Acta Oncologica最新文献

Background and purpose: Radiotherapy for head and neck cancer must balance tumour control with late toxicities such as dysphagia and xerostomia. Recent retrospective studies suggest that the margin from the gross tumour volume (GTV) to the high-dose clinical target volume (CTV1) may not be critical for local control, while larger irradiated volumes increase the risk of toxicity. The study quantifies potential reductions in dose to organs at risk (OARs) and predicted dysphagia risk when the standard 5 mm GTV-to-CTV1 margin is eliminated in oropharyngeal cancer. Patient/material and methods: Retrospectively 30 oropharyngeal cancer patients treated consecutively during 2023 according to the DAHANCA guidelines (5 mm GTV-to-CTV1 margin) were selected. For each patient, a standard plan and a modified experimental plan (CTV1 = GTV, and CTV2 reduced by 5 mm accordingly) were generated using Pinnacle3 Auto-Planning. All plans met the DAHANCA target coverage and OAR dose constraints. Dose-volume data for relevant OARs were extracted and compared in MATLAB. Normal tissue complication probability (NTCP) model for dysphagia was applied.

Results: Margin elimination reduced high-dose CTV volumes by 70%, yielding significant dose reductions to multiple OARs. Mean doses to the upper/middle pharyngeal constrictors decreased by around 4-5 Gy (p < 0.001) and to the contralateral submandibular gland by ~5 Gy (p < 0.001). These dosimetric gains correspond to an estimated median ΔNTCP of 6.0% of late grade ≥ 2 dysphagia. Target coverage and conformity were maintained in all plans.

Interpretation: Omitting the high-risk CTV margin can substantially reduce the dose to dysphagia--associated OAR without compromising target coverage. This approach shows promise for improving patient-reported swallowing outcomes and warrants clinical evaluation.

Background and purpose: Magnetic resonance imaging-guided radiotherapy (MRIgRT) enables precise tumour targeting through adaptive planning, which is particularly relevant for glioblastoma due to its dynamic morphology. Gadolinium-based contrast agents (GBCAs) enhance tumour visibility, but frequent use during MRIgRT raises safety concerns related to cumulative gadolinium exposure. This study investigated the feasibility of a reduced GBCA dose protocol for patients with glioblastoma undergoing MRIgRT, aiming to balance tumour conspicuity with minimisation of GBCA-related risks. Patient/material and methods: Nine patients with glioblastoma received hypo-fractionated MRI-Linac radiotherapy (10 × 3.4 Gy) with MRI performed with either full-dose, half-dose or no GBCA enhancement. Online gross tumour volume (GTV) delineation was performed by radiation oncologists, while offline GTV delineation was independently conducted by an expert neuroradiologist on GBCA-enhanced scans. Objective assessment using automatic thresholding and a structured Likert-scale evaluation were also performed.

Results: During online adaptation, GTV volumes generally remained stable or increased, whereas offline expert assessments revealed a general volume reduction and systematic volume underestimation with half-dose scans (~18%). Relative delineation volume discrepancies were most pronounced in small tumours. Structured radiologist feedback reported lower confidence, tumour conspicuity and image quality in half-dose scans, particularly for small lesions. Otsu's thresholding revealed reduced edge definition with decreasing contrast dose. No signs of GBCA retention were observed between fractions.

Interpretation: Reduced-dose GBCA-protocols are feasible. Full-dose contrast is recommended at key fractions (e.g. baseline and mid-treatment) and for small tumours, with half-dose imaging reserved for selected intervals or larger tumours. This hybrid approach may balance safety and imaging precision in adaptive MRIgRT.

Background and purpose: Ensuring the reliability and accuracy of artificial intelligence (AI)-generated contours is paramount, as discrepancies could lead to inadequate protection of healthy tissues. With increasing clinical workload, the aim of this study was to assess the time-saving potential of AI-assisted organs at risk (OAR) contouring in head and neck cancer (HNC) treatment planning, while also evaluating geometric accuracy, variability, and dosimetric impact. Patient/material and methods: Twenty patients had 12 OAR contoured by 11 certified dosimetrists and ARTplan (Therapanacea), including the brainstem, cochleas, larynx, mandible, oral cavity, parotid glands, pharynx constrictor muscles, spinal cord, right submandibular gland and thyroid gland. Comparisons were made using geometrical metrics, including Mean Surface Distance, Dice Similarity Coefficient (DSC), Hausdorff Distance, Volume Difference, and Centre of Mass Difference, as well as relevant dose-volume metrics, and total contouring time.

Results: Median manual contouring time of the OARs was 55 (range: 17-151) minutes per patient, while adjusted AI-based structures required 17 (7-42), resulting in 69% time saved. For manual, adjusted and AI-contours, the mean DSC were generally high, averaging 0.85, 0.86, and 0.81 respectively across the evaluated structures. Notably, variability was lowest for the AI and adjusted contours. Average mean and max dose differences were acceptably low (<3.2 Gy) for all OARs.

Interpretation: The results support the integration of AI-based contouring in HNC treatment planning. With minor adjustments, the contours achieve very good clinical quality and demonstrate improved consistency compared to manual contours, while significantly reducing contouring time.

Background and purpose: Photon-counting computed tomography (PCCT) offers enhanced image quality, including improvements in contrast, spatial resolution, and noise reduction. In radiotherapy (RT), optimal image quality is critical for accurate tumor and organ-at-risk delineation. However, reconstruction parameter selection often relies on subjective assessment. This study investigates whether quantitative image quality metrics, particularly contrast-to-noise ratio (CNR), can systematically guide PCCT reconstruction parameter optimization for prostate cancer RT planning.

Material and methods: An anthropomorphic abdomen phantom (QRM, Möhrendorf, Germany) and five patients with prostate cancer undergoing RT were scanned on a Naeotom Alpha PCCT (Siemens Healthineers, Forchheim, Germany). Reconstructions were performed across a range of kernel types, sharpness levels, and virtual monoenergetic image (VMI) energies, with the CNR calculated for each reconstruction. Additionally, a multidisciplinary expert panel qualitatively assessed a subset of reconstructions for two patients to compare with the quantitative findings.

Results: Softer kernels, particularly Br36 and Qr36, combined with lower VMI energies of 40 keV, consistently produced the highest CNR values in both phantom and patient datasets. The qualitative assessment generally supported the quantitative results, with minor deviations likely reflecting the experts' preference for a more familiar image appearance.

Interpretation: Quantitative metrics such as CNR can reliably identify optimal PCCT reconstruction settings for prostate cancer RT, favoring lower VMI energies and softer reconstruction kernels. These findings were consistent across phantom and patient data and were supported by expert evaluations, indicating that a quantitative approach can effectively guide protocol development and reduce reliance on subjective image assessment.

Background and purpose: Accurate stopping-power ratio (SPR) estimation is crucial for proton therapy planning. In brain cancer patients with metal clips, SPR accuracy may be affected by high-density materials and imaging artefacts. Dual-energy CT (DECT)-based methods have been shown to improve SPR accuracy. This study evaluated the consistency between two SPR estimation methods in brain cancer patients: (1) a Hounsfield look-up table (HLUT) for DECT-generated virtual monoenergetic images (VMIs) and (2) the DirectSPR algorithm (Siemens Healthineers). Patient/material and methods: DECT scans were acquired for 11 brain cancer patients. Two SPR maps were generated: one using a 90 keV VMI with a HLUT and the other via the DirectSPR algorithm. The VMI HLUT was adjusted in high-density regions to align with the SPR of titanium. Clinically applied proton therapy plans were recalculated on both SPR maps and dose distributions were compared using dose-volume histograms. Furthermore, a voxel-wise SPR comparison and a separate titanium implant analysis were performed.

Results: Dose differences between the SPR methods were minimal for organs-at-risk. DirectSPR showed strong SPR agreement with the VMI HLUT approach for CT numbers up to 1500 HU (SPR~1.9). Beyond this, especially in regions with titanium implants, DirectSPR yielded higher SPR values than the VMI HLUT, suggesting an adjustment may also be needed for DirectSPR.

Interpretation: DirectSPR was consistent with the VMI HLUT up to 1500 HU but deviated at higher CT numbers. These deviations had limited impact on dose metrics, but they should be considered when choosing beam orientations.

Background: In 2015, a proton therapy (PT) facility was established in Sweden with one aim being to ensure access for all children expected to benefit from PT. Despite potential dosimetric advantages and full subsidisation, PT is not always selected. This study explores reasons for choosing alternative radiotherapy (RT) modalities in a paediatric population.

Material and methods: RT courses delivered to patients ≤ 18 years during 2016-2023 were identified from a national registry. Medical records were retrospectively reviewed to identify reasons for not selecting PT.

Results: Only 34% (n = 275) of all courses identified were delivered with PT. Of the remaining 66% (n = 544), 90% were photon RT, 9% combined PT and photon RT, and 1% electron RT. Among photon RT courses, 97% were delivered with conventional external beam radiotherapy (EBRT), 2% with stereotactic radiotherapy (SRT), and 1% with brachytherapy. The most common reason for choosing photons was non-curative intent (35%), followed by equal or superior expected outcome compared to PT (23%), total body irradiation (TBI) (15%), and uncertainties due to air, organ motion, or metal in field (15%). Dosimetric comparison led to the selection of a favourable or equal photon plan in 8%. Logistical, social, and technical reasons constituted 4%.

Conclusion: While PT can reduce radiation exposure to healthy tissues, particularly important in children, clinical, logistical, and technical factors often necessitate alternative RT modalities. This study highlights the importance of individualised RT planning and multidisciplinary collaboration to balance medical, technical, and practical considerations to ensure optimal treatment approach in every child.

Background and purpose: Quantification/mapping of tumor hypoxia may guide pretreatment decision-making in radiation oncology. Hypoxia-selective positron emission tomography (PET) tracers, like 18F-fluoroazomycin arabinoside (FAZA), allow assessment of hypoxia, but since hypoxia stimulates glycolysis, fluorodeoxyglucose (FDG) and hypoxia-PET may provide overlapping/similar information. Clinical dual-tracer PET studies are highly complex and remain inconclusive. Accordingly, we developed dual-tracer autoradiography techniques to allow high-resolution assessment of the spatial coupling of FAZA and 14C-2DG (FDG-analogue), without the time-separation and co-registration-related inaccuracies intrinsic to PET. Patient/material and methods: Orthotopic lung adenocarcinomas were induced in CRISPR/Cas9 knock-in mice. Mammary adenocarcinomas developed spontaneously in transgenic mice overexpressing ErbB2 (Her2). Patient-derived-xenografts (PDX) were established in immunocompromised mice using biopsies from oropharyngeal cancer patients. Tumor growth was followed by MRI/Caliper measurements. Mice were administered with FAZA (~40 MBq)/14C-2DG (37 kBq)/pimonidazole and sacrificed. Tumor cryosections were analyzed for FAZA/14C-2DG using dual-tracer autoradiography followed by histological stainings. Complementary autoradiograms were co-registered and covered by a square-grid (0.5 × 0.5 mm), and Pearson correlation coefficients (R) were calculated.

Results/interpretation: Hypoxic sub-volumes (FAZA/pimonidazole) were commonly present. A reasonable spatial overlap between FAZA and 14C-2DG was observed in most lung and oropharyngeal tumors with R typically exceeding 0.55. In the breast tumor model, the extent of overlap between FAZA and 14C-2DG varied widely with R ranging from 0.03 to 0.82, which may relate to intertumor mutational differences in this Her2+ oncogene-driven model. Our results suggest a putative role for FDG-PET to identify hypoxic foci and guide dose-escalation.

Background and purpose: Delta biomarkers that reflect changes in tumour burden over time can support personalised follow-up in head and neck cancer. However, their clinical use can be limited by the need for manual image segmentation. This study externally evaluates a deep learning model for automatic determination of volume change from serial 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) scans to stratify patients by loco-regional outcome. Patient/material and methods: An externally developed deep learning algorithm for tumour segmentation was applied to pre- and post-radiotherapy (RT, with or without concomitant chemoradiotherapy) PET/CT scans of 50 consecutive head and neck cancer patients from The Christie NHS Foundation Trust, UK. The model, originally trained on pre-treatment scans from a different institution, was deployed to derive tumour volumes at both time points. The AI-derived change in tumour volume (ΔPET-Gross tumour volume (GTV)) was calculated for each patient. Kaplan-Meier analysis assessed loco-regional control based on ΔPET-GTV, dichotomised at the cohort median. In a separate secondary analysis confined to the pre‑treatment scans, a radiation oncologist qualitatively evaluated the AI‑generated PET‑GTV contours.

Results: Patients with higher ΔPET-GTV (i.e. greater tumour shrinkage) had significantly improved loco-regional control (log-rank p = 0.02). At 2 years, control was 94.1% (95% CI: 83.6-100%) vs. 53.6% (95% CI: 32.2-89.1%). Only one of nine failures occurred in the high ΔPET-GTV group. Clinician review found AI volumes acceptable for planning in 78% of cases. In two cases, the algorithm identified oropharyngeal primaries on pre-treatment PET-CT before clinical identification.

Interpretation: Deep learning-derived ΔPET-GTV may support clinically meaningful assessment of post-treatment disease status and risk stratification, offering a scalable alternative to manual segmentation in PET/CT follow-up.