Background: Pulmonary part-solid nodules (PSNs) reportedly have a high possibility of malignancy, while benign PSNs are common. This study aimed to reveal the differences between benign and malignant PSNs by comparing their thin-section computed tomography (CT) features.Methods: Patients with PSNs confirmed by postoperative pathological examination or follow-up (at the same period) were retrospectively enrolled from March 2016 to January 2020. The clinical data of patients and CT features of benign and malignant PSNs were reviewed and compared. Binary logistic regression analysis was performed to reveal the predictors of malignant PSNs.Results: A total of 119 PSNs in 117 patients [age (mean ± standard deviation), 56±11 years; 70 women] were evaluated. Of the 119 PSNs, 44 (37.0%) were benign, and 75 (63.0%) were malignant (12 adenocarcinomas in situ, 22 minimally invasive adenocarcinomas, and 41 invasive adenocarcinomas). There were significant differences in the patients' age and smoking history between benign and malignant PSNs.In terms of CT characteristics, malignant and benign lesions significantly differed in the following CT features: whole nodule, internal solid component, and peripheral ground-glass opacity. The binary logistic regression analysis revealed that well-defined border [odds ratio (OR), 4.574; 95% confidence interval (CI),

Background: Some pulmonary ground-glass nodules (GGNs) are benign and frequently misdiagnosed due to lack of understanding of their CT characteristics. This study aimed to reveal the CT features and corresponding pathological findings of pulmonary benign GGNs to help improve diagnostic accuracy. Patients and Methods: From March 2016 to October 2019, patients with benign GGNs confirmed by operation or follow-up were enrolled retrospectively. According to overall CT manifestations, GGNs were classified into three types: I, GGO with internal high-attenuation zone; II, nodules lying on adjacent blood vessels; and other type, lesions without obvious common characteristics. CT features and pathological findings of each nodule type were evaluated. Results: Among the 40 type I, 25 type II, and 14 other type GGNs, 24 (60.0%), 19 (76.0%), and 10 (71.4%) nodules were resected, respectively. Type I GGNs were usually irregular (25 of 40, 62.5%) with only one high-attenuation zone (38 of 40, 95.0%) (main pathological components: thickened alveolar walls with inflammatory cells, fibrous tissue, and exudation), which was usually centric (24 of 40, 60.0%), having blurred margin (38 of 40, 95.0%), and connecting to blood vessels (32 of 40, 80.0%). The peripheral GGO (main pathological component: a small amount of inflammatory cell infiltration with fibrous tissue proliferation) was usually ill-defined (28 of 40, 70.0%). Type II GGNs (main pathological components: focal interstitial fibrosis with or without inflammatory cell infiltration) lying on adjacent vessel branches were usually irregular (19 of 25, 76.0%) and well defined (16 of 25, 64.0%) but showed coarse margins (15 of 16, 93.8%). Other type GGNs had various CT manifestations but their pathological findings were similar to that of type II. Conclusion: For subsolid nodules with CT features manifested in type I or II GGNs, follow-up should be firstly considered in further management.

Purpose: Tumors with high mutation load tend to have a stronger immune response in some tumors. The correlation between expression of programmed death ligand-1 (PD-L1), a biomarker of immune response in tumors, and p53, accepted as the most frequently mutated gene in many cancers, in triple-negative breast cancer (TNBC) has not been fully investigated in cancer patients. Materials and methods: 132 cases of TNBC and 32 cases of non-TNBC paraffinembedded tissue sections were selected to detect the expression of PD-L1 and p53 by immunohistochemistry, and results were correlated with clinical data and survival outcomes. The staining of PD-L1 in tumor cells (TCs) and tumor-associated immune cells (TAICs) was assessed separately. Results: Strong positive correlations were observed between expression of p53 and PD-L1 both in TCs (r=0.338, P=0.000) and TAICs (r=0.186, P=0.033). The same positive correlation was found in the expression of PD-L1 in TCs and TAICs (r=0.764, P=0.000). Like p53 (P=0.024), positive rate of PD-L1 in TCs was significantly higher in TNBC than in non-TNBC (P=0.02). PD-L1 and p53 in TCs staining were significantly associated with histological grade, tumor size and Ki67 index (P<0.05). PD-L1 in TCs staining was also associated with lymphatic metastasis status (P=0.000). However, PD-L1 in TAICs was only related to histological grade in statistically (P=0.012). Kaplan-Meier survival analysis showed that positive groups of p53, PD-L1 in TCs and TAICs had a worse overall survival and a worse progression-free survival as compared with the negative groups, but marginal significance was found only in overall survival of PD-L1 in TCs and TAICs, and progression-free survival of PD-L1 in TAICs (P=0.074, 0.097, 0.068, respectively). Conclusion: Our findings suggest that positive correlation between p53 and PD-L1 in TNBC and the higher expression rates are closely correlated with some key prognostic factors and worse survival outcomes. These findings would lay the foundation for further study on the relationship of p53 and PD-L1 and the combination of mutated p53 inhibitors and PD-1/PD-L1 antibodies in TNBC.

Compared with other breast cancer subtypes, triple-negative breast cancer (TNBC) has poorer responses to therapy and lower overall survival rates. The use of an inhibitor of immune checkpoint programmed cell death ligand 1 (PD-L1) is a promising treatment strategy and is approved for malignant tumors, especially for TNBC. p53 regulates various biological processes, but the association between p53 and immune evasion remains unknown. miR-34a is a known tumor suppressor and p53-regulated miRNA that is downregulated in several cancers; however, it has not been reported in TNBC. Herein, we aimed to explore the regulatory signaling axis among p53, miR-34a and PD-L1 in TNBC cells in vivo and in tissue and to improve our understanding of immunotherapy for TNBC. p53-EGFP, p53-siRNA and miR-34a mimics were transfected into TNBC cell lines, and the interaction between miR-34a and PD-L1 was analyzed via dual-luciferase reporter assays. We found that p53 could inhibit the expression of PD-L1 via miR-34a and that miR-34a could inhibit both cell activity and migration and promoted apoptosis and cytotoxicity in TNBC. Furthermore, miR-34a agomir was injected into MDA-MB-231 tumors of nude mice. The results showed that miR-34a could inhibit tumor growth and downregulate the expression of PD-L1 in vivo. A total of 133 TNBC tissue samples were analyzed by immunochemistry; the proportion of positive expression of PD-L1 was 57.14% (76/133), and the proportion of samples with negative expression of PD-L1 was 42.86% (57/133). Our research may provide a novel potential target for TNBC.

BackgroundHistopathological diagnosis of bone tumors is challenging for pathologists. We aim to classify bone tumors histopathologically in terms of aggressiveness using deep learning (DL) and compare performance with pathologists.MethodsA total of 427 pathological slides of bone tumors were produced and scanned as whole slide imaging (WSI). Tumor area of WSI was annotated by pathologists and cropped into 716,838 image patches of 256 × 256 pixels for training. After six DL models were trained and validated in patch level, performance was evaluated on testing dataset for binary classification (benign vs. non-benign) and ternary classification (benign vs. intermediate vs. malignant) in patch-level and slide-level prediction. The performance of four pathologists with different experiences was compared to the best-performing models. The gradient-weighted class activation mapping was used to visualize patch’s important area.ResultsVGG-16 and Inception V3 performed better than other models in patch-level binary and ternary classification. For slide-level prediction, VGG-16 and Inception V3 had area under curve of 0.962 and 0.971 for binary classification and Cohen’s kappa score (CKS) of 0.731 and 0.802 for ternary classification. The senior pathologist had CKS of 0.685 comparable to both models (p = 0.688 and p = 0.287) while attending and junior pathologists showed lower CKS than the best model (each p &lt; 0.05). Visualization showed that the DL model depended on pathological features to make predictions.ConclusionDL can effectively classify bone tumors histopathologically in terms of aggressiveness with performance similar to senior pathologists. Our results are promising and would help expedite the future application of DL-assisted histopathological diagnosis for bone tumors.