MRI-based deep learning model predicts distant metastasis and chemotherapy benefit in stage II nasopharyngeal carcinoma.

Chemotherapy remains controversial for stage II nasopharyngeal carcinoma because of its considerable prognostic heterogeneity. We aimed to develop an MRI-based deep learning model for predicting distant metastasis and assessing chemotherapy efficacy in stage II nasopharyngeal carcinoma. This multicenter retrospective study enrolled 1072 patients from three Chinese centers for training (Center 1, n = 575) and external validation (Centers 2 and 3, n = 497). The deep learning model significantly predicted the risk of distant metastases for stage II nasopharyngeal carcinoma and was validated in the external validation cohort. In addition, the deep learning model outperformed the clinical and radiomics models in terms of predictive performance. Furthermore, the deep learning model facilitates the identification of high-risk patients who could benefit from chemotherapy, providing useful additional information for individualized treatment decisions.

Metabolic profiling of normal hepatocyte and hepatocellular carcinoma cells via 1 H nuclear magnetic resonance spectroscopy.

Hepatocellular carcinoma (HCC) causes death mainly by disseminated metastasis progression from the organ being confined. Different metastatic stages are closely related to cellular metabolic profiles. Normal hepatocyte and HepG2 cell line from low metastatic HCC were studied by NMR-based metabolomic techniques. Multivariate and univariate statistical analyses were utilized to identify characteristic metabolites from cells and cultured media. Elevated levels of acetate, creatine, isoleucine, leucine, and phenylalanine were observed in HepG2 cells, suggesting more active in gathering nutrient components along with altered amino acid metabolisms and enhanced lipid metabolism. High glucose consumption was significantly different in low metastatic cells. A series of characteristic metabolites were identified and served as biomarkers. Relative metabolic pathway analysis shows that low metastatic HepG2 cell line exhibits active behaviors in metabolisms and biosynthesis of specific amino acids and energy metabolism. Moreover, characteristic metabolites-based classification models executed by support vector machines algorithm perform robustly to classify normal hepatocyte and HepG2 cell line. It is concluded that NMR-based metabolomic analyses of cell lines can provide a powerful approach to understand metastasis-related biological alterations. The present study also provides a basis for metabolic markers determination of hepatic carcinoma in the future clinical study.

Prognostic nomogram for refining the prognostication of the proposed 8th edition of the AJCC/UICC staging system for nasopharyngeal cancer in the era of intensity-modulated radiotherapy.

BACKGROUND: The objective of this study was to develop a nomogram for refining prognostication for patients with nondisseminated nasopharyngeal cancer (NPC) staged with the proposed 8th edition of the American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) staging system. METHODS: Consecutive patients who had been investigated with magnetic resonance imaging, staged with the proposed 8th edition of the AJCC/UICC staging system, and irradiated with intensity-modulated radiotherapy from June 2005 to December 2010 were analyzed. A cohort of 1197 patients treated at Fujian Provincial Cancer Hospital was used as the training set, and the results were validated with 412 patients from Pamela Youde Nethersole Eastern Hospital. Cox regression analyses were performed to identify significant prognostic factors for developing a nomogram to predict overall survival (OS). The discriminative ability was assessed with the concordance index (c-index). A recursive partitioning algorithm was applied to the survival scores of the combined set to categorize the patients into 3 risk groups. RESULTS: A multivariate analysis showed that age, gross primary tumor volume, and lactate dehydrogenase were independent prognostic factors for OS in addition to the stage group. The OS nomogram based on all these factors had a statistically higher bias-corrected c-index than prognostication based on the stage group alone (0.712 vs 0.622, P <.01). These results were consistent for both the training cohort and the validation cohort. Patients with <135 points were categorized as low-risk, patients with 135 to <160 points were categorized as intermediate-risk, and patients with ≥160 points were categorized as high-risk. Their 5-year OS rates were 92%, 84%, and 58%, respectively. CONCLUSIONS: The proposed nomogram could improve prognostication in comparison with the TNM stage group. This could aid in risk stratification for individual NPC patients. Cancer 2016;122:3307-3315. © 2016 American Cancer Society.

Proposal for the 8th edition of the AJCC/UICC staging system for nasopharyngeal cancer in the era of intensity-modulated radiotherapy.

BACKGROUND: An accurate staging system is crucial for cancer management. Evaluations for continual suitability and improvement are needed as staging and treatment methods evolve. METHODS: This was a retrospective study of 1609 patients with nasopharyngeal carcinoma investigated by magnetic resonance imaging, staged with the 7th edition of the American Joint Committee on Cancer (AJCC)/International Union Against Cancer (UICC) staging system, and irradiated by intensity-modulated radiotherapy at 2 centers in Hong Kong and mainland China. RESULTS: Among the patients without other T3/T4 involvement, there were no significant differences in overall survival (OS) between medial pterygoid muscle (MP) ± lateral pterygoid muscle (LP), prevertebral muscle, and parapharyngeal space involvement. Patients with extensive soft tissue involvement beyond the aforementioned structures had poor OS similar to that of patients with intracranial extension and/or cranial nerve palsy. Only 2% of the patients had lymph nodes > 6 cm above the supraclavicular fossa (SCF), and their outcomes resembled the outcomes of those with low extension. Replacing SCF with the lower neck (extension below the caudal border of the cricoid cartilage) did not affect the hazard distinction between different N categories. With the proposed T and N categories, there were no significant differences in outcome between T4N0-2 and T1-4N3 disease. CONCLUSIONS: After a review by AJCC/UICC preparatory committees, the changes recommended for the 8th edition include changing MP/LP involvement from T4 to T2, adding prevertebral muscle involvement as T2, replacing SCF with the lower neck and merging this with a maximum nodal diameter > 6 cm as N3, and merging T4 and N3 as stage IVA criteria. These changes will lead not only to a better distinction of hazards between adjacent stages/categories but also to optimal balance in clinical practicability and global applicability.

Optimization of magnetic resonance sequences in lymph node staging of nasopharyngeal carcinoma.

BACKGROUND: Detection rate of retropharyngeal lymph node metastasis in patients with nasopharyngeal carcinoma (NPC) needs to be improved. The purpose of this study was to compare three magnetic resonance (MR) sequences for detecting lymph nodes in patients with NPC. METHODS: Between July 2007 and March 2008, MR staging of pre-treated tumor was conducted on 120 patients with pathologically confirmed NPC. The outcome of three different sequences for MR NPC staging were compared: coronal short TI inversion recovery (STIR), axial proton density fat-suppressed (PDWI fs), and coronal contrast enhanced fast spin echo T1 weighted fat-suppressed (CE FSE T1WI fs). Nodal classification method (1999) was applied to count the number of retropharyngeal and cervical lymph nodes discovered by each MR sequence. Paired t tests were used for statistical analysis. RESULTS: A total of 2575 lymph nodes were found using coronal STIR sequence; 1816 lymph nodes for coronal CE FSE T1WI fs sequence and 2638 lymph nodes for axial PDWI fs sequence. Significant differences existed in the number of lymph nodes detected by axial PDWI fs and coronal CE FSE T1WI fs sequence (paired t test, P < 0.05), with the former sequence getting higher numbers. Statistical differences also existed between coronal STIR and coronal CE FSE T1WI fs sequence (paired t test, P < 0.05), with the former sequence getting higher numbers. No significant difference was found between coronal STIR sequence and axial PDWI fs sequence (paired t test, P > 0.05). CONCLUSIONS: For the detection of retropharyngeal and cervical lymph nodes, coronal STIR sequence and axial PDWI fs sequence have similar performance and both sequences showed better detection than CE FSE T1WI fs sequence. Furthermore, by combining coronal STIR sequence and axial PDWI fs sequence, we can improve the detection of lymph nodes in NPC N-staging before treatment, especially for lymph nodes located in the thoracic entrance.