Mastroleo, Federico

Background Stereotactic body radiotherapy (SBRT) is increasingly used for patients with oligometastatic prostate cancer, yet the optimal target definition, dose, and toxicity profile for rib metastases remain uncertain. Objectives We assessed local control (LC), toxicity, and the risk of radiation-induced rib fracture (RF) after SBRT for rib metastases. Design Retrospective multicenter cohort study. Methods We analyzed prostate cancer patients with radiologically diagnosed rib metastases treated with SBRT and >=12 months of follow-up. Outcomes were descriptively compared between gross tumor volume (GTV)-only and GTV plus an intraosseous clinical target volume (CTV) margin. Analyses included Kaplan-Meier survival estimates, cumulative incidence functions, and exploratory univariable Cox and logistic regression; no formal equivalence testing or multivariable validation was attempted because event numbers were very low. Results We included 102 patients with a median follow-up of 20 months. The 1- and 2-year LC rates were 97.9% and 96.3%, respectively. At 3 years, local progression was 2.1% in the GTV-only group, absent in patients treated with >=10 mm margins, and numerically highest with 3-5 mm CTV margins (13.5%, p=0.022). Among 13 symptomatic patients, pain control was excellent, with complete or partial response in all patients within 6 months. No grade >=3 toxicities occurred. RFs were rare (n=3, 2.9%), comprising one symptomatic Grade 2 event (0.9%) and two asymptomatic Grade 1 events (2.0%). Baseline symptoms and extraosseous infiltration were associated with RF in exploratory univariable analysis, but confidence intervals were wide. Higher PTV dose exposure, particularly PTV D50% (36.9 vs. 31.1 Gy, p=0.041), was also observed among RF cases. Conclusions SBRT for rib metastases from prostate cancer was associated with high LC, pain relief, and low observed toxicity in this retrospective real-world cohort. GTV-only contouring may be feasible in carefully selected patients staged and planned with PSMA-PET/CT, but the non-randomized design, baseline imbalances, short follow-up, and very low event count preclude drawing conclusions about equivalence or independent validation. RF predictor analyses should be considered exploratory and hypothesis-generating.

Background Interventional radiotherapy (IRT), also known as brachytherapy, has become a cornerstone of radiation oncology due to its ability to maximize therapeutic efficacy while minimizing side effects to surrounding tissues. This bibliometric analysis aims to examine the evolution of IRT over the past decade. Material and Methods A string comprising IRT-related keywords was developed, and data were extracted from the Scopus database for English-language documents published between January 2014 and December 2024. The analysis was performed using Bibliometrix package Results A total of 15 755 documents from 2150 sources were analyzed, revealing an average annual growth rate of 1.03%. Publication trends were stable from 2014 to 2018, surged from 2019, and peaked in 2021, with subsequent levels remaining above pre-2019 figures. Original articles accounted for 72.7% of the literature, with key contributions from journals such as “Brachytherapy” and the “Journal of Contemporary Brachytherapy”. The analysis highlighted strong collaborative networks, with significant input from North American and European institutions, and identified focal research themes including uterine cervical cancer and prostate brachytherapy. Emerging topics, such as deep learning, artificial intelligence, and personalized dosimetry, underscore the field's shift towards precision medicine. Conclusions The findings reflect IRT's dynamic research landscape and its growing integration into modern oncologic practice, emphasizing the need for continued interdisciplinary efforts and clinical trials to advance this minimally invasive treatment modality further. Highlights • A decade-long bibliometric analysis of 15 755 publications on interventional radiotherapy (IRT) reveals stable growth, with a notable surge in research activity post-2019. • Uterine, cervical, and prostate cancers remain the most studied clinical indications for IRT globally. • North American and European institutions dominate collaborative networks, reflecting the strength of transatlantic research partnerships. • Emerging research themes include artificial intelligence, deep learning, and personalized dosimetry—marking a transition toward precision oncology. • The field's evolving landscape highlights the increasing clinical and scientific significance of IRT in contemporary cancer care.

Purpose: To evaluate the impact of PSMA PET versus conventional imaging for staging, and the effect of dose escalation (DE) to metastatic lymph nodes in cN1 prostate cancer patients treated with EBRT and ADT. Methods: A retrospective multicenter analysis included 197 propensity score-matched patients (77 staged with PSMA PET, 120 with conventional imaging) treated with EBRT and ADT, with or without DE to lymph node metastases. Matching criteria were T stage, nodal involvement, imaging modality, Gleason score, and ADT duration. Primary endpoint was metastasis-free survival (MFS); secondary endpoints included overall survival (OS), locoregional control (LRC), and toxicity. Results: Median follow-up was 42 months (IQR: 25–59). Median EQD2 (alpha/beta=1.5) to elective nodes was 46 Gy, and to metastatic nodes, 66 Gy (PSMA PET) and 63 Gy (conventional imaging). The 3-year MFS was 80.5% (95% CI: 71.3–90.8%) for PSMA PET and 77.2% (95% CI: 69.5–85.8%) for conventional imaging; 5-year MFS was 59.3% versus 65.8% ( P =0.57). OS at 5 years was 82.2% (PSMA PET) and 77.4% (conventional imaging) ( P =0.72). On multivariable analysis, ISUP grade group 3–5 predicted earlier metastasis (HR: 4.03, 95% CI: 1.61–10.06; P =0.003); higher prostate dose (EQD2 ≥113 Gy) improved MFS (HR: 0.53, 95% CI: 0.30–0.95; P =0.033); long-term ADT (≥24 mo) improved OS (HR: 0.38, 95% CI: 0.19–0.78; P =0.009). No significant difference in toxicity was observed between groups. Conclusions: PSMA PET staging and nodal dose escalation >60 Gy were safe but did not improve MFS, OS, or LRC. ISUP grade group 3–5 and higher prostate dose significantly influenced outcomes.

Retrieval-augmented generation (RAG) is rapidly emerging as a transformative paradigm for large language models (LLMs), especially in high-stakes domains like oncology that demand precision, factual grounding, and up-to-date knowledge. By pairing LLMs with external knowledge repositories, RAG systems explicitly ground model outputs in relevant retrieved documents, helping to reduce hallucinations and ensure responses reflect current evidence. In oncology, where clinical knowledge evolves continually with new research and drug approvals, RAG offers a way to integrate the latest data (e.g., trial results, guidelines, genomic databases) into decision-making. This review synthesizes the technical foundations of RAG, including its architecture and key components, and examines current applications in oncology such as clinical decision support, patient education, radiology reporting, pathology analysis, and genomics-driven precision medicine. We highlight recent studies that demonstrate RAG’s potential—for instance, improving treatment recommendations by incorporating genetic profiles and literature, and enhancing diagnostic accuracy by integrating guidelines. We also discuss emerging developments like multimodal RAG (combining text with imaging or other data), ensemble model approaches, and new explainability tools that trace model outputs to sources. Finally, we critically analyze the limitations and challenges of deploying RAG in healthcare, including computational costs, retrieval errors, noise or conflicts in retrieved information, and ethical and regulatory considerations. While RAG-based systems show promise in augmenting oncologists’ expertise with timely knowledge, careful implementation, high-quality curation of knowledge bases, and human oversight will be crucial for safe and effective adoption in clinical practice.

PURPOSE Cancer recurrence is a critical outcome for patients and physicians. Retrospective cancer recurrence data can evaluate recurrence-directed treatment and generate novel interventions targeting recurrent cancers. However, large cancer databases do not provide recurrence-related information, stymying study at scale, and consequently require significant manual record review. Automated evaluation of records may allow for the rapid generation of easily analyzed data sets, accelerating the evaluation of recurrence altogether. METHODS Patients treated with radiation therapy at one tertiary referral center from 2010 to 2018 with a verified cancer status (cancer recurrence v no cancer recurrence) were identified. Patients with recurrent disease were initially identified through manual record review, and the associated pathology report was collected. Google Automated Machine Learning with Natural Language Processing (AutoNLP) and Google Gemini 1.5 Pro were used to generate a model for binary classification, with comparison to the gold-standard manually developed data set. RESULTS A total of 7,054 patients were identified. 3,431 (48.6%) were female, with a median age of 64 years. Head and neck (1,482, 21%), breast (1,480, 21%), upper GI (1,307, 18.5%), and lung/thorax (1,107, 15.7%) were the most common disease sites. Recurrence was verified for 1,546 patients (21.9%) using pathology reports, of which 1,249 positive cases were paired with 651 negative pathology reports for model development. Google Gemini 1.5 Pro consistently outperformed AutoNLP across all measurements of accuracy, generating a greater absolute difference in precision, recall, negative predictive value, and specificity, and a higher likelihood of correct classification at the individual level, rendering Gemini superior in recurrence status extraction. CONCLUSION AutoNLP and Google Gemini 1.5 Pro are promising tools for identifying recurrence from pathology reports, with the latter demonstrating superior overall performance, making it particularly suitable for clinical translation.

357 Background: Prostate cancer (PCa) is the most incident cancer in adult males in the United States, and the second highest in the world. CTCAE is the standardized method to grade the severity of treatment-related adverse events (AEs), but are tedious to collect and subject to inter-observer variability. This study aimed to evaluate the performance of LLMs in automatically extracting and grading CTCAE toxicities from clinical notes and patient-reported outcomes of PCa patients from a clinical trial. Methods: A total of 55 patients from NCT02874014 trial, undergoing proton therapy (PT) for localized PCa, were included. The ground truth of CTCAE graded toxicities was obtained from the trial’s results. Clinical notes within 3 days, and the most recent patient questionnaires within 90 days of toxicity assessment were retrieved from the electronic health record. A comprehensive prompt was developed to extract and CTCAE grade toxicities from the retrieved records. LLMs used for this analysis were OpenAI GPT-4o and Gemini 2.0 Flash. Phase 1 evaluated both LLMs' performance against the ground truth. To ensure the metrics reflected the models’ performance based only on the information available in the text provided, cases where both LLMs contradicted the ground truth were manually reviewed. Phase 2 recalculated LLMs' performance metrics using this adapted ground truth. A conservative stepwise ensemble model was developed to leverage the complementary strengths of both LLMs. Results: In the conservative stepwise ensemble model, Gemini 2.0 Flash served as the initial screening platform to identify AEs in step 1. All cases with identified toxicity events were then reviewed and confirmed using OpenAI GPT-4o in step 2. This effectively reduced false positive rates while preserving high sensitivity. The ensemble model achieved strong performance metrics: binary accuracy (0.954), sensitivity (0.961), specificity (0.953), F1 score (0.836), and grade accuracy (0.932). Conclusions: This study demonstrates the feasibility of the use of LLMs for extraction and CTCAE toxicity grading in PCa patients treated with RT. The ensemble model achieved strong performance metrics. Further validation is needed for other cancer diagnosis and treatment modalities. Phase Model Binary Accuracy Precision Sensitivity Specificity F1 Score Grade Accuracy 1 Gemini 0.946 0.672 1.000 0.939 0.804 0.919 1 OpenAI 0.940 0.785 0.637 0.978 0.703 0.895 2 Gemini 0.956 0.735 1.000 0.950 0.848 0.927 2 OpenAI 0.954 0.917 0.679 0.991 0.780 0.903