6 results about "Tumor size" patented technology

This invention discloses a two-stage controllable liver CT image generation method combining GAN and Diffusion models. The first step involves segmenting a real CT image using a pre-trained nnU-Net to obtain liver and tumor masks. The second step involves training a VQ-VAE model to establish a mapping between pixel space and latent space; training a conditional LDM to perform incisive generation of liver regions: predicting noise under the guidance of multiple hot tags of tumor size, unmasked region features, and the mask. The third step involves rapidly generating a full-frame CT image with global anatomical consistency using GAN; refining liver regions lacking texture detail using the trained LDM; and replacing the liver regions in the full-frame CT image after VQ-VAE decoding. This method combines the global synthesis efficiency of GAN with the local fine-grained generation capability of the diffusion model, effectively improving the realism and diversity of the final synthesized image while significantly reducing computational resources.

The invention discloses an intelligent prediction system and method for early recurrence risk after liver cancer thermal ablation, and aims to solve the technical problem that prediction is not timely and inaccurate due to the fact that an existing prediction model ignores intraoperative dynamic information. According to the method, dynamic time sequence data such as tissue impedance and power in the thermal ablation process are collected in real time, and preoperative static data such as tumor size and alpha fetoprotein level are combined, so that a double-branch fusion prediction model is constructed. The model comprises a time sequence processing branch used for learning dynamic data time dependence and a static processing branch used for extracting static data features. By fusing the features of the two branches, the model can output a probability value representing the postoperative early recurrence risk. According to the method and the system, instant accurate risk stratification after operation is realized, and further, intraoperative intervention can be guided through real-time feedback, so that the early recurrence risk is actively reduced, and the prognosis of a patient is improved.

The invention discloses an artificial intelligence platform for comprehensive radiological image analysis across CT, MRI, and X-ray modalities. The system automatically detects and quantifies cardiac, pulmonary, and oncological abnormalities through multiple AI modules including heart localization, cardiac MRI quantification, pneumonia classification, and tumor segmentation / grading. The platform provides automated tumor detection, 3D volumetric segmentation, and malignancy grading and treatment planning. The system integrates hybrid CNN-Transformer networks, attention-based 3D segmentation, and multi-modal feature fusion, offering precise quantitative metrics for tumor size, volume, shape, and malignancy probability. It also a cardiac MRI analysis module measures atrial and ventricular volumes and detects atrial fibrillation or mitral valve stenosis. It calculates the cardiothoracic ratio and detects heart displacement. DICOM / PACS integration and federated learning enable clinical deployment and cross-institutional adaptability while preserving patient data privacy and assists clinicians in accurate and rapid diagnosis while minimizing human error. This first-in-world platform facilitates early detection, staging, grading, and treatment planning with high reproducibility and explainable outputs.

A method to develop one or more morphometric classifiers to identify a mismatch repair deficiency (MMRD). The method provides a non-invasive method of characterizing MMRD that is responsive to a tumor in its early stages of development and irrespective of the tumor size. The method allows targeting cancer therapy to the specific characteristics of the cancer that the patient may have, allowing more efficient cancer management with far fewer side effects.

Device (2) and method for determining a value quantifying a risk of relapse of breast cancer for a patient A device for determining a value quantifying a risk of relapse of breast cancer for a patient comprises a data storage (114) adapted to store patient data associating, for each given patient at a least one whole histological slide image and patient clinical data comprising at least a tumor size, a number of positive lymph nodes, a tumor grade, and progesterone receptor (PR) positivity value, a feature extractor (1210, 1220) arranged to receive a whole histological slide image and associated patient clinical data, to derive tumor architecture features comprising a tumor area, an invasive tumor nest density, a tumor density at the invasive front, and an in situ tumor nest density, tumor microenvironment features comprising a border composition of in situ tumor, a variance of healthy gland size, an inflammatory stroma area, and an average nuclei size of stromal cells, and mitosis features comprising a mitotic hotspot count and a mitotic density, and to return a patient feature vector comprising said tumor architecture features, said tumor microenvironment features, said mitosis features, and features derived from said associated patient clinical data, a risk classifier (1230) using a machine learning module which is arranged to receive a patient feature vector as an input and to return a value quantifying a risk of relapse of breast cancer for a patient, said machine learning module having been trained with a dataset of labelled patient feature vectors and being arranged to return a risk value. This device is arranged to receive a set of patient data of a given patient, to provide at least some of said given patient's whole histological slide image and said given patient's clinical data as inputs to the feature extractor, to provide at least some of the resulting patient feature vectors to the risk classifier, and to return a value quantifying a risk of relapse of breast cancer for a patient for said given patient based on the outputs of the risk classifier.

The invention discloses a real-time intelligent measurement method for the size of tumors under gastrointestinal mucosa based on an ultrasonic endoscope, intelligent measurement of the size of SMTs under EUS is realized by using an artificial intelligence technology, and the method has the characteristics of real-time performance, standardization and homogeneity. According to the method, the manual measurement deviation is reduced, the energy and time investment of an endoscopic physician is reduced, and a quantitative and objective decision basis is provided for diagnosis, treatment and follow-up visit of SMTs. According to the method, a multi-scale feature fusion and self-attention enhanced EUS image tumor segmentation model is utilized, the same tumor region under different EUS cross sections is identified, the area of the tumor cross section is quantified, the maximum tumor cross section is automatically screened in real time, and the difference of subjective selection of an endoscopic physician is reduced; and the long diameter and the transverse diameter of the tumor section under the EUS are determined by using a standardized pixel-level dynamic cutting method, so that the standardization of the measured diameter is realized, and the manual operation steps of an endoscopic physician are reduced.