Survival prediction of glioblastoma patients using modern deep learning and machine learning techniques.

In this study, we utilized data from the Surveillance, Epidemiology, and End Results (SEER) database to predict the glioblastoma patients' survival outcomes. To assess dataset skewness and detect feature importance, we applied Pearson's second coefficient test of skewness and the Ordinary Least Squares method, respectively. Using two sampling strategies, holdout and five-fold cross-validation, we developed five machine learning (ML) models alongside a feed-forward deep neural network (DNN) for the multiclass classification and regression prediction of glioblastoma patient survival. After balancing the classification and regression datasets, we obtained 46,340 and 28,573 samples, respectively. Shapley additive explanations (SHAP) were then used to explain the decision-making process of the best model. In both classification and regression tasks, as well as across holdout and cross-validation sampling strategies, the DNN consistently outperformed the ML models. Notably, the accuracy were 90.25% and 90.22% for holdout and five-fold cross-validation, respectively, while the corresponding R2 values were 0.6565 and 0.6622. SHAP analysis revealed the importance of age at diagnosis as the most influential feature in the DNN's survival predictions. These findings suggest that the DNN holds promise as a practical auxiliary tool for clinicians, aiding them in optimal decision-making concerning the treatment and care trajectories for glioblastoma patients.

Quality of in-hospital care in traumatic spinal column and cord injuries (TSC/SCI) in I.R Iran.

PURPOSE: This study aimed to implement the Quality of Care (QoC) Assessment Tool from the National Spinal Cord/Column Injury Registry of Iran (NSCIR-IR) to map the current state of in-hospital QoC of individuals with Traumatic Spinal Column and Cord Injuries (TSCCI). METHODS: The QoC Assessment Tool, developed from a scoping review of the literature, was implemented in NSCIR-IR. We collected the required data from two primary sources. Questions regarding health system structures and care processes were completed by the registrar nurse reviewing the hospital records. Questions regarding patient outcomes were gathered through patient interviews. RESULTS: We registered 2812 patients with TSCCI over six years from eight referral hospitals in NSCIR-IR. The median length of stay in the general hospital and intensive care unit was four and five days, respectively. During hospitalization 4.2% of patients developed pressure ulcers, 83.5% of patients reported satisfactory pain control and none had symptomatic urinary tract infections. 100%, 80%, and 90% of SCI registration centers had 24/7 access to CT scans, MRI scans, and operating rooms, respectively. Only 18.8% of patients who needed surgery underwent a surgical operation in the first 24 h after admission. In-hospital mortality rate for patients with SCI was 19.3%. CONCLUSION: Our study showed that the current in-hospital care of our patients with TSCCI is acceptable in terms of pain control, structure and length of stay and poor regarding in-hospital mortality rate and timeliness. We must continue to work on lowering rates of pressure sores, as well as delays in decompression surgery and fatalities.

Post-discharge follow-up of patients with spine trauma in the National Spinal Cord Injury Registry of Iran during the COVID-19 pandemic: Challenges and lessons learned.

PURPOSE: The purpose of the National Spinal Cord Injury Registry of Iran (NSCIR-IR) is to create an infrastructure to assess the quality of care for spine trauma and in this study, we aim to investigate whether the NSCIR-IR successfully provides necessary post-discharge follow-up data for these patients. METHODS: An observational prospective study was conducted from April 11, 2021 to April 22, 2022 in 8 centers enrolled in NSCIR-IR, respectively Arak, Rasht, Urmia, Shahroud, Yazd, Kashan, Tabriz, and Tehran. Patients were classified into three groups based on their need for care resources, respectively: (1) non-spinal cord injury (SCI) patients without surgery (group 1), (2) non-SCI patients with surgery (group 2), and (3) SCI patients (group 3). The assessment tool was a self-designed questionnaire to evaluate the care quality in 3 phases: pre-hospital, in-hospital, and post-hospital. The data from the first 2 phases were collected through the registry. The post-hospital data were collected by conducting follow-up assessments. Telephone follow-ups were conducted for groups 1 and 2 (non-SCI patients), while group 3 (SCI patients) had a face-to-face visit. This study took place during the COVID-19 pandemic. Data on age and time interval from injury to follow-up were expressed as mean ± standard deviation (SD) and response rate and follow-up loss as a percentage. RESULTS: Altogether 1538 telephone follow-up records related to 1292 patients were registered in the NSCIR-IR. Of the total calls, 918 (71.05%) were related to successful follow-ups, but 38 cases died and thus were excluded from data analysis. In the end, post-hospital data from 880 patients alive were gathered. The success rate of follow-ups by telephone for groups 1 and 2 was 73.38% and 67.05% respectively, compared to 66.67% by face-to-face visits for group 3, which was very hard during the COVID-19 pandemic. The data completion rate after discharge ranged from 48% to 100%, 22%-100% and 29%-100% for groups 1 - 3. CONCLUSIONS: To improve patient accessibility, NSCIR-IR should take measures during data gathering to increase the accuracy of registered contact information. Regarding the loss to follow-ups of SCI patients, NSCIR-IR should find strategies for remote assessment or motivate them to participate in follow-ups through, for example, providing transportation facilities or financial support.

Potential drug-drug interactions in a multi-center study of death cases of COVID-19: The significance of appropriate drug choice.

OBJECTIVES: The prevalence, types, severity, risk ratings, and common pairs of involved drugs, and the most important potential drug-drug interactions (pDDIs) in coronavirus disease 2019 (-COVID-19) deceased cases were evaluated. MATERIALS AND METHODS: We reviewed the medical records of 157 confirmed COVID-19 deceased cases hospitalized in 27 province-wide hospitals. Patients' demographics and clinical data (including comorbidities, vital signs, length of in-hospital survival, electrocardiograms (ECGs), medications, and lab test results) were extracted. The online Lexi-interact database and Stockley's drug interactions reference were used to detect pDDIs retrospectively. The QTc interval and total Tisdale risk score were also calculated. Descriptive analysis, analysis of variance, Fisher exact test, and multivariate analysis were conducted for data analysis. RESULTS: Of 157 study cases, 63% were male, had a mean age of 68 years, and 55.7% had one or more underlying diseases. All patients had polypharmacy, with 69.2% having ≥ 15 drugs/day. We detected 2,416 pDDIs in patients' records, of which 658 (27.2%) were interactions with COVID drugs. Lopinavir/ritonavir among -COVID drugs and fentanyl among non-COVID drugs were commonly involved in the interactions. pDDIs was significantly higher in the polypharmacy group of ≥ 15 medications (p < 0.001). A majority (83%) had received drug(s) with the QTc prolongation effect, of whom 67% had actual QTc prolongations in their ECGs. The regression analysis showed that by increasing 6.7% in polypharmacy, one day increase in-hospital survival can be expected. Moreover, an increase of 2.3% in white blood cells or 10.5% in serum potassium level decreased in-hospital survival by 1%. CONCLUSION: The findings underscored the importance of careful drug choice, especially in the hectic search for early treatments in pandemics of novel diseases. Close monitoring of patients' drug choice is warranted for reducing pDDIs and their adverse effects in any new disease outbreak.

Association between Socioeconomic Factors and the Risk of Gastric Cancer Incidence: Results from an Ecological Study.

Background: Gastric cancer (GC), one of the most common cancer worldwide, remains the third leading cause of cancer-related mortality. The etiology of GC may arise from genetic and environmental factors. This study aimed to determine the association between GC incidence and socioeconomic status in Iran. Methods: An ecological study was designed to investigate the relationship between socioeconomic factors and the risk of GC incidence. The data of socioeconomic variables such as income changes, unemployment rate, urbanization ratio, inflation rate, and air pollution changes in 31 provinces were collected from the Statistical Center of Iran, and the data of GC of 31 provinces were provided from the Iranian National Population-based Cancer Registry (INPCR). Data from 2014 to 2017 was analyzed using panel data analysis, the fixed effects model by EViews software. Results: Panel data model was suitable for the present study. Results showed that there was a positive and significant relationship between GC incidence and socioeconomic factors including income changes (P≤ 0.001), unemployment rate (P≤0.01), inflation rate (P≤ 0.05), and air pollution changes (P≤ 0.001). The urbanization ratio showed a negative relationship and was not statistically associated with GC incidence (P> 0.05). Conclusion: Our findings suggest a positive and significant association between socioeconomic status and GC incidence, proposing a GC risk factor. The key public health policies and welfare policies' priority should therefore be to schedule for the GC management.

Clinical Breast Cancer Registry of IR. Iran (CBCR-IR): Study Protocol and First Results.

BACKGROUND: Breast cancer (BC), as a significant global health problem, is the most common cancer in women. Despite the importance of clinical cancer registries in improving the quality of cancer care and cancer research, there are few reports on them from low- and middle-income countries. We established a multicenter clinical breast cancer registry in Iran (CBCR-IR) to collect data on BC cases, the pattern of care, and the quality-of-care indicators in different hospitals across the country. METHODS: We established a clinical cancer registry in 12 provinces of Iran. We defined the organizational structure, developed minimal data sets and data dictionaries, verified data sources and registration processes, and developed the necessary registry software. During this registry, we studied the clinical characteristics and outcomes of patients with cancer who were admitted from 2014 onwards. RESULTS: We registered 13086 BC cases (7874 eligible cases) between 1.1.2014 and 1.1.2022. Core needle biopsy from the tumor (61.25%) and diagnostic mammography (68.78%) were the two most commonly used diagnostic methods. Stage distribution was 2.03% carcinoma in situ, 12% stage I, 44.65% stage II, 21.32% stage III, and 4.61% stage IV; stage information was missing in 1532 patients (19.46%). Surgery (95.01%) and chemotherapy (79.65%) were the most common treatments for all patients. CONCLUSION: The information provided by this registry can be used to evaluate and improve the quality of care for BC patients. It will be scaled up to the national level as an important resource for measuring quality of care and conducting clinical cancer research in Iran.

Application of machine learning techniques for predicting survival in ovarian cancer.

BACKGROUND: Ovarian cancer is the fifth leading cause of mortality among women in the United States. Ovarian cancer is also known as forgotten cancer or silent disease. The survival of ovarian cancer patients depends on several factors, including the treatment process and the prognosis. METHODS: The ovarian cancer patients' dataset is compiled from the Surveillance, Epidemiology, and End Results (SEER) database. With the help of a clinician, the dataset is curated, and the most relevant features are selected. Pearson's second coefficient of skewness test is used to evaluate the skewness of the dataset. Pearson correlation coefficient is also used to investigate the associations between features. Statistical test is utilized to evaluate the significance of the features. Six Machine Learning (ML) models, including K-Nearest Neighbors , Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost), and Extreme Gradient Boosting (XGBoost), are implemented for survival prediction in both classification and regression approaches. An interpretable method, Shapley Additive Explanations (SHAP), is applied to clarify the decision-making process and determine the importance of each feature in prediction. Additionally, DTs of the RF model are displayed to show how the model predicts the survival intervals. RESULTS: Our results show that RF (Accuracy = 88.72%, AUC = 82.38%) and XGBoost (Root Mean Squad Error (RMSE)) = 20.61%, R2 = 0.4667) have the best performance for classification and regression approaches, respectively. Furthermore, using the SHAP method along with extracted DTs of the RF model, the most important features in the dataset are identified. Histologic type ICD-O-3, chemotherapy recode, year of diagnosis, age at diagnosis, tumor stage, and grade are the most important determinant factors in survival prediction. CONCLUSION: To the best of our knowledge, our study is the first study that develops various ML models to predict ovarian cancer patients' survival on the SEER database in both classification and regression approaches. These ML algorithms also achieve more accurate results and outperform statistical methods. Furthermore, our study is the first study to use the SHAP method to increase confidence and transparency of the proposed models' prediction for clinicians. Moreover, our developed models, as an automated auxiliary tool, can help clinicians to have a better understanding of the estimated survival as well as important features that affect survival.

Prediction of future gene expression profile by analyzing its past variation pattern.

A number of initial Hematopoietic Stem Cells (HSC) are considered in a container that are able to divide into HSCs or differentiate into various types of descendant cells. In this paper, a method is designed to predict an approximate gene expression profile (GEP) for future descendant cells resulted from HSC division/differentiation. First, the GEP prediction problem is modeled into a multivariate time series prediction problem. A novel method called EHSCP (Extended Hematopoietic Stem Cell Prediction) is introduced which is an artificial neural machine to solve the problem. EHSCP accepts the initial sequence of measured GEPs as input and predicts GEPs of future descendant cells. This prediction can be performed for multiple stages of cell division/differentiation. EHSCP considers the GEP sequence as time series and computes correlation between input time series. Two novel artificial neural units called PLSTM (Parametric Long Short Term Memory) and MILSTM (Multi-Input LSTM) are designed. PLSTM makes EHSCP able to consider this correlation in output prediction. Since there exist thousands of time series in GEP prediction, a hierarchical encoder is proposed that computes this correlation using 101 MILSTMs. EHSCP is trained using 155 datasets and is evaluated on 39 test datasets. These evaluations show that EHSCP surpasses existing methods in terms of prediction accuracy and number of correctly-predicted division/differentiation stages. In these evaluations, number of correctly-predicted stages in EHSCP was 128 when as many as 8 initial stages were given.

Cancer incidence in Iran in 2014: Results of the Iranian National Population-based Cancer Registry.

BACKGROUND: We aimed to report, for the first time, the results of the Iranian National Population-based Cancer Registry (INPCR) for the year 2014. METHODS: Total population of Iran in 2014 was 76,639,000. The INPCR covered 30 out of 31 provinces (98% of total population). It registered only cases diagnosed with malignant new primary tumors. The main sources for data collection included pathology center, hospitals as well as death registries. Quality assessment and analysis of data were performed by CanReg-5 software. Age standardized incidence rates (ASR) (per 100,000) were reported at national and subnational levels. RESULTS: Overall, 112,131 new cancer cases were registered in INPCR in 2014, of which 60,469 (53.9%) were male. The diagnosis of cancer was made by microscopic confirmation in 76,568 cases (68.28%). The ASRs of all cancers were 177.44 and 141.18 in male and female, respectively. Cancers of the stomach (ASR = 21.24), prostate (18.41) and colorectum (16.57) were the most common cancers in men and the top three cancers in women were malignancies of breast (34.53), colorectum (11.86) and stomach (9.44). The ASR of cervix uteri cancer in women was 1.78. Our findings suggested high incidence of cancers of the esophagus, stomach and lung in North/ North West of Iran. CONCLUSION: Our results showed that Iran is a medium-risk area for incidence of cancers. We found differences in the most common cancers in Iran comparing to those reported for the World. Our results also suggested geographical diversities in incidence rates of cancers in different subdivisions of Iran.

Evaluating the quality of intravenous regional anesthesia following adding dexamethasone to lidocaine.

OBJECTIVES: The quality of anesthesia in intravenous regional anesthesia (IVRA) has been evaluated in many studies so far. This study was designed to evaluate the effects of adding the dexamethasone to lidocaine on the quality of IVRA. MATERIALS AND METHODS: A double-blind clinical trial was set up involving 50 hand surgery candidates, 20 to 55 years old, and with American Society of Anesthesiologists class of I and II. Patients were randomly allocated into two groups of 25 cases and received either 3 mg/kg of lidocaine (control group) or 3 mg/kg of lidocaine plus 8 mg of dexamethasone (study group). The onset and recovery times from sensory and motor blocks, the starting time of tourniquet pain, the amount of narcotics needed during patients' recovery, and probable side-effects were all compared between the two groups. RESULTS: No significant differences were detected concerning age, gender, length of surgery and the mean time of starting of tourniquet pain between the two groups. The mean times of both sensory (P = 0.002) and motor (P = 0.004) blocks onset were significantly shorter in the study group. The mean time of recovery from sensory block was significantly longer in the study group (P = 0.01). The average amount of narcotics needed during the recovery was significantly lower in the study group (P = 0.01). No side-effect was detected. CONCLUSION: We conclude that adding the dexamethasone to lidocaine can improve the quality of anesthesia in IVRA.

Improving chemotherapy processes with a protocol-based information system: a pre and post-implementation study.

BACKGROUND: The medical application domain has been a great challenge for information technology solutions for decades, especially when the target process has been complex and multidisciplinary such as chemotherapy processes. OBJECTIVE: To evaluate the impact of a homegrown protocol based information system on the efficiency of chemotherapy workflow processes in an outpatient setting. METHODS: A day care unit of the Hematology/Oncology outpatient clinic of Erasmus Medical Center was the setting for this study. The study consisted of comparison of pre- and post-implementation of four workflow efficiency related external indicators: turn-around times of a commonly administered chemotherapy course (Paclitaxel-Carboplatin), chemotherapy course administration postponing rate, the rate of recording course administration time, and patient admission rate of the outpatient clinic. The data was gathered retrospectively from patient charts and information systems' log files. For the purpose of turn-around-time 109 Paclitaxel-Carboplatin chemotherapy courses of pre-implementation were compared to 118 those of post-implementation. For the other indicators: 247 chemotherapy courses pre-implementation were compared to 324 courses post-implementation. The process maps of pre- and post-implementation were also compared to each other. RESULTS: The implementation of the system improved the process by removing repetition and sequencing of the tasks. Following the implementation, chemotherapy postponing decreased by 17.2% (Z = -4.723, P = .000) and there were 5.7% less records with missing administration time (Z =-3.047, P = .002). The admission rate increased 1.9 patient per working day (t(94) = -5.974, P = .000). The overall turn-around-time reduced 18.9 min following the implementation (t(169) = 3.48, P = .001). In a multivariate multiple regression model the reduction in turn-around time was related to the implementation of the system (Pillai's Trace = 0.159, F(4,161) = 7.613, P = .000). CONCLUSION: Information systems based on treatment protocols can reduce communication and synchronization needs between the stakeholders in a complex workflow process. These systems can help reengineering the process and improve workflow efficiency by removing unnecessary sequencing and repetitions of tasks.

What makes an information system more preferable for clinicians? a qualitative comparison of two systems.

Two different information systems with respect to their ability to afford clinicians' needs in the chemotherapy medication process were implemented in a large Dutch academic hospital. A commercially available Computerized Physician Order Entry (CPOE) system was not appreciated because clinicians believed that it could not support complex chemotherapy process. Later, a home-grown IT system with the capability of prescribing chemotherapy medications based on standard care protocols was appreciated and fully used by clinicians. We evaluated both systems from their users' perspective to find the sources of clinicians' preference and to trace them back to their Systems Development Life Cycle (SDLC).

The nature of unintended effects of health information systems concerning patient safety: a systematic review with thematic synthesis.

In order to understand the nature and causes through which Health Information Systems (HIS) can affect patient safety negatively, a systematic review with thematic synthesis of the qualitative studies was performed. 26 papers met our criteria and were included into content analysis. 40 error contributing factors in working with HIS were recognized. Upon which, 4 main categories of contributing factors were defined. Analysis of the semantic relation between contributing reasons and common types of errors in healthcare practice revealed 6 mechanisms that can function as secondary contributing reasons. Results of this study can support care providers, system designers, and system implementers to avoid unintended negative effects for patient safety.