Neoadjuvant chemotherapy is standard for stage IB-III triple-negative breast cancer (TNBC), with pathological complete response (pCR) strongly associated with survival. Although escalation with platinum and immune checkpoint inhibitors (ICI) improves pCR and long-term outcomes, patients with pCR in control arms of pivotal trials also show favorable outcomes. Whether the regimen leading to pCR impacts long-term survival is largely unknown. We conducted a systematic review and meta-analysis, searching phase II and III trials including early-stage TNBC patients with pCR. A pooled analysis of Kaplan–Meier-derived individual patient data was performed for event-free survival (EFS) and overall survival (OS), with subgroup analyses by treatment regimens. Of 2830 identified publications, 18 trials comprising 3430 patients were included. Neoadjuvant ICI with chemotherapy improved EFS (HR 0.67; 95
Delays in diagnosis are common in rare cancers, but their frequency and impact in gastrointestinal stromal tumors (GISTs)—one of the most prevalent sarcoma subtypes—are unknown. This study aimed to (1) quantify patient- and diagnostic-related intervals, (2) identify factors influencing interval length, and (3) assess their impact on health-related quality of life (HRQoL). Patients with GIST diagnosed in the Netherlands between 2008 and 2018 completed a questionnaire assessing time to diagnosis, impact of the diagnostic trajectory, and HRQoL. Univariable logistic regression analyses identified associated patient and tumor characteristics. Among 328 patients (median 5.9 years after diagnosis), 28
The cross-sectional AILEEN study evaluated long-term cognitive function after immune checkpoint inhibition (ICI) for melanoma and compared outcomes between neoadjuvant combination therapy (ipilimumab/nivolumab) versus adjuvant monotherapy (nivolumab or pembrolizumab). Patients with stage III/IV melanoma were invited approximately two years after ICI initiation. Patients completed the online Amsterdam Cognition Scan and questionnaires on self-reported cognitive problems, fatigue, anxiety, and depression. Scores were evaluated at group- and individual-level against cancer-free norm groups and well-established cutoffs. Prevalences of cognitive impairment were compared to estimated false-positive rates accounting for the number and correlation of tests. Patients treated with ICI (N = 107) showed significantly worse verbal memory, processing speed, and executive functioning and reported clinically meaningful lower cognitive function at group-level compared to normative data. At individual-level, 34
Radiogenomics aims to non-invasively predict tumour genotypes from imaging, but most studies assume molecular homogeneity by assigning a single biopsy-derived label to all lesions within a patient. This approach risks substantial label noise given well-documented interlesional heterogeneity. We investigated whether anchoring training to biopsy-confirmed lesions improves radiogenomic model performance and generalisability. We retrospectively analysed 1646 patients (11473 segmented lesions) with contrast-enhanced CT and EGFR mutation status from next-generation sequencing at the Netherlands Cancer Institute, alongside an external NSCLC radiogenomics cohort (n = 158). All visible lesions were segmented, and the exact biopsy site was matched to its segmentation. Radiomic features were extracted, and machine learning models were trained with three lesion selection strategies: all lesions, non-biopsied lesions only, and biopsy-confirmed lesions only. To disentangle label quality from sample size, we created size-matched variants (one lesion per patient) for all-lesion and non-biopsied strategies. All models achieved significant discrimination of EGFR status on internal validation (AUC = 0.62–0.68). However, performance of the all-lesion and non-biopsied models declined on external validation (AUC = 0.55–0.63), while the biopsy-anchored model maintained stable performance (AUC = 0.62), despite having only 1/10th of the training sample size. When training sets were size-matched, the biopsy-anchored approach significantly outperformed a model trained on all available lesions on external validation (p = 0.037). Radiogenomic models trained on biopsy-confirmed lesions outperform conventional all-lesion strategies in external validation, despite using an order of magnitude fewer samples. Prioritising lesion-level label fidelity can mitigate heterogeneity-driven noise, enhancing robustness and clinical translation of imaging-based genomic prediction. Question Does assigning biopsy-derived molecular labels to all lesions introduce heterogeneity-driven label noise that reduces the generalisability of radiogenomic models? Findings Models trained exclusively on biopsy-confirmed lesions demonstrated superior external generalisability compared with all-lesion approaches, despite being trained on substantially fewer samples. Clinical relevance Biopsy-anchored radiogenomics improves the reliability of non-invasive mutation prediction by accounting for tumour heterogeneity, potentially supporting clinical decision-making when tissue sampling is limited or molecular results are discordant across lesions.
Resecting pediatric tumors is often surgically challenging caused by insufficient tumor localization, due to limited visibility and palpability. Surgical navigation systems may potentially improve intraoperative tumor localization. This study evaluates the accuracy and precision of an in-house developed navigation setup using tracked ultrasound under standardized conditions. Bone surface-based registrations, using automatic bone segmentation on tracked ultrasound images, were conducted on phantoms to validate the performance of the in-house developed surgical navigation setup. Registration was conducted on different phantoms, consisting of tumors near bones of an extremity (n = 50), the pelvis (n = 5) and the thoracic wall (n = 5). In addition, the same registration framework was evaluated in the case of kidney tumors, where the kidney surface was used instead of the bone surface for registration. Target registration error (TRE) was used as the primary outcome measure. For tumors localized with bone surface-based registration, the setup achieved a median TRE of 1.3 mm with an interquartile range (IQR) of 0.9–2.1 mm. The robustness of the bone surface registration method was demonstrated with consistent results across anatomical regions. For kidney tumor localization with kidney surface-based registration, the setup achieved a median TRE of 3.3 mm with an IQR of 2.7–3.6 mm. Under controlled circumstances, the navigation setup demonstrated reproducible < 2 mm accuracy for tumor localization using the bone surface registration. For kidney tumors, the navigation setup showed < 4 mm accuracy. These findings establish a performance benchmark that can guide interpretation of larger inaccuracies encountered during clinical use and support future development toward clinical implementation.
