Over the past two decades, targeted anti-HER2 therapies have transformed management of HER2-positive early breast cancer (HER2 + eBC), enabling chemotherapy de-escalation without compromising efficacy. This review assesses the rationale for incorporating carboplatin into (neo)adjuvant regimens and critically examines its routine use in unselected patient populations with HER2-positive eBC. While initial translational and early clinical studies suggested that adding carboplatin to taxane-based chemotherapy combined with anti-HER2 therapy might be beneficial, more recent randomized trials have failed to demonstrate incremental advantage and have raised concerns about increased toxicity. Emerging prognostic and predictive biomarkers, most notably the HER2DX genomic assay, hold promise for identifying high-risk patient subgroups who may derive meaningful benefit from carboplatin. We conclude that carboplatin should not be routinely added in unselected HER2-positive early breast cancer, but rather, a biomarker-driven, personalized strategy is recommended to enhance patient selection, refine the treatment algorithm, and maximize the cost-to-benefit ratio.
BACKGROUND:Bone-only HR+/HER2 - metastatic breast cancer generally has a favourable prognosis, but some patients develop visceral metastases. We aimed to quantify visceral conversion and identify reproducible baseline correlates during CDK4/6 inhibitor (CDK4/6i) therapy. METHODS:This multicentre retrospective cohort included 692 patients treated with palbociclib, ribociclib, or abemaciclib plus endocrine therapy as first- or second-line treatment across 24 Italian centres. Visceral conversion was analysed in a competing-risks framework, with skeletal progression or death without prior visceral conversion as competing events. Eighteen baseline candidate predictors derived from a 35-variable dataset were evaluated using Fine-Gray modelling, ridge penalisation, bootstrap stability assessment, and cause-specific Cox sensitivity analyses. RESULTS:Over a median follow-up of 31 months, 162 patients (23.4%) developed visceral conversion after a median of 17.0 months. Cumulative incidence was 8.8%, 17.5%, and 24.5% at 12, 24, and 36 months, respectively; median overall survival after visceral conversion was 18.0 months. Progesterone receptor (PR) expression was the most stable tumour-biological correlate, with complete sign concordance and a median absolute coefficient rank of 1 across 500 bootstrap resamples. In the mutually adjusted Fine-Gray model, higher PR expression was associated with lower visceral-conversion risk (sHR per 10-percentage-point increase, 0.929; 95% CI, 0.889-0.971; p = 0.0012), with consistent direction across complementary analyses. However, stand-alone PR prediction remained modest (24-month IPCW AUC, 0.585; Brier score, 0.142; IPA, 1.1%; O:E, 1.00), limiting individual-level clinical utility. CONCLUSIONS:Broad baseline-variable profiling identified PR expression as the most reproducible tumour-biological correlate of visceral conversion, although its stand-alone predictive performance was modest.
BACKGROUND:Experiencing cancer can be physically and psychosocially challenging in the short- and long-term. Integrating Patient-Reported Outcome Measurements (PROMs) to capture these challenges can support patient-centered care and may reduce morbidity and mortality. However, age-appropriate tools remain scarce for children and adolescents. We report on the development phases I/II of the EORTC QLQ-CHI questionnaire tailored for children 8-14 years. METHODS:Following the EORTC QLG module development guidelines, qualitative semi-structured interviews were conducted in children aged 8-14 years undergoing active cancer treatment, parents and healthcare professionals (HCPs). Participants rated the relevance of issues identified in a previous systematic review (Phase Ia) and provided preferences on response format and recall period (Phase Ib). During expert round tables, key issues were converted into items (Phase II). RESULTS:Interviews were completed with 47 children, 45 parents, and 22 HCPs from six countries. Among all children (mean age=10.8 ± 1.9 years; 57.4% female), 55.3% had haematological cancers. Of the initially identified issues, 57 were identified as key issues by comparing children's, parents' and HCPs feedback. Children preferred four response options (70.0%) and a shorter recall period (43.3%). Through expert round tables (Phase II), items were assigned to a more symptom-oriented core scale and an additional scale for psychosocial concerns. The provisional questionnaire consists of 50 items. CONCLUSION:This study represents the first step in developing the EORTC QLQ-CHI (8-14 years). Next steps include pilot testing, validation (Phase III and IV), co-development of a corresponding measure for children < 8 years and an observer-rating version.
BACKGROUND:Aggressive angiomyxoma (AA) is a rare, locally infiltrative mesenchymal tumor with frequent local recurrence. Optimal management remains poorly defined, particularly the roles of surgery, endocrine therapy, and observation. METHODS:We retrospectively analyzed patients with histologically confirmed AA treated at a single institution from 2000 to 2024. Clinical, pathologic, and treatment data were collected. Overall survival (OS), relapse-free survival (RFS), duration of treatment (DOT), and radiographic response by quasi-RECIST 1.1 were assessed descriptively. RESULTS:Forty patients were included; median age was 42 years, 87.5% were female, and median tumor size was 9.9 cm. Initial management was surgery in 30 patients, systemic therapy in 8, and observation in 2. Among patients initially managed with surgery, median OS from the time of first resection was not reached at a median follow-up of 68.6 months; estimated 10-year OS was 95% (95% CI, 68%-99%). Among patients with R0/R1 resection, median RFS was 49.8 months (95% CI, 21.0-64.7), with 3-year RFS of 56% (95% CI, 31%-75%). Tumor size and margin status were not significantly associated with recurrence. Thirteen patients received systemic therapy, predominantly endocrine-based; ORR was 25% and median DOT was 15.4 months. Two patients were initially observed, and three with residual macroscopic disease after R2 resection remained free from further intervention for prolonged periods. Radiation therapy was rarely used. CONCLUSIONS:AA was associated with excellent long-term survival despite frequent recurrence or subsequent intervention after surgery. Endocrine therapy showed objective activity in a subset of patients. Observation or delayed intervention may be feasible in selected, clinically stable patients with untreated or residual macroscopic disease, supporting individualized, function-preserving management focused on symptoms, morbidity, and patient preferences.
PURPOSE:In some situations, the pathological diagnosis of neuroendocrine neoplasms (NEN) remains challenging, with direct consequences for treatment decisions. ENDOCAN-TENpath, the French national expert network for NEN pathology, established monthly virtual "third-reading" sessions for collegial review of difficult cases. We evaluated the diagnostic outcomes and clinical impact of this process. METHODS:All cases submitted to third-reading sessions from March 2022 to December 2023 were retrospectively reviewed. We analyzed diagnostic difficulties, diagnostic modifications, and their expected therapeutic impact. RESULTS:Of 3239 cases received by ENDOCAN-TENpath, 126 (3.9%) were submitted to third-reading by 22 of 33 network pathologists (median turnaround: 20 days). Diagnostic difficulties fell into four categories: high-grade NEN classification (33.3%), mixed neuroendocrine/non-neuroendocrine neoplasm diagnosis (36.5%), incomplete neuroendocrine phenotypes (23.0%), and unfamiliar entities (7.1%). Consensus or majority diagnosis was reached in 102 cases (81%). Unresolved cases were due to lack of consensus (n = 13) or insufficient material (n = 10). Critically, the process yielded decisions with major therapeutic consequences: 20/46 presumed NENs were reclassified as non-neuroendocrine malignancies and 5/25 presumed adenocarcinomas as NENs, fundamentally altering treatment strategy. Furthermore, 32/42 ambiguous high-grade NENs were definitively classified as tumor or carcinoma, directly determining first-line chemotherapy regimen. These findings also prompted internal guidelines to reduce diagnostic variability. CONCLUSION:The ENDOCAN-TENpath third-reading process yields a high rate of clinically actionable diagnostic decisions directly altering treatment strategy, while identifying unmet needs and providing a framework for reducing diagnostic variability in complex NEN cases.
PURPOSE:Using real-world data (RWD), we assessed the impact of diagnostic regulatory pathways on activation timelines and biomarker testing implementation in biomarker-driven clinical trials across Europe. METHODS:RWD from 132 Clinical Performance Study Applications (PSAs) and Ethics Committee (EC) submissions across 18 European Union Member States (EU-MS) and 2 non-EU countries that operate under the In Vitro Diagnostic Medical Devices Regulation 2017/746 (IVDR) were pooled for analysis. Statistical, outlier assessments and comparative analyses were performed to evaluate differences in timelines between submission pathways. Statistical significance was assessed using Welch's t-test and the Mann-Whitney U test. RESULTS:Mean combined approval time for all submissions (EC and PSA) was 129.39 days (median, 116.5; range 32-346). Sequential submissions averaged 139.97 days versus 121.11 days for parallel submissions, corresponding to a ∼15.6% longer timeline than sequential pathways (mean difference 18.86 days; p = 0.0411). Mean country-level combined approval timelines varied substantially across countries, ranging from 75.66 to 175 days among countries (n = 17) with ≥ 3 submissions. Substantial cross-country variability in operational implementation challenges was observed. CONCLUSIONS:Diagnostic approvals represent underrecognized bottlenecks in precision medicine (PM) trial activation, with approval timelines frequently exceeding those of Clinical Trial Applications (CTAs), median 108 days. Fragmented IVDR implementation has created substantial cross-country variability and uncertainty for trial activation. Biomarker strategies and diagnostic testing approaches should be thoughtfully constructed and considered pragmatically, as inefficient implementation may directly affect trial interpretability, efficacy assessment and regulatory success, substantially compromising trial delivery in Europe.
Gastroenteropancreatic neuroendocrine carcinomas (GEP-NEC) are aggressive, poorly differentiated neoplasms representing the most common subtype of extra-pulmonary neuroendocrine carcinomas. Despite increasing recognition, they remain biologically and therapeutically elusive, with limited understanding of their molecular drivers and few established treatment paradigms. Most patients present with advanced disease, making systemic therapy the cornerstone of management. Platinum-based chemotherapy remains the standard first-line approach, though outcomes are suboptimal and consensus is lacking regarding maintenance strategies and second-line options. Later lines of therapy are ineffective, and survival is short after progression on first-line therapy. In this review, we synthesize current evidence on systemic therapies for GEP-NECs, highlighting clinical trials of cytotoxic agents, immune checkpoint inhibitors, molecularly targeted therapies, and novel modalities such as bispecific T-cell engagers. We also discuss the role of molecular profiling and emerging biomarkers in guiding therapy, with an emphasis on precision-driven approaches for this highly heterogeneous and challenging group of malignancies.
Background ETERNITY was a retrospective and prospective cohort study investigating long-term (≥5 years) survival in patients with glioblastoma. We assessed the longitudinal course of neurocognitive function (NCF) and its determinants in a subgroup of patients with glioblastoma, IDH-wildtype, or astrocytoma, IDH-mutant, CNS WHO grade 4. Methods NCF was assessed at baseline and every 6 months using the Hopkins Verbal Learning Test-Revised, Controlled Oral Word Association Test, and Trail Making Test. Scores were converted to age-, sex-, and education-adjusted Z-scores and classified as impaired or unimpaired (Z ≤ -1.5). Linear mixed models were used to analyze NCF trajectories and associations with patient and tumor characteristics. Results At baseline (mean 9 years post-diagnosis, range 5-21 years), 145 of 185 patients (78%) were impaired on ≥1 test outcome. Impairment rates varied between 17.4% (HVLT-R delayed recognition) and 58.7% (TMT B). NCF remained largely stable over time, with a small decline in HVLT-R delayed memory. Left-sided and temporal tumor location were negatively associated with poorer NCF (p’s.01 to.03). Frontal tumor location was associated with higher psychomotor speed and cognitive flexibility (p’s <.001). Patients with IDH-mutant tumors performed worse on delayed recognition, whereas IDH mutation and MGMT promoter methylation were linked to improved phonemic fluency over time. Conclusions The majority of long-term survivors with astrocytoma, IDH-mutant, CNS WHO grade 4, and glioblastoma IDH-wildtype show neurocognitive impairment. Nonetheless, NCF is generally stable, with tumor location and molecular features associated with specific outcomes. These insights can help guide patient counseling and personalized care.
Background Generative AI is increasingly being explored in cancer care, yet little is known about whether these tools align with the priorities and concerns of the people they are intended to support. We conducted a scoping review to identify studies that examined the perspectives of people affected by cancer towards generative AI. Methods PubMed, Embase, Web of Science, and MEDLINE were searched to July 2026. Two independent reviewers screened identified records for peer-reviewed studies reporting primary empirical data on the perspectives of people affected by cancer regarding generative AI in cancer care contexts. Study characteristics, applications evaluated, and perspectives assessed were extracted and synthesised descriptively and narratively. Results Of 2,441 records identified, 32 studies comprising 4,919 participants were included. Of these, 29 assessed specific applications: 21 Patient Communication and Education, 6 Clinical Note Generation involving lay summaries, 2 Clinical Decision Support; and 3 assessed general perspectives on AI without a defined task. Where perspectives were assessed, studies predominantly measured usefulness (21/32), comprehension (20/32), and trust (14/32). People affected by cancer were generally favourable when generative AI improved the accessibility and readability of health information; however, trust was conditional on personalisation, emotional appropriateness, human oversight, and perceived accuracy. These perspectives may help inform the design and implementation of generative AI, including interface design, patient education, consent processes, workflow integration, governance, monitoring, and feedback mechanisms. Conclusions Perspectives towards generative AI involving people affected by cancer remain concentrated in a narrow subset of applications. Considerations for responsible AI deployment, including safety, transparency, equity, and autonomy, have not been assessed from the perspectives of people affected by cancer.
BACKGROUND:Tobacco use is a major preventable cause of cancer, and social media offers scalable opportunities for prevention. AI-generated virtual influencers may reduce production costs, but the effects of message framing on campaign delivery remain unclear. OBJECTIVE:The objective was to compare AI-generated smoking prevention videos emphasising long-term cancer consequences versus short-term consequences under real-world conditions on Instagram. METHODS:Ten Instagram Reels featuring the same AI-generated character were deployed in a controlled field experiment, with five videos emphasising short-term consequences of smoking and five videos emphasising long-term consequences of smoking. Budget, campaign duration, and targeting were held constant, with each video treated as an independent unit of analysis. Outcomes included reach, cost per 1000 accounts reached, link clicks, cost per link click, interactions, and cost per interaction. Conditions were compared using exact Mann-Whitney U tests with Holm adjustment and exact permutation tests for sensitivity analyses. RESULTS:Videos emphasising short-term consequences of smoking achieved greater reach than videos with long-term cancer framing (median 14,117 vs 11,820; Holm-adjusted p = .048) and lower cost per 1000 accounts reached (€3.15 vs €3.86; Holm-adjusted p = .048). No significant differences were found for link clicks, cost per link click, interactions, or cost per interaction. CONCLUSIONS:In this prevention campaign, short-term framing of smoking-related adverse effects was associated with broader and less costly platform-mediated distribution, but with similar engagement. Future research should examine whether increased reach translates into meaningful cancer prevention outcomes.
Brain cancers, especially glioblastoma, remain among the deadliest adult cancers, with outcomes largely unchanged despite multimodal treatments. This review summarizes cutting-edge artificial intelligence (AI) and machine learning advances transforming neuro-oncology in diagnostics, molecular profiling, treatment planning, and monitoring. Key findings show AI-driven radiomics and deep learning reaching over 90% accuracy in tumour segmentation and grading from MRI and whole-slide images, non-invasive IDH/MGMT prediction through liquid biopsy (ctDNA/EVs) analysis, and augmented reality-guided resection that maximizes tumour removal while safeguarding expressive cortex. Treatment planning benefits from hybrid CNN-Transformer models for immunotherapy stratification and blood-brain barrier penetrant drug repurposing, while real-time progression detection via multimodal integration helps differentiate true progression from pseudoprogression. Despite these advances, significant challenges remain, including data scarcity and imbalance in rare subtypes, domain shift due to imaging variability, black-box model behaviour eroding trust, regulatory requirements for prospective validation, and workflow fragmentation. Emerging solutions include federated and transfer learning for scalable model development, explainable AI (such as SHAP and attention interpretation for vision transformers) to foster clinician-AI collaboration, and foundation models pretrained on large-scale neuro-oncology datasets to facilitate personalization. Achieving this potential will depend on harmonized multi-omics registries, strong ethical and regulatory governance, and deliberate human-AI collaboration frameworks to integrate these tools into precision neuro-oncology.
PURPOSE:Given the early recurrence of brain metastasis (BM), identifying factors that drive BM progression is of clinical interest. This study investigates genetic, epigenetic, and inflammatory signatures in progressive BM following different therapeutic approaches. METHODS:A total of 153 patients who underwent surgical resection for progressive BM were grouped according to the therapeutic strategies prior to the first BM resection: prior radiation (n = 43), systemic therapy (n = 37), combined radiation and systemic treatment (n = 10), and treatment-naive patients (n = 63). Among the treatment-naive patients, 35/63 (55.5%) experienced another intracranial relapse and underwent a second resection (=relapse group), enabling paired analyses. Of these, 23/35 (65.7%) received no therapy between resections; 12/35 (34.3%) received CNS-directed radiotherapy. Tissue samples were analysed using whole-exome sequencing, DNA methylation profiling, and immunohistochemistry. RESULTS:BM resected after progression following prior cranial radiotherapy (43/153, 28.1%) showed significantly lower densities of CD3 + , CD8 + , and CD45RO + cells together with increased FOXP3 + cell density compared with treatment-naïve BM (63/153, 41.2%; median CD3 +: 71 vs. 494 cells/mm²; CD8 +: 44 vs. 187 cells/mm²; CD45RO+: 104 vs. 302 cells/mm²; FOXP3 +: 215 vs. 41 cells/mm²). In the paired analyses, progressive specimen after prior radiation were likewise associated with significantly reduced CD3 + , CD8 + , and CD45RO + and increased FOXP3 + cell densities compared with the matched baseline specimen. In contrast, no genetic alterations or differences in DNA methylation patterns between irradiated and non-irradiated matched samples were identified. CONCLUSION:Progressive BM following cranial radiotherapy demonstrated a distinct immune marker profile consistent with a more immunoregulatory, rather immunosuppressive tumour microenvironment. No therapy-associated genetic or epigenetic alterations were identified. Further prospective studies are warranted to determine whether these immune alterations reflect treatment-related effects or biological features associated with resistance following radiotherapy.
BACKGROUND:Tumor microenvironment (TME) including stromal composition and antitumor immune responses, plays an important role in cancer progression. Tumor-stroma ratio (TSR) and tertiary lymphoid structures (TLS) both shows prognostic value, but their combined significance across urothelial carcinoma (UC) remains unclear. METHODS:We retrospectively analyzed H&E-stained whole-slide images from 884 UC patients from five cohorts (FAHZU, n = 76; Emory, n = 94; TCGA, n = 291; QDPH, n = 330; TRRC, n = 93), encompassing both localized and advanced disease settings. TLS were identified using deep learning-based nuclei classification followed by identifying clusters of aggregated immune cells, while TSR was quantified using automated stromal segmentation. An integrated TLS-TSR risk score was trained using a Cox proportional hazards model in the FAHZU cohort and externally validated in four independent test cohorts. In the TRRC cohort, associations between these biomarkers and response to immune checkpoint inhibitor (ICI) were assessed. FINDINGS:Automated TLS density and TSR individually showed good concordance with pathologist assessments (TLS: ICC=0.852; TSR: Spearman ρ=0.826; both p < 0.001). High TLS density and low TSR were each associated with improved progression-free survival (PFS) across cohorts. The TLS-TSR score outperformed either biomarker alone (C-index: 0.65-0.69) and remained an independent predictor of PFS in multivariable analysis. TLS density showed strong association with ICI response (AUC 0.745, 95% CI 0.607-0.864). Transcriptomic analysis further revealed enrichment of extracellular matrix organization and stromal-related pathways in the high-risk group, accompanied by selective alterations in immune cell composition, indicating coordinated remodeling of both stromal and immune components of TME. INTERPRETATIONS:Integrating stromal composition and immune-related morphology may improve prognostic stratification across UC patients and identify histopathologic features associated with response to ICI.
BACKGROUND:Colorectal cancer is the third most common malignancy worldwide. Although current screening relies on faecal occult blood testing (FOBT), its sensitivity for advanced adenomas (key precursor lesions) remains limited. Indeed, colonoscopy remains the gold standard for definitive diagnosis. Hence, development of novel blood-based screening tools in FOBT-positive patients is needed to optimize colonoscopy referral. METHODS:We prospectively recruited 104 FOBT-positive patients undergoing colonoscopy. Based on endoscopic and histopathological assessment, patients were classified into no polyps (NP, n = 46), non-advanced polyps (NA, n = 33) and advanced polyps (AP, n = 25). Seventy-five immune cell subsets and their homing, activation and exhaustion profiles were characterized by spectral cytometry, yielding 900 variables. Differentially expressed variables were used to train five supervised machine learning models including random forest, decision tree, multinomial regression, polynomial kernel support vector machine (SVM) and Kernel K-nearest neighbours. RESULTS:Among 91 differentially expressed immune variables, Boruta-based feature selection identified seven key cell populations. Th1-like cells emerged as the dominant predictive variable. The decision tree model achieved the best overall performance, with total classification of AP patients (AUC=1.0, 100% sensitivity and specificity). Global model accuracy reached 75% (p < 0.01 vs. no-information rate), with a macro-AUC of 0.85. CONCLUSIONS:This study demonstrates that spectral-cytometry immunotyping of the circulating immunome, combined with machine learning, can identify patients harbouring advanced colorectal adenomas from a blood sample. Hence, Th1-like cells emerge not just as key cells, but also as a promising biomarker that could complement current FOBT screening to prioritize colonoscopy in patients at highest risk of pre-malignant lesions.
BACKGROUND:Bladder cancer (BlCa) surveillance requires lifelong invasive monitoring, imposing substantial patient burden and healthcare costs. This underscores the need for minimally-invasive precision medicine approaches. Herein, we investigated the circulating miRNome of BlCa patients to assess its clinical and translational utility. METHODS:miRNA-seq was performed on tumors and matched plasma samples from 15 patient-derived xenograft (PDX) mouse models. The screening cohort of the study consisted of 216 patients, and circulating miRNAs were quantified by RT-qPCR following 3'-end poly(A)-tailing. Cox regression models were used to develop liquid biopsy-based miRNA risk-scores (miRisk-scores), for non-muscle invasive (NMIBC) and muscle invasive BlCa (MIBC). RESULTS:miRNA-seq identified six miRNAs (miR-100, miR-1246, miR-1290, miR-193b, miR-375, miR-4488) that were concurrently upregulated in tumors and plasma, and correlated with advanced stage/grade and early progression in PDX mouse models. In our screening cohort, elevated miR-375 and miR-4488 were associated with poorer NMIBC prognosis, whereas miR-1290 and miR-193b with worse survival in MIBC. Higher miRisk-scores were associated with short-term relapse and progression in NMIBC, while miRisk score-fitted multivariate models improved TaT1 risk-stratification. Moreover, higher miRisk-scores correlated with worse clinical outcomes in MIBC, and enhanced the prognostic performance of clinical markers. CONCLUSION:Overall, liquid biopsy miRNA signatures offer a minimally-invasive tool to complement biopsy-based assessments and refine BlCa risk-stratification and prognosis.
BACKGROUND:The tumor microenvironment (TME) in cervical cancer may undergo immunological changes during neoadjuvant chemotherapy (NACT). The primary aim of this study was to elucidate changes in the TME of cervical cancer during platinum-containing NACT and to evaluate whether these immunological changes correlated with survival. Characterizing these changes may help identify the optimal timing for future integration of immunotherapy. This study examined alterations in lymphocyte subsets in patients with FIGO 2009 stage IB2-IIA2 cervical cancer treated with cisplatin-based NACT. METHODS:Tissue samples from 64 patients enrolled in EORTC trial 55994 were analyzed. Baseline biopsies (n = 55) and post-NACT surgical specimens (n = 43) underwent multiplex immunohistochemistry with fluorescent labels to quantify B cells, CD8 and helper T-cells, regulatory T-cells (Tregs), and tumor cells. Proliferation was assessed using Ki67 expression. ImmuNet, a validated machine-learning-based image analysis platform, was used for cell identification and density measurements. FINDINGS:Following NACT, Treg density significantly decreased in stromal (p < 0.0001) and intratumoral regions (p = 0.0007). A lower proportion of proliferating Tregs after treatment was associated with better OS (p = 0.018). Changes in other lymphocyte subsets were less consistent. INTERPRETATION:NACT appears to reduce Treg-mediated immunosuppression in the cervical cancer TME. Patients with low post-treatment Treg proliferation demonstrated more favourable survival. These findings are hypothesis-generating and suggest that the immune microenvironment following NACT warrants further investigation, including studies evaluating the optimal timing of immunotherapy.
