Death rate has been age-adjusted by the 2000 U.S. standard population. Single-year data are only available for Los Angeles County overall, Service Planning Areas, Supervisorial Districts, City of Los Angeles overall, and City of Los Angeles Council Districts.Lung cancer is a leading cause of cancer-related death in the US. People who smoke have the greatest risk of lung cancer, though lung cancer can also occur in people who have never smoked. Most cases are due to long-term tobacco smoking or exposure to secondhand tobacco smoke. Cities and communities can take an active role in curbing tobacco use and reducing lung cancer by adopting policies to regulate tobacco retail; reducing exposure to secondhand smoke in outdoor public spaces, such as parks, restaurants, or in multi-unit housing; and improving access to tobacco cessation programs and other preventive services.For more information about the Community Health Profiles Data Initiative, please see the initiative homepage.

This map service portrays the number of deaths per 100,000 people per square mile from lung and colon cancer. It displays the distribution of lung and colon cancer across the United States. Pop-ups show attributes such as state name, county name, number of colon or lung cancer deaths, and square miles per area.Lung Cancer: Death due to malignant neoplasm of the trachea, bronchus and lung.Colon Cancer: Death due to malignant neoplasm of the colon, rectum and anus.This data was sourced from: Community Health Status Indicators_Other Health Datapalooza focused content that may interest you: Health Datapalooza Health Datapalooza

To reduce deaths from lung cancer. For information on the definitions of what these indicators include, please see the relevant specification. From 2016 onwards, mortality counts within the Compendium Mortality Indicator set are based on a bespoke extract taken from the Primary Care Mortality Database (PCMD) maintained by NHS Digital. PCMD is updated monthly using a file of death records from ONS and is continually subject to amendment. It is already well established that late registrations have a small impact on counts. This bespoke extract may be taken at a different time to that of the mortality data published by ONS and as such this may cause some small differences between ONS and NHS Digital mortality figures for a given year.

https://www.cancerimagingarchive.net/wp-content/uploads/nctn-logo-300x108.png" alt="" width="300" height="108" />

Demographic Summary of Available Imaging

Characteristic	Value (N = 26254)
Age (years)	Mean ± SD: 61.4± 5 Median (IQR): 60 (57-65) Range: 43-75
Sex	Male: 15512 (59%) Female: 10742 (41%)
Race	White: 23969 (91.3%) Black: 1135 (4.3%) Asian: 547 (2.1%) American Indian/Alaska Native: 88 (0.3%) Native Hawaiian/Other Pacific Islander: 87 (0.3%) Unknown: 428 (1.6%)
Ethnicity	Not Available

Background: The aggressive and heterogeneous nature of lung cancer has thwarted efforts to reduce mortality from this cancer through the use of screening. The advent of low-dose helical computed tomography (CT) altered the landscape of lung-cancer screening, with studies indicating that low-dose CT detects many tumors at early stages. The National Lung Screening Trial (NLST) was conducted to determine whether screening with low-dose CT could reduce mortality from lung cancer.

Methods: From August 2002 through April 2004, we enrolled 53,454 persons at high risk for lung cancer at 33 U.S. medical centers. Participants were randomly assigned to undergo three annual screenings with either low-dose CT (26,722 participants) or single-view posteroanterior chest radiography (26,732). Data were collected on cases of lung cancer and deaths from lung cancer that occurred through December 31, 2009. This dataset includes the low-dose CT scans from 26,254 of these subjects, as well as digitized histopathology images from 451 subjects.

Results: The rate of adherence to screening was more than 90%. The rate of positive screening tests was 24.2% with low-dose CT and 6.9% with radiography over all three rounds. A total of 96.4% of the positive screening results in the low-dose CT group and 94.5% in the radiography group were false positive results. The incidence of lung cancer was 645 cases per 100,000 person-years (1060 cancers) in the low-dose CT group, as compared with 572 cases per 100,000 person-years (941 cancers) in the radiography group (rate ratio, 1.13; 95% confidence interval [CI], 1.03 to 1.23). There were 247 deaths from lung cancer per 100,000 person-years in the low-dose CT group and 309 deaths per 100,000 person-years in the radiography group, representing a relative reduction in mortality from lung cancer with low-dose CT screening of 20.0% (95% CI, 6.8 to 26.7; P=0.004). The rate of death from any cause was reduced in the low-dose CT group, as compared with the radiography group, by 6.7% (95% CI, 1.2 to 13.6; P=0.02).

Conclusions: Screening with the use of low-dose CT reduces mortality from lung cancer. (Funded by the National Cancer Institute; National Lung Screening Trial ClinicalTrials.gov number, NCT00047385).

Data Availability: A summary of the National Lung Screening Trial and its available datasets are provided on the Cancer Data Access System (CDAS). CDAS is maintained by Information Management System (IMS), contracted by the National Cancer Institute (NCI) as keepers and statistical analyzers of the NLST trial data. The full clinical data set from NLST is available through CDAS. Users of TCIA can download without restriction a publicly distributable subset of that clinical data, along with the CT and Histopathology images collected during the trial. (These previously were restricted.)

What is Lung Cancer Dataset?

The effectiveness of the cancer prediction system helps people to know their cancer risk at a low cost and it also helps the people to take the appropriate decision based on their cancer risk status. The data is collected from the website online lung cancer prediction system.

.

https://user-images.githubusercontent.com/36210723/182395183-ef7519e3-9c18-47ac-b7a6-a00e234f3949.png" alt="2022-08-02_170741">

.

Acknowledgments

When we use this dataset in our research, we credit the authors as :

• License : CC BY 4.0.

• Hong, Z.Q. and Yang, J.Y. "Optimal Discriminant Plane for a Small Number of Samples and Design Method of Classifier on the Plane", Pattern Recognition, Vol. 24, No. 4, pp. 317-324, 1991 and it is published t to reuse in google research dataset

The main idea for uploading this dataset is to practice data analysis with my students, as I am working in college and want my student to train our studying ideas in a big dataset, It may be not up to date and I mention the collecting years, but it is a good resource of data to practice

This dataset consists of CT and PET-CT DICOM images of lung cancer subjects with XML Annotation files that indicate tumor location with bounding boxes. The images were retrospectively acquired from patients with suspicion of lung cancer, and who underwent standard-of-care lung biopsy and PET/CT. Subjects were grouped according to a tissue histopathological diagnosis. Patients with Names/IDs containing the letter 'A' were diagnosed with Adenocarcinoma, 'B' with Small Cell Carcinoma, 'E' with Large Cell Carcinoma, and 'G' with Squamous Cell Carcinoma.

The images were analyzed on the mediastinum (window width, 350 HU; level, 40 HU) and lung (window width, 1,400 HU; level, –700 HU) settings. The reconstructions were made in 2mm-slice-thick and lung settings. The CT slice interval varies from 0.625 mm to 5 mm. Scanning mode includes plain, contrast and 3D reconstruction.

Before the examination, the patient underwent fasting for at least 6 hours, and the blood glucose of each patient was less than 11 mmol/L. Whole-body emission scans were acquired 60 minutes after the intravenous injection of 18F-FDG (4.44MBq/kg, 0.12mCi/kg), with patients in the supine position in the PET scanner. FDG doses and uptake times were 168.72-468.79MBq (295.8±64.8MBq) and 27-171min (70.4±24.9 minutes), respectively. 18F-FDG with a radiochemical purity of 95% was provided. Patients were allowed to breathe normally during PET and CT acquisitions. Attenuation correction of PET images was performed using CT data with the hybrid segmentation method. Attenuation corrections were performed using a CT protocol (180mAs,120kV,1.0pitch). Each study comprised one CT volume, one PET volume and fused PET and CT images: the CT resolution was 512 × 512 pixels at 1mm × 1mm, the PET resolution was 200 × 200 pixels at 4.07mm × 4.07mm, with a slice thickness and an interslice distance of 1mm. Both volumes were reconstructed with the same number of slices. Three-dimensional (3D) emission and transmission scanning were acquired from the base of the skull to mid femur. The PET images were reconstructed via the TrueX TOF method with a slice thickness of 1mm.

The location of each tumor was annotated by five academic thoracic radiologists with expertise in lung cancer to make this dataset a useful tool and resource for developing algorithms for medical diagnosis. Two of the radiologists had more than 15 years of experience and the others had more than 5 years of experience. After one of the radiologists labeled each subject the other four radiologists performed a verification, resulting in all five radiologists reviewing each annotation file in the dataset. Annotations were captured using Labellmg. The image annotations are saved as XML files in PASCAL VOC format, which can be parsed using the PASCAL Development Toolkit: https://pypi.org/project/pascal-voc-tools/. Python code to visualize the annotation boxes on top of the DICOM images can be downloaded here.

Two deep learning researchers used the images and the corresponding annotation files to train several well-known detection models which resulted in a maximum a posteriori probability (MAP) of around 0.87 on the validation set.

Number and rate of new cancer cases diagnosed annually from 1992 to the most recent diagnosis year available. Included are all invasive cancers and in situ bladder cancer with cases defined using the Surveillance, Epidemiology and End Results (SEER) Groups for Primary Site based on the World Health Organization International Classification of Diseases for Oncology, Third Edition (ICD-O-3). Random rounding of case counts to the nearest multiple of 5 is used to prevent inappropriate disclosure of health-related information.

PurposeCurrently the screening for lung cancer for risk groups is based on Computed Tomography (CT) or low dose CT (LDCT); however, the lung cancer death rate has not decreased significantly with people undergoing LDCT. We aimed to develop a simple reliable blood test for early detection of all types of lung cancer based on the immunogenicity of aberrant forms of BARD1 that are specifically upregulated in lung cancer.MethodsELISA assays were performed with a panel of BARD1 epitopes to detect serum levels of antibodies against BARD1 epitopes. We tested 194 blood samples from healthy donors and lung cancer patients with a panel of 40 BARD1 antigens. Using fitted Lasso logistic regression we determined the optimal combination of BARD1 antigens to be used in ELISA for discriminating lung cancer from healthy controls. Random selection of samples for training sets or validations sets was applied to validate the accuracy of our test.ResultsFitted Lasso logistic regression models predict high accuracy of the BARD1 autoimmune antibody test with an AUC = 0.96. Validation in independent samples provided and AUC = 0.86 and identical AUCs were obtained for combined stages 1–3 and late stage 4 lung cancers. The BARD1 antibody test is highly specific for lung cancer and not breast or ovarian cancer.ConclusionThe BARD1 lung cancer test shows higher sensitivity and specificity than previously published blood tests for lung cancer detection and/or diagnosis or CT scans, and it could detect all types and all stages of lung cancer. This BARD1 lung cancer test could therefore be further developed as i) screening test for early detection of lung cancers in high-risk groups, and ii) diagnostic aid in complementing CT scan.

About Dataset 📌 Overview This dataset has been carefully synthesized to support research in lung cancer survival prediction, enabling the development of models that estimate:

Whether a patient is likely to survive at least one year post-diagnosis (Binary Classification). The probability of survival based on clinical and lifestyle factors (Regression Analysis). The dataset is designed for machine learning and deep learning applications in medical AI, oncology research, and predictive healthcare.

📜 Dataset Generation Process The dataset was generated using a combination of real-world epidemiological insights, medical literature, and statistical modeling. The feature distributions and relationships have been carefully modeled to reflect real-world clinical scenarios, ensuring biomedical validity.

📖 Medical References & Sources The dataset structure is based on well-established lung cancer risk factors and survival indicators documented in leading medical research and clinical guidelines:

World Health Organization (WHO) Reports on lung cancer epidemiology. National Cancer Institute (NCI) & American Cancer Society (ACS) guidelines on lung cancer risk factors and treatment outcomes. The IASLC Lung Cancer Staging Project (8th Edition): Standard reference for lung cancer staging. Harrison’s Principles of Internal Medicine (20th Edition): Provides an in-depth review of lung cancer diagnosis and treatment. Lung Cancer: Principles and Practice (2022, Oxford University Press): Clinical insights into lung cancer detection, treatment, and survival factors. 🔬 Features of the Dataset Each record in the dataset represents an individual’s clinical condition, lifestyle risk factors, and survival outcome. The dataset includes the following features:

1️⃣ Patient Demographics Age → A key risk factor for lung cancer progression and survival. Gender → Male and female lung cancer survival rates can differ. Residence → Urban vs. Rural (impact of environmental factors). 2️⃣ Risk Factors & Lifestyle Indicators These factors have been linked to lung cancer risk in epidemiological studies:

Smoking Status → (Current Smoker, Former Smoker, Never Smoked). Air Pollution Exposure → (Low, Moderate, High). Biomass Fuel Use → (Yes/No) – Associated with household air pollution. Factory Exposure → (Yes/No) – Industrial exposure increases lung cancer risk. Family History → (Yes/No) – Genetic predisposition to lung cancer. Diet Habit → (Vegetarian, Non-Vegetarian, Mixed) – Nutritional impact on cancer progression. 3️⃣ Symptoms (Primary Predictors) These are key clinical indicators associated with lung cancer detection and severity:

Hemoptysis (Coughing Blood) Chest Pain Fatigue & Weakness Chronic Cough Unexplained Weight Loss 4️⃣ Tumor Characteristics & Clinical Features Tumor Size (mm) → The size of the detected tumor. Histology Type → (Adenocarcinoma, Squamous Cell Carcinoma, Small Cell Carcinoma). Cancer Stage → (Stage I to Stage IV). 5️⃣ Treatment & Healthcare Facility Treatment Received → (Surgery, Chemotherapy, Radiation, Targeted Therapy). Hospital Type → (Private, Government, Medical College). 6️⃣ Target Variables (Predicted Outcomes) Survival (Binary) → 1 (Yes) if the patient survives at least 1 year, 0 (No) otherwise. Survival Probability (%) (Can be derived) → Estimated probability of survival within one year. ⚡ Why This Dataset is Valuable? ✅ Balanced Data Distribution Designed to ensure a representative distribution of lung cancer survival cases. Prevents model bias and improves generalization in predictive models. ✅ Medically-Inspired Feature Engineering Features are derived from real-world lung cancer risk factors, validated through medical literature. Incorporates both lifestyle and clinical indicators to enhance predictive accuracy.(no real person data is used,just have made an biomedical environment) ✅ Diverse Risk Factors Considered Smoking, air pollution, and genetic history as primary lung cancer contributors. Symptom severity and tumor histology influence survival rates. ✅ Scalability & ML Suitability Ideal for classification and regression tasks in machine learning. Can be used with deep learning (TensorFlow, PyTorch), ML models (XGBoost, Random Forest, SVM), and explainable AI techniques like SHAP and LIME. 📂 Dataset Usage & Applications This dataset is highly useful for multiple healthcare AI applications, including:

🩺 Predictive Analytics → Early detection of high-risk lung cancer patients. 🤖 Healthcare Chatbots → AI-powered risk assessment tools.

A measure of the number of adults diagnosed with breast, lung or colorectal cancer in a year who are still alive one year after diagnosis.

ONS still publish survival percentages for individual types of cancers. These can be found at: http://www.ons.gov.uk/ons/rel/cancer-unit/cancer-survival/cancer-survival-in-england--patients-diagnosed-2007-2011-and-followed-up-to-2012/index.html

A time series for one-year survival figures for breast, lung and colorectal cancer individually (previous NHS Outcomes Framework indicators 1.4.i, 1.4.iii and 1.4.v) is still published and can be found under the link 'Indicator data - previous methodology (.xls)' below.

Purpose

This indicator attempts to capture the success of the NHS in preventing people from dying once they have been diagnosed with breast, lung or colorectal cancer.

Current version updated: Feb-14

Next version due: To be confirmed

Analysis of ‘ lung cancer’ provided by Analyst-2 (analyst-2.ai), based on source dataset retrieved from https://www.kaggle.com/h13380436001/h-lung-cancer on 28 January 2022.

--- Dataset description provided by original source is as follows ---

Gender: M(male), F(female) Age: Age of the patient Smoking: YES=2 , NO=1. Yellow fingers: YES=2 , NO=1. Anxiety: YES=2 , NO=1. Peer_pressure: YES=2 , NO=1. Chronic Disease: YES=2 , NO=1. Fatigue: YES=2 , NO=1. Allergy: YES=2 , NO=1. Wheezing: YES=2 , NO=1. Alcohol: YES=2 , NO=1. Coughing: YES=2 , NO=1. Shortness of Breath: YES=2 , NO=1. Swallowing Difficulty: YES=2 , NO=1. Chest pain: YES=2 , NO=1. Lung Cancer: YES , NO.

Content

What's inside is more than just rows and columns. Make it easy for others to get started by describing how you acquired the data and what time period it represents, too.

Acknowledgements

We wouldn't be here without the help of others. If you owe any attributions or thanks, include them here along with any citations of past research.

Inspiration

Your data will be in front of the world's largest data science community. What questions do you want to see answered?

--- Original source retains full ownership of the source dataset ---

A measure of the number of adults diagnosed with breast, lung or colorectal cancer in a year who are still alive five years after diagnosis. ONS still publish survival percentages for individual types of cancers. These can be found at: http://www.ons.gov.uk/ons/rel/cancer-unit/cancer-survival/cancer-survival-in-england--patients-diagnosed-2007-2011-and-followed-up-to-2012/index.html A time series for five-year survival figures for breast, lung and colorectal cancer individually (previous NHS Outcomes Framework indicators 1.4.ii, 1.4.iv and 1.4.vi) is still published and can be found under the link 'Indicator data - previous methodology (.xls)' below. Purpose This indicator attempts to capture the success of the NHS in preventing people from dying once they have been diagnosed with breast, lung or colorectal cancer. Current version updated: May-14 Next version due: To be confirmed

ObjectiveAutoantibodies have been reported to be associated with cancers. As a biomarker, autoantibodies have been widely used in the early screening of lung cancer. However, the correlation between autoantibodies and the prognosis of lung cancer patients is poorly understood, especially in the Asian population. This retrospective study investigated the association between the presence of autoantibodies and outcomes in patients with lung cancer.MethodsA total of 264 patients diagnosed with lung cancer were tested for autoantibodies in Henan Provincial People’s Hospital from January 2017 to June 2022. The general clinical data of these patients were collected, and after screening out those who met the exclusion criteria, 151 patients were finally included in the study. The Cox proportional hazards model was used to analyze the effect of autoantibodies on the outcomes of patients with lung cancer. The Kaplan-Meier curve was used to analyze the relationship between autoantibodies and the overall survival of patients with lung cancer.ResultsCompared to lung cancer patients without autoantibodies, those with autoantibodies had an associated reduced risk of death (HRs: 0.45, 95% CIs 0.27~0.77), independent of gender, age, smoking history, pathological type, and pathological stage of lung cancer. Additionally, the association was found to be more significant by subgroup analysis in male patients, younger patients, and patients with small cell lung cancer. Furthermore, lung cancer patients with autoantibodies had significantly longer survival time than those without autoantibodies.ConclusionThe presence of autoantibodies is an independent indicator of good prognosis in patients with lung cancer, providing a new biomarker for prognostic evaluation in patients with lung cancer.

Among many types of cancers, to date, lung cancer remains one of the deadliest cancers around the world. Many researchers, scientists, doctors, and people from other fields continuously contribute to this subject regarding early prediction and diagnosis. One of the significant problems in prediction is the black-box nature of machine learning models. Though the detection rate is comparatively satisfactory, people have yet to learn how a model came to that decision, causing trust issues among patients and healthcare workers. This work uses multiple machine learning models on a numerical dataset of lung cancer-relevant parameters and compares performance and accuracy. After comparison, each model has been explained using different methods. The main contribution of this research is to give logical explanations of why the model reached a particular decision to achieve trust. This research has also been compared with a previous study that worked with a similar dataset and took expert opinions regarding their proposed model. We also showed that our research achieved better results than their proposed model and specialist opinion using hyperparameter tuning, having an improved accuracy of almost 100% in all four models.

This table contains 33048 series, with data for years 2000/2002 - 2010/2012 (not all combinations necessarily have data for all years), and was last released on 2016-03-16. This table contains data described by the following dimensions (Not all combinations are available): Geography (36 items: Total, census metropolitan areas; St. John's, Newfoundland and Labrador; Halifax, Nova Scotia;Moncton, New Brunswick; ...), Sex (3 items: Both sexes; Males; Females), Indicators (2 items: Mortality; Potential years of life lost), Selected causes of death (ICD-10) (17 items: Total, all causes of death; All malignant neoplasms (cancers); Colorectal cancer; Lung cancer; ...), Characteristics (9 items: Number; Low 95% confidence interval, number; High 95% confidence interval, number; Rate; ...).

One-year and five-year net survival for adults (15-99) in England diagnosed with one of 29 common cancers, by age and sex.

This dataset presents the footprint of cancer mortality statistics in Australia for all cancers combined and the 6 top cancer groupings (colorectal, leukaemia, lung, lymphoma, melanoma of the skin and pancreas) and their respective ICD-10 codes. The data spans the years 2009-2013 and is aggregated to Greater Capital City Statistical Areas (GCCSA) from the 2011 Australian Statistical Geography Standard (ASGS).

Mortality data refer to the number of deaths due to cancer in a given time period. Cancer deaths data are sourced from the Australian Institute of Health and Welfare (AIHW) 2013 National Mortality Database (NMD).

For further information about this dataset, please visit:

• Australian Institute of Health and Welfare - Cancer Incidence and Mortality Across Regions (CIMAR) books.

• Australian Institute of Health and Welfare - 2013 National Mortality Database.

Please note:

• AURIN has spatially enabled the original data.

• Due to changes in geographic classifications over time, long-term trends are not available.

• Values assigned to "n.p." in the original data have been removed from the data.

• The Australian and jurisdictional totals include people who could not be assigned a GCCSA. The number of people who could not be assigned a GCCSA is less than 1% of the total.

• The Australian total also includes residents of Other Territories (Cocos (Keeling) Islands, Christmas Island and Jervis Bay Territory).

• Cause of Death Unit Record File data are provided to the AIHW by the Registries of Births, Deaths and Marriages and the National Coronial Information System (managed by the Victorian Department of Justice) and include cause of death coded by the Australian Bureau of Statistics (ABS). The data are maintained by the AIHW in the NMD.

• Year refers to year of occurrence of death for years up to and including 2012, and year of registration of death for 2013. Deaths registered in 2011 and earlier are based on the final version of cause of death data; deaths registered in 2012 and 2013 are based on revised and preliminary versions, respectively and are subject to further revision by the ABS.

• Cause of death information are based on underlying cause of death and are classified according to the International Classification of Diseases and Related Health Problems (ICD). Deaths registered in 1997 onwards are classified according to the 10th revision (ICD-10).

• Colorectal deaths presented are underestimates. For further information, refer to "Complexities in the measurement of bowel cancer in Australia" in Causes of Death, Australia (ABS cat. no. 3303.0).

BackgroundNon-small-cell lung cancer (NSCLC) patients with ipsilateral pleural dissemination are defined as M1a in the eighth of American Joint Committee on Cancer (AJCC) TNM staging. We aimed to build a nomogram to predict lung cancer specific survival (LCSS) of NSCLC patients with ipsilateral pleural dissemination and to compare the impact of primary tumor resection (PTR) on LCSS among patients with different features.MethodsA total of 3,918 NSCLC patients with ipsilateral pleural dissemination were identified from the Surveillance, Epidemiology, and End Results (SEER) database. We selected and integrated significant prognostic factors based on competing risk regression to build a nomogram. The model was subjected to internal validation within SEER cohort and external validation with the cohort of 97 patients from Peking University People’s Hospital.ResultsAge (P < 0.001), gender (P = 0.037), T stage (P = 0.002), N stage (P < 0.001), metastasis pattern (P = 0.005), chemotherapy (P < 0.001), and PTR (P < 0.001) were independent prognostic factors. The calibration curves presented a good consistency and the Harrell’s C-index of nomogram were 0.682 (95%CI: 0.673–0.691), 0.687 (95%CI: 0.670–0.704) and 0.667 (95%CI: 0.584–0.750) in training, internal, and external validation cohort, respectively. Interaction tests suggested a greater LCSS difference caused by PTR in patients without chemotherapy (P < 0.001).ConclusionsWe developed a nomogram based on competing risk regression to reliably predict prognosis of NSCLC patients with ipsilateral pleural dissemination and validated this nomogram in an external Chinese cohort. This novel nomogram might be a practical tool for clinicians to anticipate the 1-, 3- and 5-year LCSS for NSCLC patients with pleural dissemination. Subgroup analysis indicated that patients without chemotherapy could get more benefit from PTR. In order to assess the role of PTR in the management of M1a patients more accurately, further prospective study would be urgently required.

ObjectiveTo compare the performance of three machine learning algorithms with the tumor, node, and metastasis (TNM) staging system in survival prediction and validate the individual adjuvant treatment recommendations plan based on the optimal model.MethodsIn this study, we trained three machine learning madel and validated 3 machine learning survival models-deep learning neural network, random forest and cox proportional hazard model- using the data of patients with stage-al3 NSCLC patients who received resection surgery from the National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) database from 2012 to 2017,the performance of survival predication from all machine learning models were assessed using a concordance index (c-index) and the averaged c-index is utilized for cross-validation. The optimal model was externally validated in an independent cohort from Shaanxi Provincial People’s Hospital. Then we compare the performance of the optimal model and TNM staging system. Finally, we developed a Cloud-based recommendation system for adjuvant therapy to visualize survival curve of each treatment plan and deployed on the internet.ResultsA total of 4617 patients were included in this study. The deep learning network performed more stably and accurately in predicting stage-iii NSCLC resected patients survival than the random survival forest and Cox proportional hazard model on the internal test dataset (C-index=0.834 vs. 0.678 vs. 0.640) and better than TNM staging system (C-index=0.820 vs. 0.650) in the external validation. The individual patient who follow the reference from recommendation system had superior survival compared to those who did not. The predicted 5-year-survival curve for each adjuvant treatment plan could be accessed in the recommender system via the browser.ConclusionDeep learning model has several advantages over linear model and random forest model in prognostic predication and treatment recommendations. This novel analytical approach may provide accurate predication on individual survival and treatment recommendations for resected Stage-iii NSCLC patients.

This dataset presents the footprint of male cancer mortality statistics in Australia for all cancers combined and the 11 top cancer groupings (bladder, colorectal, head and neck, kidney, leukaemia, lung, lymphoma, melanoma of the skin, pancreas, prostate and stomach) and their respective ICD-10 codes. The data spans the years 2009-2013 and is aggregated to Greater Capital City Statistical Areas (GCCSA) from the 2011 Australian Statistical Geography Standard (ASGS). Mortality data refer to the number of deaths due to cancer in a given time period. Cancer deaths data are sourced from the Australian Institute of Health and Welfare (AIHW) 2013 National Mortality Database (NMD). For further information about this dataset, please visit: Australian Institute of Health and Welfare - Cancer Incidence and Mortality Across Regions (CIMAR) books. Australian Institute of Health and Welfare - 2013 National Mortality Database. Please note: AURIN has spatially enabled the original data. Due to changes in geographic classifications over time, long-term trends are not available. Values assigned to "n.p." in the original data have been removed from the data. The Australian and jurisdictional totals include people who could not be assigned a GCCSA. The number of people who could not be assigned a GCCSA is less than 1% of the total. The Australian total also includes residents of Other Territories (Cocos (Keeling) Islands, Christmas Island and Jervis Bay Territory). Cause of Death Unit Record File data are provided to the AIHW by the Registries of Births, Deaths and Marriages and the National Coronial Information System (managed by the Victorian Department of Justice) and include cause of death coded by the Australian Bureau of Statistics (ABS). The data are maintained by the AIHW in the NMD. Year refers to year of occurrence of death for years up to and including 2012, and year of registration of death for 2013. Deaths registered in 2011 and earlier are based on the final version of cause of death data; deaths registered in 2012 and 2013 are based on revised and preliminary versions, respectively and are subject to further revision by the ABS. Cause of death information are based on underlying cause of death and are classified according to the International Classification of Diseases and Related Health Problems (ICD). Deaths registered in 1997 onwards are classified according to the 10th revision (ICD-10). Colorectal deaths presented are underestimates. For further information, refer to "Complexities in the measurement of bowel cancer in Australia" in Causes of Death, Australia (ABS cat. no. 3303.0).