to remedy the problem. This systematic scoping review investigates the application of XAI in breast cancer detection and risk prediction. We conducted a comprehensive search on Scopus, IEEE Explore, PubMed, and Google Scholar (first 50 citations) using a systematic search strategy. The search spanned from January 2017 to July 2023, focusing on peer-reviewed studies implementing XAI methods in breast cancer datasets. Thirty studies met our inclusion criteria and were included in the analysis. The results revealed that SHapley Additive exPlanations (SHAP) is the top model-agnostic XAI technique in breast cancer research in terms of usage, explaining the model prediction results, diagnosis and classification of biomarkers, and prognosis and survival analysis. Additionally, the SHAP model primarily explained tree-based ensemble machine learning models. The most common reason is that SHAP is model agnostic, which makes it both popular and useful for explaining any model prediction. Additionally, it is relatively easy to implement effectively and completely suits performant models, such as tree-based models. Explainable AI improves the transparency, interpretability, fairness, and trustworthiness of AI-enabled health systems and medical devices and, ultimately, the quality of care and outcomes.

Amit K Chattopadhyay, Aimee Pascaline N Unkundiye, Gillian Pearce et al.

The study explores Artificial Intelligence (AI) powered modeling to predict the evolution of cancer tumor cells in mice under different forms of treatment. The AI models are analyzed against varying ambient and systemic parameters, e.g. drug dosage, volume of the cancer cell mass, and time taken to destroy the cancer cell mass. The data required for the analysis have been synthetically extracted from plots available in both published and unpublished literature (primarily using a Matlab architecture called "Grabit"), that are then statistically standardized around the same baseline for comparison. Three forms of treatment are considered - saline (multiple concentrations used), magnetic nanoparticles (mNPs) and fluorodeoxyglycose iron oxide magnetic nanoparticles (mNP-FDGs) - analyzed using three Machine Learning (ML) algorithms, Decision Tree (DT), Random Forest (RF), Multilinear Regression (MLR), and a Deep Learning (DL) module, the Adaptive Neural Network (ANN). The AI models are trained on 60-80% data, the rest used for validation. Assessed over all three forms of treatment, ANN consistently outperforms other predictive models. Our models predict mNP-FDG as the most potent treatment regime that kills the cancerous tumor completely in ca 13 days from the start of treatment. The models can be generalized to other forms of cancer treatment regimens.

R. Kanishka

Treating cancer is one of the most challenging task in medical sciences. Only limited types of cancer treatments are available as their study is still ongoing. The earlier therapies like radiotherapy with x-rays, chemotherapy are associated with lot of side-effects. One of the most desirable cancer treatment is using particle beam therapies. These therapies are quite less risky than other types of cancer and tumor treatments. In this paper we present the proton and carbon-12 ion beam simulation that can help in tumor and cancer treatment. We simulated the proton and carbon-12 ion beam in water and soft tissue using geant4 toolkit. The protons are observed to have much better energy deposition in the water and soft tissue than carbon-12 ion and gamma photon beams.