Oncology最新文献

Introduction: Atezolizumab plus bevacizumab (ATZ/BEV) is used for the treatment of unresectable hepatocellular carcinoma (u-HCC). However, there are few reports investigating the predictors of whether or not disease control can be achieved for long periods.

Methods: We enrolled 106 u-HCC patients who were treated with ATZ/BEV as the first-line systemic chemotherapy and were evaluated as nonprogressive disease (non-PD) in their initial assessments. They were divided into two groups: patients who had achieved disease control (non-PD) for more than 6 months, Group-L, and patients who had progressed (PD) within 6 months, Group-S. We investigated the predictors of long-term therapeutic efficacy.

Results: Patients in Group-L had significantly lower neutrophil-to-lymphocyte ratios (NLR) (the ratios of NLR at the start of the second course of ATZ/BEV to their pretreatment levels) than Group-S (p = 0.044), and the optimal cutoff value was 1.129. The patients with an NLR ratio <1.129 had longer progression-free survival (PFS) and overall survival (OS) than those with an NLR ratio ≥1.129 (median: 16 vs. 6 months, p = 0.016 for PFS; not reached vs. 22 months, p = 0.007 for OS). Univariate and multivariate analyses showed that an NLR ratio ≥1.129 was an independent predictor for unfavorable PFS and OS (hazard ratio [HR] 2.03, p = 0.022 for PFS; HR 2.37, p = 0.019 for OS).

Conclusions: In u-HCC patients treated with ATZ/BEV, a lower NLR ratio was associated with a durable response, which led to favorable PFS and OS.

Introduction: Claudins are essential for tight junctions, maintaining cell adhesion, regulating intercellular molecule movement, and preserving cellular polarity. Altered claudin expression can lead to dysfunctions, potentially contributing to oncogenesis in epithelial cancers. The role of CLDN18.2 in rectal cancer is not well understood.

Methods: We analyzed tissue samples from 343 rectal cancer patients who underwent concurrent chemoradiotherapy (CCRT) followed by proctectomy.

Results: Upregulated CLDN18.2 expression was associated with older age (p = 0.016), higher pre-CCRT tumor N stage (p = 0.014), higher post-CCRT tumor T stage (p = 0.005), more vascular invasion (p = 0.008), and worse tumor regression (p < 0.001). Univariate analysis showed high CLDN18.2 expression correlated with worse disease-free survival (p < 0.0001), local recurrence-free survival (p < 0.0001), and metastasis-free survival (p < 0.0001). Multivariate analysis indicated high CLDN18.2 expression was associated with inferior disease-specific survival (p < 0.001) and metastasis-free survival (p < 0.001).

Conclusion: Elevated CLDN18.2 expression is associated with adverse clinical outcomes and pathological features, yet suggests a more unfavorable treatment response in rectal cancer patients undergoing CCRT, indicating its potential as a biomarker.

Introduction: The automated classification of clinical trials and key categories within the medical literature is increasingly relevant, particularly in oncology, as the volume of publications and trial reports continues to expand. Large language models (LLMs) may provide new opportunities for automating diverse classification tasks. They could be used for general-purpose text classification, retrieving information about oncological trials.

Methods: A general text classification framework with adaptable prompt, model and categories for the classification was developed. The framework was tested with four datasets comprising nine binary classification questions related to oncological trials. Evaluation was conducted using a locally hosted Mixtral-8x7B-Instruct v0.1-GPTQ model and three cloud-based LLMs: Mixtral-8x7B-Instruct v0.1, Llama3.1-70B-Instruct, and Qwen-2.5-72B.

Results: The system consistently produced valid responses with the local Mixtral-8x7B-Instruct model and the Llama3.1-70B-Instruct model. It achieved a response validity rate of 99.70% and 99.88% for the cloud-based Mixtral and Qwen models, respectively. Across all models, the framework achieved an overall accuracy of >94%, precision of >92%, recall of >90%, and an F1-score of >92%. Question-specific accuracy ranged from 86.33% to 99.83% for the local Mixtral model, 85.49%-99.83% for the cloud-based Mixtral model, 90.50%-99.83% for the Llama3.1 model, and 77.13%-99.83% for the Qwen model.

Conclusion: The LLM-based classification framework exhibits robust accuracy and adaptability across various oncological trial classification tasks. While there remain some challenges such as strong prompt dependence and high computational and hardware demands, LLMs will play a crucial role in automating the classification of oncological trials and literature as the technology continues to advance.

Introduction: Immunotherapy-modified Positron Emission Tomography Response Criteria in Solid Tumors (imPERCIST) using FDG-PET/CT and immune-related Response Evaluation Criteria in Solid Tumors (irRECIST) using CT were compared for evaluation of response, prognosis prediction, and effect on patient management in non-small-cell lung cancer (NSCLC) patients treated with first-line immunotherapy, or immunotherapy and chemotherapy combined.

Methods: Thirty-three patients (26 males, median 68.0 years old) with histologically NSCLC without EGFR/ALK/KRAS alterations underwent pembrolizumab treatment only or immunotherapy combined with chemotherapy. FDG-PET/CT and diagnostic CT scanning was performed at the baseline and after two to four cycles (2 in 8, 3 in 12, 4 in 13 patients). Findings for therapeutic response obtained with imPERCIST and irRECIST were compared. Progression-free survival (PFS) and overall survival (OS) were examined using log-rank and Cox methods.

Results: imPERCIST indicated that 10 patients had complete metabolic response (CMR), eight partial metabolic response (PMR), three stable metabolic disease (SMD), and 12 progressive metabolic disease (PMD), while irRECIST showed that two had complete response (CR), 14 PR, nine SD, and eight PD, indicating substantial concordance (κ = 0.615). Twenty-eight patients showed progression and 18 died from NSCLC after a median 16.2 months. Patients showing response based on imPERCIST and irRECIST (CMR/PMR, CR/PR, respectively) showed significantly longer PFS and OS than nonresponders (SMD/PMD, SD/PD, respectively) (imPERCIST: p < 0.0001 and p = 0.0001, respectively; irRECIST: p = 0.0018 and p = 0.011, respectively). imPERCIST resulted in a change of management in 12 of the 33 patients (36.3%) with an additional effect on patient management in 4 patients (12.1%) evaluated by irRECIST.

Conclusion: For evaluation of NSCLC patients for tumor response and patient management following first-line immunotherapy or that combined with chemotherapy, and predicting prognosis, both FDG-PET/CT and CT findings are accurate, with FDG-PET/CT being superior.

Introduction: Obesity has been associated with poorer outcomes in conventional cancer therapies. However, its impact on chimeric antigen receptor (CAR) T-cell therapy for patients with diffuse large B-cell lymphoma (DLBCL) remains unclear. This study aims to evaluate how obesity affects in-hospital outcomes in patients with DLBCL receiving CAR-T therapy.

Methods: Using the National Readmission Database (NRD), we included adults' age ≥18 with DLBCL who received CAR-T therapy between 2018 and 2020. We performed propensity score matching (caliper of 0.2, 1:1 ratio) and adjusted the following confounding variables: age, comorbidities, and social factors including smoking, alcohol use, and illicit drug use. Data analysis was conducted using R studio software.

Results: A total of 1,874 patients with DLBCL who received CAR-T therapy (9.1% with obesity and 90.9% without) were included in the study. After propensity score matching, there were 160 patients with obesity (50.0%, 59.7 ± 12.2 years of age, 41.9% female) and 160 patients without obesity (50.0%, 58.6 ± 13.0 years of age, 41.3% female). Patients with obesity had significantly higher rates of early mortality (10.6% vs. 4.4%, p = 0.03) and non-home discharge (18.8% vs. 8.1%, p = 0.01) compared to those without obesity. There were no significant differences in 30-day readmission (18.1% vs. 21.3%, p = 0.48).

Conclusion: The rates of acute kidney injury, cardiac complications, leukopenia, neurotoxicity, pulmonary embolism, and infection were comparable between these two groups. Our findings underscore the need for tailored management strategies to improve outcomes following CAR-T therapy for patients with obesity.

Introduction: Most tools trying to automatically extract information from medical publications are domain agnostic and process publications from any field. However, only retrieving trials from dedicated fields could have advantages for further processing of the data.

Methods: We trained a small transformer model to classify trials into randomized controlled trials (RCTs) versus non-RCTs and oncology publications versus non-oncology publications. In addition, we used two large language models (GPT-4o and GPT-4o mini) for the same task. We assessed the performance of the three models and then developed a simple set of rules to extract the tumor entity from the retrieved oncology RCTs.

Results: On the unseen test set consisting of 100 publications, the small transformer achieved an F1 score of 0.96 (95% CI: 0.92-1.00) with a precision of 1.00 and a recall of 0.92 for predicting whether a publication was an RCT. For predicting whether a publication covered an oncology topic, the F1 score was 0.84 (0.77-0.91) with a precision of 0.75 and a recall of 0.95. GPT-4o achieved an F1 score of 0.94 (95% CI: 0.90-0.99) with a precision of 0.89 and a recall of 1.00 for predicting whether a publication was an RCT. For predicting whether a publication covered an oncology topic, the F1 score was 0.91 (0.85-0.97) with a precision of 0.91 and a recall of 0.91. The rule-based system was able to correctly assign every oncology RCT in the test set to a tumor entity.

Conclusion: Classifying publications depending on whether they were randomized controlled oncology trials or not was feasible and enabled further processing using more specialized tools such as rule-based systems and potentially dedicated machine learning models.

Introduction: Oral cancer (OC) is the second most common type of cancer in Pakistan, and early diagnosis can improve survival rates. However, 75% of these cases are diagnosed at an advanced stage. To reduce the incidence and mortality associated with OC, emphasis should be placed on the detailed screening of the oral cavity by dentists. Prevention and early detection practices are important because they increase 5-year survival and reduce recurrence. This study aimed to assess the early detection practices of OC by dentists and determine the factors influencing their screening practices.

Methods: This cross-sectional study was conducted from June 1, 2022, to October 31, 2022, among 235 dentists from the public and private sectors of Islamabad, Pakistan. A validated survey questionnaire with 62 closed-ended items was used. The outcome variable "early detection practices of oral carcinomas" was measured on a scale of 0-8 based on the steps involved in the screening practices. Knowledge scores related to clinical presentation (0-14) and risk factors (0-16) of OC were generated by assigning one point to each correct response. Multivariate linear regression analysis was used to evaluate factors influencing early detection practice scores.

Results: A total of 81.7% of participants had low knowledge related to clinical presentation (mean score = 7.5, SD = 2.2), and 55.3% had high knowledge (mean score = 10.4, SD = 2.06) related to the risk factors of OCs. The mean score for practices related to the early detection of oral carcinomas was 5.4 (SD = 2.04). The duration since attending professional development courses significantly predicted early detection practice scores (p < 0.001). With increasing duration, the early detection practice scores decreased. Moreover, "guidelines about OC examinations" significantly predicted practice scores (higher scores where written guidelines were present) (p < 0.001).

Conclusion: This study offers new evidence on dentists' early detection practices in Pakistan and factors that influence them supporting future training and policy initiatives. It highlights that the early detection of OC improves prognosis, reduces mortality, and minimizes treatment costs. These results emphasize regular professional development courses for the early detection of OC among dentists. This indicates that early detection practices are better in facilities with written guidelines, further emphasizing the need for comprehensive educational materials and training of dentists to improve screening practices. Healthcare authorities should put efforts into implementing a continuous dental education system tailored to the needs of dentists in Pakistan.

Introduction: Human epidermal growth factor receptor 2 (HER-2) in the development of gastric cancer has been widely discussed in the literature. This meta-analysis aimed to assess the relationship between HER-2 and lymph node metastasis in gastric cancer using systematic evaluation and meta-analysis, which can be of some clinical value for treatment and prognosis.

Methods: Relevant databases were searched according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart. PubMed, Embase, Cochrane Library, and Web of Science databases were used to systematically search the literature published up to May 1, 2024. Twenty-one articles were screened and included for quantitative analysis, a data collection form was created, and the quality of the literature was evaluated using the Newcastle-Ottawa Scale (NOS). The results were represented by a Forest plot. I2 was used to assess heterogeneity, Begg's funnel plot was used to test for possible publication bias, and the source of heterogeneity was clarified by the Galbraith plot. Subgroup analyses were performed according to country and test method after excluding studies with significant heterogeneity. All statistical analyses were calculated using stata17.0. This study followed the PRISMA 2020 reporting specifications and the full checklist is available in the online supplementary material (online suppl. Table 4).

Results: A total of 624 articles were retrieved, and 21 were selected and included. Meta-analysis showed that there was heterogeneity between HER-2 positive and HER-2 negative mediators in the study (I2 = 68.0%, p < 0.000). Therefore, the 95% CI (3.12, 95% CI: 2.10, 4.65) of the combined OR calculated by the random-effects model was statistically significant. By comparing the incidence of lymph node metastasis between the HER-2 positive group and the HER-2 negative group, it was found that the incidence of lymph node metastasis between the two groups was significantly different. It is suggested that patients with HER-2 overexpression in cancer tissues may be more prone to lymph node metastasis.

Conclusion: This meta-analysis showed an association between HER-2 and positive lymph node status in gastric cancer patients. This could be valuable for selecting treatment options for gastric cancer patients.

Introduction: This study aimed to evaluate AI-based chatbots (GPT, DeepSeek, Copilot, Gemini) in disseminating information on liver cancer, emphasizing content quality, adherence to established guidelines, and ease of comprehension.

Methods: Between January and February 2025, four chatbots were examined using publicly accessible free versions lacking independent reasoning capabilities. Three frequently searched Google Trends questions ("What is liver cancer awareness?," "What are the symptoms of liver cancer?", and "Is liver cancer treatable?") were posed. Their responses were assessed via the DISCERN instrument, Coleman-Liau Index, Patient Education Materials Assessment Tool for Print, and alignment with American Association for the Study of Liver Diseases, National Comprehensive Cancer Network, and European Society for Medical Oncology recommendations. Statistical analysis was performed using SPSS 22.

Results: All chatbots largely provided relevant and impartial information. GPT and DeepSeek scored lower on specifying information sources and update timelines, whereas Copilot omitted local therapies (e.g., radiofrequency ablation, transarterial chemoembolization, transarterial radioembolization), resulting in reduced scientific accuracy. Gemini and Copilot performed better in "understandability," while GPT and DeepSeek excelled in "actionability." Although GPT demonstrated consistency across multiple treatment options, it did not explicitly reference international guidelines. Study limitations included language constraints, variations in chatbot updates, and reliance on a single inquiry round.

Conclusions: AI chatbots show potential as initial informational tools for liver cancer but cannot replace professional medical consultation. In complex diseases requiring multidisciplinary management, frequent guideline-based updates, expert validation, and diverse data sources are critical to enhancing clinical relevance and patient outcomes.

Introduction: Radiofrequency ablation (RFA) generally involves the insertion of a radiofrequency electrode into the hepatocellular carcinoma (HCC) nodule under ultrasonography (US) guidance. However, the procedure is often not feasible for patients whose HCC is undetectable on conventional US. Advances in imaging technology, such as fusion imaging (FI) and cone-beam computed tomography (CBCT), may enhance treatment precision and efficacy for these challenging cases. This study assessed the efficacy of RFA guided by FI and CBCT in managing HCC poorly visualized on US.

Methods: HCC nodules were classified into GOOD (clearly delineated), POOR (poorly delineated), and NONE (undetectable) based on US visualization. All nodules underwent RFA guided by FI and CBCT either in combination with transcatheter arterial chemoembolization (TACE) or without TACE. The technical success rate and local tumor progression post-RFA were evaluated using dynamic contrast-enhanced imaging. Between-group differences were analyzed retrospectively.

Results: A total of 420 patients with 595 HCC nodules were enrolled. Complete ablation rates were 91.4%, 94.9%, and 86.2% in the GOOD, POOR, and NONE groups, respectively. For nodules with over 50% lipiodol accumulation, the complete ablation rates were 91.5%, 96.5%, and 88.8%; for those with less than 50% lipiodol accumulation, they were 95.5%, 100%, and 62.5%; and for those without lipiodol accumulation, they were 89.5%, 77.8%, and 82.4% in the GOOD, POOR, and NONE groups, respectively. Significant factors associated with complete ablation included larger nodule size and lipiodol accumulation. Cumulative local tumor progression rates at 1 year were 4.5%, 0%, and 3.8%, with no significant differences among groups.

Conclusion: FI and CBCT guidance effectively achieve local control for HCC, including nodules poorly visualized on US, with outcomes comparable to US-visible nodules, especially for those with lipiodol accumulation.