AI and Cancer Research ── 2026-10-18
New papers on cancer and AI, collected from journal feeds, PubMed (searched across all journals) and OpenAlex (a public index used instead of Google Scholar), judged by Jev (an AI that makes language judgements): original research, AI central, which of 13 themes, treatment or trials, and then ranked by importance up to a daily limit. Journals are not filtered by impact factor; each paper shows its journal's IF-equivalent (OpenAlex two-year mean citedness) and the date of that value. Each entry shows the full original abstract.
Read by cancer type: Cancer biology（4） · Breast（3） · Lung（2） · Pancreatic（1） · Prostate（2） · ACC（0） · Brain（2） · Ovarian（0） · Endometrial（0） · Gastric（1） · Liver（0） · Kidney（2） · Bladder（0）
Cancer molecular biology and AI
- Abstract
Accurate and rapid prediction of transcranial magnetic stimulation (TMS)-induced electric fields is essential for personalized clinical applications, from individualized therapy in major depressive disorder to presurgical functional mapping in neuro-oncology. Yet conventional finite element methods are often too computationally expensive for interactive use. Existing deep learning surrogate models partly address this limitation but are largely constrained to fixed voxel-grid representations that cannot naturally accommodate the irregular geometry of subject-specific head meshes. We present the first physics-informed neural operator (PINO) for subject-specific prediction of vector TMS electric fields, built on a mesh-native dual-encoder architecture that jointly encodes head anatomy and coil configuration. By incorporating a curl-based regularization term aligned with the curl-free nature of the TMS secondary field, the PINO model promotes physical consistency as a soft regularizer alongside data-driven supervision. Trained on 75 subjects from the IXI dataset, the framework was evaluated across four independent held-out test datasets, including healthy subjects from two additional MRI cohorts and, for the first time in learning-based TMS electric-field prediction, tumor patients. The resolution-agnostic formulation generalizes across subjects, coil types, and spatial resolutions without retraining, predicting vector electric fields on approximately 300,000 gray matter volume nodes in under 100 ms on a consumer-grade GPU. Across all test datasets within the focality-50% region, the PINO model achieved median center-of-gravity errors of 0.39–0.43 mm, RMSE of 0.12–0.16 V/m, and mean angular deviations of 6.7°–7.3°, with performance within or below the error ranges reported by recent deep-learning approaches. At the selected 5% physics-loss weighting, curl loss decreased by 9.1% relative to the data-only model, while RMSE and MAD increased by only 1.6% and 1.3%, respectively. These results suggest that physics-informed neural operators provide a scalable technical foundation for rapid gray-matter electric-field approximation, supporting future studies on real-time electric-field-guided TMS workflows
Journal IF-equivalent: 5.2 (OpenAlex 2-year mean citedness, value as of 2026-10-04, retrieved 2026-10-05; not the official Clarivate IF)Reference: Moser P, Greifeneder F, Freinberger D, Reishofer G, Ritter P, Frank L, et al. Physics-informed neural operators for rapid electric field prediction in transcranial magnetic stimulation of healthy and tumor subjects. Sci Rep. 2026 Oct 16 [Epub ahead of print]. doi:10.1038/s41598-026-74936-z.Checked: Abstract only - Abstract
Missed genetic deletions can prevent cancer patients from receiving effective targeted therapy. Standard pipelines, such as TSO500, use bin count averaging for copy number variation detection but often overlook focal events. To address this, we developed a machine learning (ML) pipeline for deletion detection and zygosity evaluation within the MASTERsg precision oncology program. We focused on the tumor suppressor gene PTEN, whose loss confers sensitivity to mTOR/Akt inhibition. An Extreme Gradient Boosting model trained on PTEN-deleted and wild-type tumor samples identified focal deletions missed by the TSO500 pipeline and CNVkit that were associated with concordant PTEN protein loss by immunohistochemistry. Importantly, the model detected PTEN deletions in 33/204 profiled samples that were classified as PTEN wild-type by TSO500; 75% (15/20) of stainable cases showed absent or reduced PTEN protein expression. These findings demonstrate the potential of ML to enhance panel sequencing, uncovering therapeutically relevant alterations missed by established diagnostic methods.
Journal IF-equivalent: 8.9 (OpenAlex 2-year mean citedness, value as of 2026-10-04, retrieved 2026-10-05; not the official Clarivate IF)Reference: Kerle IA, Husak O, Gross T, Hartig A, Prause R, Richter D, et al. Machine learning uncovers previously undetectable genetic deletions in cancer panel sequencing. npj Precis Onc. 2026 Oct 16;10(1):397. doi:10.1038/s41698-026-01756-9.Checked: Abstract only
See all 4 Cancer biology papers →
Breast cancer: treatment, trials and AI
- Abstract
Distant metastasis is a major cause of breast cancer mortality. Early prediction may support risk-adapted surveillance and treatment planning. We evaluated imaging-based machine learning (ML) and deep learning (DL) models for predicting future distant metastasis in patients without baseline metastatic disease. PubMed, Scopus, Web of Science, and supplementary Google Scholar searches were performed through June 2025. Eligible studies evaluated imaging-based AI prediction models and reported quantitative performance. The primary outcome was the area under the receiver operating characteristic curve (AUC). Risk of bias was assessed using PROBAST + AI. Random-effects meta-analysis, exploratory subgroup analyses, and univariable meta-regression were performed. Nine studies contributed 43 model evaluations. The pooled model-level AUC was 0.80 (95% CI: 0.77–0.84). A secondary analysis restricted to the best-performing model from each study yielded an AUC of 0.89 (95% CI: 0.83–0.95). CT-based models were associated with higher AUCs than MRI-based models (0.83 vs. 0.71), and models incorporating clinicopathological features had higher AUCs than imaging-only models (0.91 vs. 0.74). Exploratory meta-regression identified associations with imaging modality, follow-up duration, and inclusion of clinicopathological features. Heterogeneity was substantial, and five of nine studies had high risk of bias. Imaging-based ML and DL models show potential for predicting future distant metastasis. In the adjuvant setting, externally validated models may support risk stratification and individualized surveillance, but current evidence is insufficient to justify surveillance or treatment changes based solely on AI predictions. Prospective external validation and transparent reporting are needed before clinical implementation.
Journal IF-equivalent: 2.5 (OpenAlex 2-year mean citedness, value as of 2026-10-17, retrieved 2026-10-18; not the official Clarivate IF)Reference: Dehghani SS, Zakeri A, Zahedinejad S, Nafisi M, Habibi S, Vardanjani HM. Performance of machine learning (ML) and deep learning (DL) algorithms in prediction of breast cancer distant metastasis via imaging: a systematic review and meta‑analysis. Langenbecks Arch Surg. 2026 Oct 16 [Epub ahead of print]. doi:10.1007/s00423-026-04344-4.Checked: Abstract only - Abstract
Objective: To evaluate the incremental predictive value of pretreatment tri-modal ultrasound deep learning integrating two-dimensional ultrasound (2DUS), color Doppler flow imaging (CDFI), and contrast-enhanced ultrasound (CEUS) for predicting breast pathologic complete response (bpCR) after neoadjuvant chemotherapy (NAC), and to assess performance stability across independent centers.
Methods: This retrospective multicenter study included 1089 patients with primary invasive breast cancer. Center 1 provided the training cohort (n = 572) and internal validation cohort (n = 246), whereas centers 2 (n = 160) and 3 (n = 111) served as independent external validation cohorts. Using pretreatment 2DUS, CDFI, and CEUS within a standardized intratumoral and 10-mm peritumoral framework, single-modality, all three dual-modality combinations, and tri-modal models were developed. The best-performing dual-modality model was identified and locked during training as the primary imaging comparator. A Tri-modal + Clinical model further integrated pretreatment clinical and biomarker variables. Performance was assessed using AUC, Brier score, calibration, decision curve analysis, and ablation experiments.
Results: Overall, 383 patients (35.2%) achieved bpCR. CEUS was the strongest single modality, and 2DUS + CEUS was the best-performing dual-modality model. Compared with 2DUS + CEUS, the tri-modal model increased the AUC from 0.895 to 0.919 in the combined external cohort (ΔAUC, +0.023; 95% CI, +0.010 to +0.038; p = 0.001), with corresponding improvements in Brier score (0.124 to 0.105) and F1 score (0.747 to 0.786). The AUC point estimate increased in both external centers, although the 95% CI for the incremental AUC included zero in external validation cohort 2. Despite modest standalone discrimination, CDFI provided measurable incremental information when added to 2DUS + CEUS. Further integration of pretreatment clinical and biomarker variables increased the combined external AUC to 0.935. Reduced-input retraining showed performance decreases after omission of any modality, with the largest reduction after CEUS omission.
Conclusions: Pretreatment tri-modal ultrasound deep learning provided incremental predictive value for bpCR beyond the best dual-modality model, with a consistent direction of benefit across two independent external centers. Clinical and biomarker integration provided additional predictive information, supporting further evaluation of this multimodal framework for individualized pretreatment bpCR prediction.
Journal IF-equivalent: 5.3 (OpenAlex 2-year mean citedness, value as of 2026-10-07, retrieved 2026-10-08; not the official Clarivate IF)Reference: Wei C, Pan T, Li T, Zheng C, Duan Y, Nie F. Incremental Value and Cross-Center Validation of Pretreatment Tri-Modal Ultrasound Deep Learning for Predicting Breast Pathologic Complete Response to Neoadjuvant Chemotherapy: A Multicenter Study. Cancers. 2026 Oct 15;18(20):3326. doi:10.3390/cancers18203326.Checked: Abstract only
Lung cancer: treatment, trials and AI
- Abstract
Chemoimmunotherapy (CIT) is the standard first-line regimen for advanced lung squamous cell carcinoma (LUSC), yet clinical benefit varies substantially across patients, and available biomarkers remain inadequate for treatment selection. We aimed to establish and externally validate a whole-slide image (WSI)-based graph convolutional network (GCN) pathomics model integrated with clinical variables for objective response prediction and clinically relevant stratification of CIT benefit.
Journal IF-equivalent: 1.0 (OpenAlex 2-year mean citedness, value as of 2026-10-16, retrieved 2026-10-17; not the official Clarivate IF)Reference: Wang D, Mu S, Zhang M, Tao G, Wang S, Shen Y, et al. P3.062: GCN Pathomics-Clinical Model Predicts Response to First-Line Chemoimmunotherapy in Advanced LUSC: A Multicenter Study. Journal of Thoracic Oncology. 2026 Oct;21(10):105231. doi:10.1016/j.jtho.2026.105231.Checked: Abstract only - Abstract
Purpose: To develop and evaluate an interpretable, multimodal deep learning model to improve postoperative risk stratification and predict adjuvant therapy benefit in stage I non-small cell lung cancer (NSCLC). Materials and Methods In this multicenter retrospective study of patients with stage I NSCLC, a multimodal model called the Deep Learning-Radiomic-Text Transformer (DLRTransformer) was developed by integrating multiview 2.5-dimensional deep learning features from CT scans, radiomics, clinical variables, and prognostic embeddings extracted from pathology reports using a large language model. A central transformer architecture fuses these features to predict disease-free survival. Model performance was evaluated on independent external datasets and compared with guideline-based risk stratification. Associations between model-derived risk scores and tumor microenvironment features were analyzed to assess biologic interpretability.
Results: A total of 1073 patients (mean age ± SD, 60.8 years ± 9.8; 550 male) were enrolled from five centers. On external datasets, DLRTransformer demonstrated robust prognostic performance (concordance index, 0.77-0.78), outperforming the guideline-based risk stratification (concordance index, 0.60-0.65). Risk scores were associated with tumor microenvironment phenotypes, including increased neoplastic cellularity and reduced tertiary lymphoid structures. A treatment-by-risk interaction was observed (interaction hazard ratio [HR], 0.14; 95% CI: 0.06, 0.32; P P = .02) but not in low-risk patients (HR, 1.70; 95% CI: 0.97, 2.99; P = .06).
Conclusion: The DLRTransformer improved postoperative risk stratification compared with guideline-based risk stratification and identified patients with stage I NSCLC who benefited from adjuvant therapy. Keywords: Deep Learning, Large Language Models LLMs, CT, Prognosis & Prediction, Non-Small Cell Lung Cancer (NSCLC), Multimodal Deep Learning, Risk Stratification, Adjuvant Therapy, Tumor Microenvironment (TME) Supplemental material is available for this article. © The Author(s) 2026. Published by the Radiological Society of North America under a CC BY 4.0 license.
Journal IF-equivalent: 2.2 (OpenAlex 2-year mean citedness, value as of 2026-10-16, retrieved 2026-10-17; not the official Clarivate IF)Reference: Wang L, Zong Y, Li X, Tang H, Yao J, Cao Y, et al. Multimodal Transformer Integrating CT and Pathology for Postoperative Risk Stratification in Stage I Non–Small Cell Lung Cancer. Radiology: Imaging Cancer. 2026 Nov 1;8(6):e250617. doi:10.1148/rycan.250617.Checked: Abstract only
Pancreatic cancer: treatment, trials and AI
- Abstract
Pancreatic cancer (PaC), the third leading cause of cancer related deaths in the United States, is a highly heterogenous malignancy with a multitude of potential subtypes. Each PaC subtype has unique molecular pathways and therapeutic characteristics, which can be used in downstream patient risk stratification and tailored treatment design. Conventional wet‐lab approaches for PaC subtyping (microdissection, histopathological studies, etc.) are laborious, costly, and time‐consuming. To address these concerns, we present MetaPaCS, a novel meta‐learning framework to accurately identify PaC subtypes based on transcriptomics data. Specifically, after preprocessing, the transcriptome‐based feature vectors are classified by multiple base machine learning classifiers, whose diverse prediction outputs are combined with the initial preprocessed feature vectors to constitute a new set of ensemble feature vectors for a meta‐learning classifier. Leave‐one‐out cross‐validation demonstrates that MetaPaCS performs remarkably better optimized individual base classifiers. We believe that MetaPaCS is a promising tool for downstream risk stratification and personalized treatment design.
Journal IF-equivalent: 5.6 (OpenAlex 2-year mean citedness, value as of 2026-10-09, retrieved 2026-10-10; not the official Clarivate IF)Reference: Peterson NB, Sun M, Wu X, Wang J, Wan S. MetaPaCS: A Novel Meta‐Learning Framework for Pancreatic Cancer Subtype Identification. Advanced Intelligent Systems. 2026 Oct 15 [Epub ahead of print]. doi:10.1002/aisy.70576.Checked: Abstract only
Prostate cancer: treatment, trials and AI
- Abstract
Accurate assessment of different types of prostate cancer (PCa) lesions is a crucial part of treatment planning. Following a computed tomography (CT) or a positron emission tomography (PET)/CT imaging scan, a significant amount of time is required from the clinicians to detect lesions and characterize them by their anatomical locations. To facilitate the development of new image analysis applications, we investigated the potential of automated classification of PCa lesions between bone metastases, lymph node metastases, local lesions, and other PCa lesions in CT images via machine learning (ML) and deep learning models. We annotated a total of 4949 PCa lesions by their type in a private dataset of 18 F-prostate-specific membrane antigen (PSMA) PET/CT images of 532 PCa patients and an external test set of publicly available PSMA PET/CT images from 30 patients. Using information of coordinates of the detected lesions, we automatically extracted 15 different features from the CT and the results of the automated 3D anatomical segmentation tool TotalSegmentator for each lesion, and trained several traditional ML models to classify lesions by their anatomical type. Additionally, we experimented with convolutional neural networks (CNNs) of different dimensionalities with CT and TotalSegmentator inputs. According to our results, the most accurate model was XGBoost with an overall accuracy of 96.0%. It identified bone metastases, lymph node metastases, and local lesions with sensitivities of 97.9±1.2%, 95.2±1.3%, and 93.1±2.8% in the five-fold cross-validation experiments of the primary data and 96.3%, 94.0%, and 80.0% on the external test set, respectively. Conversely, the CNNs underperformed the best ML models in distinguishing local lesions. To conclude, a simple ML model can accurately classify PCa lesions between bone, lymph node, and local lesions based on CT images, lesion locations, and automated anatomical segmentation data, but more research is required for distinguishing PCa lesions outside these categories.
Journal IF-equivalent: 3.9 (OpenAlex 2-year mean citedness, value as of 2026-10-13, retrieved 2026-10-15; not the official Clarivate IF)Reference: Rainio O, Vehmanen A, Li A, Jaakkola MK, Kemppainen J, Klén R. Investigation on Automatic Classification of Bone, Lymph Node, and Local Lesions in CT Images of Prostate Cancer Patients. J Digit Imaging Inform med. 2026 Oct 16 [Epub ahead of print]. doi:10.1007/s10278-026-02402-3.Checked: Abstract only - Abstract
Large language models show great promise for clinical decision support but generally lack medical imaging integration. Therefore, vision–language models have been proposed, but there is little evidence on their performance, particularly compared with conventional machine learning. Prostate-specific membrane antigen (PSMA)–targeted radiopharmaceutical therapy is a well-established treatment for metastatic castration-resistant prostate cancer (mCRPC), but it is still not known which patients will be good long-term responders. Therefore, this study proposes a physician-inspired visual question answering (VQA) framework that emulates the structured integration of PSMA-directed imaging and clinical data used in routine clinical assessment for overall survival prognostication.
Methods: As a proof of concept, we conducted a retrospective, multicenter study across 2 academic institutions that included 110 patients with mCRPC who were treated with either [177Lu]Lu-PSMA I&T or [177Lu]Lu-PSMA-617. Our VQA approach incorporates a 3-dimensional image encoder, the domain-specific language model RadBERT, and a cross-modal transformer to process baseline and interim (after 2 cycles) PET and SPECT images, clinical data, and clinical query. The model was trained to predict 24-mo survival, and performance was compared with state-of-the-art approaches. External validation was performed with the [177Lu]Lu-PSMA-617 cohort.
Results: The VQA model outperformed all state-of-the-art approaches while showing accuracy comparable to the query-free attention multimodal fusion model. Specifically, it achieved 81.8% accuracy (95% CI, 0.52–0.95) when integrating SPECT or PET images with prostate-specific antigen measurements. External validation of the VQA showed 71.9% accuracy (95% CI, 0.53–0.86) after few-shot fine-tuning, with further improvement over all state-of-the-art approaches in survival stratification. The VQA approach showed the closest calibration with the ground-truth survival distribution, with no significant differences observed (log-rank test, P = 0.76). Receiver operating characteristic analysis showed the highest area under the curve (0.75) among the evaluated models. Together, these findings suggest that the VQA approach demonstrated adaptability across different PSMA ligands while maintaining slightly improved yet comparable overall survival prognostic performance.
Conclusion: By adopting a physician-inspired assessment workflow, the VQA framework shows advantages in integrating imaging and laboratory data for clinical decision support compared with conventional machine learning. The preliminary results indicate that VQA-based approaches warrant future evaluation for survival prognostication in patients with mCRPC who are receiving PSMA radiopharmaceutical therapy.
Journal IF-equivalent: 6.9 (OpenAlex 2-year mean citedness, value as of 2026-10-17, retrieved 2026-10-18; not the official Clarivate IF)Reference: Kassar M, Hu J, Rauscher I, Gomes CV, Wendler T, Ziegler S, et al. Multimodal Visual Question Answering System for Survival Prognosis in PSMA-Targeted Radiopharmaceutical Therapy. J Nucl Med. 2026 Oct 15 [Epub ahead of print]. doi:10.2967/jnumed.125.271912.Checked: Abstract only
Adenoid cystic carcinoma: treatment, trials and AI
No new qualifying paper for this issue.
Brain tumours: treatment, trials and AI
- Abstract
Introduction: High-grade gliomas (HGGs) are aggressive primary brain tumors with high recurrence rates and poor prognosis. Survival benefit from adjuvant temozolomide (TMZ) varies substantially across patients, and reliable preoperative biomarkers for treatment stratification remain limited. This study aims to develop an interpretable multimodal deep learning model integrating whole-slide histopathological images and multiparametric MRI to preoperatively predict survival benefit from adjuvant TMZ in HGG patients.
Methods: A Transformer-based multimodal fusion model was proposed. Key tumor regions were identified from whole-slide pathology images via weakly supervised clustering, and pathological features were extracted using attention-based multiple instance learning. Concurrently, 2.5D tumor subregional features were derived from multiparametric MRI sequences. These multimodal features were deeply fused using a Transformer encoder. The model was developed and validated on a multicenter cohort (504 patients), and its biological basis was explored through transcriptomic analysis, gene set enrichment analysis, protein-protein interaction networks, and immune deconvolution.
Results: The fusion model achieved AUCs of 0.910, 0.861, and 0.833 in the training, internal validation, and external validation cohorts, respectively. Transcriptomic analysis implicated DNA damage repair processes in high-benefit patients, suggesting that enhanced DNA repair capacity may contribute to TMZ sensitivity. Immune deconvolution showed higher M1 macrophage infiltration and lower regulatory T cell abundance in the high-benefit group, consistent with a less immunosuppressive tumor microenvironment. Importantly, low-benefit patients who completed TMZ chemotherapy had shorter survival than those treated with concurrent chemoradiotherapy alone or surgery alone.
Conclusion: This model identifies patients unlikely to benefit from standard TMZ chemotherapy, providing a tool for postoperative personalized treatment stratification and optimization of therapeutic strategies in HGGs.
Journal IF-equivalent: 5.5 (OpenAlex 2-year mean citedness, value as of 2026-10-02, retrieved 2026-10-03; not the official Clarivate IF)Reference: Li Z, Duan X, Lian J, Li X, Liang Q, Liu X, et al. Pathology-guided multimodal transformer predicts postoperative survival benefit from adjuvant chemotherapy in glioma: from DNA damage repair to immune characterization. Front Immunol. 2026 Oct 16;17:1845251. doi:10.3389/fimmu.2026.1845251.Checked: Abstract only - Abstract
Diffuse Gliome vom adulten Typ sind die häufigsten malignen Gehirntumoren bei Erwachsenen und mit einer sehr schlechten Prognose assoziiert. Für eine optimale und personalisierte Diagnostik und Therapie der Gliome ist die neuroradiologische Diagnostik von elementarer Bedeutung. Die Kontrasterzeugung durch CEST in der metabolischen MRT ist eine vielversprechende Technik zur Untersuchung von Gehirntumoren, ihren molekularen Eigenschaften und zur Therapieevaluation. Für die Betrachtung und Auswertung der CEST-Bildgebung ist ein exakt segmentiertes Tumormodell erforderlich. Die Segmentierung des Tumors für jeden individuellen Patienten ist zeitintensiv und erfordert neuroradiologische Expertise. Eine automatisierte Tumorsegmentierung durch Deep learning gestützte Methoden gewinnt daher stark an Bedeutung. [AG96.1] Aus diesem Grund wurde zum einen die automatisierte Segmentierung durch das BraTS-Toolkit mit einer manuell erstellten, validierten Segmentierung mit dem Ziel verglichen, die Qualität, Einsatzbereiche und Anwendbarkeit zu evaluieren. Zum anderen wurden die Intensitätslevel in der CEST-Bildgebung in den unterschiedlichen Tumorarealen (Nekrose/Zyste, KME, Ödem) betrachtet. Es sollte hierbei gezeigt werden, wie sich die APT- und NOE-gewichteten CEST-Aufnahmen in Bezug auf WHO-Grad, IDH-Status und Zeitpunkt der Bildgebung unterscheiden. Ziel war es hier zu untersuchen, ob die CEST-Aufnahmen in dieser Studiengruppe einen Beitrag zur Therapieentscheidung liefern können. Für diese Studie wurden 51 Patientendaten verwendet. Diese unterteilen sich in 36 Datensätzen von Patienten mit Gliomen und 15 von Patienten mit tumorähnlicher Läsion[AG97.1]. Die MRT-Aufnahmen dieser Patienten wurden bei einer klinischen Feldstärke von 3 T durchgeführt. Die APT-gewichteten CEST-Aufnahmen wurden bei einer Sättigung von B1 = 2,22 µT durchgeführt, die NOE-gewichteten CEST-Aufnahmen bei einer Sättigung von B1 = 0,6 µT. Die automatisierte Segmentierung der Gliome und Läsionen erfolgte durch das BraTS-Toolkit und wurden anschließend manuell mit dem Programm 3D Slicer korrigiert[AG98.1]. Für die Auswertung der Intensitätslevel der CEST-Aufnahmen wurden mit Hilfe von 3D Slicer die Segmentierungen zunächst auf den Bereich der CEST-Aufnahme zugeschnitten. Danach wurden diese Segmentierungen [AG99.1]an CEST-Artefakte angepasst, ein Referenzsegment in gesund erscheinender weißer Substanz auf der kontralateralen Hemisphäre eingefügt und für weitere Untersuchungen ein Fusionssegment (KME + Flair Hyperintensität) erstellt. Abschließend konnten die Intensitätswerte mit 3D Slicer gemessen werden. Bei der Evaluation der automatisierten Segmentierung lag der Median der mittleren Dice Koeffizienten bei posttherapeutischen Patienten bei 0,8 und der Median der m. D. c. bei tumorähnlichen Läsionen bei 0,81. Jedoch wich der Median der mittleren Dice Koeffizienten bei Patienten mit Erstdiagnose mit einem Wert von 0,64 hiervon ab. Bei den Patienten mit Gliomen traten vor allem in den nekrotischen/zystischen Arealen gehäuft niedrige Dice Koeffizienten auf (siehe Abb. 18[AG100.1]), was hauptsächlich an zu groß automatisch segmentierten nekrotisch/zystischen Arealen lag (exemplarisch in Abb. 14 zu sehen). Für die Abhängigkeit der Segmentierung von WHO-Grad und Diagnose zeigten sich bessere Dice Koeffizienten bei höhergradigen Astrozytomen und Glioblastomen mit KM-Enhancement. Bei der Betrachtung der Intensitätslevel des APT zeigte sich, dass die Intensitätslevel grundsätzlich höher waren als die der Referenzsegmente (Abb. 54). Die APT-Intensitätslevel waren zudem in der Gruppe Erstdiagnose höher als in der posttherapeutischen Gruppe und bei Wildtyp-Tumoren tendenziell höher als bei IDH-mutieren Tumoren (Abb. 57). In der posttherapeutischen Gruppe konnte beobachtet werden, dass sich die APT-Intensitätslevel analog zum WHO-Grad steigend verhalten (Abb. 34). Die Intensitätslevel des NOE waren durchgängig niedriger als die der Referenzsegmente (Abb. 55). Bei IDH-mutierten Tumoren zeigten sich die NOE-Intensitätslevel niedriger als beim Wildtyp [AG101.1](Abb. 48). Eine Tendenz der NOE-Intensitätslevel konnte in Bezug auf WHO-Grad oder Therapie-Status nicht gezeigt werden. Schlussfolgernd lässt sich feststellen, dass die automatisierte Segmentierung gute Ergebnisse erzielt [AG102.1]hat, jedoch bei morphologisch atypischen Tumoren ohne KM-anreicherndes Areal mit hohem T2 Signal Probleme bei der Segmentierung des nekrotischen/zystischen Areals hatte. Die CEST-Bildgebung stellt vielversprechende und interessante Kontraste bereit. Vor allem die APT-gewichteten Aufnahmen zeigten eindeutige Tendenzen bezüglich Therapie- und Mutationsstatus der Patienten, sind jedoch durch die vorhandene Streuung noch nicht zur individuellen Therapieentscheidung geeignet.
Journal IF: unknown (could not be matched)Reference: [No authors listed]. Qualität und Anwendbarkeit von automatisierten Segmentierungen von Hirntumoren bei der Evaluation des Nutzens von APT- und NOE-CEST bei 3 T bei individuellen Patienten. Universitätsbibliothek Tübingen. [Epub ahead of print]. doi:10.15496/publikation-125797.Checked: Abstract only
Ovarian cancer: treatment, trials and AI
No new qualifying paper for this issue.
Endometrial cancer: treatment, trials and AI
No new qualifying paper for this issue.
Gastric cancer: treatment, trials and AI
- Abstract
Deep learning (DL) applied to haematoxylin and eosin (H&E)-stained pathology images can predict microsatellite instability (MSI) in gastric cancer (GC). We evaluated its diagnostic accuracy and discriminative performance. PubMed, Embase, Cochrane Library and Web of Science were searched for studies applying DL to H&E images for MSI prediction in GC. The protocol was prospectively registered in PROSPERO (CRD420261503140). Risk of bias was assessed with QUADAS-2. Pooled sensitivity and specificity were estimated with a bivariate random-effects model, with subgroup analyses by validation strategy and data source; slide-level areas under the curve (AUCs) were pooled after logit transformation. Prespecified scenario analyses assessed robustness. Ten studies were included; seven independent datasets from six studies entered the primary synthesis. Pooled sensitivity, specificity and summary receiver operating characteristic AUC were 0.78 (95% CI 0.63–0.88), 0.85 (95% CI 0.77–0.90) and 0.89 (95% CI 0.86–0.91); nine studies contributed to the secondary AUC synthesis (pooled AUC 0.85, 95% CI 0.80–0.89). Hospital-derived cohorts showed numerically higher sensitivity (0.86) and diagnostic odds ratio (39.8) than TCGA-derived cohorts (0.71 and 11.5), but subgroup differences were limited by the small number of studies and substantial heterogeneity. DL models show good diagnostic accuracy for MSI prediction in GC, but clinical utility still requires confirmation in standardised multicentre, independent external and prospective studies.
Journal IF: unknown (could not be matched)Reference: Du Y, Liao H, Li J, He Y, Di J. Diagnostic accuracy of deep learning for microsatellite instability prediction in gastric cancer: A systematic review and meta-analysis. medRxiv. 2026 Oct 15 [Epub ahead of print]. doi:10.64898/2026.10.12.26365351.Checked: Abstract only
Liver cancer: treatment, trials and AI
No new qualifying paper for this issue.
Kidney cancer: treatment, trials and AI
- Abstract
This study aimed to develop a deep learning model, a peritumoral radiomics model, a habitat radiomics model, and a conventional clinical-imaging model, and to explore their value in the preoperative differentiation of common pathological subtypes of small renal cell carcinoma (sRCC). A total of 143 patients with pathologically confirmed sRCC (92 clear cell RCC and 51 non-clear cell RCC) were retrospectively included. Patients were randomly divided into a training set ( n = 101) and a validation set ( n = 42) at a ratio of 7:3. Corticomedullary phase images were used for analysis. Tumor regions of interest (ROIs) were manually delineated using ITK-SNAP software, and peritumoral ROIs were generated by uniform 2-mm expansion. Unsupervised K-means clustering was performed on the intratumoral ROI based on local features to identify intratumoral subregions, and the optimal number of subregions was determined using the maximum Silhouette Coefficient. Radiomics features were extracted from peritumoral region and each subregion. The maximum cross-section of the tumor was obtained. Transfer learning was then applied to initialize the ResNet18 model with pre-trained parameters from the ImageNet dataset, and deep transfer learning features were subsequently extracted. Feature selection and dimensionality reduction were conducted using the max-relevance and min-redundancy (mRMR) algorithm and the least absolute shrinkage and selection operator (LASSO) method, with cross-validation applied within the training set. Univariate and multivariate logistic regression analyses were used to identify the independent clinical predictors. Based on the identified independent clinical predictors and the selected features, clinical, peritumoral, habitat, and deep learning models were developed. The model with the highest performance was combined with independent clinical predictors to build the combined model. The performance of the models was evaluated using the area under the curve (AUC), accuracy, sensitivity, and specificity, and clinical utility was assessed using decision curve analysis. The optimal number of clusters was 3. Logistic regression analysis identified cystic or necrotic areas and enhancement characteristics as independent clinical predictors. A total of 8 peritumoral radiomics features, 12 habitat radiomics features, and 7 deep learning features were ultimately selected. The deep learning model (AUC: 0.960, 95% CI: 0.927–0.993 in the training set; AUC: 0.968, 95% CI: 0.926–1.000 in the validation set) outperformed the clinical model (AUC: 0.798, 95% CI: 0.714–0.883; AUC: 0.794, 95% CI: 0.647–0.940), the peritumoral model (AUC: 0.885, 95% CI: 0.815–0.955; AUC: 0.886, 95% CI: 0.776–0.997), and the habitat model (AUC: 0.951, 95% CI: 0.904–0.998; AUC: 0.943, 95% CI: 0.881–1.000). The performance of the combined model, established by integrating the deep learning model with the clinical model, was further improved (AUC: 0.968, 95% CI: 0.939–0.997 in the training set; AUC: 0.973, 95% CI: 0.934–1.000 in the validation set). Decision curve analysis demonstrated that the combined model provided the highest net benefit. The radiomics and deep learning models based on contrast-enhanced CT enabled effective noninvasive differentiation of sRCC subtypes. Habitat radiomics can capture intratumoral heterogeneity and shows favorable diagnostic performance, while deep learning demonstrates superior discriminative ability. These approaches may assist in preoperative decision-making, although further validation in larger external cohorts is warranted.
Journal IF-equivalent: 2.4 (OpenAlex 2-year mean citedness, value as of 2026-10-16, retrieved 2026-10-17; not the official Clarivate IF)Reference: Liu Q, Wang Y, Duan B, Liu M, Ye Y, Tang Q, et al. Deep learning and habitat radiomics based on contrast-enhanced CT for differentiating pathological subtypes of small renal cell carcinoma. BMC Urol. 2026 Oct 15 [Epub ahead of print]. doi:10.1186/s12894-026-02417-6.Checked: Abstract only - Abstract
Clear cell renal cell carcinoma (ccRCC), the most common renal cell carcinoma subtype (60%-85%), remains challenging due to suboptimal TNM-based risk stratification and limited precision in adjuvant targeted therapy (5.4%-10.5% survival benefit with overtreatment risk). Perirenal fat, linked to ccRCC prognosis via volume and density, is underexplored in integration with radiogenomics (tumor molecular features) for comprehensive prognosis. Retrospective analysis included a training cohort (Xijing Hospital, n = 226) and testing cohort (TCIA/TCGA, n = 134) of resectable ccRCC patients with preoperative contrast-enhanced CT and exome sequencing data. CT-based tumor/perirenal fat radiomic features were extracted via 3D Slicer/PyRadiomics, with scores built using LASSO and stepwise Cox regression. Genetic scores derived from 639 cancer-related genes (top 20 mutated genes) and clinical scores (e.g., TNM stage) were constructed via Cox regression. A multimodal nomogram integrating these scores was validated using C-index, time-dependent AUC, and SHAP analysis. The nomogram outperformed unimodal scores, with C-indices of 0.856 (training) and 0.859 (testing). Time-dependent AUCs for 1-5-year PFS were consistently higher. SHAP identified tumor/perirenal fat scores, TP53 mutation, and grade G4 as key contributors.Multiple machine learning algorithms were used to predict four genes with prognostic value, and their correlations in multimodal data were analyzed. Subgroup analyses confirmed robust risk stratification. Patients with targeted therapy risk difference rate ≥ 47% had longer PFS with therapy (HR = 0.545, p = 0.038), while those with < 10% fared better without therapy (HR = 2.051, p = 0.035), with similar trends in high-risk subgroups. Targeted therapy-related analyses are exploratory and require prospective validation. This CT-based multimodal nomogram integrates radiomic, genomic, and clinical data to characterize ccRCC across scales, providing auxiliary reference for screening patients who may gain potential survival benefit from adjuvant targeted therapy, reducing overtreatment/undertreatment.
Journal IF-equivalent: 3.8 (OpenAlex 2-year mean citedness, value as of 2026-10-02, retrieved 2026-10-03; not the official Clarivate IF)Reference: Zhang YF, Ji YH, Wen DD, Liu C, Shi PW, Hou HZ, et al. CT-based multimodal radiogenomic model integrating tumor and perirenal fat features for prognosis in clear cell renal cell carcinoma. BMC Med Imaging. 2026 Oct 15 [Epub ahead of print]. doi:10.1186/s12880-026-02920-0.Checked: Abstract only
Bladder cancer: treatment, trials and AI
No new qualifying paper for this issue.