Research

Jump to: Review

13 pages, 3783 KiB

Open AccessArticle

Automated Classification of Idiopathic Pulmonary Fibrosis in Pathological Images Using Convolutional Neural Network and Generative Adversarial Networks

by Atsushi Teramoto, Tetsuya Tsukamoto, Ayano Michiba, Yuka Kiriyama, Eiko Sakurai, Kazuyoshi Imaizumi, Kuniaki Saito and Hiroshi Fujita

Diagnostics 2022, 12(12), 3195; https://0-doi-org.brum.beds.ac.uk/10.3390/diagnostics12123195 - 16 Dec 2022

Cited by 2 | Viewed by 1254

Abstract

Interstitial pneumonia of uncertain cause is referred to as idiopathic interstitial pneumonia (IIP). Among the various types of IIPs, the prognosis of cases of idiopathic pulmonary fibrosis (IPF) is extremely poor, and accurate differentiation between IPF and non-IPF pneumonia is critical. In this [...] Read more.

Interstitial pneumonia of uncertain cause is referred to as idiopathic interstitial pneumonia (IIP). Among the various types of IIPs, the prognosis of cases of idiopathic pulmonary fibrosis (IPF) is extremely poor, and accurate differentiation between IPF and non-IPF pneumonia is critical. In this study, we consider deep learning (DL) methods owing to their excellent image classification capabilities. Although DL models require large quantities of training data, collecting a large number of pathological specimens is difficult for rare diseases. In this study, we propose an end-to-end scheme to automatically classify IIPs using a convolutional neural network (CNN) model. To compensate for the lack of data on rare diseases, we introduce a two-step training method to generate pathological images of IIPs using a generative adversarial network (GAN). Tissue specimens from 24 patients with IIPs were scanned using a whole slide scanner, and the resulting images were divided into patch images with a size of 224 × 224 pixels. A progressive growth GAN (PGGAN) model was trained using 23,142 IPF images and 7817 non-IPF images to generate 10,000 images for each of the two categories. The images generated by the PGGAN were used along with real images to train the CNN model. An evaluation of the images generated by the PGGAN showed that cells and their locations were well-expressed. We also obtained the best classification performance with a detection sensitivity of 97.2% and a specificity of 69.4% for IPF using DenseNet. The classification performance was also improved by using PGGAN-generated images. These results indicate that the proposed method may be considered effective for the diagnosis of IPF. Full article

(This article belongs to the Special Issue Lung Cancer and Lung Injury: Diagnosis, Differential Diagnosis, and Management)

► Show Figures

Figure 1

13 pages, 1449 KiB

Open AccessArticle

Survival Rates of Patients with Non-Small Cell Lung Cancer Depending on Lymph Node Metastasis: A Focus on Saliva

by Lyudmila V. Bel’skaya, Elena A. Sarf and Victor K. Kosenok

Diagnostics 2021, 11(5), 912; https://0-doi-org.brum.beds.ac.uk/10.3390/diagnostics11050912 - 20 May 2021

Cited by 6 | Viewed by 1859

Abstract

The aim of this study was to compare overall survival (OS) rates at different pN stages of NSCLC depending on tumor characteristics and to assess the applicability of saliva biochemical markers as prognostic signs. The study included 239 patients with NSCLC (pN₀ [...] Read more.

The aim of this study was to compare overall survival (OS) rates at different pN stages of NSCLC depending on tumor characteristics and to assess the applicability of saliva biochemical markers as prognostic signs. The study included 239 patients with NSCLC (pN₀-120, pN₁-51, pN₂-68). Saliva was analyzed for 34 biochemical indicators before the start of treatment. For pN₀, the tumor size does not have a prognostic effect, but the histological type should be taken into account. For pN₁ and pN₂, long-term results are significantly worse in squamous cell cancer with a large tumor size. A larger volume of surgical treatment reduces the differences between OS. The statistically significant factors of an unfavorable prognosis at pN₀ are the lactate dehydrogenase activity <1294 U/L and the level of diene conjugates >3.97 c.u. (HR = 3.48, 95% CI 1.21–9.85, p = 0.01541); at pN₁, the content of imidazole compounds >0.296 mmol/L (HR = 6.75, 95% CI 1.28–34.57, p = 0.00822); at pN₂ levels of protein <0.583 g/L and Schiff bases >0.602 c.u., as well as protein >0.583 g/L and Schiff bases <0.602 c.u. (HR = 2.07, 95% CI 1.47–8.93, p = 0.04351). Using salivary biochemical indicators, it is possible to carry out stratification into prognostic groups depending on the lymph node metastasis. Full article

(This article belongs to the Special Issue Lung Cancer and Lung Injury: Diagnosis, Differential Diagnosis, and Management)

► Show Figures

Figure 1

Review

Jump to: Research

13 pages, 1436 KiB

Open AccessReview

A Novel Strategy for the Diagnosis of Pulmonary High-Grade Neuroendocrine Tumor

by Kentaro Miura, Kimihiro Shimizu, Shogo Ide, Shuji Mishima, Shunichiro Matsuoka, Tetsu Takeda, Takashi Eguchi, Kazutoshi Hamanaka and Takeshi Uehara

Diagnostics 2021, 11(11), 1945; https://0-doi-org.brum.beds.ac.uk/10.3390/diagnostics11111945 - 20 Oct 2021

Cited by 5 | Viewed by 2807

Abstract

Correctly diagnosing a histologic type of lung cancer is important for selecting the appropriate treatment because the aggressiveness, chemotherapy regimen, surgical approach, and prognosis vary significantly among histologic types. Pulmonary NETs, which are characterized by neuroendocrine morphologies, represent approximately 20% of all lung [...] Read more.

Correctly diagnosing a histologic type of lung cancer is important for selecting the appropriate treatment because the aggressiveness, chemotherapy regimen, surgical approach, and prognosis vary significantly among histologic types. Pulmonary NETs, which are characterized by neuroendocrine morphologies, represent approximately 20% of all lung cancers. In particular, high-grade neuroendocrine tumors (small cell lung cancer and large cell neuroendocrine tumor) are highly proliferative cancers that have a poorer prognosis than other non-small cell lung cancers. The combination of hematoxylin and eosin staining, Ki-67, and immunostaining of classic neuroendocrine markers, such as chromogranin A, CD56, and synaptophysin, are normally used to diagnose high-grade neuroendocrine tumors; however, they are frequently heterogeneous. This article reviews the diagnostic methods of lung cancer diagnosis focused on immunostaining. In particular, we describe the usefulness of immunostaining by Stathmin-1, which is a cytosolic phosphoprotein and a key regulator of cell division due to its microtubule depolymerization in a phosphorylation-dependent manner, for the diagnosis of high-grade neuroendocrine tumors. Full article

(This article belongs to the Special Issue Lung Cancer and Lung Injury: Diagnosis, Differential Diagnosis, and Management)

► Show Figures