Stereotactic body radiotherapy (SBRT/SABR) is a curative treatment for early-stage lung cancer in patients unsuitable for surgery. While tumor laterality is routinely considered in surgical decision-making, its relevance to outcomes after SBRT remains unclear. This study evaluated clinical outcomes after SBRT and explored the association between tumor laterality and treatment outcomes. This single-institution retrospective study included consecutive patients treated with SBRT for peripheral pulmonary lesions between 2011 and 2024. Prescribed ablative doses corresponded to a median BED₁₀ > 100 Gy. Local recurrence (LR), loco-regional recurrence (LRR), distant metastases (DM), and cancer-specific death (CSD) were analyzed using competing-risk models; progression-free survival (PFS) and overall survival (OS) using Kaplan–Meier method. A total of 115 patients (median age 75 years) were treated with SBRT. The most common regimens were 42–54 Gy in three or 50–55 Gy in five fractions (median BED₁₀=116 Gy). Toxicities were mild and transient. At a median follow-up of 39 months, 5-year OS and PFS were 55
Internal margins for thoracic lymph nodes in radiotherapy are commonly applied uniformly despite variability in nodal motion during respiration across patients and node stations. Four-dimensional CT (4DCT) has demonstrated movement of mediastinal and hilar lymph nodes and discordance with motion of the primary tumor, yet translation into practical margin strategies remains limited. We hypothesized that the amount of movement during respiration differs between upper and lower thoracic lymph nodes within the same patient and evaluated location-dependent motion to characterize regional and interpatient variability in respiratory motion. This retrospective cohort study included patients undergoing 4DCT simulation for thoracic radiotherapy. Eligible patients had complete respiratory phase datasets and at least one upper and one lower thoracic lymph node identified on reference-phase CT. Nodes were classified relative to the carina. Lymph node centroids were tracked across respiratory phases to quantify directional displacement and three-dimensional vector magnitude relative to the reference phase. Differences in motion between upper (stations 1–6) and lower (stations 7–14) nodes were evaluated using linear mixed-effects models accounting for within-patient clustering. Forty-one patients with 229 lymph nodes were analyzed (130 upper, 99 lower). Motion was predominantly cranio–caudal. Mean cranio–caudal displacement was 3.3 ± 1.9 mm for upper nodes and 5.2 ± 3.0 mm for lower nodes. Within-patient analysis demonstrated greater motion for lower nodes (mean difference 1.9 mm; p < 0.001). Vector displacement was also greater for lower nodes (5.7 ± 2.9 mm vs. 4.1 ± 2.0 mm; p < 0.001). Motion exceeding 5 mm in three-dimensional vector magnitude occurred more frequently with lower nodes (51
To develop and validate a multimodal deep learning model that integrates post-neoadjuvant chemoradiotherapy (nCRT) rectal MRI and endoscopic images for prediction of pathologic complete response (pCR) and disease-free survival (DFS) in patients with locally advanced rectal cancer (LARC). This multicenter retrospective study included patients with LARC from two institutions who underwent nCRT followed by total mesorectal excision. Post-nCRT, preoperative multiparametric MRI (T2-weighted imaging, diffusion-weighted imaging, and contrast-enhanced T1-weighted imaging) and endoscopic images were analyzed using a multimodal deep learning framework incorporating graph neural networks and cross-attention to integrate complementary imaging features. Model performance was evaluated using AUC, accuracy, F1 score, calibration curves, Brier score, and decision curve analysis. DFS risk stratification was assessed using Kaplan-Meier analysis. A total of 288 patients were included. For pCR prediction, the multimodal model achieved AUCs of 0.876, 0.848, and 0.818 in the training, internal validation, and external validation cohorts, respectively, outperforming the MRI-only and endoscopy-only models. For DFS prediction, the corresponding AUCs were 0.853, 0.823, and 0.794. The model showed acceptable calibration and higher net clinical benefit across clinically relevant threshold probabilities. Kaplan-Meier analysis showed significant separation of DFS between the model-defined high-risk and low-risk groups (log-rank p < 0.01). A multimodal MRI-endoscopy deep learning model showed robust performance for predicting pCR and DFS after nCRT in patients with LARC. By integrating intratumoral and mucosal information, this approach may support preoperative response assessment and individualized risk stratification. Not applicable. Preoperative prediction of pathologic complete response and disease-free survival remains challenging in locally advanced rectal cancer. MRI–endoscopy multimodal deep learning consistently outperformed single-modality models in predicting pathologic complete response and disease-free survival risk. Multimodal deep learning may guide individualized treatment and organ-preserving strategies by improving preoperative response assessment and risk stratification in rectal cancer patients.
Microbeam radiation therapy (MRT) delivers spatially fractionated, high-dose radiation in parallel beam arrays, creating spatial fractionation by alternating peak and valley dose regions. While this approach offers significant potential for tumor control with reduced normal tissue toxicity, the precise biological dose and damage distribution, particularly the relative contributions of direct and non-targeted (bystander) effects, remain incompletely understood. We employed complementary biodosimetric approaches to quantify DNA damage responses in two-dimensional CHO-K1 cell monolayers and three-dimensional epidermis tissue models following MRT. Biological dosimetry in CHO-K1 monolayers was performed using the cytokinesis-block micronucleus (CBMN) assay for micronuclei (MN) and nucleoplasmic bridges (NPB). In epidermis models, DNA double-strand break (DSB) damage was analyzed by immunofluorescence (IF) with γH2AX+53BP1 DSB markers. Spatial and temporal mapping of damage in tissue sections was performed across microbeam peak and valley regions and compared with physical dosimetry and broadbeam irradiation. In 2D CHO-K1 cells, MN- and NPB-based biological dosimetry closely matched physical dose measurements, with accurate dose reconstruction in valleys but substantial underestimation in peaks due to extensive cell death and loss of scorable nuclei. In 3D epidermis models, MRT induced distinct spatial patterns of DNA damage with pan-γH2AX–positive cells indicating excessive DSB formation at peak regions at early time points, whereas 53BP1 DSB foci predominated in dose valleys. Over 24 h, pan-γH2AX signals declined to sham levels, while 53BP1 foci persisted, particularly in peak regions, indicating DNA damage regions refractory to repair. Comparative analyses revealed that valley cells exposed to 1 Gy MRT radiation exhibited in average significantly more 53BP1 foci at 0.5 h than after uniform 1 Gy broadbeam irradiation, suggesting a bystander contribution from adjacent high-dose peaks. By 24 h, however, damage levels converged between MRT valleys and broadbeam, indicating that bystander-associated excess damage was transient. CBMN-based biological dosimetry reliably quantifies absorbed dose in MRT valleys but is limited in peak regions. 3D epidermis models enable spatial and temporal mapping of DNA damage, revealing persistent complex damage in peak-region cells and provide evidence for transient bystander effects in valley regions. Extending previously established biodosimetric and DNA damage mapping approaches to X-ray MRT, these findings, obtained in one cell line and one tissue model, contribute to the understanding of DSB induction, repair dynamics, and non-targeted effects in MRT.
