Deep Learning-Based Classification of Cancer Cell in Leptomeningeal Metastasis on Cytomorphologic Features of Cerebrospinal Fluid.

Background: It is a critical challenge to diagnose leptomeningeal metastasis (LM), given its technical difficulty and the lack of typical symptoms. The existing gold standard of diagnosing LM is to use positive cerebrospinal fluid (CSF) cytology, which consumes significantly more time to classify cells under a microscope. Objective: This study aims to establish a deep learning model to classify cancer cells in CSF, thus facilitating doctors to achieve an accurate and fast diagnosis of LM in an early stage. Method: The cerebrospinal fluid laboratory of Xijing Hospital provides 53,255 cells from 90 LM patients in the research. We used two deep convolutional neural networks (CNN) models to classify cells in the CSF. A five-way cell classification model (CNN1) consists of lymphocytes, monocytes, neutrophils, erythrocytes, and cancer cells. A four-way cancer cell classification model (CNN2) consists of lung cancer cells, gastric cancer cells, breast cancer cells, and pancreatic cancer cells. Here, the CNN models were constructed by Resnet-inception-V2. We evaluated the performance of the proposed models on two external datasets and compared them with the results from 42 doctors of various levels of experience in the human-machine tests. Furthermore, we develop a computer-aided diagnosis (CAD) software to generate cytology diagnosis reports in the research rapidly. Results: With respect to the validation set, the mean average precision (mAP) of CNN1 is over 95% and that of CNN2 is close to 80%. Hence, the proposed deep learning model effectively classifies cells in CSF to facilitate the screening of cancer cells. In the human-machine tests, the accuracy of CNN1 is similar to the results from experts, with higher accuracy than doctors in other levels. Moreover, the overall accuracy of CNN2 is 10% higher than that of experts, with a time consumption of only one-third of that consumed by an expert. Using the CAD software saves 90% working time of cytologists. Conclusion: A deep learning method has been developed to assist the LM diagnosis with high accuracy and low time consumption effectively. Thanks to labeled data and step-by-step training, our proposed method can successfully classify cancer cells in the CSF to assist LM diagnosis early. In addition, this unique research can predict cancer's primary source of LM, which relies on cytomorphologic features without immunohistochemistry. Our results show that deep learning can be widely used in medical images to classify cerebrospinal fluid cells. For complex cancer classification tasks, the accuracy of the proposed method is significantly higher than that of specialist doctors, and its performance is better than that of junior doctors and interns. The application of CNNs and CAD software may ultimately aid in expediting the diagnosis and overcoming the shortage of experienced cytologists, thereby facilitating earlier treatment and improving the prognosis of LM.

Single-Beam Acoustic Tweezer Prepared by Lead-Free KNN-Based Textured Ceramics.

Acoustic tweezers for microparticle non-contact manipulation have attracted attention in the biomedical engineering field. The key components of acoustic tweezers are piezoelectric materials, which convert electrical energy to mechanical energy. The most widely used piezoelectric materials are lead-based materials. Because of the requirement of environmental protection, lead-free piezoelectric materials have been widely researched in past years. In our previous work, textured lead-free (K, Na)NbO3 (KNN)-based piezoelectric ceramics with high piezoelectric performance were prepared. In addition, the acoustic impedance of the KNN-based ceramics is lower than that of lead-based materials. The low acoustic impedance could improve the transmission efficiency of the mechanical energy between acoustic tweezers and water. In this work, acoustic tweezers were prepared to fill the gap between lead-free piezoelectric materials research and applications. The tweezers achieved 13 MHz center frequency and 89% -6 dB bandwidth. The -6 dB lateral and axial resolution of the tweezers were 195 µm and 114 µm, respectively. Furthermore, the map of acoustic pressure measurement and acoustic radiation calculation for the tweezers supported the trapping behavior for 100 µm diameter polystyrene microspheres. Moreover, the trapping and manipulation of the microspheres was achieved. These results suggest that the KNN-based acoustic tweezers have a great potential for further applications.

Establishment and Verification of Neural Network for Rapid and Accurate Cytological Examination of Four Types of Cerebrospinal Fluid Cells.

Fast and accurate cerebrospinal fluid cytology is the key to the diagnosis of many central nervous system diseases. However, in actual clinical work, cytological counting and classification of cerebrospinal fluid are often time-consuming and prone to human error. In this report, we have developed a deep neural network (DNN) for cell counting and classification of cerebrospinal fluid cytology. The May-Grünwald-Giemsa (MGG) stained image is annotated and input into the DNN network. The main cell types include lymphocytes, monocytes, neutrophils, and red blood cells. In clinical practice, the use of DNN is compared with the results of expert examinations in the professional cerebrospinal fluid room of a First-line 3A Hospital. The results show that the report produced by the DNN network is more accurate, with an accuracy of 95% and a reduction in turnaround time by 86%. This study shows the feasibility of applying DNN to clinical cerebrospinal fluid cytology.