Progression-free survival (PFS) is a critical end point in oncology, yet real-world applications of individualized, explainable machine learning (ML) predictions remain limited.

This study aimed to develop and validate explainable ML models to predict PFS using retrospective data from a national prostate cancer cohort in Brunei Darussalam.

We analyzed a retrospective cohort of 212 patients (478 longitudinal observations) treated at the Brunei Cancer Centre (January 2018 to December 2024). Clinical, laboratory, and treatment data were harmonized, with missing values imputed via extremely randomized trees. Longitudinal patterns were captured using a recurrent autoencoder to generate latent representations. We compared 4 modeling approaches: Cox proportional hazards, random survival forest (RSF), gradient boosting survival (GBS), and deep neural network survival models. Performance was evaluated using time-dependent area under the receiver operating characteristic curve (AUC), Harrell C-index, and integrated Brier score (IBS), with Shapley additive explanations (SHAP) used for interpretability.

RSF demonstrated improved discriminative performance and balanced calibration, achieving a C-index of 0.906 and AUCs of 0.941 and 0.917 at 4 and 5 years (IBS=0.0698). In contrast, the traditional CPH model performed poorly (C-index 0.531 and AUC 0.706 at 4 years and 0.833 at 5 years). Deep survival (AUCs of 0.941 at 4 years and 0.917 at 5 years, C-index 0.719, and IBS=0.0887) and GBS (AUCs of 0.765 at 4 years and 0.833 at 5 years, C-index 0.844, and IBS=0.0590) models showed moderate performance. SHAP analysis identified sodium, alanine aminotransferase, mean corpuscular hemoglobin, platelet count, and specific treatment categories as key drivers of increased progression risk.

Tree-based ensemble approaches, particularly RSF integrated with SHAP, offer high accuracy for personalized risk stratification in prostate cancer. These findings highlight the potential of explainable ML to enhance clinical decision-making. However, external validation in a larger multi-institutional, multiomics dataset is required before routine clinical implementation.

We evaluated the performance of two scoring systems-the Bladder Cancer Italian Active Surveillance (BIAS) criteria and the International Bladder Cancer Group (IBCG) score-for predicting high-grade (HG) disease at recurrence in patients with primary low-grade (LG) non-muscle-invasive bladder cancer eligible for active surveillance (AS).

We retrospectively included patients with primary LG-NMIBC (2013-2022), eligible for AS according to BIAS criteria at time of diagnosis. All subsequent recurrences were managed with TURBT, allowing histopathological confirmation of tumor grade across 383 recurrences. Clinical data necessary to assess IBCG criteria and adherence to BIAS protocol were systematically collected at each recurrence, and their diagnostic performance in predicting HG progression was evaluated using C-statistics. Multivariable logistic regression was used to assess which individual BIAS and IBCG criteria were most predictive of progression.

At a median (IQR) follow-up of 55 (18-83) months, HG progression occurred in 45/383 (12%) of recurrences. BIAS criteria were met in 59% (226/383) of recurrences, of which 5.3% showed upgrading to HG. Conversely, only 16% (62/383) of recurrences had 0 IBCG risk factors, of which 8 (13%) upgraded. BIAS showed higher discriminative ability than IBCG (C-statistics 0.68 vs. 0.47, p<0.001) and higher negative predictive value (95% vs. 87%).

In patients with primary LG-NMIBC initially eligible for AS, adherence to BIAS criteria at recurrence identifies a subgroup with a lower prevalence of HG disease at immediate TURBT and may help inform shared decision-making regarding surveillance versus immediate intervention.

To validate the MRI-based soft tissue tumor scoring system by comparing the accuracy of the Soft Tissue-RADS to qualitative assessment.

In this cross-sectional study, adult MRIs of soft-tissue tumors and tumor-like lesions with known reference standards, obtained from June 2015 to May 2024, were included. Patients with incomplete imaging, no final diagnosis, and visceral tumors were excluded. The scoring categories were 1-6. The cases were presented to ten expert musculoskeletal (MSK) radiologists from ten institutions under curated image-review conditions. Inter-reader agreement and MRI accuracy were calculated and compared to the accuracy of final qualitative diagnoses and respective confidence scales.

Among 150 MRI scans from 150 patients (68 males; 82 females) from one center, 6 cases were category 1; 60, category 2; 32, category 3; 14, category 4; 22, category 5; and 16, category 6. Good-to-excellent inter-reader agreement was observed overall (Kappa: 0.79; 95% CI: 0.74, 0.83). The mean accuracy of Soft-Tissue RADS between benign (categories 1-3) versus suspicious (categories 4 and 5) tumors was 82% (95% CI: 80%, 84%) versus 61% (95% CI: 58%, 63%) for qualitative diagnoses (p < 0.001). The mean true positive rate for RADS 6 was 97%. AUC for ten readers was 0.85 to 0.91. Average confidence level using Soft-Tissue RADS was greater than qualitative diagnoses (4.35 ± 0.81 vs. 4.14 ± 0.99; p < 0.001).

Soft Tissue-RADS is a reliable and accurate system for standardized terminology and recommended management of MRIs of soft-tissue tumors and tumor-like lesions among experts. Further validation in routine clinical workflows is necessary.

Question Given the limited accuracy of radiologists' rendered qualitative diagnoses, what is the diagnostic performance of Soft Tissue-RADS versus qualitative diagnoses? Findings Soft Tissue-RADS scoring is reliable (Kappa: 0.79; 95% CI: 0.74, 0.83) and more accurate (82% vs. 61%, p < 0.001) than qualitative diagnosis with better confidence levels (p < 0.001). Clinical relevance Soft Tissue-RADS system can potentially lead to consistent management recommendations and longitudinal data collection in the domain of sarcoma diagnosis.

Canine mammary tumor (CMT) is the most prevalent neoplasms in female dogs and frequently undergo malignant progression and metastasis, leading to poor clinical outcomes and highlighting a crucial need for effective therapies beyond surgery. The Porcine Recombinant NK-lysin (prNK-lysin), a recombinant peptide produced from our lab, has previously shown inhibitory effects on orthotopic murine breast cancer growth and lung metastasis. However, its efficacy against CMT remains unclear. In this study, the CMT cell lines (CHMp and CHMm) and Orthotopic CMT model with CHMp in BALB/c nude mice were employed to investigate the anticancer effects of prNK-lysin. Mechanistic studies were performed using quantitative proteomics and pull-down assays, followed by PPI network analysis, molecular docking, CETSA, SPR, and gene overexpression. Our results revealed that prNK-lysin inhibits the proliferation, migration, adhesion, and invasion of both CHMp and CHMm cell lines and suppresses the tumor growth of CMT model. Mechanistically, Elongation Factor 1-Alpha (EEF1A1) and T-Complex Protein 1 Subunit Gamma (CCT3) were identified as high-affinity direct targets of prNK-lysin. The direct engagement of prNK-lysin with these two targets resulted in the perturbation of the MEK/ERK/ROCK2 signaling axis, the subsequent downregulation of a core set of actin-binding regulatory proteins, ultimately leading to cytoskeletal collapse and inhibition of CMT progression. Our findings strongly indicate that prNK-lysin is a promising therapeutic candidate for CMT by dismantling the cytoskeleton.

Neoplasms greatly increase the risk of infection-the main causes of death are pneumonia and influenza infection. Streptococcus pneumoniae, the most common pathogen causing community-acquired pneumonia, is of great significance in patients suffering from hematological neoplasms, which have a 30 times higher incidence of invasive pneumococcal disease than the general population. Vaccination plays a key role in prevention. The Standing Committee on Vaccination (STIKO) recommends vaccination for adults over 60 and immunosuppressed adults. Only about 20% of these individuals are vaccinated.

Relevance of pneumococcal vaccination in cancer patients and challenges in practical implementation.

Case study of the clinical course of a hematologic patient with recurrent pneumococcal infections.

Invasive pneumococcal diseases are associated with high morbidity and mortality in cancer patients and are potentially preventable by vaccination. There are adapted recommendations for this group.

Increased awareness of the risk of pneumococcal disease in cancer patients can help to increase vaccination rates. Coordination between health care sectors is essential in order to implement the recommendations.

Gastrin regulates gastric acid secretion and mucosal homeostasis via the cholecystokinin-2 receptor (CCK2R). Aberrant signaling via CCK2R (cholecystokinin family) and the distinct class A GPCR, gastrin-releasing peptide receptor (GRPR; bombesin family), drives oncogenesis across diverse GI and non-GI malignancies through calcium mobilization, MAPK/PI3K activation, EGFR transactivation, and constitutive splice variants like CCK2Ri4sv. Furthermore, precursor forms such as progastrin remodel the tumor microenvironment by interacting with non-CCK binding sites (p78, Annexin A2), converting fibroblasts into pro-invasive CAFs and sustaining cancer stem cells. Chronic hypergastrinemia, induced by Helicobacter pylori infection, autoimmune gastritis, or long-term proton pump inhibitor (PPI) therapy, drives enterochromaffin-like (ECL) cell hyperplasia, promoting gastric neuroendocrine tumors and diffuse-type carcinomas. Beyond gastroenteropancreatic physiology, selective expression of CCK2R on ECL and thyroid C-cells, combined with widespread ectopic GRPR overexpression in solid tumors, establishes versatile platforms for molecular imaging and radiotheranostics. Radiolabeled peptide antagonists and stabilized minigastrin/bombesin analogs conjugated to diagnostic (⁶⁸Ga, ⁶⁴Cu) or therapeutic radionuclides (¹⁷⁷Lu,²²⁵Ac, ²¹²Pb) enable high-resolution PET imaging and targeted radionuclide therapy (PRRT/TAT). These strategies demonstrate high diagnostic accuracy and localized cytotoxicity in medullary thyroid carcinoma, small cell lung cancer, prostate neoplasms, and ER+ breast cancers. Concurrently, novel circulating biomarkers, including multi-genomic liquid biopsies (NETest®) and ProGRP, significantly improve non-invasive disease tracking. Unraveling these receptor pathways bridges tumor biology with translational medicine, offering transformative potential for personalized diagnosis, risk stratification, and targeted radionuclide therapy.

Pulmonary metastases from pleomorphic adenoma (PA)-spectrum salivary gland neoplasms are rare and may be difficult to diagnose, particularly in limited pulmonary specimens. We aimed to characterize their clinicopathologic and molecular features and highlight practical diagnostic pitfalls.

We conducted a retrospective single-institution study of three patients with pathologically confirmed pulmonary metastases from PA-spectrum salivary gland neoplasms. Clinical history, histomorphology, immunophenotype, and PLAG1 break-apart fluorescence in situ hybridization findings were reviewed. Targeted next-generation sequencing was performed on the two metastasizing pleomorphic adenoma (MPA) lesions, and a focused literature-based comparison was undertaken.

Two pulmonary lesions were classified as histologically bland MPA, whereas the third was classified as metastatic carcinoma ex pleomorphic adenoma (CXPA) with a myoepithelial carcinoma component. In one MPA, apparent focal TTF-1 and Napsin A immunoreactivity in a core biopsy was attributable predominantly to entrapped pneumocytes rather than convincing expression in the neoplastic component, creating a potential pitfall for misdiagnosis as primary pulmonary adenocarcinoma. PLAG1 rearrangement was detected in all three pulmonary lesions. Targeted sequencing identified an ARID1A loss-of-function alteration and a TSHR variant considered potentially of germline origin in one MPA; the sequencing findings in the other MPA were of uncertain clinical significance.

Pulmonary PA-lineage metastases may be histologically bland or overtly malignant. Entrapped pneumocytes may cause misleading TTF-1 and Napsin A immunoreactivity in limited pulmonary specimens. PLAG1 FISH provides useful support for PA-lineage origin but does not independently distinguish MPA from metastatic CXPA. Integrated assessment of clinical history, morphology, immunophenotype, and molecular findings is essential for accurate classification.

Myeloproliferative neoplasms (MPNs) are clonal hematopoietic neoplasms characterized by a substantial risk of arterial and venous thrombosis, which remains a major cause of disease-related morbidity and mortality. Beyond abnormalities in blood-cell counts and rheology, MPN-associated thrombosis is increasingly recognized as a consequence of chronic thromboinflammation arising from dysregulated interactions among leukocytes, platelets, endothelial cells, inflammatory mediators, and the coagulation system. The canonical driver mutations JAK2, CALR, and MPL converge on constitutive JAK-STAT signaling but are associated with differential thrombotic risk, with JAK2V617F being most consistently linked to an increased risk of thrombosis. Growing evidence suggests that thrombotic complications in MPNs are not solely attributable to clonal expansion and quantitative hematologic abnormalities but also involve qualitative alterations in immune-cell function, platelet activation, endothelial dysfunction, inflammatory signaling, and coagulation. This review summarizes the current mechanistic and clinical evidence linking MPN driver mutations and thromboinflammatory processes to thrombosis and discusses the effects of disease-directed therapies on these interconnected pathways.

KRAS-mutant lung adenocarcinoma (LUAD) shows limited benefit from immune checkpoint blockade (ICB), and the mechanisms by which oncogenic KRAS signaling promotes immune resistance remain poorly understood. Here, we sought to identify tumor-intrinsic mediators of KRAS-driven immune evasion and therapeutic resistance in LUAD.

Integrated analyses of human LUAD cohorts, genetically engineered mouse models, tumor cell models, and patient-derived organoids were performed to define the role of PRTN3 in immune regulation. PRTN3 regulation and function were investigated using transcriptional analyses, chromatin immunoprecipitation assays, biochemical cleavage assays, and immune cell co-culture systems. The therapeutic effect of PRTN3 inhibition was evaluated using flow cytometric immune profiling, functional immune assays, and syngeneic tumor models.

We identified PRTN3 as a downstream effector of KRAS signaling that promotes immune evasion in LUAD. Mechanistically, KRAS signaling transcriptionally upregulated PRTN3 through c-Myc activation. Secreted PRTN3 directly cleaved and inactivated CXCL9 in the tumor microenvironment, disrupting CXCL9/CXCR3 signaling and impairing the recruitment and activation of cytotoxic CD8⁺ T cells and natural killer cells. In patients with LUAD, elevated PRTN3 expression was associated with reduced cytotoxic immune signatures and decreased CXCL9 activity. Genetic inhibition of PRTN3 enhanced antitumor immunity and improved responses to anti-programmed cell death protein 1 therapy in vivo. Furthermore, a hydrolysis-resistant CXCL9 variant restored cytotoxic lymphocyte activation and enhanced immune-mediated tumor cell killing in patient-derived organoids.

Our findings reveal a KRAS/c-Myc/PRTN3 axis that promotes immune suppression through extracellular proteolytic inactivation of CXCL9. Targeting PRTN3-mediated immune evasion may represent a promising strategy to enhance ICB efficacy in KRAS-driven LUAD.

A woman in her 70s presented with severe abdominal pain following a mechanical fall. A CT of the abdomen revealed a massive cystic lesion with evidence of rupture, while MRI localised the lesion to the left retroperitoneum; however, its aetiology remained unclear. She underwent laparotomy with extensive adhesiolysis and excision of a large multiloculated retroperitoneal cyst containing approximately 6500 mL of serous and haemorrhagic fluid. Histopathological examination demonstrated a benign serous cystadenoma lined by ciliated serous epithelium with tubal metaplasia and no evidence of atypia. This case highlights the diagnostic challenges associated with primary retroperitoneal serous cystadenomas and emphasises the importance of surgical excision and histopathological assessment for definitive diagnosis.