ABSTRACT
Mobile crowd sensing (MCS) is an application that collects data from a network of conscientious volunteers and implements it for the common or personal benefit. This contribution proposes an implementation that collects the data from hypertensive patients, thus creating an experimental database using the cloud service Platform as a Service (PaaS). The challenge is to perform the analysis without the main diagnostic feature for hypertension-the blood pressure. The other problems consider the data reliability in an environment full of artifacts and with limited bandwidth and battery resources. In order to motivate the MCS volunteers, a feedback about the patient's current status is created, provided by the means of machine-learning (ML) techniques. Two techniques are investigated and the Random Forest algorithm yielded the best results. The proposed platform, with slight modifications, can be adapted to the patients with other cardiovascular problems.
Subject(s)
Hypertension/diagnosis , Mobile Applications , Algorithms , Artifacts , Electrocardiography , Heart Rate , Humans , Hypertension/physiopathology , ROC Curve , Signal Processing, Computer-AssistedABSTRACT
PURPOSE: Presence of metastasis in sentinel lymph node (SLN) is considered to be the most important factor in terms of patient survival. The main aim of this study was to identify predictors of positive SLN in Serbian patients with melanoma. METHODS: This retrospective study was conducted on 147 patients. Univariate chi-square and univariate logistic regression analyses were used to identify the association between prognostic factors and positive SLN. Receiver Operating Characteristics (ROC) was conducted to find the Breslow thickness cutoff point at which to perform SLN biopsy (SLNB). Kaplan-Meier analysis was used to evaluate disease-free survival (DFS), and log rank test was applied to compare differences between groups. RESULTS: Breslow thickness and Clark level (p≤0.05), presence of ulceration and a high mitotic rate (>6 mitoses/mm2) (p<0.001) were significant independent predictors of SLN metastasis. ROC curve showed that Breslow thickness of 2.8 mm was the most suitable cutoff point for SLN positivity (sensitivity 86%, specificity 67%). Furthermore, Breslow thickness and presence of ulceration were found to be associated with DFS (p<0.05). CONCLUSIONS: Patients with Breslow thickness ≥2.8 mm, ulceration, and high mitotic rate are at higher risk for SLN metastasis. In addition, high Breslow thickness and presence of ulceration are associated with decreased DFS. These results indicate that multiple selection criteria should be used when performing and predicting SLN metastasis and disease recurrence.
Subject(s)
Melanoma/diagnosis , Neoplasm Recurrence, Local/diagnosis , Prognosis , Sentinel Lymph Node/pathology , Skin Neoplasms/diagnosis , Adult , Female , Humans , Kaplan-Meier Estimate , Lymphatic Metastasis , Male , Melanoma/epidemiology , Melanoma/pathology , Middle Aged , Neoplasm Recurrence, Local/epidemiology , Neoplasm Recurrence, Local/pathology , Sentinel Lymph Node Biopsy , Serbia/epidemiology , Skin Neoplasms/epidemiology , Skin Neoplasms/pathology , Melanoma, Cutaneous MalignantABSTRACT
A female patient aged 42, started chemotherapy for advanced ovarian carcinoma in June 2016. Considering intraoperative findings, cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (HIPEC) were performed, followed by adjuvant chemotherapy. In March 2018, computed tomography (CT) examination showed disease progression in the form of pleural carcinomatosis with increased levels of tumor markers. In April 2018, total parietal pleurectomy, partial visceral pleurectomy, and then hyperthermic intrathoracic chemotherapy (HITHOC) with cisplatin were performed. The procedure was uneventful, as was the postoperative course. The patient was discharged on the 13th postoperative day with no major postoperative complications. Three months after surgery, CT showed no signs of disease relapse. Since this is a relatively new method of treating pleural carcinomatosis, real results are to be expected with larger series of patients and longer postoperative follow-up.
Subject(s)
Antineoplastic Combined Chemotherapy Protocols/therapeutic use , Chemotherapy, Cancer, Regional Perfusion/methods , Hyperthermia, Induced/methods , Ovarian Neoplasms/therapy , Peritoneal Neoplasms/therapy , Adult , Combined Modality Therapy , Female , Humans , Ovarian Neoplasms/pathology , Peritoneal Neoplasms/secondary , PrognosisABSTRACT
PURPOSE: To analyze the correlation of primary tumor (PT) pathological characteristics (size, stage, type and grade) and the extent of initial surgical treatment of non small cell lung cancer (NSCLC) with the incidence and time to local recurrence (LR) and disease-free survival (DFS), as well as to determine in what way these parameters and LR localizations affect the possibility for surgical retreatment. METHODS: The research was conducted on 114 patients with NSCLC and LR that had initial surgery in two reference institutes in Serbia from January 2002 to December 2010. PT size and disease stage were defined according to the revised 2004 WHO classification. PTs were grouped by size into 3 categories. Due to great diversity, surgical procedures were sorted into 6 operation types. Standard statistical methods and tests were used for data analysis. RESULTS: Statistical analyses showed significant difference in DFS and LR reoperability that were related to PT size, disease stage and the extent of initial surgery. LR localization on the chest wall was favorable for secondary surgery due to LR. CONCLUSIONS: Squamous cell lung carcinoma relapses locally more frequently than other lung tumor types, and the commonest LR site is the chest wall. This localization provides high possibility for surgical retreatment. Adequate staging, proper indications for surgical treatment and quality surgery provide longer DFS in patients with NSCLC. All these suggest that the surgeon may be considered as the most significant factor of prognosis.
