2017年 · 第51卷第12期

Imaging examination plays an important role in clinical decision-making of diagnosis and treatment of modern diseases. X-ray, CT, MRI, ultrasound and nuclear medicine examination methods are indispensable means of clinical practice and have become a part of clinical diagnosis and treatment routine. They are widely used in lesion detection, characterization, efficacy evaluation and prognosis judgment, etc., and can provide important evidence for diagnosis and treatment decision-making. For example, clinical staging of tumors based on TNM staging system is an important basis for the formulation of clinical treatment plans for tumors.

artificial intelligence (AI) will become the most important technological revolution in the field of information technology, and all walks of life will benefit from its rapid development. Therefore, there will be corresponding changes and even challenges in concepts, working habits and work connotations. Medical imaging will be the first field where artificial intelligence based on deep learning will achieve breakthroughs and explosions in the medical industry. How to correctly treat, actively participate and accept this new thing is worth thinking about by every imaging worker.

Implantable intravenous infusion port technology provides safe and reliable reserved venous access for patients undergoing long-term chemotherapy after breast cancer surgery. Because of its advantages of safety, simplicity, economy and less impact on daily life, it has gradually been recognized by many patients and clinicians. However, the occurrence of intracorporeal catheter rupture complications affects the popularization and use of this technology to a certain extent. Once the intracorporeal catheter ruptures, the disconnected catheter can enter the vena cava, right atrium, right ventricle and pulmonary artery along the blood flow, which can lead to pulmonary artery embolism, thrombosis, vascular tear, arrhythmia and other serious complications, and even endanger the patient's life^[1]。 The rupture of venous port catheter is affected by many factors, but there is still a lack of relevant reports of bulk cases at home and abroad. Through the follow-up observation of 3 102 patients with venous port implantation through internal jugular vein, we explored the related factors of intracorporeal catheter rupture in venous port, and provided the basis and preventive measures for the safe use of clinical venous port.

A 67-year-old female was admitted to hospital on June 30, 2016 due to dizziness and numbness in her left limb for more than half a month. The patient underwent radical colonectomy in 2012. On admission physical examination, the left upper and lower extremities muscle strength was reduced (grade IV), and the corners of the mouth twitched without skew; Laboratory tests showed no definite abnormalities. After admission, the patient was given neurotrophic anti-epileptic therapy, and his condition improved slightly. He was discharged from the hospital on 5 July 2016 due to unclear diagnosis. Later, due to the aggravation of the patient's condition, he was re-admitted to the hospital on August 12, 2016. After general practice discussion, the patient's symptoms were considered to be severe, and then he was operated on August 19, 2016.

The patient was a 59-year-old female. He was admitted to hospital due to severe headache for 2 d and unconsciousness for 3 h. The patient underwent a CT scan of the head in an external hospital 2 days ago due to severe headache, which showed subarachnoid hemorrhage (Figure 1), cerebral angiography revealed an aneurysm in the communicating segment of the right internal carotid artery, and no abnormalities in the intracranial vertebrobasilar artery system were found. After angiography, the patient appeared irritable, then unconscious and gradually aggravated, and was transferred to our hospital for further treatment. Emergency physical examination: The patient was in a coma, with bilateral pupils of equal size and round, and light reflection existed. Both upper limbs were flexed, both lower limbs were straightened, and there was no voluntary movement. Emergency head CT showed no new subarachnoid hemorrhage (Figure 2), considering acute posterior circulation ischemic stroke complicated with intracranial aneurysm rupture, emergency interventional treatment. The preoperative NIH Stroke Scale (NIHSS) score was 28.

The patient was a 49-year-old female. Admitted for upper abdominal discomfort. Physical examination and laboratory tests showed no abnormalities.

With the development of medical image equipment, the influence and role of medical image processing technology on medical scientific research and clinical practice is increasing day by day. The observation of human internal diseases is more direct and clear, and the diagnosis of diseases is more accurate. Image texture analysis is a new image post-processing technology in recent years. It can analyze the distribution of pixels in medical images mathematically, and obtain a series of relevant parameters to quantify the heterogeneity of lesions. It is helpful to judge the characteristics of diseases (especially tumors), evaluate the prognosis, predict and monitor the response to tumor treatment, etc. It is a useful additional tool in tumor imaging research. How to find the optimal texture features and texture parameters according to different kinds and characteristics of medical images is the focus and difficulty of current research^[1]。 The author reviews the application and progress of texture analysis in human tumor imaging.

Medical imaging examination has become a routine examination item because of its non-invasive, repeatable and local advantages^[1]X-ray, CT, PET, MRI, ultrasound, etc. have been widely used in the examination and diagnosis of various tissues^[2]。 Due to the strong temporal and spatial heterogeneity of solid tumors, different conditions may have similar imaging findings. In clinical work, imaging physicians diagnose imaging signs mainly according to personal experience, and missed diagnosis and misdiagnosis will inevitably occur. In the current process of diagnosis and treatment, some image features that cannot be distinguished by naked eye in medical images have not been fully utilized. How to make more effective use of medical image data, extract more valuable information from tumor phenotype, and make clinical diagnosis more scientific, objective and accurate is a technical problem to be solved urgently.

Lung cancer is the most common malignant tumor of the lungs, affecting about 1.8 million people worldwide every year^[1]。 According to the latest statistics of cancer in China in 2015, lung cancer ranks first in both prevalence and mortality^[2]。 radiomics refers to the application of a large number of image features to comprehensively quantify tumor heterogeneity, and provide all tumor information in a non-invasive and three-dimensional way. The author will review the current lung cancer imagomics research from CT and PET, and discuss the progress and challenges of imagomics in lung cancer research, in order to improve the diagnosis of lung cancer and accurately stratify patients, and provide evidence-based support for precision medicine.

Breast cancer is the most prevalent tumor in women^[1,2]。 Newly diagnosed breast cancer cases and deaths in China account for 12.2% and 9.6% of new and deaths in the world, respectively. Breast cancer is currently the sixth leading cause of cancer-related deaths among Chinese women^[3]。 Breast cancer is a tumor with multiple subtypes. Patients with different subtypes have different clinical manifestations, rate of progression, response to treatment and prognosis. With the development of genetic testing technology, breast cancer is divided into Luminal A, Luminal B, human epidermal growth factor receptor-2 (HER2) -enriched and Basal-like types at the genetic level. Accurate judgment of breast cancer subtypes helps to select the best treatment plan for patients. Although genetic testing can accurately distinguish breast cancer subtypes, due to cost and technical limitations, the current commonly used alternative method is pathological immunohistochemical analysis. Immunohistochemical results are limited by tumor heterogeneity and volume, and pathological histology cannot analyze tumors comprehensively, objectively and quantitatively. In contrast, imaging examination can reflect tumor characteristics non-invasively and as a whole. On the one hand, the differences within each subtype are further analyzed at the genetic level; On the other hand, the therapeutic effect can be dynamically evaluated multiple times. radiogenomics, as a product of the combination of imaging and genes, has become a potential biomarker for tumor diagnosis, treatment, evaluation of treatment response and prognosis, creating a new approach for personalized treatment and precision medicine^{[4,5,6,7,8,9]}。 At present, imaging genomics of breast cancer is a hot research topic, and the author reviews the related research progress.