[Aprospective study of detection and clinical significance of bone marrow tumor cells in small cell lung cancer].

Objective: To investigate the detection of bone marrow tumor cells in small cell lung cancer (SCLC) patients and their relationship with clinical features, treatment response and prognosis. Methods: A total of 113patients with newly diagnosed SCLC from January 2018 to October 2022 at Beijing Chest Hospital were prospectively enrolled. Before treatment, bone marrow was aspirated and separately submitted for tumor cells detection by liquid-based cytology and disseminated tumor cells (DTCs) detection by the substrction enrichment and immunostaining fluorescence in situ hybridization (SE-iFISH) platform. The correlation between the detection results of the two methods with patients' clinical features and treatment response was evaluated by Chi-square. Kaplan-Meier method was applied to create survival curves and the Cox regression model was used for multivariate analysis. Results: The positive rate of bone marrow liquid-based cytology in SCLC was 15.93% (18/113). The liver and bone metastases rates were significantly higher (55.56% vs 11.58% for liver metastasis, P＜0.001; 77.78% vs 16.84% for bone metastasis, P＜0.001) and thrombocytopenia was more common (16.67% vs 2.11%, P=0.033) in patients with tumor cells detected in liquid-based cytology than those without detected tumor cells. As for SE-iFISH, DTCs were detected in 92.92% of patients (105/113), the liver and bone metastasis rates were significantly higher (37.93% vs 11.90% for liver metastasis, P=0.002; 44.83% vs 20.23 % for bone metastasis, P=0.010), and the incidence of thrombocytopenia was significantly increased (13.79% vs 1.19%, P=0.020) in patients with DTCs≥111 per 3 ml than those with DTCs＜111 per 3 ml. The positive rates of bone marrow liquid-based cytology in the disease control group and the disease progression group were 12.00% (12/100) and 46.15% (6/13), respectively, and the difference was statistically significant (P=0.002). However, the result of SE-iFISH revealed the DTCs quantities of the above two groups were 29 (8,110) and 64 (15,257) per 3 ml, and there was no statistical difference between the two groups (P=0.329). Univariate analysis depicted that the median progression-free survival (PFS) and median overall survival (OS) of liquid-based cytology positive patients were significantly shorter than those of tumor cell negative patients (6.33 months vs 9.27 months for PFS, P=0.019; 8.03 months vs 19.50 months for OS, P=0.019, P=0.033). The median PFS and median OS in patients with DTCs≥111 per 3 ml decreased significantly than those with DTCs＜111 per 3 ml (6.83 months vs 9.50 months for PFS, P=0.004; 11.2 months vs 20.60 months for OS, P=0.019). Multivariate analysis showed that disease stage (HR=2.806, 95%CI:1.499-5.251, P=0.001) and DTCs quantity detected by SE-iFISH (HR=1.841, 95%CI:1.095-3.095, P=0.021) were independent factors of PFS, while disease stage was the independent factor of OS (HR=2.538, 95%CI:1.169-5.512, P=0.019). Conclusions: Both bone marrow liquid-based cytology and SE-iFISH are clinically feasible. The positive detection of liquid-based cytology or DTCs≥111 per 3 ml was correlated with distant metastasis, and DTCs≥111 per 3 ml was an independent prognostic factor of decreased PFS in SCLC.

[Pathological diagnosis of lung cancer based on deep transfer learning].

Objective: To establish an artificial intelligence (AI)-assisted diagnostic system for lung cancer via deep transfer learning. Methods: The researchers collected 519 lung pathologic slides from 2016 to 2019, covering various lung tissues, including normal tissues, adenocarcinoma, squamous cell carcinoma and small cell carcinoma, from the Beijing Chest Hospital, the Capital Medical University. The slides were digitized by scanner, and 316 slides were used as training set and 203 as the internal test set. The researchers labeled all the training slides by pathologists and establish a semantic segmentation model based on DeepLab v3 with ResNet-50 to detect lung cancers at the pixel level. To perform transfer learning, the researchers utilized the gastric cancer detection model to initialize the deep neural network parameters. The lung cancer detection convolutional neural network was further trained by fine-tuning of the labeled data. The deep learning model was tested by 203 slides in the internal test set and 1 081 slides obtained from TCIA database, named as the external test set. Results: The model trained with transfer learning showed substantial accuracy advantage against the one trained from scratch for the internal test set [area under curve (AUC) 0.988 vs. 0.971, Kappa 0.852 vs. 0.832]. For the external test set, the transferred model achieved an AUC of 0.968 and Kappa of 0.828, indicating superior generalization ability. By studying the predictions made by the model, the researchers obtained deeper understandings of the deep learning model. Conclusions: The lung cancer histopathological diagnostic system achieves higher accuracy and superior generalization ability. With the development of histopathological AI, the transfer learning can effectively train diagnosis models and shorten the learning period, and improve the model performance.

The significance of Siglec-15 expression in resectable non-small cell lung cancer.

Siglec-15 (S15) is another important mechanism of tumor immune escape besides the PD-L1/PD-1 pathway and represents a new kind of immune checkpoint inhibitor. However, the associations of tumor Siglec-15 expression with clinicopathological characteristics and outcomes of non-small cell lung cancer (NSCLC), and tumor-infiltrating lymphocytes (TILs) in a tumor microenvironment (TME) have so far been unclear. A total of 324 NSCLC surgical samples on tumor microarray were used in this study for investigating the association of S15 expression with clinicopathological characteristics and overall survival (OS) as well as correlation with TILs using multiplex immunofluorescence staining and PD-L1. Results showed that the expression of S15 in adenocarcinoma was significantly higher than that in squamous cell carcinoma. S15 expression was positively correlated with CD8+ T cell density in the stroma. The expression rate of PD-L1 in lung squamous cell carcinoma was higher than that in lung adenocarcinoma. S15 expression was not associated with the prognosis of early NSCLC. The pathological mechanism of the co-expression of S15 and PD-L1 in resectable NSCLC remains to be further studied.

The joint impact of smoking plus alcohol drinking on treatment of pulmonary tuberculosis.

Tuberculosis, smoking, and alcohol drinking are major public health and social issues worldwide. We investigated the joint effect of smoking plus alcohol drinking on TB treatment. Retrospective study was conducted among TB patients in 49 units from eight provinces in China. All patients enrolled were classified into four groups according to their smoking and/or alcohol status. Current smokers plus drinkers belonged to group 1; ex-smokers plus ex-drinkers were in group 2; current smokers and ex-drinkers, current smokers and never drinkers, ex-smokers and current drinkers, ex-smokers and never drinkers, never smokers and current drinkers, and never smokers and ex-drinkers belonged to group 3; while the never smokers plus never drinkers were in group 4. We used a chi-square test to compare adverse drug reaction, lesions absorption and cavities of lung, sputum culture at the end of the second month, and treatment outcomes among the four groups. Among the 1256 participants enrolled in the study, 6.1% (76/1256) were current smokers plus drinkers; 25.9% (325/1256) were ex-smokers plus drinkers; 29.1% (366/1256) were current/never/ex-smokers and/or drinkers, and 38.9% (489/1256) were never smokers plus drinkers, respectively. Compared to the never smokers and drinkers, smoker plus drinker TB patients were more likely to experience adverse drug reaction (x2 = 8.480, P = 0.037), less proportion of lesions absorption in lungs (x2 = 10.330, P = 0.016), lower proportion of culture conversion (x2 = 18.83, P = 0.04), and more unfavorable outcomes. Smoking plus alcohol drinking adversely affect response against TB treatment, which increase adverse drug reactions, sputum culture-positive rate at the end of the second month, and failure rate of pulmonary tuberculosis patients.