Background: N staging process was essential for evaluation of outcome and indication of following adjuvant therapies in Non-smallcell lung cancer treatment. Various clinical observations on the potential transition patterns of lymph node drainage are reported, however, most of the previous conclusions were made by clinical physicians and focused on specific empirical transition patterns. The fact that there is no definitive and holistic map for lymphatic metastasis transition patterns, and the patients were suffering from either excessive nodes collection along with more damage, or insufficient nodes collection with potential recurrent risks. Method: We perform complete lymph node examination of a total of 936 subjects diagnosed with NSCLC lung cancer. Lymph nodes sampling or dissection are performed according to NCCN guidelines. A probabilistic model is developed due to the presence of these missing values. Using the maximum likelihood estimation and proximal gradient algorithm, the summarization of dataset is obtained, which were several explicit metastases and their corresponding probabilities. The metastasis graph is constructed from the summarization result with greedy algorithm and a given threshold. Besides, numerical simulation experiments are conducted to validate the stability of algorithms. Result: Lymph node sites are shown as round circles according to their anatomical locations. The inferred transition paths are shown as edges connecting them. Edges colored in red are those consistently found in the left and right lung, and blue for unique nodes at each side thus cannot be compared, and black for different patterns between left and right lungs. Closely connected intralobar (N1) nodes: strong connections among intra-lobar nodes (10w14). Over 78% among all the patients have more than 7 edges connecting the 5 intra-lobar nodes. Jumping metastasis from N1 to N2 stations: We found that there exists several jumping metastasis at both sides of the lobes which has not been well-studied yet posing a challenge for the diagnosis and accurate staging (eg. 12 to 4, 13 to7, 11 to 2), revealing potential long-range transition pathways. Correlation among N2 lymph nodes: We discovered the presence of certain metastatic groups, including node 5/6, 5/9, 7/9, 6/8 for left lung, and 2/4,2/ 7,7/9, 2/3p, 3p/7 for right lung. Conclusion: So we drew a map precisely to make a better understanding for metastatic pathways and provide a potential tool for the prediction of involved nodes pre/ or intro-operatively, so that an individualized surgical planning strategy could be made.

No abstract

Result: Of the 179 eligible patients, 103 were male and 76 were female. The median age was 70 years. The median whole tumor size, solid component size, and consolidation/tumor ratio on CT were 2.1 cm (range, 0.3-5.8 cm)，1.7 cm (range, 0-3.0 cm)，and 87% (range, 0-100%)，respectively. The median SUV max on FDG-PET/CT was 4.20 (range, 0.78-28.07). According to the postoperative pathological examination, 147 patients (82%) had adenocarcinoma, 25 (14%) had squamous cell carcinoma, and 7 (4%) had other types of NSCLC. Pleural involvement, pulmonary metastasis, lymph node metastasis, lymphatic permeation, and vascular invasion were identified in 26 (15%), 6 (3%), 15 (8%), 11 (6%), and 19 patients (11%), respectively, and in total, 46 patients (26%) developed the pathological metastasis and/or involvement. The univariate analysis identified SUV max , sex, carcinoembryonic antigen, Sialyl Lewis x-1, whole tumor size, solid component size, consolidation/tumor ratio, diseased side, tumor location, and tumor histology as significant predictors. A multivariate analysis revealed SUV max (OR: 1.197, p<0.001) and consolidation/tumor ratio (OR: 1.052, p ¼ 0.001) as significant independent predictors. The solid component size and whole tumor size were not identified as significant independent predictors. By the ROC analysis, the optimal cutoff point for SUV max was determined as 6.10. Thirty-seven of the 67 patients (55%) with 6.10 of SUV max developed the pathological metastasis and/or involvement whereas 9 of the 112 patients (8%) with <6.10 of SUV max developed pathological metastasis and/or involvement. Conclusion: Our results suggested the predictive effect of high SUV max on pathological metastasis and involvement in clinical stage IA NSCLC patients. Thus, FDG-PET/CT should be utilized for the preoperative precise evaluation of early stage NSCLC, and we may consider SUV max on FDG-PET/CT to decide surgical procedure for these patients, such as the extent of pulmonary resection and lymphadenectomy.

No abstract

No abstract