To correlate response evaluation after neoadjuvant chemotherapy (NAC), assessed by magnetic resonance imaging (MRI) and pathology, with disease-free survival (DFS) in breast cancer patients, according to immunophenotype.

Single-center, IRB-approved retrospective cohort study included consecutive breast cancer patients who underwent NAC and preoperative breast MRI. Pathologic response was evaluated using the residual cancer burden (RCB) system, with pathological complete response (pCR) defined as the absence of invasive carcinoma. Radiological complete response (rCR) was defined as the absence of abnormal enhancement on MRI. The Kaplan-Meier method estimated DFS and Cox regression analysis calculated hazard ratios (HR).

571 patients were included (mean age 46 years, range 26-90). The most common immunophenotype was Luminal (42.3%), followed by triple-negative (TNBC, 31.5%) and HER2-overexpressed (26.3%). Radiological and pathological responses were concordant in 71.5%. Overall, 35.2% achieved rCR and 37.5% achieved pCR. DFS curves did not differ significantly according to radiologic-pathologic response combinations in Luminal or HER2 groups (LogRank p = 0.505 and p = 0.257). In the TNBC group, patients without pCR or rCR had significantly worse DFS compared to those achieving either response (LogRank p = 0.001). Cox regression revealed that TNBC patients with both non-rCR and non-pCR had a markedly higher risk of recurrence or death (HR 7.728; 95%CI 2.696-22.149; p < 0.001).

Integrating MRI and pathological response assessments after NAC may enhance risk stratification and prognostication, especially in triple-negative breast cancer.

Patients with both non-rCR and non-pCR have significantly worse DFS, underscoring the prognostic value of combining imaging and pathological findings, particularly in TNBC.

A 20-year-old neutered male ring-tailed lemur (Lemur catta) diagnosed with an intrathoracic mass died during surgical excision. Morphological characteristics of the mass during necropsy were consistent with a mediastinal abscess in the left hemithorax. Histological examination identified a disseminated histiocytic sarcoma involving spleen, kidneys, multiple lymph nodes, diaphragm and parietal pleura, as well as a concurrent follicular lymphoma in a separate lymph node. A mixed metastatic infiltrate composed of both neoplasms was found within the capsule of the intrathoracic abscess. Immunohistochemical analysis (Iba-1, CD3, PAX-5, IRF-4, pancytokeratin AE1/AE3 and Ki-67) confirmed both neoplasms and revealed mild IRF-4 immunoreactivity for histiocytic sarcoma cells. The histomorphologic and immunohistochemical features of the histiocytic sarcoma in this Lemur catta were characteristic of this condition in other species. This report documents the first case of histiocytic sarcoma in Lemur catta, providing novel histopathological and immunohistochemical insights of this condition in prosimians.

In computed tomography (CT)-guided cervical cancer brachytherapy, the manual contouring for the high-risk clinical target volume (HR-CTV) is a time-consuming and expertise-dependent process. Furthermore, automated approaches struggle with ambiguous boundaries of HR-CTV.

We aimed to develop a clinically efficient interactive segmentation framework integrating deep learning with clinician expertise.

We propose a slice-prompted interactive segmentation method (SPSeg) for HR-CTV delineation in CT-guided cervical cancer brachytherapy. Clinicians provided sparse prompts by manually outlining HR-CTV on key slices, which were then encoded into a 3D U-Net architecture to guide full-volume segmentation. We investigated two architectural variants: SPSeg-Mono, which jointly processes the CT images and the prompt masks with a single encoder; and SPSeg-Dual, which employs two separate encoders for image and prompt, fusing their features at a deeper level. The model was trained on 640 CT scans (from 160 patients) and validated on 160 scans (40 patients) from a single center, and externally tested on three multi-center cohorts: 400 scans (100 patients), 115 scans (40 patients), and 150 scans (30 patients), respectively. Evaluation included Dice Similarity Coefficient (DSC), 95% Hausdorff Distance (HD95), a 5-point Likert scale for clinical acceptability, time efficiency, and inter-observer agreement.

Performance consistently improved with the addition of prompt slices, with SPSeg-Dual outperforming SPSeg-Mono. Without prompts, the model yielded DSCs of 0.83, 0.76, and 0.76, and HD95s of 7.5, 10.1, and 11.6 mm for Test Sets 1, 2, and 3, respectively. With the addition of just three prompt slices, DSCs increased significantly to 0.95, 0.92, and 0.91, while HD95s decreased to 2.1, 3.1, and 3.2 mm, respectively (all p < 0.001). Qualitative scores confirmed high clinical acceptability (mean Likert scores > 3), and the interactive method substantially reduced contouring time for both clinicians (from 11.7 to 1.7 min for Clinician A, and from 9.9 to 1.5 min for Clinician B). It also improved inter-observer agreement, with DSC increasing from 0.88 to 0.93 and HD95 decreasing from 3.2 to 2.5 mm (p < 0.001).

The proposed SPSeg method effectively integrates clinical expertise with deep learning, offering a highly precise and efficient solution for HR-CTV delineation in cervical cancer brachytherapy.

Segmentation is the most effort-consuming step for magnetic resonance imaging guided adaptive radiotherapy (MRIgART). Although the segment anything model (SAM) exhibits impressive capabilities, its application in medical imaging necessitates clicks, bounding boxes, or providing mask prompts on each target image, which would still require complex human interactions.

This study introduces SAM-ART, a large vision model that integrates personalized information to enhance the segmentation accuracy of MRIgART.

This study utilized planning computed tomography (pCT), approved contours, and daily MRI (dMRI) from 38 patients with prostate cancer and 10 patients with rectal cancer. SAM-ART comprises an image encoder, a prompt encoder, and a mask decoder. Using mask and box prompts, SAM-ART propagates contours from pCT to dMRI using deformable image registration (DIR) and employs them as mask prompts, providing patient-specific information. The box prompts are used in slices prone to false negative (FN) predictions. A 5-fold cross-validation was then conducted, comparing SAM-ART with DIR, traditional deep learning (tDL), and SAM-ART using other manual prompts (point or box).

The proposed SAM-ART exhibited a mean dice similarity coefficient of 0.934 ± 0.023 for the regions of interest, surpassing DIR (0.873 ± 0.063) and tDL (0.887 ± 0.056). Moreover, the proposed mask/box prompts also outperformed the other modes (point: 0.910 ± 0.027, and box: 0.921 ± 0.025). Mask/box prompts effectively mitigated FN predictions with minimal manual intervention. The ratio of acceptable slices (using the criteria of dice ≥ 0.85, 95th percentile of Hausdorff distance ≤ 5 mm, and mean distance to agreement ≤ 1.5 mm) was 89.38% with the proposed method, which means that the segmentations on about 90% of the slices did not require manual modification.

This study proposed a novel method that integrates personalized information and manual prompts into a SAM-based segmentation model. It outperformed the baseline methods, with only a few contours needing to be revised for clinical use.

Accurate histopathological and molecular characterization of lung cancer is essential for optimal treatment selection in the era of precision medicine. While conventional biopsy techniques are widely available and safe, they often yield small tissue samples with crush artifacts that may be insufficient for comprehensive molecular testing. Cryobiopsy has emerged as a promising diagnostic technique that addresses these limitations. For endobronchial and peripheral pulmonary lesions, cryobiopsy demonstrates superior diagnostic yields compared to conventional forceps biopsy, with enhanced capability for molecular diagnostics and programmed death-ligand 1 assessment. Endobronchial ultrasound (EBUS)-guided mediastinal cryobiopsy shows particular promise for lymph node sampling, achieving higher diagnostic yields than EBUS-guided transbronchial needle aspiration, especially for lymphoma and metastatic disease. Moreover, thoracoscopic cryobiopsy provides larger pleural specimens with preserved architecture, improving diagnosis of challenging cases including malignant mesothelioma. The safety profile remains favorable across all applications, with bleeding as the primary complication that is typically manageable with standard techniques. Cryobiopsy represents a significant advancement in thoracic oncology diagnostics, providing high-quality tissue specimens essential for contemporary cancer management.

Volumetric modulated arc therapy (VMAT) machine parameter optimization (MPO) is a complex, high-dimensional problem typically solved with inverse planning solutions that are both temporally and computationally expensive. While machine learning techniques have been explored to automate this process, they often supplement rather than replace conventional optimizers and are fundamentally limited by the quality and diversity of training data. Reinforcement learning (RL) offers a promising alternative, finding optimal strategies through trial-and-error by maximizing a narrowly tailored reward function, which can potentially discover novel solutions beyond mimicking features present in existing plans.

The purpose of this study was to develop and validate a deep reinforcement learning-based VMAT MPO algorithm capable of automatically generating clinically comparable treatment plans for prostate cancer that meet machine constraints, entirely independent of a commercial treatment planning system (TPS) optimizer.

A dataset comprised of 100 prostate cancer patients planned using the criteria from PACE-B SBRT arm serve as the basis for network training using a 70-10-20 training/validation/testing split. An RL framework using a Proximal Policy Optimization (PPO) algorithm was developed to train two tandem convolutional neural networks that sequentially optimize multi-leaf collimator (MLC) positions and monitor units (MUs) using current dose, contoured structure masks, and current machine parameters as inputs. Training was designed to predict MLC positions and MUs that maximize a dose-volume histogram (DVH)-based reward function tailored to prioritize meeting clinical objectives. The fully trained networks were executed on a test set of 20 patients and compared to reference plans optimized with a commercial TPS.

The RL algorithm generated plans in an average of 6.3 ± 4.7 s. Compared to the reference plans, the RL-generated plans demonstrated improved sparing for both the bladder and rectum across their respective dosimetric endpoints. When normalizing to 95% coverage, the RL generated plans resulted in a statistically significant increase in the PTV D 2 % ${{D}_{2\% }}$ , while achieving a significantly reduced D m e a n ${{D}_{mean}}$ for the rectum. All RL plans successfully satisfied all clinical objectives used to optimize the reference plans.

We successfully developed and validated a deep RL framework for VMAT MPO. The algorithm rapidly generates VMAT prostate cancer treatment plans that meet clinical constraints and are dosimetrically comparable to manually optimized plans without the use of a commercial TPS optimizer. This work demonstrates the feasibility of RL as a tool to fully automate the VMAT planning process, offering the potential to decrease planning times while maintaining plan quality.

Colorectal cancer (CRC) is the third most commonly diagnosed malignancy worldwide and remains a major challenge in contemporary oncology, where early detection is critical for improving treatment outcomes and survival. Despite significant progress in diagnostics and therapy, the epidemiology, risk factors, and molecular mechanisms driving CRC development continue to be intensively investigated. This paper provides an overview of current trends in CRC diagnosis and management, with particular emphasis on advances in molecular medicine and biological sciences. Screening recommendations in Poland are discussed, comparing invasive methods-such as colonoscopy, sigmoidoscopy, and CT colonography-with non-invasive stool-based tests (FOBT, FIT, sDNA-FIT), and evaluating their sensitivity, specificity, and impact on mortality reduction. Key tumor markers with diagnostic, prognostic, and predictive value, including CEA, CA19-9, mSEPT9, ctDNA, TPS, TAG-72, CTCs, and circulating microRNAs, as well as p53 and PTEN proteins, are reviewed in the context of their clinical utility in early detection, disease monitoring, and treatment response assessment. The analysis also highlights the epidemiological situation in Poland and underscores the growing importance of integrating molecular biomarkers with traditional diagnostic methods, which may ultimately support the development of more precise and individualized clinical management strategies in the future.

Artificial intelligence, particularly machine learning, has great potential to improve health outcomes, including predicting adverse conditions. In breast cancer, machine learning models can help personalize prevention strategies for radiation-induced cutaneous toxicity.

This scoping review aimed to explore machine learning models for predicting radiation dermatitis in women with breast cancer. Data collection was conducted in November 2023 from 7 electronic databases and gray literature, with no restrictions on publication year. Publication selection was supported by the RAYYAN reference manager, and ResearchRabbit software expanded the search.

A total of 22 publications were included. The reviewed models primarily predicted acute radiation dermatitis using clinical predictors. Most studies used cross-validation, and class imbalance was observed. The predominant models were developed using the Random Forest algorithm, with the Bayesian Network emerging as the top-performing model, incorporating clinical, clinicopathological, demographic, radiomic, and dosimetric predictors.

This review underscores the importance of further investigation into multiomic biomarkers and the establishment of minimum nursing databases to support predictive model development for radiation dermatitis in breast cancer patients.

Aromatase inhibitors (AIs) are vital in the treatment of estrogen-dependent breast cancer, especially in postmenopausal women. In this study, a series of steroidal glycosides (SGs) derived from trans-androsterone (tAND), estrone (E1), and estradiol (E2) were synthesized using a one-pot multi-enzyme glycosylation approach and structurally characterized via HPLC, MS, and NMR. Among the synthesized compounds, E1-α-2DG (2b) and E2-α-2DG (3b) demonstrated the most potent aromatase inhibition, with IC₅₀ values of 0.101 ± 0.001 μM and 0.159 ± 0.009 μM, respectively. Molecular docking revealed that these glycosides form key hydrogen bonds with catalytic residues and the heme group of CYP19A1. In vitro cytotoxicity showed that E1-α-2DG selectively inhibited the growth of MCF-7 and MDA-MB-231 breast cancer cells in a dose-dependent manner, with the highest potency observed against triple-negative MDA-MB-231 cells (IC₅₀ = 20.46 ± 2.92 μM), while exhibiting no toxicity toward non-cancerous HEK293 cells. These findings suggest that glycosylation enhances the pharmacological potential of steroidal scaffolds and highlights E1-α-2DG as a promising lead compound for the development of safer, dual-function breast cancer therapies.

Statins, traditionally used for managing hypercholesterolemia, have emerged as promising agents for cancer therapy. By targeting the mevalonate pathway-a cornerstone of cellular metabolism and tumorigenesis-statins disrupt critical processes for cancer cell survival and proliferation. Some of these processes include cholesterol biosynthesis, protein prenylation, and post-translational modifications. This review discusses repurposing statins for cancer treatment given their anti-tumoral effects across many cancers, including breast, prostate, colorectal and hepatocellular carcinoma. Despite statins' ability to induce apoptosis or autophagy, arrest cell cycle, or modulate favorable epigenetic reprogramming, their efficacy is highly context-dependent, influenced by cancer type, molecular subtype and genetic variations. Challenges such as statin resistance, low bioavailability and pharmacokinetic variability further complicate their application in oncology. Nonetheless, emerging strategies, including nanoparticle-based drug delivery systems and combination therapies with chemotherapy, radiotherapy or immunotherapy, appear to help overcome these limitations. Despite encouraging preclinical findings, clinical evidence remains tantalizingly inconsistent. Future research should prioritize identifying biomarkers of statin sensitivity and optimizing nanoformulations to enhance tumor targeting while minimizing toxicity. Ultimately, statins represent an attractive opportunity to expand the anti-tumor armamentarium and highlight innovative treatment paradigms integrating metabolic modulation to precision oncology.