Search

Cancer cachexia is increasingly recognized as a poor prognostic marker for various tumor types. Weight loss in esophageal cancer is multifactorial, as patients with bulky tumors also have reduced ability to eat. We aimed to investigate the relationship between prediagnosis weight loss and mortality in esophageal cancer and to determine whether these associations vary with tumor stage. We conducted a prospective cohort study of esophageal cancer patients at two tertiary centers. We recorded baseline patient characteristics including medications, smoking, body mass index, and weight loss in the year prior to diagnosis, and collected data on treatment and outcomes. We used Cox regression modeling to determine the associations between percent weight loss and outcomes. The main outcome of interest was all-cause mortality; secondary endpoints were esophageal cancer-specific mortality and development of metastases. We enrolled 134 subjects, the majority of whom had adenocarcinoma (82.1%); median percent weight loss was 4.7% (IQR: 0%–10.9%). Increasing percent weight loss was not associated with all-cause mortality (ptrend = 0.36). However, there was evidence of significant interaction by tumor stage (p = 0.02). There was a strong and significant association between prediagnosis weight loss and mortality in patients with T stages 1 or 2 (adjusted HR 8.26 for highest versus lowest tertile, 95%CI 1.11–61.5, ptrend = 0.03) but not for T stages 3 or 4 (ptrend = 0.32). Body mass index one year prior to diagnosis was not associated with mortality. Prediagnosis weight loss was associated with increased all-cause mortality only in patients with early stage esophageal cancer. This suggests that tumor-related cachexia can occur early in esophageal cancer and represents a poor prognostic marker.

Background: Esophageal cancer remains associated with poor outcomes, yet little is known regarding factors that influence survival. Aspirin use prior to cancer diagnosis may influence outcomes. We aimed to assess the effects of prediagnosis aspirin use in patients with esophageal cancer. Methods: We conducted a prospective cohort study of newly-diagnosed esophageal cancer patients at two tertiary care centers. We assessed history of prediagnosis aspirin use, and prospectively followed patients and assessed mortality, cause of death, and development of metastases. Results: We enrolled 130 patients, the majority of whom were male (81.5%) and had adenocarcinoma (80.8%). Overall, 57 patients (43.9%) were regular aspirin users. In unadjusted analyses, we found no difference in all-cause mortality between aspirin users and nonusers. In multivariate analyses, prediagnosis aspirin use was not associated with all-cause mortality [hazard ratio (HR) 0.86, 95% confidence interval (CI) 0.48–1.57] or esophageal cancer-specific mortality (HR 1.07, 95% CI 0.52–2.21). Prediagnosis aspirin use was associated with a significantly increased risk of interval metastasis (HR 3.59, 95% CI 1.08–11.96). Conclusions: In our cohort of esophageal cancer patients, prediagnosis aspirin use was not associated with all-cause or cancer-specific mortality. However, risk of interval metastatic disease was increased among those who took aspirin regularly prediagnosis. Future studies are warranted to assess whether aspirin influences the molecular characteristics of esophageal tumors, with potential prognostic and therapeutic implications.

Objective: Accurate classification methods are critical in computer-aided diagnosis (CADx) and other clinical decision support systems. Previous research has reported on methods for combining genetic algorithm (GA) feature selection with ensemble classifier systems in an effort to increase classification accuracy. In this study, we describe a CADx system for pulmonary nodules using a two-step supervised learning system combining a GA with the random subspace method (RSM), with the aim of exploring algorithm design parameters and demonstrating improved classification performance over either the GA or RSM-based ensembles alone. Methods and materials: We used a retrospective database of 125 pulmonary nodules (63 benign; 62 malignant) with CT volumes and clinical history. A total of 216 features were derived from the segmented image data and clinical history. Ensemble classifiers using RSM or GA-based feature selection were constructed and tested via leave-one-out validation with feature selection and classifier training executed within each iteration. We further tested a two-step approach using a GA ensemble to first assess the relevance of the features, and then using this information to control feature selection during a subsequent RSM step. The base classification was performed using linear discriminant analysis (LDA). Results: The RSM classifier alone achieved a maximum leave-one-out Az of 0.866 (95% confidence interval: 0.794-0.919) at a subset size of s=36 features. The GA ensemble yielded an Az of 0.851 (0.775-0.907). The proposed two-step algorithm produced a maximum Az value of 0.889 (0.823-0.936) when the GA ensemble was used to completely remove less relevant features from the second RSM step, with similar results obtained when the GA-LDA results were used to reduce but not eliminate the occurrence of certain features. After accounting for correlations in the data, the leave-one-out Az in the two-step method was significantly higher than in the RSM and the GA-LDA. Conclusions: We have developed a CADx system for evaluation of pulmonary nodule based on a two-step feature selection and ensemble classifier algorithm. We have shown that by combining classifier ensemble algorithms in this two-step manner, it is possible to predict the malignancy for solitary pulmonary nodules with a performance exceeding that of either of the individual steps.

Objective Computer-aided diagnosis (CADx) systems can be utilized to estimate the likelihood of malignancy of pulmonary nodules detected on CT scans. Such systems often require an initial segmentation step. We report on the impact of segmentation variations on the CADx output and describe methods for reducing this effect.

Background: Esophageal cancer survival rates remain extremely low and factors influencing outcomes for this malignancy are not well understood. Tumor cachexia is a poor prognostic factor for certain tumor types, but is not well-studied in esophageal cancer. Weight loss in esophageal cancer is likely multifactorial; it can be due to tumor cachexia as well as dysphagia from obstructing tumors. In this present study, we aimed to investigate the relationship between weight loss and overall survival in a cohort of esophageal cancer patients and to determine whether these associations differed with tumor size. Methods: We prospectively enrolled subjects with recently diagnosed esophageal cancer at two tertiary care centers. Using a baseline questionnaire, we assessed demographics, medical history, medication use, and lifestyle factors. We recorded self-reported height and weight one year prior to and at diagnosis, which we used to calculate body mass index (BMI) and percent weight change, categorized by tertile. We ascertained from the medical records tumor characteristics including T stage (T1/2 or T3/4), location, pathology, and metastasis status, and collected follow-up data on treatment, imaging, and death. We used Cox regression to assess the association between percent weight loss and all-cause mortality. Results: We included 134 subjects in the analyses, the majority of whom were male (81.3%) and had adenocarcinoma (82.1%). The median BMI one year prior to diagnosis was 28.3 (IQR 24.4 - 31.0) and median percent weight loss was 4.7% (IQR 0 - 10.9%). There was no association between BMI one year prior and all-cause mortality (HR 0.98, 95% CI 0.93 - 1.03). Increasing percent weight loss was associated with increased risk of all-cause mortality (unadjusted HR 2.74 for highest vs. lowest tertile, 95% CI 1.34 - 5.58, Ptrend = 0.005) and this remained significant when adjusted for BMI one year prior (HR 2.87 for highest vs. lowest tertile, 95% CI 1.40 - 5.89, Ptrend = 0.003). We found significant interaction between weight loss and T stage. Percent weight loss was significantly associated with all-cause mortality, adjusted for BMI one year prior, among patients with T stages 1 or 2 (HR 6.49 for highest vs. lowest tertile, 95% CI 1.30 - 32.4, Ptrend = 0.012), but not T stages 3 or 4 (HR 1.42 for highest vs. lowest tertile, 95% CI 0.54 - 3.71). In the final multivariable model, there remained a significant association between percent weight loss and all-cause mortality among patients with T stages 1 or 2 (HR 6.34 for highest vs. lowest tertile, 95% CI 1.22 - 33.1, Ptrend = 0.022). Conclusions: In this cohort of esophageal cancer patients, we found that pre-diagnosis weight loss was associated with increased risk of all-cause mortality in patients with earlier stage tumors, independent of baseline BMI. We suspect that weight loss in early stage esophageal cancer may be due to tumor cachexia, a potential marker of more aggressive disease and worse prognosis. Citation Format: Sherry Shen, James L. Araujo, Nasser K. Altorki, Joshua R. Sonett, Adriana Rodriguez, Kivilcim Sungur-Stasik, Alfred I. Neugut, Julian A. Abrams. Impact of pre-diagnosis weight loss on outcomes in a prospective cohort of esophageal cancer patients. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2637.

Accurate classification methods are critical in computer-aided diagnosis and other clinical decision support systems. Previous research has studied methods for combining genetic algorithms for feature selection with ensemble classifier systems in an effort to increase classification accuracy. We propose a two-step approach that first uses genetic algorithms to reduce the number of features used to characterize the data, then applies the random subspace method on the remaining features to create a set of diverse but high performing classifiers. These classifiers are combined using ensemble learning techniques to yield a final classification. We demonstrate this approach for computer-aided diagnosis of solitary pulmonary nodules from CT scans, in which the proposed method outperforms several previously described methods.